Archives: NNT Reviews

Source

Twelve high-quality trials have been completed, see list below

Efficacy Endpoints

Mortality, neurologic improvement

Harm Endpoints

Intracranial hemorrhage, mortality

Narrative

Thrombolysis has been rigorously studied in >60,000 patients for acute thrombotic myocardial infarction, and is proven to reduce mortality. It is theorized that thrombolysis may similarly benefit ischemic stroke patients, though a much smaller number (8120) has been studied in relevant, large scale, high quality trials thus far.

There are 12 such trials ^1-12. Despite the temptation to pool these data the studies are clinically heterogeneous. While statistical heterogeneity refers to variation in the results of different trials, and is quantifiable using p- and I2 values, clinical heterogeneity refers to variations in study design, setting, and population characteristics.* Data from multiple trials must be clinically and statistically homogenous to be validly pooled.¹⁴ Large thrombolytic studies demonstrate wide variations in anatomic stroke regions, small- versus large-vessel occlusion, clinical severity, age, vital sign parameters, stroke scale scores, and times of administration. Even within a prominent and recent meta-analysis examining a single agent this continues to be true, as supported by clear statistical heterogeneity in the most important outcomes of such trials (see Table 2).¹⁵

Examining each study individually is therefore, in our opinion, both more valid and more instructive. The reference list below includes information from each study including size, primary outcomes, results, and times to treatment. Studies in green conclude a benefit. Those in blue conclude no benefit. Those in red were stopped early due to harm.

Two of twelve studies suggest a benefit, NINDS-2 and ECASS-3. Unfortunately both use nonstandard data analysis and reporting, and there is a lack of clarity in both. To wit, the ‘NINDS’ publication is two studies. The first (NINDS-1) tested neurologic improvement at 24 hours and found no benefit. NINDS-2 then sought “a difference of 20 percentage points" at 90 days between groups, and it is unclear whether this meant a difference between groups in degree of improvement (i.e. the outcome in NINDS-1) or in 'the chance of a good outcome’. The paper reported only the latter. This is unfortunate because subjects in the thrombolytic arm experienced milder strokes than those in the placebo arm. Patients with milder strokes are obviously more likely to achieve a ‘good outcome’, making degree of improvement a much better comparison.

Ultimately NINDS-2 reported that 12% more subjects experienced ‘good’ outcomes in the thrombolytic group. Despite being not close to the 20% goal this is reported as a statistically significant difference. The choice not to analyze the data as they had in the first study remains unexplained and non-intuitive.

The second study suggesting benefit is ECASS-3, which again exhibited severity imbalances that favored the thrombolytic arm, and again dichotomized outcomes as good/bad. Moreover, a mathematical error of p-value calculations in ECASS-3 calls into question the published result.¹⁶

Regardless of the irregularities in these two industry-funded investigations, they remain the only two claiming benefit. In comparison, twice as many studies showed harm and these were stopped early. This early stoppage means that the number of subjects in studies demonstrating harm would have included over 2400 subjects based on originally intended enrollments. Pooled analyses are therefore missing these phantom data, which would have further eroded any aggregate benefits. In their absence, any pooled analysis is biased toward benefit. Despite this, there remain five times as many trials showing harm or no benefit (n=10) as those concluding benefit (n=2), and 6675 subjects in trials demonstrating no benefit compared to 1445 subjects in trials concluding benefit.

Finally, the issue of time windows remains open as an important confounding variable. Are thrombolytic agents, for instance, beneficial in the setting of the first 3h from onset? This argument would be the strongest time window argument to make based on NINDS-2 and subgroup results from IST-3. However, presuming that early (0-3h) administration is better than later administration (3-4.5h or 4.5-6h) the subgroup results of IST-3 suggest an implausible biological effect in which early administration is beneficial, 3-4.5h administration is harmful, and 4.5-6h administration is again beneficial. Incidentally, each of these subgroups in IST-3 represents a larger 'n' than any existing single trial of thrombolytics prior to IST-3, making the comparison extremely powerful. Based on these time window effects the p-value calculation testing for the hypothesis that time to administration was a significant predictor of effect in IST-3 was 0.61 (see figure 3 in the original publication). This essentially removes time as an important variable, an argument first proposed by the Cochrane Collaboration review based on their mathematical calculations of data generated prior to IST-3.¹⁷

Thrombolytics for ischemic stroke may be harmful or beneficial. The answer remains elusive. We struggled therefore, debating between a 'yellow' or 'red' light for our recommendation. However, over 60,000 subjects in trials of thrombolytics for coronary thrombosis suggest a consistent beneficial effect across groups and subgroups, with no studies suggesting harm. This consistency was found despite a very small mortality benefit (2.5%), and a very narrow therapeutic window (1% major bleeding). In comparison, the variation in trial results of thrombolytics for stroke and the daunting but consistent adverse effect rate caused by ICH suggested to us that thrombolytics are dangerous unless further study exonerates their use.

* There are ongoing debates about how much clinical heterogeneity is too much, but there are extremes of agreement. It would be possible to combine, for instance, prayer and surgery into a review of 'The Effectiveness of Therapy for Back Pain'. The two will have about the same effect at 6 months and therefore will be 'statistically homogeneous'. But the two therapies are physiologically and mechanically different, and should not be pooled. Thus low statistical heterogeneity, in isolation, is inadequate justification for pooling data, because clinical heterogeneity must be low as well.

** Trials that were stopped early due to harm would, presumably, have showed just as much or more harmful effect had they continued to completion. Had the full complement of subjects been enrolled for instance in the ATLANTIS-B trial there would have been 968 subjects rather than 613. The additional 355 subjects would have made the results of the trial statistically more powerful and therefore would have more forcefully counterbalanced the benefits from trials suggesting benefit (both of which were run to completion). The lack of completion of 5 trials because of harms therefore creates an irreversible bias in any data review, whereby the trials that suggest benefit are completed, and become more statistically powerful, while the impact of trials that show harmful effects are weakened.

Caveats

There is a Cochrane review that pooled estimates of effect. ¹⁷ We do not endorse this choice because of clinical heterogeneity. However, we present the NNT’s from the pooled analysis for the reader's benefit. The Cochrane review suggested a 6% reduction in disability (based on a dichotomy in which a modified Rankin score of 0 or 1 is 'favorable' and 2 or greater is 'unfavorable') with thrombolytics. This would mean that 17 were treated for every 1 avoiding an unfavorable outcome. The review also noted a 1% increase in mortality (1 in 100 patients die because of thrombolytics) and a 5% increase in nonfatal intracranial hemorrhage (1 in 20), for a total of 6% harmed (1 in 17 suffers death or brain hemorrhage).

1. MAST-Italy, 1995 (n = 622) Lancet 1995; 346: 1509-14
   
   
   
   • Time to treatment:  < 6 hours
   
   
   • Results:  Increased early death (OR 2.7), Slight decrease 6 m disability (OR 0.5)
   
   
   • Overall:  No benefit
   
   
   
   


2. ECASS, 1995 (n = 620) JAMA 1995; 274(13): 1017-25
   
   
   
   • Time to treatment: < 6 hours, Notable inclusion: Moderate - severe hemispheric stroke
   
   
   • Results: No difference in disability or death
   
   
   • Overall: No benefit
   
   
   
   


3. NINDS-1, 1995 (n = 291) New Engl J Med 1995; 333:
   
   
   
   • Time to treatment: < 3 hours
   
   
   • Results: No difference NIHSS improvement by 4 pts or resolution of deficits at 24 hours
   
   
   • Overall: No benefit
   
   
   
   


4. NINDS-2, 1995 (n = 333) N Engl J Med 1995; 333
   
   
   
   • Time to treatment: < 3 hours
   
   
   • Results: Planned primary outcome was to be improvement in functional and stroke scores (of 20%). This was changed, post-hoc, to dichotomous favorable vs. unfavorableoutcome. Original primary outcome result not reported.
   
   
   • Overall: Benefit = NNT of 8 for post-hoc ‘favorable’ outcome measure
   
   
   
   


5. MAST-Europe, 1996 (n = 310***) N Engl J Med 1996;335:145.
   
   
   
   • Time to treatment: < 6 hours   Notable Inclusion: Mod-severe stroke, MCA territory only Outcome: No difference combined disability/death at 6 months; increased ICH (21% vs. 3%) and statistically nonsignificant increased mortality (47% vs. 38%)
   
   
   • Overall: Harmful
   
   
   • ***STOPPED EARLY DUE TO ICH AND MORTALITY, n of 600 planned
   
   
   
   


6. ASK, 1996 (n = 340***) JAMA 1996; 279: 961.
   
   
   
   • Time to treatment: <4h (small % 4-5h)
   
   
   • Outcome: No difference combined disability/death at 3 months; slightly decreased disability and increased mortality at 3 months, OR 1.83 (1.14-2.93)
   
   
   • Overall: Harmful
   
   
   • ***STOPPED EARLY DUE TO MORTALITY, n of 600 planned
   
   
   
   


7. ECASS-II, tPA 1998 (n = 800) Lancet 1998; 352: 1245-51
   
   
   
   • Time to treatment: <6h (20% < 3 hours)
   
   
   • Outcome: No difference in favorable outcomes (modified Rankin Scale) at 3 months
   
   
   • Overall: No benefit
   
   
   
   


8. ATLANTIS-B, tPA 1999 (n = 613*) JAMA 1999; 282: 2019-26
   
   
   
   • Time to treatment:  20% 3-4 h, 70% 4-5
   
   
   • Outcome: No difference, favorable outcome at 3 months; increased ICH (7% vs. 1%) and  nonsignificant increase in mortality (11% vs. 7%)
   
   
   • Overall: Harmful
   
   
   • ***STOPPED EARLY BECAUSE “unlikely to prove beneficial” - n of 968 planned
   
   
   
   


9. ATLANTIS-A, tPA 2000 (n = 142*) Stroke 2000; 31: 811-16 
   
   
   
   • Time to treatment: Initially 0-6 hours, stopped enrolling 0-3 based on NINDS result
   
   
   • Outcome: Improvement in NIHSS score (4 pts) at 24h favored lytics (40% vs. 21%, p=0.02) but at 1 month favored placebo (75% vs. 60%, p=0.05). Increased ICH with tPA (11% vs. 0%) and increased mortality at 3 months (23% vs. 7%)
   
   
   • Overall: Harmful
   
   
   • ***STOPPED EARLY FOR HARM - n of 300 planned
   
   
   
   


10. DIAS-2, 2008 (n = 193)
    
    
    
    • Time to treatment: 3-9 hours  Notable inclusion: reversible ischemic penumbra on MR or CT
    
    
    • Outcome: No difference in favorable outcomes; Statistically nonsignificant increase in mortality for high dose desmoteplase group (stopped early for harm, analyzed, restarted)
    
    
    • Overall: No benefit
    
    
    
    


11. ECASS-III, tPA 2008 (n = 821) N Engl J Med 2008; 359: 1317-29
    
    
    
    • Time to treatment: 3-4.5 h
    
    
    • Outcome: More favorable outcomes with tPA (OR 1.34, 1.02-1.76), mortality no difference
    
    
    • Overall: Benefit = NNT of 15 for ‘favorable’ outcomes
    
    
    
    


12. IST-3, tPA 2012 (n = 3035) Lancet 2012; 379(9834): 2352-63
    
    
    
    • Time to treatment: 0-6 h
    
    
    • Outcome: Short term mortality increase of 4% (1st week); no difference found in primary outcome (% alive and independent at 6 months); authors report a secondary ordinal analysis that showed a "shift" in outcomes favoring thrombolytics
    
    
    • Overall: No benefit
    
    
    
    

• 13. Hoffman JR, Schriger DL. A graphic re-analysis of the NINDS trial. Ann Emerg Med. 2009; 54(3): 329-36


• 14. Thompson SG. Systematic Review: Why sources of heterogeneity in meta-analysis should be investigated. BMJ. 1994; 309(6965): 1351-1355
  


• 15. Wardlaw JM, Murray V, Berge E, del Zoppo G, Sandercock P, Lindley RL, Cohen G. Recombinant tissue plasminogen activator for acute ischaemic stroke: an updated systematic review and meta-analysis. Lancet. 2012;379(9834):2364-72. PubMed PMID: 22632907; PubMed Central PMCID: PMC3386494.


• 16. Shy BD. Implications of ECASS III Error on Emergency Department Treatment of Ischemic Stroke. J Emerg Med. 2012 Nov 7. doi:pii: S0736-4679(12)00655-5. 10.1016/j.jemermed.2012.05.014. [Epub ahead of print] PubMed PMID: 23141561.


• 17. Wardlaw JM, Murray V, Berge E, del Zoppo GJ. Thrombolysis for acute ischaemic stroke. Cochrane Database of Systematic Reviews 2009, Issue 4. Art. No.: CD000213. DOI: 10.1002/14651858.CD000213.pub2.
  

It has been argued that the difference between studies suggesting benefit and all other studies is the timing of the intervention. The Cochrane findings do not support this view, nor does one extensive post-hoc analysis.¹³ More importantly, the most recent study (ECASS-3) suggested benefits when thrombolytics were administered between 3 and 4.5 hours. This is outside the window that purportedly distinguished the NINDS trials, and effectively neutralizes time as the defining factor that separates trials concluding benefit from those that do not.

It has also been noted that differing agents have been utilized in trials, including streptokinase and urokinase and t-Pa. There is no theoretical basis, nor any clinical data from the thrombolytics for MI, that would suggest that t-Pa is less likely to cause ICH or more likely to demonstrate benefit, than any other agent.

Author

David Newman, MD

Published/Updated

March 25, 2013

Source

Ratilala BO, Costa J, Sampaio C. Antibiotic prophylaxis for surgical introduction of intracranial ventricular shunts. Cochrane Database of Systematic Reviews 2009, Issue 1. Art No.: CD005365.

Efficacy Endpoints

All-cause mortality, infection of the device

Harm Endpoints

Adverse events of antibiotics

Narrative

Often patients develop a very dangerous condition, where brain swelling compresses brain matter within the skull. This can cause a patient to experience anything from mild symptoms of headache to rapid death. Devices are often surgically placed into the ventricles of the brain to ‘decompress’ the swelling. In certain situations when the pressure is dangerously high, the devices may be life saving. Ventricular shunts work by draining cerebrospinal fluid (CSF) to another part of the body or to the outside environment. As with any surgical procedure, placement of a ventricular shunt poses risks including the introduction of infections, which are typically devastating and often fatal. As a result researchers have studied the use of systemic antibiotics and also antibiotic-impregnated devise to reduce the infection risk. This Cochrane review examines the best evidence for using systemic antibiotics and catheter impregnated antibiotics to prevent infections when placing intraventricular devices.
The Cochrane review included seventeen randomized control trials with 2134 patients. For the outcome of all-cause mortality, there were insufficient data. In terms of infection of the device the review found that both systemic antibiotics and antibiotic-impregnated catheters decreased infections with an NNT of 20 and 11, respectively. Additionally, a preplanned subgroup analysis by the Cochrane reviewers found that additional antibiotics (more than just around the time of the procedure) did not confer additional benefit.

Caveats

All the included trials were randomized, however details of the study protocol and methods (randomization, concealment, etc.) for some trials were unknown. Additionally, all studies were analyzed per-protocol and not intention-to-treat, which presents a serious bias in the individual results. As a result, this calls into question the validity of the results found by the Cochrane Reviewers. Furthermore, the trials mostly included patients receiving internal shunts and analyzed data for both adults and children in aggregate. Only one study looked at a very small subset of patients that received an external shunt and found no difference in infection rates with the use of antibiotics, however it was a small sample and no conclusions can be drawn.

Unfortunately, the review reported that only two of the included trials had data on adverse events from the antibiotics. However, the two studies used different drugs and different patient populations. One was stopped early for harm from the antibiotics, which may indicate an under-reported major problem, and the second article did not report their harm data in the published paper.

Overall, while there are serious methodologic limitations, and we would like to see a large, high quality randomized trial on the subject, it appears that a current best guess suggests that antibiotics (either systemically or impregnated into the device) will reduce shunt infections.

Author

Jarone Lee, MD, MPH

Published/Updated

February 11, 2013

Source

Hoenig MR, Aroney CN, Scott IA. Early invasive versus conservative strategies for unstable angina and non-ST elevation myocardial infarction in the stent era. Cochrane Database of Systematic Reviews 2010, Issue 3. Art. No.: CD004815

Efficacy Endpoints

Death, myocardial infarction (heart attack), angina (chest pain)

Harm Endpoints

Bleeding complications, peri-procedural MI (heart attack), death

Narrative

Reperfusion therapy for major heart attacks (i.e. ST-elevation myocardial infarction) is beneficial1. Controversy exists regarding the balance of harms and benefits of revascularization therapy for other types of acute coronary syndromes. This Cochrane review aggregates the best available data from five trials of 7818 subjects randomized to either undergo immediate invasive management with coronary catheterization and stent placement (as necessary) or else be treated with medications and no immediate invasive strategy.

The overall results were mixed. Patients included in the studies were predominantly those with biomarker evidence of ischemia, and most often with electrocardiographic changes indicating ischemia as well (though not ST-elevation). Ultimately, deaths were statistically unaffected, although there appeared to be what the authors refer to as a ‘trend’ toward increased mortality in the group managed with an early invasive strategy (a statistically nonsignificant difference of 0.5% in deaths, or roughly 1 in every 200 subjects). Heart attacks were reduced by 2% over the next year, but increased by 3% in the peri-procedural period. Refractory angina (ongoing chest pain) was present in 22% of those randomized to an invasive strategy and 34% of those treated conservatively, and was thus reduced.

Harms of the procedure were apparent, both in the heart attack rate noted above and also in the rate of bleeding events, although these were typically minor and usually involved the site of arterial puncture for the procedure.

Caveats

The cohort here includes a preponderance of patients suffering overt coronary ischemia including positive serum biomarkers. When the authors analyzed only studies of patients without positive biomarkers there was a statistically significant increase in deaths (RR 1.71, NNT=200). Indeed, the single study in the review designed and powered to test the question of whether benefits are confined to biomarker-positive patients2 found that while death was not reduced overall, MI was slightly reduced (2.1%), However, MIs were reduced only in the subgroup with positive biomarkers. These results suggest that the reduction in heart attacks is likely to be confined to those with a positive biomarker, while there is an increase in deaths among those who are biomarker negative. Furthermore, the Cochrane authors note that when confining their analysis to studies in which 100% of subjects had biomarker elevations they were still unable to detect a mortality benefit to early invasive therapy.

This raises two concerning possibilities: 1) that there may be an increase in death overall, as reflected in the 0.5%, nearly statistically significant difference in the review, and 2) that there is mortal danger to invasive therapy among patients without biomarker elevation.

There are clear limitations to these data. In many of the studies a significant proportion of patients, typically those who fail medical therapy due to ongoing pain or new biomarker elevations, ultimately underwent invasive therapy (typically 20-30%), making it impossible to discern whether having invasive therapy provides benefit over strictly non-invasive therapy. In addition, there is a decrease in rehospitalization among those undergoing early invasive therapy. For a variety of reasons the decision to hospitalize is subjective, and biased in favor of invasive strategies (invasive strategies are a ‘step-up’ therapy, thus the subjects randomized to noninvasive therapy who have ongoing symptoms are far more likely to be readmitted for this). Therefore we have not included this outcome in the final numbers. While we consider it a biased outcome, we offer here the NNT for rehospitalization for those readers who disagree with our assessment of bias (NNT for rehospitalization=10).


Finally, the suggested benefits in refractory angina (i.e. ongoing chest pain) are difficult to interpret. First, there is often a considerable meaning response (i.e. placebo effect) associated with undergoing invasive management, even in the case of sham procedures, and this has classically been apparent for anginal pain associated with coronary ischemia.3 4 Second, while it is clear that most patients prioritize avoiding heart attacks and death, the patient-centered value of pain reduction when it is achieved in just 1 out of 9 patients and entails an invasive procedure with considerable risk and discomfort, is on much shakier ground.

Because of our concerns about a potential increase in mortality, alongside a nominal reduction in nonfatal MI that is equaled or eclipsed by peri-procedural MIs and bleeding events, and despite a reduction in symptoms, we have classified the early invasive strategy for coronary syndromes as ‘Red’. This is largely based on our belief that patients will typically value the avoidance of a 1 in 200 risk of death more than they will value a 1 in 50 chance of preventing a future nonfatal MI, and a 1 in 9 chance of preventing future pain.

Thus we would counsel that, in aggregate, early invasive therapy is not a beneficial intervention. This is clearly true for patients without biomarker elevation. If there is a patient for whom the benefits outweigh the harms it would likely be the high risk patient with biomarker elevations who is judged by the treating physician to be ‘on the ladder’, which is to say showing dynamic changes and evolution that would suggest that STEMI is imminent or active in an early stage.

Additional caveat: for acute coronary syndromes not associated with positive biomarker testing we classify an early invasive strategy as ‘Black’, as current data strongly suggest that this leads to an increase in mortality without a significant decrease in MI.

We note that the Cochrane authors conclude that they believe that an early invasive strategy is superior to a noninvasive strategy. This seems likely to be based on the 2% overall reduction in MI at 6 months. We respectfully disagree that a 1 in 50 chance of an avoided nonfatal MI is adequate justification for this resource-intensive, invasive, risky procedure that appears to be potentially associated with an increase in mortality.

Author

David H. Newman, MD

Published/Updated

Source

Diao D, Wright JM, Cundiff DK, Gueyffier F. Pharmacotherapy for mild hypertension. Cochrane Database of Systematic Reviews 2012, Issue 8. Art. No.: CD006742

Efficacy Endpoints

Mortality, stroke, coronary artery disease, cardiovascular events

Harm Endpoints

Stopping medication due to adverse events

Narrative

Hypertension affects almost 29% of adults in the United States, most of whom are taking medication to lower their blood pressure1. Blood pressure control has been shown to reduce the chances of developing cardiovascular problems and stroke (Mancia et al, 2009), however these reductions are derived from studies of patients with moderate or severe hypertension, and those with a history of prior cardiovascular events such as heart attack or stroke. However, evidence has been unclear on whether pharmacological treatment for previously healthy patients with ‘mild’ hypertension is beneficial.

This review included four randomized-controlled trials enrolling 8,912 subjects with mild elevations in blood pressure (systolic blood pressure 140-159 or diastolic blood pressure 90-99) without preexisting cardiovascular disease. Patient data for individuals satisfying the inclusion criteria were obtained from three studies; pooled data was used from the fourth study since it met the a priori inclusion criteria of having less than 20% of its total subjects with moderately elevated blood pressure.

At a period of four to five years follow up, no differences were seen in mortality, cardiovascular events, CAD, or stroke. Approximately 9% more patients in the treatment arms withdrew due to medication side effects.

Caveats

Studies included in this analysis were of variable quality, some with questionable randomization, incomplete blinding, or partial follow up. Additionally, antihypertensive agents were often older, or used in higher doses than in current practice. Subjects often received non-thiazide diuretics and beta blockers, rather than low-dose thiazides, ACE inhibitors, and calcium-channel blockers, drugs which appear to confer a slight advantage in outcomes2.

Included trials were often powered to detect composite endpoints and underpowered to evaluate individual outcome measures. For low-risk patients with mild hypertension, a study powered to detect differences in mortality or cardiovascular outcomes would require more subjects followed for more time. To account for this some authors have called for trials utilizing intermediate/surrogate endpoints such as left ventricular hypertrophy or microalbuminuria3. We disagree, because of the misleading results that such surrogate markers can often generate.

While data for higher risk patients do suggest a benefit from treatment of hypertension, and the lack of statistically significant benefits in low risk patients may be due to inadequate power, there are important notes of caution here. The high rate of drug discontinuation is concerning. Moreover, it is well known that occasional serious, potentially life threatening adverse events occur with antihypertensives (angioedema with ACE inhibitors, toxicity and bradycardia with beta blockers and calcium channel blockers, electrolyte disturbances with diuretics). These risks are reasonable when there is a proven benefit. In the absence of proof of benefit, the risk of inficting serious harm with the drugs becomes ethically dubious. In one study that represents nearly 80% of the data included, for instance, women experienced higher mortality in the treatment group than in the placebo group4, thus understanding impact on subgroups may be critical. In the final analysis this low-risk population may be a perfect example of a group in whom, despite documented relative risk reductions, extremely small absolute reductions (in rare outcome events) are trumped by increases in harm.

Ultimately, while we require more and higher-quality research to answer this question definitively in a contemporary milieu, these data strike us as being of adequate power and quality to label this intervention as ‘Red’, indicating no benefit. The reversible nature of the harms and the weakness of the data, however, suggest to us that a rating of ‘Black’ (harmful) would convey more certainty than is justified. Although we are aware of clinical guidelines that have come to a different conclusion5, 6, 7, 8, the data analyzed here do not support the treatment of mild hypertension for primary prevention.

Author

Gary Green, MD

Published/Updated

February 2, 2013

Source

Pinto RZ, Maher CG, Ferreria ML, Hancock M, Oliveira VC, McLachlan AJ, Koes B, Ferreira PH. Epidural corticosteroid injections in the management of sciatica: a systematic review and meta-analysis. Annals of Internal Medicine 2012.

Efficacy Endpoints

Pain, disability

Harm Endpoints

Infection, bleeding, nerve damage

Narrative

Sciatica, or the symptomatic pain believed to be associated with lumbar nerve root compression, affects up to 15% of the population. While many individuals achieve resolution with conservative therapy, a sizable minority go on to chronic pain. Epidural steroid injections have been utilized since the 1950s for pain not responsive to conservative treatment. By 2004, almost 1.5 million Medicare patients received such injections.1

The authors of this systematic review included 23 randomized, placebo-controlled trials involving over 2,000 subjects with symptoms of sciatica who had not undergone previous back surgery. The studies examined the efficacy of epidural steroid injections, delivered through one of three approaches (caudal, interlaminar, and transforaminal).

In the short term after injection (2 weeks – 3 months), while back pain was unaffected, leg pain was reduced by 6 points on a pain scale and disability was reduced by 3 points (both scales total 100 points). Validation studies and consensus statements regarding pain and disability scales suggest that a minimum change of 10 to 30 points is required for these effects to be considered clinically meaningful. There was no significant difference seen for back pain in the short term and no significant differences in leg pain, back pain, or disability in the long term (>12 months).

Serious adverse effects of epidural steroid injections are quite rare, but case reports include epidural abscesses, bacterial meningitis, and epidural hematomas. More common adverse effects include intravascular placement, headache, vasovagal symptoms, and local pain.

Caveats

While this review found short-term, statistically-significant reductions in leg pain and disability, the effect size was small and did not meet proposed thresholds for clinical significance. Moreover, steroids were compared to sham injections and not necessarily compared to contemporary conservative management, which may offer benefit, particularly in the short term. No differences were seen in the long term, a finding confirmed by a recent study examining two-year follow up.

Adverse events are rare but potentially serious, as witnessed by a recent cohort of patients who contracted fungal meningitis from contaminated epidural injections.2

The included studies encompassed a large number of patients and varied from fair to high quality. A minority of trials fulfilled complete blinding of providers, allocation concealment, and intention to treat analyses.

Notably, this study examined pain and disability due to sciatica, and not spinal stenosis or undifferentiated low back pain.

Given the clear lack of any clinically significant benefit in both the short and long term, the cost and the small, but real potential for procedural complications, we have chosen to classify this intervention as Red.

Author

Gary Green, MD

Published/Updated

January 11, 2013

Source

Crowther CA, Middleton P. Anti-D administration in pregnancy for preventing Rhesus alloimmunisation. Cochrane Database of Systematic Reviews 2009, Issue 1.
Clarke C, Hussey RM. Decline in deaths from rhesus haemolytic disease of the newborn. J R Coll Physicians Lond. 1994 Jul-Aug;28(4):310-1.
Cooper N. Intravenous immunoglobulin and anti-RhD therapy in the management of immune thrombocytopenia. Hematol Oncol Clin North Am. 2009 (6):1317-27.
Crowther CA, Middleton P. Anti-D administration after childbirth for preventing Rhesus alloimmunisation. Cochrane Database of Systematic Reviews 1997, Issue 2.
Duckett, J. R. A., & Constantine, G. (1997). The Kleihauer Technique: an accurate method of quantifying fetomaternal haemorrhage?. BJOG: An International Journal of Obstetrics & Gynaecology, 104(7), 845-846.
Duguid, J. K., & Bromilow, I. (1994). Value of Kleihauer testing after administration of anti-D immunoglobulin. BMJ, 309(6949), 240.
Kim YA, Makar RS. Detection of fetomaternal hemorrhage. Am J Hematol. 2012 (4):417-23.
Letsky, E. A., & De Silva, M. (1994). Preventing Rh immunisation. BMJ: British Medical Journal, 309(6949), 213.
MacKenzie IZ, Bichler J, Mason GC, Lunan CB, Stewart P, Al-Azzawi F, De Bono M, Watson N, Andresen I. Efficacy and safety of a new, chromatographically purified rhesus (D)immunoglobulin. Eur J Obstet Gynecol Reprod Biol. 2004 (2):154-61

Efficacy Endpoints

Rhesus Rh D alloimmunization, Rhesus disease

Harm Endpoints

Anemia requiring transfusion, renal failure, DIC, death

Narrative

• Alloimmunization, in which a Rh negative mother develops anti-Rh antibodies after being exposed to Rh positive blood during pregnancy or in the immediate intra-partum period, can lead to severe complications in future pregnancies. Rh antibodies are capable of crossing the placenta, entering the fetal circulation, and attacking fetal (Rh-positive) red blood cells. This can lead to massive hemolysis, profound anemia, and high output fetal cardiac failure (hydrops fetalis). Post-partum administration of Rh immunoglobulin prophylaxis has made these immune-mediated complications incredibly rare. Routine ante-partum administration of Rh immunoglobulin has the additional hypothesized role of preventing alloimmunization during pregnancy due to occult fetal-maternal hemorrhage.

This review included two relatively low quality trials involving over 4,500 Rh-negative unsensitized women. The studies examined the efficacy of IgG anti-D antibodies given at 28 and 34 weeks gestation to prevent transplacental fetal blood exposure and maternal alloimmunization. Infusion of anti-D in Rh-negative pregnant women significantly reduced the chance of a positive Kleihauer-Betke test at delivery. There was a non-significant trend towards reducing alloimmunization during pregnancy, at delivery, and up to 12 months post partum.

Adverse effects of Rh immunoglobulin, including clinically-significant anemia, renal failure, pulmonary edema, DIC, and death, appear to be exceedingly rare.

Caveats

• Maternal alloimmunization (the development of anti-Rh antibodies) has been widely accepted as a surrogate marker for potential fetal morbidity and mortality. This is supported by population-based studies showing a marked decrease in rhesus disease since the routine administration of post-partum anti-D immunoglobulin. While we feel that maternal alloimmunization has been validated as a surrogate marker and outcome measure, the clinical accuracy and importance of a positive Kleihauer-Betke or similar test is much less certain. The test is at best a weak proxy for potential alloimmunization,, and as such, is two steps removed from a patient-important outcome.

One of the review’s two cited studies showed non-significant trends toward reduced alloimmunization using a 100 mcg dose, while the other failed to show benefit with a 50 mcg dose. This dose response mirrors what has been seen in postpartum administration. Reporting data for the higher-dose study separately suggests a NNT for preventing alloimmunization of 213. This is roughly consistent with other non-randomized trials excluded by the Cochrane review but often cited as supporting data in obstetric guidelines. In general, the data for antepartum prophylaxis is less robust than that for postpartum women, with the systematic review showing only a trend toward benefit. Additionally, the review’s included studies suffer from relatively poor quality, lacking placebo controls, true blinding, and full randomization.

Adverse events from anti-D are exceedingly rare. Nearly all reports of significant events occur in the treatment of idiopathic thrombocytopenia (ITP). It should be noted that the doses used in ITP (up to 200 mcg/kg) are much higher than those used for the prevention of rhesus disease, and many ITP regimens require multiple doses. Therefore, it is not clear how well estimates of harm equate in these two populations. The available data on the use of anti-D in pregnancy suggests that significant adverse events are vanishingly rare, though these outcomes were not recorded in the trials involved in this study.

When administered to Rh-negative women during the third trimester, anti-D immunoglobulin may have a role in reducing Rhesus disease in future pregnancies. Given the excellent safety profile of the intervention, the risk to benefit ratio suggests that the use of anti-D immunoglobulin in the third trimester may be a reasonable option in regions with sufficient resources and access to the drug. However, we believe that the lack of conclusive quantitative evidence has brought the cost effectiveness of unproven antepartum therapy into question,. High-quality, randomized, placebo-controlled trials should be performed to assess whether either routine or event-based administration of anti-D immunoglobulin adds importantly to the routine use of the drug in the post-partum period.

Author

Gary Green, MD and Dan Runde, MD

Published/Updated

January 8, 2013

Source

Crowther CA, Middleton P. Anti-D administration after childbirth for preventing Rhesus alloimmunisation. Cochrane Database of Systematic Reviews 1997, Issue 2.
Clarke C, Hussey RM. Decline in deaths from rhesus haemolytic disease of the newborn. J R Coll Physicians Lond. 1994 Jul-Aug;28(4):310-1.
Cooper N. Intravenous immunoglobulin and anti-RhD therapy in the management of immune thrombocytopenia. Hematol Oncol Clin North Am. 2009 (6):1317-27.
MacKenzie IZ, Bichler J, Mason GC, Lunan CB, Stewart P, Al-Azzawi F, De Bono M, Watson N, Andresen I. Efficacy and safety of a new, chromatographically purified rhesus (D)immunoglobulin. Eur J Obstet Gynecol Reprod Biol. 2004 (2):154-61

Efficacy Endpoints

Rhesus Rh D alloimmunization at six months in a future pregnancy

Harm Endpoints

Maternal anemia requiring transfusion, renal failure, DIC, death

Narrative

• Alloimmunization, in which a Rh negative mother develops anti-Rh antibodies after being exposed to Rh positive blood during pregnancy or in the immediate intra-partum period, can lead to severe complications in future pregnancies. Rh antibodies are capable of crossing the placenta, entering the fetal circulation, and attacking fetal (Rh-positive) red blood cells. This can lead to massive hemolysis, profound anemia, and high output fetal cardiac failure (hydrops fetalis). Post-partum administration of Rh immunoglobulin prophylaxis has made these immune-mediated complications incredibly rare.

This review included six, randomized-controlled trials involving over 10,000 Rh-negative women without detectable anti-D antibodies who gave birth to Rh positive infants. The studies examined the efficacy of post-partum administration of IgG anti-D antibodies in preventing alloimmunization at six months, and in subsequent pregnancies.

Infusion of anti-D in at risk post-partum women significantly reduced detectable alloimmunization at six months and in future pregnancies. Adverse effects of Rh immunoglobulin, including clinically-significant anemia, renal failure, pulmonary edema, DIC, and death, appear to be exceedingly rare.

Caveats

Decreased rates of detectable anti-Rh antibodies do not directly prove that administration of anti-D immunoglobulin reduces actual neonatal mortality and morbidity. The reduction in alloimmunization, has, however, been widely accepted as a surrogate marker for potential fetal morbidity and mortality. This is supported by population-based studies showing a marked decrease in rhesus disease since the administration of post-partum anti-D immunoglobulin has become common practice. We therefore feel that this surrogate marker has been validated, and that it is a reasonable outcome measure for the efficacy of anti-D immunoglobulin in regions with resources and access to the drug.

As with any review, inference for clinical practice is limited by the quality of the included studies. Here, there was moderate variation in the quality of the included studies in terms of randomization, blinding, placebo control, and documented follow up. One trial, for instance, accounted for 47% of the total weight in the overall meta-analysis, and the combination of these studies found significant heterogeneity in the degree of benefit.

Adverse events from anti-D appear to be rare, and nearly all relevant case reports occur in the treatment of idiopathic thrombocytopenic purpura (ITP), during which doses are typically much higher than those used for postpartum women and many regimens require multiple doses.

Author

Gary Green, MD and Dan Runde, MD

Published/Updated

Source

Stergiopoulos K, Brown DL. Initial coronary stent implantation with medical therapy vs medical therapy alone for stable coronary artery disease: meta-analysis of randomized controlled trials. Arch Intern Med. 2012;172(4):312-9

Efficacy Endpoints

Death, nonfatal myocardial infarction, angina symptoms

Harm Endpoints

Procedural complications such as bleeding, stroke, kidney damage, death

Narrative

Percutaneous coronary interventions (typically stents) are used hundreds of thousands of times each year to open narrowed coronary arteries. Most such patients are not in the midst of an acute coronary syndrome such as a heart attack. This meta-analysis aimed to determine whether stenting (i.e. opening) as an initial approach to narrowed arteries is more beneficial than simply taking medicines to prevent future attacks or death.

Over 7000 patients with coronary narrowing were included from eight trials in which patients were randomly assigned to receive either stents plus optimal medical therapy, or optimal medical therapy alone (i.e. without stents). The studies were generally high quality and the results appear robust, suggesting that medical therapy without stenting is as effective at preventing deaths, heart attacks, and symptoms as the placement of coronary stents.

Caveats

Prior meta-analyses examining this topic have come to differing conclusions on this and similar questions, in some cases suggesting benefits to stenting,1 2 and in other cases suggesting no benefit.3 The differences are likely attributable to variations in the included studies. A strength of this review, and the reason we have given it primacy in our summary, is that studies chosen for this review best represent both stenting and medical therapy in their current form. In other words many prior studies used older methods (ballon angioplasty, etc.) for coronary opening, and in many studies patients receiving medical therapy did not receive the recommended pill regimens that contemporary CAD patients receive. Therefore this review appears to best represent a proper comparison of current PCI versus current medical therapy.

PCI and the coronary angiography that necessarily precedes PCI are both invasive procedures with harms. Morbidity from these procedures has been poorly documented and inadequately studied in a contemporary milieu, thus harm numbers are best-guess estimates, however it is widely accepted that major complications include stroke, kidney failure, heart attack, and death.4 5 The American Heart Association suggests that 2% of patients, or 1 in 50, suffers an important complication.6

It is also important to recall that coronary artery bypass graft surgery appears to benefit a small number of patients and PCI has in many cases become a less risky replacement for this surgery. Such cases are often patients with specific, severe patterns of coronary narrowing, or patients who are poor surgical candidates. However the studies in this review excluded such patients. Thus there will be cases of nonacute coronary disease for which PCI is appropriate or beneficial, and this review does not apply to such patients.

For the patients in this group of studies, however, who appear to represent the majority of patients currently eligible for PCI, there was no identifiable benefit to the procedure and there are established harms. While the frequency of these harms is not clear, their existence is, thus we have chosen to classify this intervention as ‘Black’.

Author

David Newman, MD

Published/Updated

June 18, 2012

Source

Duley L, Gülmezoglu AM, Henderson-Smart DJ, Chou D. Magnesium sulphate and other anticonvulsants for women with pre-eclampsia. Cochrane Database Syst Rev. 2010 Nov 10;(11):CD000025. Review. PubMed PMID: 21069663.
Weeks AD, Ononge S. The magpie trial. Lancet. 2002 Oct 26;360(9342):1331; author reply 1331-2. PubMed PMID: 12414232.

Efficacy Endpoints

Mother: Mortality, eclampsia, serious maternal morbidity related to preeclampsia (renal failure, liver failure, stroke, coagulopathy) Child: Mortality, preterm birth, NICU stay> 7 day

Harm Endpoints

Mother: Respiratory Depression, risk of cesarian section, postpartum hemorrhage (up to 2 years) Child: Mortality, neurosensory disability/ cerebral palsy (up to 18 months)

Narrative

Preeclampsia is a multisystem disorder usually associated with hypertension and proteinuria. Eclampsia, or seizures in the setting of pre-eclampsia, is a rare but known neurologic complication of preeclampsia accounting for 50,000 deaths worldwide (10% direct maternal death). Although the mechanism is unclear magneisum sulfate has been used to prevent eclampsia since the 1950s.

In this cochrane review, 11,444 women in 15 randomized trials were recruited from a mix of high, middle, and low income countries. The largest is the 33-country MagPie Trial comparing intravenous magnesium to placebo. The study was published in 2002 and at >10,000 subjects accounts for more than 87% of those in the review.

The use of magnesium sulfate was associated with a greater than 50% relative reduction in the risk of eclampsia (RR 0.41). As noted by the reviewers, a similar reduction in maternal mortality (RR 0.54) was also found, though the absolute number of deaths was small which may have kept this reduction from reaching statistical significance. A reduction in placental abruption was also noted (RR 0.64) as well as a small increase in the rate of cesarian section. No differences were seen in child outcomes.

As a secondary outcome, magnesium sulfate was more effective than phenytoin for reducing risk of eclampsia in 3 trials (RR 0.08) but also increased caesarian section in 2 trials (RR 1.21).

Toxicity with magnesium sulfate is rare with 0.5% of women having respiratory depression (RR 1.98). More minor side effects such as flushing (20%) were also noted.

Caveats

This review strongly supports the current use of magnesium sulfate as the drug of choice in the prevention of eclampsia, particularly in those with severe preeclampsia, and provides reasonable reassurance that this benefit in the short term is not associated with long-term negative sequelae for either women (up to 2 years) or children (up to 18 months). Although few women in these studies had mild preeclampsia, it would seem plausible that magnesium sulfate would also reduce the risk of eclampsia in these women. In addition, given the safety profile of magnesium sulfate and the ability to monitor toxicity clinically with limited training, serum monitoring is not required, allowing for wide applicability in hospital settings.

Author

Nadia Shaukat, MD

Published/Updated

March 13, 2012

Source

Rodrigo et al. A Meta-analysis of the Effects of Ipratropium Bromide in Adults with Acute Asthma. American Journal of Medicine. 1999;107:363–370
Stoodley. The Role of Ipratropium Bromide in the Emergency Management of Acute Asthma Exacerbation: A Metaanalysis of Randomized Clinical Trials. Annals of Emergency Medicine. 1999; 34(1) 8-18.
Rodrigo GJ, Rodrigo C. First-line therapy for adult patients with acute asthma receiving a multiple-dose protocol of ipratropium bromide plus albuterol in the emergency department. American Journal of Respiratory and Critical Care Medicine 2000; 161:1862-8.
Cydulka. RK et al. Levalbuterol versus levaalbuterol plus ipatropium in the treatment of severe acute asthma. Journal of Asthma. 2010 Dec; 47(10):1094-100.
Salo, D. A randomized, clinical trial comparing the efficacy of continuous nebulized albuterol (15 mg) versus continuous nebulized albuterol (15 mg) plus ipratropium bromide (2 mg) for the treatment of acute asthma. Journal of Emergency Medicine. 2006 Nov; 31(4):371-6.

Efficacy Endpoints

Need for hospitalization

Harm Endpoints

Medication side effects

Narrative

Acute asthma exacerbation is responsible for an estimated 1.6 million emergency department visits and over 440,000 hospital admissions per year1. The cornerstone of management remains aerosolized, short acting, β2-agonists and systemic corticosteroids2. Anticholinergic agents such as ipratropium bromide, are effective bronchodilators, but comparative studies suggest that they are not as effective as β2-agonists3. The studies reviewed here examined whether adding ipratropium reduced hospitalizations for adults with acute asthma exacerbation.

The review included seven randomized trials of almost 1500 subjects comparing the addition of ipratropium or the addition of placebo to β2-agonists therapy (albuterol, levalbuterol, or salmeterol). There may be a small benefit to adding ipratropium with a decrease in hospital admissions by ~9% (range 1%-39%). This means that for every 11.5 patients treated with combination therapy, there will be one less hospitalization admission. There was no increased incidence of adverse events reported.

Caveats

These data should be interpreted with caution as most studies did not specifically study hospitalization as their primary outcome. In fact, many recorded hospitalization but did not report this as a study outcome, focusing instead on surrogate markers of lung function (FEV1 or PEF). Admission decisions were often made after additional treatment, separate from the study protocol, at the discretion of the treating physician.4 5 6 7 8 9. Furthermore, there was significant clinical heterogeneity in these trials: β2-agonists and ipratropium doses varied, and use of systemic corticosteroids was not standardized.

Lastly, though hospitalization is a patient-oriented outcome, it is rarely based on a priori criteria10. Admission rates varied from 8.4-30%, which may be due to setting, patient, or physician characteristics. Therefore these data should be considered hypothesis generating rather than conclusive. However, national guidelines recommend the use of ipratropium in the management of acute asthma10, specifically to reduce hospital admissions. While this seems a reasonable ‘best guess’ based on the available evidence, we feel it is important to emphasize the somewhat conjectural nature of recommendations based on this evidence. Since ipratropium is relatively inexpensive, safe, and does suggest a benefit in a patient oriented outcome, it is a reasonable adjunct therapy in the treatment of an acute asthma exacerbation.

Author

Koustav Mukherjee, MD

Published/Updated

January 30, 2012

Source

Bellolio MF, Gilmore RM, Stead LG. Insulin for glycaemic control in acute ischaemic stroke. Cochrane Database Syst Rev. 2011;(9).

Efficacy Endpoints

Death, disability

Harm Endpoints

Symptomatic hypoglycemia

Narrative

Hyperglycemia is common after acute ischemic stroke and occurs in up to two-thirds of patients. Clinical trials have concluded that hyperglycemia predicts increased mortality. It is uncertain whether this contributes to brain injury or is merely a physiologic response to acute stroke, and animal studies have suggested that insulin may reduce stroke size by reducing glucose levels, acidosis, and cell injury.

In this 2011 Cochrane review of randomized trials ‘dependence’, a primary outcome, was defined as being severely dependent on others in activities of daily living. Three treatment comparisons were investigated: insulin vs. placebo, low dose insulin vs. high dose insulin, or tight versus liberal glycemic control, all for glucose levels greater >110 mg/dl. This review found no benefit in any comparison.

Seven trials involving 1296 participants were included. Maintaining blood sugar level between 72 and 135 mg/dl immediately after a stroke did not improve outcomes (i.e. did not reduce death or dependence). It did however significantly increase symptomatic hypoglycemic events (confusion, visual disturbances, seizures, sweating, or hunger in a patient with a glucose level lower than 54 mg/dl). The NNH for symptomatic hypoglycemia in the experimental group was 7.

Caveats

The Cochrane authors report two major subgroup analyses. The first is a comparison of diabetic and non-diabetic stroke patients that also showed no benefit. In the second analysis the authors note that studies reporting only 30-day final outcomes appeared to show more favorable results for insulin treatment than studies reporting 90-day outcomes. The latter studies were larger and accounted for 82% of all subjects, and because the natural history of ischemic stroke is improvement and stabilization through three months this appears to be a more reliable and patient-oriented outcome measure. Notably, however, stroke scores did show nearly significant differences favoring insulin treatment in the overall group. However, this did not translate into a death or dependency advantage, statistically or otherwise.

Author

Jason Bell, MD

Published/Updated

Source

Naccarato M, Chiodo Grandi F, Dennis M, Sandercock PAG. Physical methods for preventing deep vein thrombosis in stroke (Review). Cochrane Database Syst Rev. 2010. Issue 8. Art. No.: CD001922. DOI: 10.1002/14651858.CD001922.pub3
The CLOTS Trials Collaboration Effectiveness of thigh-length graduated compression stockings to reduce the risk of deep vein thrombosis after stroke (CLOTS trial 1): a multicentre, randomised controlled trial. Lancet 2009; 373: 1958–65

Efficacy Endpoints

Deep Vein Thrombosis (DVT), PE, Death

Harm Endpoints

Skin breakdown, blisters, ulcers, skin necrosis

Narrative

It has long been observed that hospitalized patients are at increased risk for developing a DVT. Venous stasis and lack of mobility are known risk factors for developing venous thromboemboli. It has been hypothesized that compression stockings, by compressing superficial veins, cause increased blood flow through the deep veins in the extremities, decreasing venous stasis and DVT.

The Cochrane review identified two trials comparing GCS (graduated compression stockings) with routine care for DVT prophylaxis including a total of 2615 patients. The larger is the 2009 CLOTS trial, with 2518 (96%) of the subjects included in the review. Approximately 12% of subjects in each group were on anticoagulants in this study and the review showed no difference in DVT, PE, or death between the control and the study group, regardless of anticoagulation status. The stocking group, however did have an increased risk of harm including skin breakdown, blisters, ulcers, or skin necrosis.

Caveats

The purpose of preventing DVT is to prevent the life threatening or physiologically uncomfortable symptoms of either DVT or pulmonary embolism. The great bulk of the included trials did not measure these outcomes, making it impossible to know if preventing DVTs has an important impact on patient-oriented outcomes. Large, high quality, well-powered trials that closely track both the adverse effects of stockings (discomfort, etc.) and clinically important events related to venous thromboembolism are badly needed.

All of the included trials involved patients with acute stroke as the reason for their hospitalization, and none of the studies included low risk patients. Consequently, firm conclusions from these data can only be drawn regarding GCS for stroke patients.

It should be noted that there is relatively robust data to suggest that GCS appear to be of benefit when used for surgical patients in the post-operative inpatient setting. For more information, please refer to related entries on this topic.

Author

Viral Patel, MD

Published/Updated

Source

Sachdeva A, Dalton M, Amaragiri SV, Lees T Elastic compression stockings for prevention of deep vein thrombosis(Review). Cochrane Database Syst Rev. 2010 April 18;(7):CD001484

Efficacy Endpoints

Deep Vein Thrombosis (DVT)

Harm Endpoints

Patient Discomfort

Narrative

It has long been observed that hospitalized patients are at increased risk for developing a DVT. Venous stasis and lack of mobility are known risk factors for development of venous thromboemboli. It has been hypothesized that graded compression stockings, by compressing superficial veins, cause increased blood flow through the deep veins in the extremities thereby decreasing venous stasis and DVT. (This summary, examining the impact of stockings when added to antithrombotic medication is an accompaniment our summary of the impact of stockings without medication, drawn from the same source review).

The Cochrane review identified 10 randomized control trials (n=1248) comparing treatment with GCS as an adjunct to medical therapy (heparin or aspirin) for DVT prophylaxis among surgical patients. The incidence went from 16% in the control group to 4% in the treatment group, with an NNT of 9 (based on an Odds Ratio of 0.25, control event rate of 16%).

The authors conclude that graduated compression stockings (GCS) are an inexpensive and noninvasive means by which DVTs can be prevented.

Caveats

The purpose of preventing DVT is to prevent the life threatening or physiologically uncomfortable symptoms of either DVT or pulmonary embolism. The great bulk of the included trials did not measure these outcomes, making it impossible to know if preventing DVTs has an important impact on patient-oriented outcomes. Large, high quality, well-powered trials that closely track both the adverse effects of stockings (discomfort, etc.) and clinically important events related to venous thromboembolism are badly needed.

All but one of the included trials involved patients undergoing surgical procedures as the reason for their hospitalization, and none of the studies included low risk patients. Consequently, firm conclusions can only be drawn regarding GCS for moderate to high risk surgical patients.

In addition 6 of the 18 trials used the other leg as the control group. It has been hypothesized that compression to one leg could have an adverse effect on the other leg. If true, this could reflect a bias toward the treatment group in these studies. In addition, 10 of the trials included industry funding of some form.

For the time being GCS seem like an inexpensive, low side effect intervention that is a reasonable choice to reduce the presence of imaging-detectable DVT, and we have therefore rated this intervention ‘Green’ (i.e. beneficial), with the caveat that patient-oriented study designs may ultimately challenge or support this conclusion.

Author

Viral Patel, MD

Published/Updated

December 8, 2011

Source

Sachdeva A, Dalton M, Amaragiri SV, Lees T Elastic compression stockings for prevention of deep vein thrombosis(Review). Cochrane Database Syst Rev. 2010 April 18;(7):CD001484

Efficacy Endpoints

Deep Vein Thrombosis (DVT)

Harm Endpoints

Patient Discomfort

Narrative

It has long been observed that hospitalized patients are at increased risk for developing a DVT. Venous stasis and lack of mobility are known risk factors for development of venous thromboemboli. It has been hypothesized that graded compression stockings, by compressing superficial veins, cause increased blood flow through the deep veins in the extremities thereby decreasing venous stasis and DVT.

The Cochrane review identified 8 randomized control trials (n=1279) comparing treatment with GCS vs. standard therapy for DVT prophylaxis among surgical patients. The incidence of DVT decreased from 26% in the control group to 13% in the treatment group, with an NNT of 8 (based on an odds ratio of 0.35 and a control event rate of 26%).

Caveats

The purpose of preventing DVT is to prevent the life threatening or physiologically uncomfortable symptoms of either DVT or pulmonary embolism. The great bulk of the included trials did not measure these outcomes, making it impossible to know if preventing DVTs has an important impact on patient-oriented outcomes. Large, high quality, well-powered trials that closely track both the adverse effects of stockings (discomfort, etc.) and clinically important events related to venous thromboembolism are badly needed.


All of the included trials involved patients undergoing surgical procedures as the reason for their hospitalization, and none of the studies included low risk patients. Consequently, firm conclusions can only be drawn regarding GCS for moderate to high risk surgical patients.

In addition 6 of the 18 trials used the other leg as the control group. It has been hypothesized that compression to one leg could have an adverse effect on the other leg. If true, this could reflect a bias toward the treatment group in these studies. In addition, 10 of the trials included industry funding of some form.

For the time being GCS seem like an inexpensive, low side effect intervention that is a reasonable choice to reduce the presence of imaging-detectable DVT, and we have therefore rated this intervention ‘Green’ (i.e. beneficial), with the caveat that patient-oriented study designs may ultimately challenge or support this conclusion.

Author

Viral Patel, MD

Published/Updated

Source

Vale N, Nordmann AJ, Schwartz GG, de Lemos J, Colivicchi F, den Hartog F, Ostadal P, Macin SM, Liem AH, Mills E, Bhatnagar N, Bucher HC, Briel M. Statins for acute coronary syndrome. Cochrane Database Syst Rev. 2011 Jun 15;(6):CD006870. Review. PubMed PMID: 21678362.

Efficacy Endpoints

Death, heart attack, stroke, acute heart failure, unstable angina

Harm Endpoints

Myopathy, rhabdomyolysis

Narrative

Patients are at highest risk of severe adverse events in the early period after an episode of acute coronary syndrome (ACS). Beyond their cholesterol lowering properties, some experimental data suggests that statins may improve endothelial function, decrease platelet aggregation, and reduce inflammation. It has been hypothesized that these effects may reduce adverse outcomes if initiated soon after an episode of ACS.

The Cochrane review identified 18 randomized control trials with useable data comparing treatment with early administration of statins following an episode of ACS to placebo or standard treatment. The studies included 14,303 patients, had a population with ages ranging from 53-69 and were predominantly male (59-88%). This large, systematic review concluded that early initiation of statin therapy (within 14 days of MI) did not significantly decrease the risk of subsequent MI, death, revascularization, heart failure or stroke at 4 months follow up. The authors of the review do note, however, that there was a small decrease in episodes of hospitalization for unstable angina at 4 months, amounting to an approximately 1.5% absolute reduction, and suggesting a NNT of 67.

Caveats

Early addition of statins is a Level 1A recommendation by the ACC/AHA and guidelines state that statins should be started prior to hospital discharge regardless of baseline LDL level. This analysis and systematic review does not support these recommendations.

While the eventual addition of statin therapy for patients with established elevations in LDL cholesterol who have suffered a prior MI has been shown to improve long-term outcomes, giving statins in the acute or subacute period following a heart attack without attention to cholesterol levels appears to provide no benefit for virtually any major patient-important outcomes. Presumably because patient-level data was not available the Cochrane authors did not perform subgroup analysis based on entry LDL levels, thus it remains possible that those with elevated cholesterol may benefit in the short term.

There was a small decrease in rehospitalization for unstable angina and some may note that at 12 months there was a decrease in the secondary outcome of ‘revascularization’ procedures (ARR = 5%, NNT = 20). As we have noted in prior reviews, since the condition that led to these hospitalizations and procedures was not an MI, did not result in death, and is based on clinician judgment rather than anatomic disease or objective pathology, the need for these measures is questionable and does not meet our standard for a patient-important outcome or clinically important end-point.

Finally, it is a point of interest that this review, like many other Cochrane reviews, notes important and large discrepancies in the results of trials performed with lesser and greater methodologic rigor. As one may expect, statistical benefits of statin medications were considerably and consistently greater in trials that did not explicitly use methods such as blinding and allocation concealment. This finding suggests that benefits of these medications may be strongly exaggerated, or in some cases created, by methods that potentially introduce bias into the conduct of randomized trials. This highlights the importance of these methods in maximizing validity, and also supports the Cochrane Collaboration practice of routinely including subgroup analyses examining results separately from trials that do and do not employ these methods.

Author

Viral Patel, MD and Daniel Runde, MD

Published/Updated

September 28, 2011

Source

Allen SJ, Martinez EG, Gregorio GV, Dans LF. Probiotics for treating acute infectious diarrhoea. Cochrane Database Syst Rev. 2010 Nov 10;(11):CD003048. Review. PubMed PMID: 21069673.

Efficacy Endpoints

Duration of diarrhea, severity of diarrhea

Harm Endpoints

None reported

Narrative

Diarrhea is a major cause of mortality and morbidity worldwide, and treatment regimens often include a oral and/or intravenous rehydration, antibiotics, and anti-motility agents. The administration of probiotics have been proposed as an add-on intervention that may be able to decrease the duration and severity of diarrhea.

The Cochrane Review included 63 randomized controlled trials (56/63 in infants and children) of 8014 subjects with proven or presumed infectious diarrhea. Control groups were given no treatment or placebo. Administration of a probiotic agent decreased the mean duration of diarrhea by 24.76 hours (95% CI 15.9 – 33.6 hours) and stool frequency on day 2 by 0.8 stools (95% CI 0.45 – 1.14). No adverse effects were reported.

Caveats

While the aggregate study population was quite large most, of the studies were primarily involving children (6489 were infants and children under age 18). Additionally, included trials did not all use the WHO defined inclusion criteria, nor did they administer the same strain or dosage of probiotics. Further studies clearly need to be done to determine the most efficacious dose and probiotic strain (or combination of strains). Many studies were also in developing nations, making extrapolation to developed settings more difficult.

The considerable heterogeneity found in the studies deserves comment. Arguably it is inappropriate to pool the data from these studies, as the heterogeneous results are an indicator that the populations, settings, probiotic agents, and methods of these studies were all highly variable. Indeed they are likely too variable to be considered biologically similar, and therefore pooling these data may represent apples-to-oranges comparisons and combinations. We have chosen to report the pooled results, however, because in this case it seems that these results are a reasonable indicator of the likely impact of probiotics on diarrheal illness in almost all of the included settings and populations. In other words even if the review had been entirely narrative (i.e. did not include a meta-analysis of the studies but rather only provided a descriptive review) we believe that this result is a plausible and reasonable inference from the available data. Individual practitioners will, as always, have to use their judgment in the application of these data. In developing settings with dehydrated children these results are likely to be robust and relevant. In developed settings among adults probiotics may be considered more experimental.

Finally, while the studies did not address decreased long term morbidity or mortality with a probiotic agent, given the low cost of treatment and a significant decrease in a major patient oriented outcome we feel this is a meaningful effect.

Author

Viral Patel, MD

Published/Updated

Source

Chavez-Tapia NC, Barrientos-Gutierrez T, Tellez-Avila FI, Soares-Weiser K, Uribe M. Antibiotic prophylaxis for cirrhotic patients with upper gastrointestinal bleeding. Cochrane Database Syst Rev. 2010 Sep 8;(9):CD002907. Review. PubMed PMID: 20824832.

Efficacy Endpoints

Mortality, infection during hospitalization (spontaneous peritoneal peritonitis, bacteremia, pneumonia, urinary tract infection)

Harm Endpoints

Not Reported

Narrative

Cirrhotic patients often develop bleeding from gastric or esophageal varices that occur secondary to portal hypertension. Gastrointestinal (GI) bleeding is fatal in approximately 20% of these episodes and bacterial infections are an important contributor to this mortality. Patients with cirrhosis are also known to have impaired immune function and increased translocation of bacteria from the gut into the bloodstream, thus the administration of prophylactic antibiotics during the bleeding event has been proposed as a treatment to help prevent such infections.

This Cochrane Review includes 12 trials (n = 1241) involving cirrhotic patients with upper GI bleeding. Of the 12 included trials, only 1 was placebo controlled, the other 11 examined antibiotics vs. no intervention. These trials demonstrated a clear decrease in overall rate of bacterial infections, with marked reductions in bacteremia, pneumonia, spontaneous bacterial peritonitis and urinary tract infections. With the exception of pneumonia, all of the infections were confirmed by cultures. The trials also noted an overall decrease in mortality. The choice of antibiotic regimen appeared to have no effect, although all antibiotics used in these trials were all chosen because of their activity against gram negative organisms (the most common infecting agents for the targeted infection types).

Caveats

None of the included trials reported on harms or adverse effects associated with the administration of antibiotics. Furthermore, only 1 trial was placebo controlled, which introduces a real risk of bias. Perhaps more importantly, as noted by the Cochrane Review authors, the data for decreased mortality in the intervention group was not as compelling as it was for preventing infection. Many of the included trials were not powered to determine a mortality benefit. Furthermore, trial sequential analysis (a statistical tool used to evaluate the strength of results found during meta-analysis) found that the 12 included trials were not enough to come a definitive conclusion about the significance of decrease mortality in patients receiving antibiotics. This indicates that even one large, high quality, methodologically rigorous randomized trial has the potential to trump the results of this review, as it may well generate results that are in disagreement with the results of this review (for all outcomes including infection rates). We would like to see a trial like this performed, and we believe that given the poor quality of existing trial data there is clinical equipoise adequate to perform such a trial. However, given that this review represents the best available data, and given the reported benefits of reducing infection and the possible decrease in mortality, it seems appropriate to recommend this intervention in cirrhotic patients with upper GI bleeding.

Author

Daniel Runde, MD

Published/Updated

Source

Perel P, Roberts I. Colloids versus crystalloids for fluid resuscitation in critically ill patients. Cochrane Database Syst Rev. 2011 Mar 16;(3):CD000567. Review. PubMed PMID: 21412866.

Finfer S, Bellomo R, Boyce N, French J, Myburgh J, Norton R; SAFE Study Investigators. A comparison of albumin and saline for fluid resuscitation in the intensive care unit. N Engl J Med. 2004 May 27;350(22):2247-56. PubMed PMID: 15163774.

Delaney AP, Dan A, McCaffrey J, Finfer S. The role of albumin as a resuscitation fluid for patients with sepsis: a systematic review and meta-analysis. Crit Care Med. 2011 Feb;39(2):386-91. Review. PubMed PMID: 21248514.

Efficacy Endpoints

Mortality

Harm Endpoints

Not reported

Narrative

In the critically ill, fluid replacement is believed to be an essential part of resuscitation. While there is consensus regarding the importance of early delivery of fluids, the evidence regarding which type of fluid has been confusing and often contradictory. Colloid solutions increase intravascular osmotic pressure, with a proposed benefit of increased volume expansion per unit of solution infused when compared to standard crystalloid solutions.

This Cochrane Review includes 65 trials (n = 11,623) involving critically ill patients undergoing fluid resuscitation. The trials were divided into 3 subgroups: colloids vs. isotonic crystalloid, colloids and hypertonic saline vs. isotonic crystalloid and colloids vs. hypertonic crystalloids. Administration of colloids yielded no mortality benefit, regardless of the type of colloid used or the manner in which it was delivered.

Caveats

Over 60% of all patients in this review come from a single trial (Finfer et al. 2004, the 'SAFE' trial). However, of all the included trials it was the most methodologically rigorous. Because of concerns about potentially important differences among the various colloids, the authors of this review also performed analyses stratified by type of colloid. Once again, they failed to find any mortality benefit, regardless of which type of colloid was used. There was also significant heterogeneity among the included trials, including patient population, type of colloid and resuscitation algorithm.

A recent meta-analysis (Delaney et. al 2011) specifically examined the role of albumin in patients undergoing treatment for sepsis, a subgroup not examined in the Cochrane review. When the authors pooled the results, they did find a borderline mortality benefit, with an OR = 0.82 (95% CI 0.67-1.0, p=0.047). Less than half of the included studies specifically enrolled patients with sepsis and all of these were small (n < 90). Finally, using a random-effects analysis of the data (a statistical tool that calculates pooled results with greater allowances for variations among studies), they difference was no longer statistically significant. This leaves the question of albumin in sepsis open. Presuming expected mortality rates of 20-30% in patients with sepsis the odds ratio of 0.82 for mortality found in the albumin groups from the Delaney review would suggest, if true, an absolute reduction of 3-4% in mortality.

Despite the limitations of the Cochrane analysis, the results suggest that there is no compelling reason to use colloid solutions during resuscitation of critically ill patients. While there were no identifiable harms associated with colloid use the fact remains that the agents cost 10-20 times more than their crystalloid counterparts and in general it appears that colloids offer no benefit. However, there remains some question about their possible benefit in septic patients and a large, high quality trial is needed to clarify the issue. For this narrow question (albumin use for sepsis patients compared to crystalloids) we would grade the issue unresolved, i.e. color-code Yellow. Overall, we see no evidence of benefit, i.e. Red.

Author

Daniel Runde, MD

Published/Updated

Source

Colman et al Paraenteral dexamethasone for acute severe migraine headache: meta-analysis of randomised controlled trials for preventing recurrence. BMJ 2008 Jun.;336(7657):1359–1361

Singh A, et al. Does the addition of dexamethasone to standard therapy for acute migraine headache decrease the incidence of recurrent headache for patients treated in the emergency department? A meta-analysis and systematic review of the literature. Academic Emergency Medicine. 2008; 15(12): 1223-33

Efficacy Endpoints

Recurrence of migraine between 24 and 72 hrs of leaving the Emergency Department

Harm Endpoints

Adverse drug events

Narrative

Migraine is a common entity and is usually successfully treated in outpatient and inpatient environments. Recurrence of headache within two days following successful termination of migraine is reportedly as high as 50%, and up to 10% of patients treated in emergency departments will return for this ‘rebound’ headache.

This review summarizes available data on the use of steroids to reduce migraine recurrence. It includes 738 patients in 7 studies. Dexamethasone was more effective than placebo in reducing recurrence rates (relative risk 0.74, 95% confidence interval 0.60 to 0.90), despite being no better than placebo for the acute migraine.

Included studies were considered high quality and there was no significant heterogeneity between them. All studies (except Baden 2006) used the International Headache Society Classification criteria to diagnose migraine and all were conducted in the emergency department setting where the migraine was initially treated with standard abortive treatment.

All trials (except Fiesseler 2006) tracked adverse events to some extent though significant differences were identified in only two of these: dizziness was more common (NNH=38) while nausea was less common (NNT = 16).

The reviews in question were published prior to one recent trial1, which found similar results (NNT of 10).

Caveats

All trials used a single dose of between 10mg and 24mg dexamethasone given intravenously, and methods were reasonably standardized. Most subjects were emergency department patients who met International Headache Society criteria for migraine, and all trials appear to be of moderate to high quality. These points argue in favor of the general validity of the results as published, and a second systematic review by a different author group (using the same studies with the exception of one) came to identical conclusions qualitatively and quantitatively.

The most important limitation to this review is its limited power, with <1000 migraine patients. Indeed only one of the individual studies showed a statistically significant benefit. The methodologic rigor of the studies and the similar conclusions of two separate author groups, however, argue in favor of the aggregated results being valid. Despite this a high quality randomized trial properly powered to detect even small potential differences would be a helpful addition to this database.

Author

Andy Neil, MD

Published/Updated

September 8, 2011

Source

Antibiotic prophylaxis for mammalian bites. Cochrane Database Syst Rev. 2001;(2):CD001738.

Efficacy Endpoints

Proven bacterial infection plus positive microbiological cultures

Harm Endpoints

Not reported

Narrative

Mammals are known to harbor a variety of bacterial species in their oral cavities, and infections in mammalian bite wounds are both potentially aggressive and common. This raises an important question: can antibiotics, when given following a bite but before signs of infection, prevent skin and soft tissue infections?

This systematic review included eight randomized trials (total n = 522) six of which (n = 463) enrolled only subjects with dog bites. In the eight trials overall there appears to be no clear benefit to antibiotic use. Among the six trials examining only dog bites there was also no reduction in rates of infection (antibiotics - 4% [10/225] versus no antibiotics - 5.5% [13/238]).

Notably, the site of the bite was the most powerful predictor of infection: bites to the hand (3 studies) had a high rate of infection in the control group (28%), and in this small subgroup a beneficial effect of antibiotics was statistically significant [2% [1/61] vs 12/43], NNT = 4).

In addition, since the Cochrane review there has been a further reasonable quality trial addressing the question1. This study also found a low rate of infection (2%) and no significant benefit of prophylaxis.

Caveats

Overall, the event rate across studies was low (6.5%) and the number of subjects (522) is inadequate to determine whether antibiotic treatment significantly lowers infection rates. In addition, the enrolled sample is somewhat selected: superficial grazes and tendon and joint capsule injuries were excluded, and presumably all patients were treated with copious irrigation and wound cleansing. These factors may raise or lower infection rates when compared to wounds seen in community practice. In addition, multiple agents were used and it is not possible to draw conclusions about specific antibiotics.

The included studies of hand bites are of moderate to poor quality. While it is reasonable in practice to consider this the best available evidence (i.e. to treat with antibiotics prophylactically for hand wounds), this remains a ‘best guess’ with inadequate data.

These data do suggest the possibility of an as yet unproved preventive effect with prophylactic antibiotics for bite wounds in general. The enrollment numbers are too small to rule out a potentially important effect. Thus large, high quality, placebo-controlled randomized trials remain important and necessary to determine whether or not prophylactic antibiotics are appropriate. These studies are all the more important when considered in the context of the considerable side effect profile of antibiotics, which includes diarrhea, allergic reactions, and increasingly incident clostridium difficile infections. Therefore clinical equipoise between antibiotics and no treatment exists, and further studies are both ethical and necessary. This includes the subgroup of bite wounds that occur on the hand, where numbers are small and from studies of poor quality.

Author

Andy Neil, MD

Published/Updated

September 5, 2011

Source

Logan AC, Yank V, Stafford RS. Off-label use of recombinant factor VIIa in U.S. Hospitals: analysis of hospital records. Ann Intern Med. 154(8): 516-522, April 2011.

Boffard KD, Riou B, Warren B, et al. Recombinant factor VIIa as adjunctive therapy for bleeding control in severely injured trauma patients: two parallel randomized, placebo-controlled, double-blind clinical trials. J Trauma. 2005;59:8-15.

Nishijima DK, Zehtabchi S. The efficacy of recombinant activated factor VII in severe trauma. Ann Emerg Med. 2009 Nov;54(5):737-744.

Hauser CJ, Bofford K, Dutton R, et al. Results of the CONTROL trial: efficacy and safety of recombinant activated factor VII in the management of refractory traumatic hemorrhage. Journal of Trauma. 2010 Sep;69(3):489-500.

Efficacy Endpoints

Primary endpoints: Mortality; Secondary endpoints: Blood products needed, multiple morbidity events (i.e. ventilator days, ICU days, etc)

Harm Endpoints

Thromboembolic events

Narrative

Uncontrollable hemorrhage is a significant problem in severe trauma. Traumatic injury leads to a relative coagulopathy through a number of mechanisms. Recombinant activated factor VIIa (rFVIIa) was developed for the treatment of bleeding episodes in patients with hemophilia who acquire antibodies to factor VII, and the Food and Drug Administration has approved it for this indication. rFVIIa has been used off-label in a number of bleeding conditions including severe trauma.1

To date, the only existing data on rFVIIa in severe trauma is from the CONTROL trial, a prospective double-blind multicenter study4: Phase 2 RCT (277 subjects)2 and Phase 3 (573 subjects)3. Severe trauma in this study was defined as the need for transfusion of 4 or more units of packed RBCs within 4 hours of admission. The trial included both blunt and penetrating trauma but excluded isolated head injury, moribund patients, or those with injuries stabilized by standard hemostatic interventions.

There was no statistically significant difference in mortality at 48 hours or at 30 days between the two groups; the phase 3 trial was stopped early due to enrollment difficulties and analysis suggesting futility. There were slight decreases in transfusion products needed though given the small numbers the confidence intervals were wide and with significant overlap. Contrary to most other data there was no detectable increase in adverse thromboembolic events with rFVIIa.

Caveats

Based on this one trial rFVIIa has not been shown to decrease mortality. There may be a modest decrease in blood product transfusion but it is unclear if or how this would translate into clinical benefits.

An interesting point to note is that the authors found a much lower mortality in both study arms than expected (13% found versus 30% expected). They theorize that this may be due to improving trauma care and add that improved survival with the administration of any one agent is unlikely.

Author

Sebastian Siadecki, MD

Published/Updated

August 25, 2011

Source

Sandercock PAG, Counsell C, Gubitz GJ, Tseng MC. Antiplatelet therapy for acute ischaemic stroke. Cochrane Database of Systematic Reviews 2008, Issue 3. Art No.: CD000029.

Efficacy Endpoints

Primary - death or dependence greater than 30 days after stroke; Secondary – death from any cause at < 30 days (treatment period) or > 30 days (follow-up period), evidence of DVT, symptomatic pulmonary embolus during the treatment period, recurrent stroke during the treatment period (ischemic or unknown), any recurrent stroke, and complete recovery from stroke

Harm Endpoints

Symptomatic intracranial hemorrhage during treatment period and major extracranial hemorrhage during the treatment period (resulting in a fatal bleed or requiring a transfusion or operation).

Narrative

Ischemic stroke is thought to occur by one of two processes - the activation of platelets leading to thromboembolic events, or plaque rupture in the cerebral arterial circulation leading to the formation of a clot. Interventions that prevent the clotting of platelets and thus the formation of thrombi or progression of ruptured plaques have been proposed to reduce the incidence and severity of acute ischemic stroke.

The Cochrane review included twelve randomized trials encompassing 43,041 stroke patients and, in aggregate, the review clearly suggests a benefit. Included trials compared an antiplatelet agent to control (placebo or no therapy with adequate blinding) for acute stroke, initiated within 14-days of onset. Antiplatelet medications evaluated by the included trials in the Cochrane review included: COX-2 inhibitors, thienopyridine derivatives, phosphodiesterase inhibitors, thromboxane A2 antagonists, and GP IIb/IIIa receptor antagonists. 94% of the data came from two trials, CAST and IST, which evaluated 160mg and 300mg of once daily aspirin for acute stroke continued for either four or two-weeks, respectively.

The OR for death or dependence was 0.95 (95% CI 0.91 – 0.99, p = 0.008), NNT = 79.The OR for death from all causes at 6-months was 0.93 (95% CI 0.87 – 0.99, p = 0.01), NNT = 108.The OR for symptomatic PE during the first 30-days was 0.71 (95% CI 0.52 – 0.95, p = 0.02), NNT 693. The OR for recurrent stroke in the first 30-days was 0.77 (95% CI 0.68 – 0.86, p < 0.00001), a reduction of 7 events per 1000 treated, NNT = 140. The OR for complete recovery was 1.06 (95% 1.02 – 1.11, p = 0.006), a benefit of 11 per 1000, NNT = 89.

Intracranial hemorrhage increased with antiplatelet agents at a rate of 2 per 1000 patients, NNH (number needed to harm) = 574. Major extracranial hemorrhage occurred in 4 per 1000 patients, NNH = 245.

Caveats

94% of the available data came from the two largest trials, CAST and IST, and unlike some of the other included trials in this meta-analysis, not all of these patients had a CT scan to exclude hemorrhagic stroke prior to randomization. Thus, some patients may have had intracranial hemorrhage rather than ischemic stroke at the time of treatment, which would lead to an underestimate of benefit when compared to a setting where imaging is generally secured before therapy is started.

Not all included trials in the meta-analysis used aspirin as the antiplatelet agent to be studied; however aspirin was the study drug used in CAST and IST, and it is likely the cheapest and most feasible medication to be used in a broad array of settings. Because of the small scale of the other trials included in the meta-analysis, there is limited evidence to recommend the use of other antiplatelet agents other than aspirin in patients who cannot tolerate aspirin. Finally, a significant interaction between aspirin and streptokinase was noted in the MAST-I trial, leading to an increase in early and late mortality, in-hospital death from all causes, neurologic cause, and increase in intracranial hemorrhage. Thus, it is unclear what role antiplatelet agents may play and when they should be administered in those patients who receive thrombolytic therapy for acute ischemic stroke.

Author

Paul Roszko, MD and Anthony M. Napoli, MD FACEP

Published/Updated

July 17, 2011

Source

Antithrombotic Trialists Collaboration. Aspirin in the primary and secondary prevention of vascular disease: collaborative meta-analysis of individual participant data from randomised trials. Lancet. 2009; 373(9678); 1849-60

Antithrombotic Trialists Collaboration. Collaborative meta-analysis of randomised trials of antiplatelet therapy for prevention of death, myocardial infarction, and stroke in high risk patients. BMJ. 2002 Jan 12;324(7329):71-86.

Efficacy Endpoints

Heart attack, stroke, death

Harm Endpoints

Bleeding, death

Narrative

Aspirin blocks the action of platelets, reducing clots and ostensibly lowering the risk of heart attacks, strokes, and deaths. This review examined and summarized the magnitude of benefits from daily aspirin when compared to placebo for 'secondary prevention', i.e. among patients who have had a recent heart attack or stroke.

Aspirin works: those taking aspirin in these studies suffered fewer heart attacks, strokes, and deaths than those taking a placebo, at the cost of a small number of bleeding events. In addition, the benefits outlined here were seen after just over two years of daily aspirin therapy, in contrast to the 4 and 5 year periods seen with many other cardiovascular preventive interventions.

Caveats

Aspirin can cause bleeding events and gastrointestinal problems (ulcers, indigestion, etc.). However the cost of generic aspirin is very low and the benefits are considerably more common than the harms, making aspirin an excellent intervention for the right patients. This intervention does not translate, however, to patients who are not at very high risk (i.e. those who have established disease). Note our related summary on the topic of aspirin for 'primary prevention'.

For those who cannot tolerate aspirin other antiplatelet agents such as clopidogrel, ticlopidine, and related drugs are probably very reasonable and effective alternatives and seem to have similar beneficial effects, despite having more side effects.

Author

David H. Newman, MD

Published/Updated

July 10, 2011

Source

Squizzato A, Keller T, Romualdi E, Middeldorp S. Clopidogrel plus aspirin versus aspirin alone for preventing cardiovascular disease. Cochrane Database of Systematic Reviews 2011, Issue 1.

Bowry A, Brookhart MA, Choudhry NK. Meta-Analysis of the Efficacy and Safety of Clopidogrel Plus Aspirin as Compared to Antiplatelet Monotherapy for the Prevention of Vascular Events. American Journal of Cardiology;101(7):960-6

Efficacy Endpoints

Heart attack, stroke, death

Harm Endpoints

Bleeding, death

Narrative

Aspirin blocks the action of platelets, reducing the chance of blood clots. This is proven to reduce heart attacks, strokes, and deaths among patients who have had a heart attack or stroke. Clopidogrel, another drug that blocks platelet activity, can be temporarily added to daily aspirin during times of high risk in an attempt to provide even more protection than aspirin alone. But blocking platelet activity also leads to an increased risk of bleeding events. This review examined whether temporarily adding clopidogrel to aspirin (i.e. taking both every day) is more helpful than harmful among people who have just suffered an acute coronary syndrome.

Clopidogrel added to aspirin was statistically better than aspirin alone in three large, high quality trials (roughly 8000 subjects) examining this question. There were bleeding harms caused by clopidogrel but they were less common than the benefits.

Caveats

Despite being ineffective for most patients with heart disease clopidogrel seems beneficial when added to aspirin before and after percutaeous coronary interventions (PCI) and acute coronary events. This use of clopidogrel is temporary, as both of these situations increase the risk of major events in the short-term. However, the window of benefit is narrow, because longer use has not been shown to confer any benefit, but will certainly increase the risk of bleeding. There is no evidence to suggest that any patients should therefore be taking clopidogrel for more than one year after a heart attack or PCI. In the case where clopidogrel is being used as a replacement for aspirin (rather than an add-on), it is likely that clopidogrel will have the same benefits seen with aspirin.

Technical note: To derive the NNT from the clopidogrel data we used the two highest quality systematic reviews of which we are aware. Neither pools the absolute risk data, therefore we calculated backwards from the odds ratios reported in our second source, and have estimated the overall control-event rate among the three trials to be roughly 12% (PCI-CLARITY 12%, PCI-CURE 12.6%, CREDO 11.5%). This is the control event rate for all major vascular events combined, and thus separating strokes from MIs was not possible. Therefore we report them as a combined outcome.

Author

David Newman, MD

Published/Updated

July 8, 2011

Source

Antithrombotic Trialists Collaboration. Collaborative meta-analysis of randomised trials of antiplatelet therapy for prevention of death, myocardial infarction, and stroke in high risk patients. BMJ. 2002 Jan 12;324(7329):71-86.

Sudlow CLM, Mason G, Maurice JB, Wedderburn CJ, Hankey GJ. Thienopyridine derivatives versus aspirin for preventing stroke and other serious vascular events in high vascular risk patients. Cochrane Database of Systematic Reviews 2009, Issue 4.

Efficacy Endpoints

Heart attack, stroke, death

Harm Endpoints

Bleeding, death

Narrative

Clopidogrel blocks the action of platelets, reducing the chance of blood clots. Aspirin uses this same mechanism to reduce heart attacks, strokes, and deaths for patients with heart disease and strokes. Because aspirin is inexpensive, effective, and safe, clopidogrel is typically considered only for patients who can't tolerate aspirin because of side effects or allergies. This review examined whether clopidogrel is a suitable replacement for aspirin in people who have known cardiovascular disease.

Clopidogrel and its class of drug were as effective as aspirin in preventing heart attacks, strokes, and deaths in a number of comparative studies (ATTC, 2002). There are minor adverse effects to consider, as clopidogrel and ticlopidine caused more rashes [6.0% vs 4.6%], more diarrhea [4.5% vs 3.4%], and occasional episodes of neutropenia compared to aspirin in one of the largest and best studies.1

Caveats

Clopidogrel was very similar to aspirin in studies comparing the two and in some narrow categories (stroke) may even be slightly better. However this is speculative and the difference between the two drugs, if real, is extremely small and not likely to be clinically important. In addition, the majority of data suggesting this finding are industry sponsored, and aspirin has proven effective and safe for decades in non-industry trials, making this a more staid and robustly supported first line choice.

*** Important technical note: Based on the very similar performance in comparative trials (see first source study above), it is widely held that aspirin and clopidogrel are essentially equals. However by the time clopidogrel was introduced aspirin was already a proven agent. Therefore it would have been considered unethical not to use aspirin for high risk patients in most settings. This means that there is very little study data comparing clopidogrel alone to a placebo alone for preventing of cardiovascular events. Instead, the addition of either clopidogrel or a placebo to aspirin is more common in studies (see related NNT summaries). Thus our estimates of clopidogrel’s efficacy are projections based on the numbers in our aspirin review, rather than based on existing evidence from studies of clopidogrel.

Author

David Newman, MD

Published/Updated

Source

Squizzato A, Keller T, Romualdi E, Middeldorp S. Clopidogrel plus aspirin versus aspirin alone for preventing cardiovascular disease. Cochrane Database of Systematic Reviews 2011, Issue 1.

Bowry A, Brookhart MA, Choudhry NK. Meta-Analysis of the Efficacy and Safety of Clopidogrel Plus Aspirin as Compared to Antiplatelet Monotherapy for the Prevention of Vascular Events. American Journal of Cardiology;101(7):960-6

Efficacy Endpoints

Heart attack, stroke prevention, death

Harm Endpoints

Bleeding, death

Narrative

Aspirin blocks the action of platelets in the blood, reducing the chance of blood clots. This is proven to reduce heart attacks, strokes, and deaths among patients with known cardiovascular disease. Clopidogrel, another drug designed to block platelet activity, is often considered when patients can’t tolerate aspirin because of side effects or allergies, and there is hope that adding clopidogrel to aspirin might provide even more protection that aspirin alone. This review examined whether adding clopidogrel to aspirin (i.e. taking both every day) is helpful among people who either have cardiovascular disease already or are at very high risk for developing it.

Clopidogrel added to aspirin was not statistically better than aspirin alone in the only large, high quality trial (over 15,000 subjects) examining this question. In addition, there were bleeding harms caused by clopidogrel in this trial and many others, suggesting that the drug does not provide a measurable or proven benefit in this role, but does cause dangerous bleeding.

Caveats

There is one other trial that compared the two drugs together to just one drug alone, but this trial compared aspirin plus clopidogrel to clopidogrel alone (rather than aspirin alone) thus we did not include these data here. However, this trial also found no benefit to combination therapy for prevention in high risk patients, but there was an increase in bleeding events.1

Note that although we have labeled this review relevant to those looking to prevent a second event, roughly 20% of the subjects in this trial had never had a heart attack or stroke, but did have multiple other very high risk features (diabetes, etc.), and were therefore considered to be at nearly equivalent risk to those who have already had a heart attack or stroke.

Finally, this is not the only use of clopidogrel. Clopidogrel does seem to provide a reliable benefit (the prevention of nonfatal heart attacks or strokes, though not the prevention of deaths) in patients having, or recovering from heart attacks and strokes, and in those undergoing 'stenting' and other procedures to open their coronary or peripheral arteries. See our relevant summaries for these data.

Author

David Newman, MD

Published/Updated

Source

Gabriel Sanchez R, Sanchez Gomez LM, Carmona L, Roqué i Figuls M, Bonfill Cosp X. Hormone replacement therapy for preventing cardiovascular disease in post-menopausal women. Cochrane Database of Systematic Reviews 2005, Issue 2. Art. No.: CD002229

Efficacy Endpoints

Heart attack, stroke, death

Harm Endpoints

Development of a blood clot, death

Narrative

Hormone replacement therapy ('HRT') was believed for many years to potentially restore cardiovascular hormonal balance for women after menopause, thus reducing heart attacks, strokes and death. While this therapy has been shown to reduce hot flashes,1 before the past ten years there were no major trial results to answer the question of whether or not HRT works to prevent heart problems and strokes.

This review examined the highest quality randomized trials that tested HRT for preventing heart attacks, strokes, and death among women who have already had a heart attack or stroke, and included over 5000 subjects from trials. The results were very surprising for many scientists and physicians who had been recommending HRT for years, because HRT did not prevent any important cardiovascular problems or deaths, and overall it did increase blood clots in the legs and lungs. The numbers of stroke and heart attacks were just slightly higher in the group on HRT, but this wasn't a large enough difference to be considered 'statistically significant'. It is also important to note that despite the increases in heart attacks, strokes, and blood clots there was no increase in deaths due to HRT.

Caveats

The results in this review come from trials that are of a decent, but not very large, size. It is likely that larger trials asking the same question would show a 'significant' increase in heart attacks and strokes due to HRT. It is also important to note however, that there are benefits of HRT as well, including reduced hot flashes and small reductions in colon cancer and fractures, and it may be worth integrating all of these in making a decision about whether or not HRT is right for a given patient.

Author

David Newman, MD

Published/Updated

Source

Hartling L, Bialy LM, Vandermeer B, Tjosvold L, Johnson DW, et al.
Epinephrine for bronchiolitis. Cochrane Database of Systematic Reviews. 2011, Issue 6.
Art No: CD003123. DOI: 10.1002/14651858.CD003123.pub3.

Efficacy Endpoints

Admission Rate (Day 1)

Harm Endpoints

None

Narrative

Bronchiolitis, a viral respiratory infection in the first 2 years of life (most commonly caused by respiratory syncytial virus) often leads to significant use of healthcare resources. Optimal management strategies remain unclear, including pharmacotherapy for wheezing and shortness of breath.

This Cochrane review identified and assessed six studies of epinephrine versus placebo
with 995 children. The meta-analysis of these trials suggested that epinephrine reduced
the rate of hospitalization on day 1, risk ratio (RR) 0.67 (95% CI, 0.50 – 0.89).

These beneficial effects were not seen by day 7, although the analysis may have been
underpowered with only 3 studies of 875 children (RR 0.81, .95CI: 0.63 - 1.03), and
included trials of predominantly lower quality.

Of note, the review also identified that combination therapy using epinephrine and steroid for outpatients lowered 7-day admission rate when compared to placebo (RR = 0.65, 95% CI, 0.44 – 0.95), and improved clinical scores and respiratory rates. These findings, however, are based on a single high quality study.1

Finally, when epinephrine was compared to salbutamol in this review the rates of
admission within one day were similar, although inadequate power weakens these results
as well.

There was no significant harm identified in the review.

Caveats

This updated Cochrane review represents an important shift, adding 7 studies and altering the definition of bronchiolitis to mean a ‘first episode of wheezing’, a modification that standardized the inclusion criteria. Outcome measures were also changed to emphasize patient-centered outcomes such as admission rate and length of stay. The authors support the use of epinephrine in the treatment of outpatient bronchiolitis because of a reduction in hospitalization noted with the first 24 hours of treatment . In the 2004 version of the review no statistically apparent benefit of epinephrine was identified, and in the current review the benefit was not apparent at 7 days. However it seems quite possible that this was due to power and quality constraints in the studies reporting the 7 day outcome, which were smaller and less rigorous than studies reporting the 24-hour outcome.

One important update was including the synergistic effects of epinephrine and steroid to lower hospitalization rate, as reported in a single study. Additional studies are needed to replicate these results and to define the optimal combinations under specific conditions.

From this review, epinephrine appears helpful when administered early in disease
progression, i.e. in the outpatient/emergency department setting, although inpatient use
did not appear to provide any significant or identifiable benefits.

Author

Jeff Hom, MD

Published/Updated

June 9, 2011

Source

Logan AC, Yank V, Stafford RS. Off-label use of recombinant factor VIIa in U.S. Hospitals: analysis of hospital records. Ann Intern Med. 154(8): 516-522, April 2011.

Al-Shahi Salman R. Haemostatic drug therapies for acute spontaneous intracerebral hemorrhage. Cochrane Database of System Rev 2009, Issue 4. Art No.: CD005951.

Diringer MN, Skolnick BE, Mayer SA, et al. Thromboembolic events with recombinant activated factor VII in spontaneous intracerebral hemorrhage: results from the factor seven for acute hemorrhagic stroke (FAST) trial. Stroke. 41(1): 48-53, 2010 Jan.

Mayer SA, Brun NC, Begtrup K, et al. Efficacy and safety of recombinant activated factor VII for acute intracerebral hemorrhage. N Engl J Med. 358(20): 2127-2137, 2008 May.

Yuan ZH, Jiang JK, Huang WD, et al. A meta-analysis of the efficacy and safety of recombinant activated factor VII for patients with acute intracerebral hemorrhage without hemophilia. Journal of Clinical Neuroscience. 17(6):685-93, 2010 Jun.

Narayan K, Maas AI, Marshall LF, et al. Recombinant factor VIIA in traumatic intracerebral hemorrhage: results of a dose-escalation clinical trial. Neurosurgery. 62(4): 776–786, 2008 April.

Levi M, Levy JH, Andersen HF, and Truloff D. Safety of Recombinant Activated Factor VII in Randomized Clinical Trials. N Engl E Med 363;19: 1791-1800, 2010 Nov.

Efficacy Endpoints

Mortality and Dependence

Harm Endpoints

Thromboembolic Events

Narrative

Spontaneous intracerebral hemorrhage (ICH) accounts for 10% of strokes. Approximately 40% die in the first month and 62% die in the first year. Larger hematomas are associated with increased mortality so a drug that limits the early expansion of ICH could theoretically improve mortality and functional outcome. Recombinant activated factor VII (rFVIIa) is believed to bind to exposed tissue factor and activated platelets, eventually generating thrombin. It is FDA approved for the treatment of bleeding in hemophilia A and B, but 97% of uses are currently off-label: 11% of these are for ICH, 29% for cardiovascular surgery, and 29% for trauma.1

The Cochrane Review analyzed all published and unpublished randomized controlled trials evaluating hemostatic drugs in ICH.2 A total of 1398 patients within 4 hours ICH onset were included: 975 subjects received rFVIIa, 2 epsilon-aminocaproic acid (EACA), and 423 placebo. The primary endpoints were mortality and dependence at 90 days. There was no significant difference in the risk of death (18.5% vs 19.4% for placebo, RR 0.85) and no difference in the combination of ‘death or dependence’ (49.6% vs 51.7%, RR 0.91). There was also no significant difference in thromboembolic events (8.5% vs 6.2%, RR 1.37). However updated publications from one of the trials suggest that there was a statistically significant increase in arterial events among subjects in the high dose rFVIIa group.3,4 Similar conclusions were made in a recent meta-analysis regarding the increased risk of arterial events with rFVIIa in both spontaneous and traumatic ICH.5,6 In addition, a recent systematic review of trials involving rFVIIa suggests that there is a roughly 2% (1 in 50) increase in arterial thrombotic complications with the drug. However, these adverse effects increased by more than 5% (1 in 20) among subjects >65 years of age, and nearly 7% (1 in 15) among those >75.7 ICH occurs predominantly in the elderly and this adverse effect profile may help to explain the lack of benefit shown in trials assessing clinical outcomes.

Adding the new 2010 data from the FAST trial to the Cochrane review, the risk for thromboembolic events for rFVIIa is 181 of 921 patients (19.7%) and for placebo is 70 of 409 patients (17.1%), but this is not a significant difference.2,3

Caveats

In the trials assessing rFVIIa adverse event reporting was suboptimal, clinical outcome assessments were often unblinded, and follow-up was generally incomplete. In addition, exclusion criteria for the FAST trial were changed mid-study and the Cochrane review authors have reported difficulty obtaining further information from the industry sponsor that controls the data (Novo Nordisk).2,6 These issues highlight the fact that available published data may be a fraction of the complete data on this drug, and that the available data are likely to represent an optimistic view of the drug’s performance.

Despite this these data do not support the use of rFVIIa for the treatment of spontaneous ICH. While off label use of rFVIIa continues to climb despite the absence of trial evidence to support this practice, use for ICH has decreased since the publication of the FAST trial.

Author

Dan Rolston, MD, MS

Published/Updated

June 4, 2011

Source

Ducharme FM, Ni Chroinin M, Greenstone I, Lasserson TJ. Addition of long-acting beta2-agonists to inhaled steroids versus higher dose inhaled steroids in adults and children with persistent asthma. Cochrane Database of Systematic Reviews 2010, Issue 4.

Efficacy Endpoints

Asthma attack prevention

Harm Endpoints

Increased asthma attacks, death

Narrative

Long-acting beta-agonists (LABAs) are designed to keep smooth muscle in the airways constantly relaxed, and they are the novel ingredient in the combination inhalers Advair® and Symbicort®. Unfortunately, these LABA medicines are proven to increase severe asthma attacks and asthma-related deaths and thus fell out of favor, most markedly after the 'SMART trial' published in 2006.1 Despite this, it is still felt by some that if the LABAs could be combined with steroids this might reduce any danger by providing steroid protection. This could theoretically allow the two medicines to work side-by-side to improve asthma control safely. While no one has suggested this as a first line therapy (inhaled steroids are first line), it has been suggested as a therapy when a first attempt to use inhaled steroids is inadequate.

This review examined 48 trials including over 33,000 subjects in an attempt to determine if there is a benefit to the combination LABA/steroid inhalers over simply increasing the dose of inhaled steroids. The combination medicine successfully reduced the chances of mild to moderate asthma attacks for about 1 in every 73 people using them. This effect depended heavily upon the severity of the asthma attack. For study subjects with a nearly 20% (1 in 5) chance of an asthma attack requiring pills in the next three months the benefit was 1 in every 45 people. But if the chance was only 1% then only 1 in every 673 people benefited. Thus the more severe the asthma, the more likely the benefit.

The combination therapy reduced general daily symptoms of asthma as well, but by very little. Quality of life was the same, and the symptom reductions were not clinically important.

However, large reviews suggest that there is a danger with using the combination of LABA and inhaled steroids for asthma maintenance. 2,3 These reviews vary in their findings only slightly, and suggest that approximately for 1 out of 140 people taking the combined therapy in trials the LABA/steroids combination caused a severe asthma attack. For 1 in every 1400 the combination seems to have caused an asthma-related death.

Caveats

While most asthma patients use their inhaler medications for months and years, the benefits of the combination inhaler were tested in predominantly 12-week trials, thus there may be less or more benefit when their use is extended.

The symptom reductions in these trials that were found with combination therapy were very small. Typically the therapy led to 0.5 pumps/day fewer uses of the rescue (beta-agonist) inhaler, and about 10% more 'symptom-free days' during the 12-week study periods. Despite these findings there was no 'Quality of Life' advantage. However, these advantages may be greater if the asthma is more severe. This may be a consideration for patients with severe relapsing asthma that leads to multiple hospital admissions. On the other hand, these patients may also be at highest risk for a fatal or severe attack, which would make LABA medicines potentially dangerous. This difficult balance will have to be struck by physicians and patients in concert based on symptoms, risk thresholds, and patient values.

Of the 48 trials represented in this review, 44 were sponsored by the pharmaceutical maker of the combination therapy. A long history of misrepresentation of data and occasionally fraudulent reporting of data suggests that industry sponsored results are often more optimistic than subsequent data produced by researchers and parties that do not have a financial stake in the results.

We have chosen to designate this therapy RED (no benefits) rather then BLACK (harms>benefits) because of the small benefit and the possibility of a trade-off of harms and benefits that may be worthwhile based on individual circumstances. In addition, while asthma-related deaths increase, all-cause mortality was unchanged in the reviewed studies, making the overall harm impact less definitive in terms of fatalities.

Author

David Newman, MD

Published/Updated

May 16, 2011

Source

Ducharme FM, Ni Chroinin M, Greenstone I, Lasserson TJ. Addition of long-acting beta2-agonists to inhaled steroids versus higher dose inhaled steroids in adults and children with persistent asthma. Cochrane Database of Systematic Reviews 2010, Issue 4.

Efficacy Endpoints

Asthma attack prevention

Harm Endpoints

Increased asthma attacks, death

Narrative

Long-acting beta-agonists (LABAs) are designed to keep smooth muscle in the airways constantly relaxed, and they are the novel ingredient in the combination inhalers Advair® and Symbicort®. Unfortunately, these LABA medicines are proven to increase severe asthma attacks and asthma-related deaths and thus fell out of favor, most markedly after the 'SMART trial' published in 2006.1 Despite this, it is still felt by some that if the LABAs could be combined with steroids this might reduce any danger by providing steroid protection. This could theoretically allow the two medicines to work side-by-side to improve asthma control safely. While no one has suggested this as a first line therapy (inhaled steroids are first line), it has been suggested as a therapy when a first attempt to use inhaled steroids is inadequate.

Surprisingly, the authors were able to find only three trials in children to include in this review. The results of these trials did not suggest a benefit and actually hinted at a potential harm (increased asthma attacks requiring pill treatment).

Caveats

The strong evidence of harm that is associated with LABA medicines2, 3 is a reason to be very cautious in their use, and the pediatric data here offer no optimism. Children under 12 years of age were tested in these studies, and there is little to no review data for children under 4 years of age. Clearly, more trials of high quality are needed to evaluate whether this therapy is safe and effective as an option for children with asthma whose current regimen of inhaled steroids is inadequate.

Author

David Newman, MD

Published/Updated

Source

Wells GA, Cranney A, Peterson J, Boucher M, Shea B, Welch V, Coyle D, Tugwell P. Etidronate for the primary and secondary prevention of osteoporotic fractures in postmenopausal women. Cochrane Database of Systematic Reviews 2008, Issue 1. Art. No.: CD003376.

Wells GA, Cranney A, Peterson J, Boucher M, Shea B, Welch V, Coyle D, Tugwell P. Alendronate for the primary and secondary prevention of osteoporotic fractures in postmenopausal women. Cochrane Database of Systematic Reviews 2008, Issue 1. Art. No.: CD001155.

Wells GA, Cranney A, Peterson J, Boucher M, Shea B, Welch V, Coyle D, Tugwell P. Risedronate for the primary and secondary prevention of osteoporotic fractures in postmenopausal women. Cochrane Database of Systematic Reviews 2008, Issue 1. Art. No.: CD004523.

Efficacy Endpoints

Fracture prevention

Harm Endpoints

Atypical fractures, jaw osteonecrosis, GI and musculoskeletal side effects (harms are uncommon but do clearly occur and are not well-studied)

Narrative

The bisphophonates (etidronate, alendronate, risedronate) are anti-resorptive medicines that block the resorptive action in bone, thus increasing the density of the bone in some areas. Bone construction and resorption are complex processes and the increase in bone density that the bisphosphonates are able to effect leads to a complicated cascade of hormonal and muscuoskeletal effects. Ideally this will lead to increased bone strength and fewer fractures. This summary and review is a combination of three different source reviews that examined the three medicines that have been tested in randomized trials.

The bisphosphonates did not appreciably reduce fractures among women who had not had a previous fracture and those with relatively normal bone density. The harms of bisphosophonates are increasingly an object of study and speculation, particularly now that the drugs have been on the market for some time and their long term use has become more common. Virtually none of the initial studies examining the drugs included treatment for more than three to five years, however serious side effects such as atypical fractures1 and osteonecrosis of the jaw2 and gastrointestinal, musculoskeletal, and other side effects leading to disproportionate discontinuation3 have been increasingly linked to use of these drugs in the long term. These side effects should strike a note of caution, and should firstly be discussed with any woman considering bisphosphonate therapy, and secondly be a warning against initiating this therapy in patient groups who have not been proven to benefit (i.e. women with very low bone mineral density and women with a history of fractures).

Caveats

The bisphosphonates do appear to reduce fractures among women with very low BMD and those who have had previous fractures, and should be considered. However in those without very low BMD or fractures no study has demonstrated a true benefit. In the Cochrane review of alendronate the authors suggest a benefit in the reduction of vertebral fractures based on the results of a single study.4 However this finding is based on a subgroup analysis of a group subjected to a post-hoc change in the BMD cut offs for primary versus secondary. Moreover, the clinical significance of vertebral fractures in this study (and others) is unclear, as these were typically detected radiographically (by routine screening x-rays of the spine) rather than clinically. Thus it is not at all obvious that these are patient-important measures.

Author

David Newman, MD

Published/Updated

Source

Wells GA, Cranney A, Peterson J, Boucher M, Shea B, Welch V, Coyle D, Tugwell P. Etidronate for the primary and secondary prevention of osteoporotic fractures in postmenopausal women. Cochrane Database of Systematic Reviews 2008, Issue 1. Art. No.: CD003376.

Wells GA, Cranney A, Peterson J, Boucher M, Shea B, Welch V, Coyle D, Tugwell P. Alendronate for the primary and secondary prevention of osteoporotic fractures in postmenopausal women. Cochrane Database of Systematic Reviews 2008, Issue 1. Art. No.: CD001155.

Wells GA, Cranney A, Peterson J, Boucher M, Shea B, Welch V, Coyle D, Tugwell P. Risedronate for the primary and secondary prevention of osteoporotic fractures in postmenopausal women. Cochrane Database of Systematic Reviews 2008, Issue 1. Art. No.: CD004523.

Efficacy Endpoints

Fracture prevention

Harm Endpoints

Atypical fractures, jaw osteonecrosis, GI and musculoskeletal side effects (harms are uncommon but do clearly occur and are not well-studied)

Narrative

The bisphophonates (etidronate, alendronate, risedronate) are anti-resorptive medicines that block the resorptive action in bone, increasing the density of the bone in some areas. Bone construction and resorption are complex processes and the increase in bone density that the bisphosphonates are able to effect leads to a complicated cascade of hormonal and muscuoskeletal effects. Ideally this will lead to increased bone strength and fewer fractures. This summary and review is a combination of three source Cochrane reviews that examined three bisphosphonate medicines that have been tested in randomized trials.

The medicines reduced fractures. Their greatest impact was on the rate of vertebral fractures, but they also demonstrated statistically demonstrable benefits in the reduction of dreaded hip fractures, and also wrist fractures. Typically for every 100 women taking the medicines six avoided a fracture of some sort over three years of therapy. Particularly based on the one hip fracture that is avoided per 100 women this benefit may be highly important in terms of reducing morbidity or disability.

It is important to recall that the harms of bisphosophonates are increasingly an object of study and speculation, particularly now that the drugs have been on the market for some time and their long term use has become more common. Virtually none of the initial studies examining the drugs included treatment for more than three to five years, however serious side effects such as atypical fractures1 and osteonecrosis of the jaw2 and gastrointestinal, musculoskeletal, and other side effects leading to disproportionate discontinuation3 have been increasingly linked to use of these drugs in the long term. These side effects should strike a note of caution, and should firstly be discussed with any woman considering bisphosphonate therapy, and secondly be a warning against initiating this therapy in patient groups who have not been proven to benefit (i.e. women with very low bone mineral density and women with a history of fractures).

Caveats

The majority of the benefit to the bisphosphonates is seen in vertebral fracture reduction, an outcome that is often made by screening radiographs rather than clinically. In other words it is not clear that it would be important to prevent subclinical vertebral fractures, nor is it clear that reducing this outcome represents an aggregate benefit when one considers the adverse effects of the medicines. However a reduction in hip fractures, even at 1 per 100, may represent a potentially important public health measure. Thus it would be helpful to have large scale population-based studies determine whether or not there are overall increases in longevity or decreases in disability with the bisphosphonates.

Finally, the studies examined in the Cochrane reviews are industry sponsored. A long history of selective outcome reporting,4 selective publication,5,6 occasionally fraudulent reporting,7,8 and dubious methodologic choices9 all indicate that there is reason to scrutinize and doubt the optimistic claims of these data. Thus these results, while adequate for making clinical decisions, should be considered preliminary unless and until they can be verified on a large scale by parties without a financial stake in the results.

Author

David Newman, MD

Published/Updated

Source

Avenell A, Gillespie WJ, Gillespie LD, O’Connell D. Vitamin D and vitamin D analogues for preventing fractures associated with involutional and post-menopausal osteoporosis. In: Avenell A, with The Cochrane Collaboration, eds. Cochrane Database of Systematic Reviews.
Sanders KM, Stuart AL, Williamson EJ, et al. Annual High-Dose Oral Vitamin D and Falls and Fractures in Older Women. JAMA: The Journal of the American Medical Association. 2010;303(18):1815 -1822.
Bolland MJ, Grey A, Avenell A, Gamble GD, Reid IR. Calcium supplements with or without vitamin D and risk of cardiovascular events: reanalysis of the Women’s Health Initiative limited access dataset and meta-analysis. BMJ. 2011;342:d2040.

Efficacy Endpoints

Prevention of fractures

Harm Endpoints

Kidney stones, kidney damage

Narrative

Vitamin D increases the absorption of calcium in the intestine and has a role in maintaining bone density. Sources of Vitamin D are limited other than exposure of the skin to ultraviolet light (sunlight) which stimulates Vitamin D production. Some authorities suggest Vitamin D as a supplement in the hopes that it might lead to reductions in fractures by increasing bone strength. The Cochrane review summarized here examined two large trials with >3850 elderly, institutionalized subjects.

In frail, elderly individuals in institutional settings there was a small reduction in hip fractures, which is an important and dreaded injury for a variety of reasons. The NNT for avoiding a hip fracture in this group group [164/2023 (8.1%) vs 199/1830 (10.9%)] was approximately 36.

There does appear to be consistent harm associated with Vitamin D, in the form of an increase in kidney problems, either kidney stones or renal insufficiency. The NNH for this was [453/22529 for 20.1% versus 388/22449 for 17.3%] 36.

Caveats

In community dwelling adults in trials there was no benefit. There are recent data to suggest that Vitamin D supplementation may increase falls and fractures though this was in non-institutionalized persons,1 and calcium supplementation has recently been implicated in heart attacks among post-menopausal women. The benefit noted here for institutionalized elderly patients was recorded in two different trials, both performed at the same institution in France, and the Cochrane authors point out that this should be confirmed in other settings before it is widely accepted as accurate.2

However, the avoidance of hip fractures is immensely important. While the prospect of kidney damage or stones and the hints of cardiac problems may turn out to be significant in future studies, hip fractures are known to be very painful and can lead to major changes in levels of independence and overall health. Thus until proven otherwise we believe that the current state of evidence supports the use of Vitamin D combined with calcium supplementation for the institutionalized elderly, and have classified this as a ‘GREEN’ intervention.

Author

David Newman, MD

Published/Updated

May 15, 2011

Source

Avenell A, Gillespie WJ, Gillespie LD, O’Connell D. Vitamin D and vitamin D analogues for preventing fractures associated with involutional and post-menopausal osteoporosis. In: Avenell A, with The Cochrane Collaboration, eds. Cochrane Database of Systematic Reviews.
Sanders KM, Stuart AL, Williamson EJ, et al. Annual High-Dose Oral Vitamin D and Falls and Fractures in Older Women. JAMA: The Journal of the American Medical Association. 2010;303(18):1815 -1822.
Bolland MJ, Grey A, Avenell A, Gamble GD, Reid IR. Calcium supplements with or without vitamin D and risk of cardiovascular events: reanalysis of the Women’s Health Initiative limited access dataset and meta-analysis. BMJ. 2011;342:d2040.

Efficacy Endpoints

Prevention of fractures

Harm Endpoints

Kidney stones, kidney damage

Narrative

Vitamin D increases the absorption of calcium in the intestine and has a role in maintaining bone density. Sources of Vitamin D are limited other than exposure of the skin to ultraviolet light (sunlight) which stimulates Vitamin D production. Some authorities suggest Vitamin D as a supplement in the hopes that it might lead to reductions in fractures by increasing bone strength. The Cochrane review summarized here is the largest review of this topic to date.1

Unfortunately, in large randomized trials (some of very high quality) there is one group that appears to have benefited and this was only true when Vitamin D was combined with calcium supplementation: frail, elderly individuals in institutional settings. Among this group there was a small reduction in hip fractures, which is an important and dreaded injury for a variety of reasons. The NNT for avoiding a hip fracture in this group [395/23123 for 17.1% vs 459/23535 for 19.5%] was approximately 42.

For all other groups there is no evidence of an identifiable benefit, and there does appear to be consistent harm associated with Vitamin D, in the form of an increase in kidney problems, either kidney stones or renal insufficiency. The NNH for this was [453/22529 for 20.1% versus 388/22449 for 17.3%] 36.

Caveats

The trials included amount to roughly 85000 participants, making it highly unlikely that there is even a tiny benefit that is, as of yet, undiscovered. However Vitamin D was delivered variably (injections, pills, infusions) and the compound itself was variable (cholecalciferol, calcitriol, alfacalcidol, etc.) therefore perhaps one of these forms and compounds may provide a benefit that was diluted out by the larger group. If present this benefit would be small, perhaps negligibly so.

There is recent data to suggest that Vitamin D supplementation may increase falls and fractures in non-institutionalized persons2 and the Cochrane data demonstrate a nearly significant increase in hip fractures (OR 1.16, 95%CI; 0.99-1.35). Moreover, calcium supplementation has recently been implicated in heart attacks among post-menopausal women.3

Finally, the benefit seen in institutionalized elderly patients was recorded in two different trials, both performed at the same institution in France, and the Cochrane authors point out that this should be confirmed in other settings before it is widely accepted as accurate.

Author

David Newman, MD

Published/Updated

Source

The role of antibiotic prophylaxis for prevention of infection in patients with simple hand lacerations. Ann Emerg Med 2007;49:682-89.

Efficacy Endpoints

Infection, Cosmesis (Not reported)

Harm Endpoints

Diarrhea (Not reported)

Narrative

There are 3 reviews addressing the utility of antibiotic prophylaxis in simple hand lacerations. The oldest was published by Cummings and Del Beccaro in 1995 and found no benefit to antibiotic administration. This meta-analysis was limited, however, by the inclusion of several trials lacking in key methodologic standards. In the two most recent reviews, Zehtabchi and BestBets, more rigorous standards were employed for study selection. All 3 reviews reach the same conclusion.

The most relevant to the outpatient emergent and urgent care settings, Zehtabchi’s review excludes studies with complicated hand lacerations managed by surgical specialties. Simple laceration was further defined by a lack of involvement of structures like nerves, tendons, bones and large vessels. The 3 included studies (n = 778) described wound care management during the initial encounter. All wounds were irrigated with normal saline and prepped with an antiseptic agent such as chlorhexadine. Antibiotic regimens varied across studies both in terms of type and delivery method. Selected antibiotics included different penicillin preparations and cephalosporins and both oral and intramuscular routes of administration were compared to placebo or nothing. Overall, administration of a prophylactic antibiotic, in any form and of any type, to patients with uncomplicated hand lacerations did not reduce the rate of infection. In two of the included studies, infection rates were surprisingly low, roughly 1%. One trial had unusually high rates of infection (8.8 vs. 12%) but still found no statistical benefit to antibiotic usage.

Summary of Individual Trials:
Grossman et al. n = 265, infection rate 1.15% vs 1.1% (RR 1.05, 95% CI 0.09-11.38)
Roberts and Teddy, n = 305, infection rate 8.8% vs. 12%, (RR 0.73, 95% CI 0.37-1.46)
Beesley et al. n = 145, infection rate 1.4% vs 1.3%, (RR 1.07, 95% CI 0.07-16.8)

Caveats

While one study described diabetes as an exclusion criterion the others did not offer explicit information about the comorbidities of enrolled patients. Accordingly, for patients with conditions placing them at higher risk for infection, clinician judgment should factor into decision-making. Also, the most recent of the 3 included studies was performed about 30 years ago. But as wound care management has not changed dramatically over this time period, this factor should not affect applicability.

*Harms were not rigorously tracked in any of the included studies, however fungal vulvovaginitis, diarrhea, and other harms are common enough there is certain to be a fraction (likely 10-20% or more) of patients that will experience these adverse effects when administered antibiotics. Less common but more dangerous harms have not been as well described and cannot be as consistently expected.

Author

Ashley Shreves, MD

Published/Updated

April 23, 2011

Source

Brouwer MC, McIntyre P, de Gans J, Prasad K, van de Beek D. Corticosteroids for Acute Bacterial Meningitis. Cochrane Database of Systematic Reviews 2010, Issue 9.

Van de Beek D, Farrar JJ, de Gans J, et al. Adjunctive dexamethsone in bacterial meningitis: a meta-analysis of individual patient data. Lancet 2010;9;254-63.

Efficacy Endpoints

Mortality, hearing loss, neurological sequelae

Harm Endpoints

Recurrent fever, persistent fever, gastrointestinal bleeding, herpes zoster, fungal infection, or reactive arthritis

Narrative

Acute bacterial meningitis (ABM) carries a high morbidity and mortality. As bacteria enter the meninges there is a dramatic inflammatory response that potentiates the infectious process and initiates a cascade of events that can culminate in cerebral edema and a rise in intracranial pressure. This can result in neurologic damage and/or death. Giving steroids to patients with bacterial meningitis would theoretically decrease this inflammatory response, reducing neurologic sequelae and death.

This review included 24 studies (4041 patients) and compared the effects of steroids with placebo. Twenty-one of the studies used dexamethasone with a range of dosing from 0.4-1.5mg/kg/day for two to four days. About two-thirds of included patients were children and a third were adults.

Benefits: Overall, there was no statistically significant difference in mortality rate between patients who received steroids and those who received placebo (18% vs 20%, p = 0.18).

Only 19/24 studies looked at hearing loss as a complication of ABM. The administration of steroids led to an absolute risk reduction of hearing loss by approximately 5% (NNT=20.5). Seventeen studies differentiated between any hearing loss and severe hearing loss, showing a three percent reduction (NNT=31).

Excluding hearing loss there were thirteen studies (1756 patients) showing a decreased risk of short term neurological sequelae (NNT=26.9), though no difference in long term sequelae was observed.

Results were also subdivided by pathogen. S. pneumoniae was the etiologic agent in about 1100 subjects and was the most deadly (mortality 32.8%). H. influenzae and N. meningitidis had mortality rates of 10.5 and 4.3 respectively. There was a decrease in mortality for the subgroup of patients with S. pneumoniae (NNT=17.5). Children with H. influenzae also had decreased rates of hearing loss (NNT=12.5).

One further subgroup analysis examined differences in high and low income countries. Results for high income countries were similar to those reported previously in the analysis. On the contrary, results for low income countries had significant heterogeneity and found no statistically significant differences.

Harms: Adverse events were recorded in nineteen studies. The steroid group had an increase in recurrent fever (NNH=16.3).

Caveats

There is an updated meta-analysis of ‘patient-level data’ for dexamethasone in bacterial meningitis, which tends to offer a slightly more nuanced and detailed look at the possibility of a benefit (Van de Beek, 2010). In patient-level data analyses data from individual patients in each trial are analyzed, rather than analyzing aggregated blocks of data from each trial. This meta-analysis included 5 large, high quality trials and found no mortality benefit to dexamethasone, results consistent with the Cochrane review. However, the authors also found no benefit in terms of severe neurologic problems, other than hearing loss among adult survivors. They conclude that the benefits of dexamethasone remain unproven. An accompanying editorial points out that there did seem to be benefit in trials from industrialized settings and that this may have been neutralized by lack of benefit in other settings where meningitis may present later or be associated with poorer outcomes. The editorialist supports steroid administration in industrialized settings but concedes small benefits and a lack of certainty (Brewer, 2010).

Author

Patricia Van Leer, MD

Published/Updated

April 22, 2011

Source

CRASH-2 trial collaborators, 2010. Effects of tranexamic acid on death, vascular occlusive events, and blood transfusion in trauma patients with significant haemorrhage (CRASH-2): a randomized, placebo-controlled trial. Lancet. 2010 Jul 3;376(9734):23-32.

Efficacy Endpoints

All-cause mortality and bleeding-related mortality

Harm Endpoints

Vascular occlusive events, requirement for transfusion of blood products

Narrative

Injuries, particularly associated with motor vehicle accidents, represent a substantial cause of premature death worldwide. Approximately one-third of in hospital deaths due to trauma are directly attributed to haemorrhage. In addition, multi-organ failure can also be partially attributed to blood loss and its sequelae. To date, studies have been carried out examining the role of anti-fibrinolytic agents in elective surgery, however little research has been carried out examining the role of these drugs in trauma.

Benefits: This trial enrolled 20211 adult trauma patients from 274 hospitals across 40 countries. The trial showed that early intervention with a standardized regimen of tranexamic acid administered intravenously can significantly reduce all cause mortality as well as mortality due to bleeding.

Harms: No increase in rates of vascular occlusive events were noted, the adverse event of most concern related to administration of this drug.

Caveats

One caveat concerns internal validity. Given the drug’s reported mechanism of action, there are inconsistent findings. There was no difference between treatment and placebo groups in the rates of vascular occlusive events. While it is reassuring that a large systematic review of anti-fibrinolytic agents in surgery1 corroborates these safety findings for tranexamic acid, the authors note that observational studies of a similar agent, aprotinin, revealed increases in thrombotic events after early trials had reported no difference. Along the same lines, while it was expected that transfusion rates should decrease in patients treated with tranexamic acid, there was no significant difference between groups. However transfusion decisions are subjective and mortality was higher in the control arm, potentially reducing the opportunities for transfusion in that group.

For external validity, the study sample was a sick trauma cohort. Enrollment was limited to those with signs of significant hemorrhage or for those in whom a high suspicion of hemorrhage existed. Accordingly, about a third of the patients were in shock on presentation, systolic BP < 90 mm Hg, and about a third had a GCS between 3 and 8. This is consistent with the high mortality (1 in 6 dead at 1 month). The NNT cited above may be be significantly higher (less benefit) in a lower risk population. Second, many of the recruiting centres lacked sophisticated trauma systems that are well resourced with imaging and high tech surgical care. It is possible that the benefits of this drug would be lost in a more modern context.

Still, the drug is inexpensive and, based on this data, demonstrated no associated adverse effects, so its use should be considered in trauma patients with hemorrhagic shock or at high risk for bleeding.

Author

Kyle McCoy

Published/Updated

March 11, 2011

Source

Zhang et al. Nebulized hypertonic saline solution for acute bronchiolitis in infants. Cochrane Database Syst Rev. 2008 Oct 8;(4):CD006458.

Kuzik et al. Effect of inhaled hypertonic saline on hospital admission rate in children with viral bronchiolitis: a randomized trial. CJEM. 2010 Nov;12(6):477-84.

Efficacy Endpoints

Duration of Hospital Stay, Decrease in Clinical Severity Scores, Decrease in Admission Rates

Harm Endpoints

Tachycardia, Hypertension, Pallor, Tremor, Nausea, and Vomiting

Narrative

Acute bronchiolitis is the one of the most common respiratory infections in infants. It affects the lower respiratory tract and is usually caused by a virus, with respiratory syncytial virus (RSV) being the most common pathogen. It is estimated that nearly all infants are infected with RSV and that between 1 and 2% of these develop symptoms severe enough to warrant admission. While the use of nebulized racemic epinephrine has been shown to cause a temporary and modest improvement of symptoms in the outpatient population, the treatment options for admitted patients are limited. The administration of nebulized 3% saline had been proposed as an intervention that may help improve mucous clearance and therefore clinical outcomes through various mechanisms. The Cochrane Review included four trials, one outpatient (n = 65) and three inpatient (n = 189) of patients who were randomized to receive either nebulized 3% saline or 0.9% saline in addition to bronchodilators (either albuterol or racemic epinephrine) and standard supportive care. We also looked at a recently completed emergency department study (Kuzik et al 2010) that randomized infants to 3% or 0.9% saline and examined whether there was a significant change in the rate of admission (n = 81).

Administration of nebulized 3% saline in the emergency department did not result in a statistically significant decrease in the rate of admission for infants presenting with acute bronchiolitis, though there was a trend towards decreased admissions in both outpatient studies. However, in the studies examining inpatients, infants in the treatment arms had an average decrease in hospital stay of approximately 1 day, with no adverse reactions or harms recorded. Subjects receiving 3% saline were also noted to have improved clinical severity scores, especially on the first day of admission, when compared to the groups receiving 0.9% saline.

Caveats

The total number of patients involved in both the inpatient (189 total) and outpatient trials (146 total) included in our analysis were quite small. In addition, given the subjectivity of calculating clinical severity scores it is difficult to give much weight to the benefit reported in the children receiving the 3% saline. The patients treated with 3% saline went home an average of one day earlier. Given the prevalence of bronchiolitis, the cost and morbidity associated with hospitalization, and the emotion and stress involved for the admitted infants and their families, we feel this is a meaningful reduction. Furthermore, this intervention is affordable, minimally invasive and does not appear to be associated with any significant potential harms or adverse reactions.

Author

Dan Runde, MD

Published/Updated

March 10, 2011

Source

Russell KF, Liang Y, O’Gorman K, Johnson DW, Klassen TP. Glucocortocoids for croup. Cochrane Database of Systematic Review. 2011, Issue 1.

Efficacy Endpoints

Return visits and/or (re)admissions; required for additional treatment (rescue therapy)

Harm Endpoints

Medication side effects

Narrative

Croup is a self-limiting upper airway disease in children. However, croup accounts for a high number of physician visits and hospital admissions. It accounts for an annual incidence of 1.5 to 6% in children less than six years of age. Admission rates vary from 1.5% to 31%.

This Cochrane review has 41 trials, all parallel design, occurring in the inpatient and outpatient setting. Dexamethosone is the most commonly evaluated glucocorticoid in this Cochrane review. Other glucocorticoids are budesonide, prednisone, and fluticasone.

Ten trials were examined for return visits and/or (re)admission rates from 1679 enrolled patients. The authors found the rates were reduced with glucocorticoids therapy, as compared to placebo (ARR = 9.2%, NNT = 11).

There were 17 trials that assessed the need for rescue therapy, such as intubation/tracheostomy, antibiotic use, additional glucocorticoid use, or mist tent use. While there was no reduction in these rescue therapy endpoints, a reduction in epinephrine as a rescue therapy was noted, with an absolute risk reduction of 10% (NNT = 10).

Caveats

In the Cochrane review, glucocorticoid has been shown to reduce the revisits and/or (re)admission rates and to decrease epinephrine use as a rescue therapy. The seven trials in the ED show the most benefit of glucocorticoid, where the absolute risk difference is 10% (NNT = 10). In the inpatient setting this absolute risk difference was 6.8% (NNT = 15). The 9.2% absolute benefit in need for readmission (NNT=11) is impressive, but this is based on a 21% overall rate of readmission in children in the control groups in this review. The actual benefit for a child will depend upon how often croup children are readmitted at baseline in one’s institution. The Cochrane authors note that in their group of studies the average readmission rate among institutions was roughly 12%, which would suggest an average NNT of 17.

The Cochrane review examined other outcome measures, such as length of stay and clinical severity. The cumulative length of stay was shortened by glucocorticoids by just over 11 hours in the combined inpatient (8 trials) and outpatient (1 trial) trials. Leipzig (1979) was the sole ED trial, however these study patients were admitted to the inpatient floors.

Finally, glucocorticoid treatment reduced clinical severity as measured by Westley scores. Children have a better clinical outcome according to the Westley score at the 6-hour and 12-hour mark (-1.2 and -1.9 points, respectively). A score reduction of 1 unit from baseline is deemed to be clinically significant.

Overall, the Cochrane review suggests that 5 patients would need to be treated with glucocorticoid for 1 patient to experience some measurable clinical benefit. The benefits are lower Westley score, fewer visits and/or (re)admission, decreased length of stay, or decreased usage of epinephrine as rescue therapy.

Author

Jeffrey Hom, MD, MPH

Published/Updated

February 15, 2011

Source

Arrich J, Holzer M, Herkner H, Müllner M. Hypothermia for neuroprotection in adults after cardiopulmonary resuscitation. Cochrane Database of Systematic Reviews 2009.

Bernard SA, Gray TW, Buist MD, Jones BM, Silvester W, Gutteridge G, et al. Treatment of comatose survivors of out-of-hospital cardiac arrest with induced hypothermia. New England Journal of Medicine 2002;346(8):557–63.

Hypothermia after Cardiac Arrest Study Group (HACA). Mild therapeutic hypothermia to improve the neurologic outcome after cardiac arrest. New England Journal of Medicine 2002;346(8):549–56.

Hachimi-Idrissi S, Corne L, Ebinger G, Michotte Y, Huyghens L. Mild hypothermia induced by a helmet device: a clinical feasibility study. Resuscitation 2001;51(3):275–81.

Efficacy Endpoints

Neurologic recovery (best outcome while in hospital, as measured by cerebral performance categories), survival to hospital discharge

Harm Endpoints

Bleeding, pneumonia, sepsis, pulmonary edema, cardiac dysrhythmias

Narrative

After the return of spontaneous circulation following cardiac arrest, neurologic injury may occur in the process of reperfusion. Past investigations have suggested that the induction of hypothermia in initially comatose survivors can help mitigate this injury through several potential mechanisms, including reduction of free radical and neurotransmitter-mediated neuron damage, improved cerebral microcirculation, and protection of the blood-brain barrier.

The Cochrane review included five randomized trials and excluded two from data pooling, one because of limited information and the other because it utilized hemofiltration as a means of hypothermia. The remaining three trials comprised 195 comatose subjects treated with mild hypothermia (32-34˚C) and 188 controls. All trials utilized external cooling. Pooled analysis of these trials revealed a relative risk of 1.55 (95% CI 1.22-1.96) for good neurologic outcome overall, and relative risk of 1.35 (95% CI 1.1-1.65) for survival to hospital discharge when considered alone. The NNT for these pooled outcomes were 5 and 7, respectively.

Adverse events were addressed in each of the 3 studies, though none were powered to detect differences. These adverse events included, but were not limited to, pneumonia, bleeding, need for transfusion, sepsis, pulmonary edema, lethal arrhythmias, and renal failure. There were no statistically significant differences between treatment groups.

Caveats

Only three trials were pertinent for inclusion, one of which was a small feasibility trial. There was substantial clinical heterogeneity among the trials, although the survival and neurologic outcome results were statistically homogenous as measured by the I2 statistic. In addition, Bernard and HACA focused on patients with ventricular arrhythmias as their first observed cardiac rhythm, while Hachimi-Idrissi only examined patients with initial asystole/PEA. The inclusion of subjects with an initial rhythm other than ventricular arrhythmia remains a subject of ongoing debate and this third trial that enrolled such subjects was the smallest (n=33), thus it is unclear the degree to which this can be considered evidence in support of this practice. Patients also reached target temperature at different times (2 hours vs. 8 hours) and average duration of cooling varied significantly (3 hours vs. 24 hours).

Finally, we chose not to report improvement in neurologic outcome and survival to discharge as separate outcomes (as the Cochrane review did) because neurologically intact survival is clinically relevant and includes overall survival. Improved neurologic outcome in survivors treated with a hypothermia protocol versus control survivors was not individually tested, is not distinguishable in the reported studies, and would require larger studies.

Author

Brandon Maughan, MD and Anthony M. Napoli, MD FACEP

Published/Updated

Source

O’Malley GF, Dominici P, Giraldo P, et al. Routine packing of simple cutaneous abscesses is painful and probably unnecessary. Acad Emerg Med 2009;16:470-473.
Barnes SM and Milsom PL. Abscess: an open and shut case. Arch Emerg Med 1988;5:200-205.

Efficacy Endpoints

Prevention of need for secondary intervention (for example, repeat incision and drainage); prevention of recurrent abscess

Harm Endpoints

Pain during packing/removal of packing

Narrative

The management of abscesses has classically included incision and drainage of the lesion followed by the insertion of packing material into the cavity. Patients are typically told to return to the ED 48 hours later, at which point the packing is removed and, occasionally, new packing inserted. Multiple textbooks support this approach and describe packing as critical in preventing recollection of pus. It is believed that packing absorbs pus and debris and prevents premature closure of the incision, leaving a route of egress for pus.

Surprisingly, there is only one randomized, controlled trial evaluating the utility of packing in simple cutaneous abscesses managed in the ED setting1. The sample size is small (n = 48), the methods are good, and outcome assessments were blinded. Of note, all subjects were prescribed antibiotics. At 48 hours, 17% in the packing group and 20% in the nonpacking group received a secondary intervention (p = 0.72), although use of these interventions was left to judgment and most were of questionable clinical utility.

In this study pain scores were evaluated, with an average difference in pain scores post-procedure and at two days of 23.8 mm (95% CI, 5-42 mm) and 16.4 mm (95% CI, 1.6-31.2 mm), respectively, in favor of not receiving packing.

Caveats

These data include only one small trial. However there are many studies in the surgical literature of varying quality indirectly assessing the utility of packing in cutaneous abscesses. An informal review of observational and trial data (n =1943) evaluating a ‘closed’ versus ‘open’ technique for abscess treatment found the closed approach yielded faster healing and no difference in recurrence. In the closed approach, incision and drainage is performed and sutures placed to close the wound. The open technique included packing. While all types of abscesses were included in this review, they were drained under general anesthesia and all patients received one dose of intravenous antibiotics during the procedure, suggesting that the abscesses in these data were larger and more complex than those seen in most outpatient settings. This may limit the applicability of these results to the ED setting, although the less complex nature of most abscesses suggests that most abscesses are even less likely to benefit from packing than those evaluated in observational data. These data therefore support the findings of the one trial, suggesting that packing after abscess drainage, a classically recommended practice that was not based on evidence but rather based on theoretical concerns, may lead to increased pain without any corresponding benefit.

Author

Ashley Shreves, MD

Published/Updated

December 23, 2010

Source

Gosselin RA, Roberts I, Gillespie WJ. Antibiotics for preventing infection in open limb fractures. Cochrane Database Syst Rev.

Efficacy Endpoints

Early Wound Infections

Harm Endpoints

None noted

Narrative

The use of early antibiotics in the treatment of open limb fractures has long been considered standard in the United States and most other developed nations. Open fractures are, by definition, considered contaminated. Use of systemic antibiotics as part of the initial treatment of these fractures is aimed at preventing this contamination from developing into more serious complications such as osteomyelitis or abscess. This review found that early antibiotic use helped prevent early infections in open limb fractures (ARR 6.5%, NNT = 16).

Caveats

The interpretation of the data presented by this review is severely limited. The primary outcome of “early infection” was poorly defined and heterogeneous across the 8 studies included (in one study early infection was defined as any infection within 6 weeks of injury, in others it was defined only as “early wound infection”). The studies were underpowered and unable to evaluate significant patient oriented outcomes such as mortality, amputation, development of osteomyelitis (chronic or acute), delayed bone union or bone non-union and length of hospital stay. The use of antibiotics for the treatment of open finger fractures was equivocal, and the little existing data appear to show no benefit (3 trials with N = 367 and 95% CI 0.26 to 1.23). None of the studies reported any data on antibiotic side effects, nor did they examine the potentially significant issue of developing antibiotic resistance. Additionally, only one study was performed within the last 20 years.

While the authors of this Cochrane Review conclude that no further study is needed on the subject of whether early antibiotic use should be standard practice in open limb fractures, they only found statistical significance in one poorly defined primary endpoint. There is a glaring lack of available evidence for a large number of patient-oriented outcomes. There is also no evidence advocating the use of early antibiotics in open finger fractures. Furthermore, the issues of antibiotic side effects and of developing antimicrobial resistance in this patient population have not been examined. In summary, while this evidence does not represent a robust or valid enough dataset to suggest a ‘standard of care’ it is reasonable to utilize antibiotic treatment for open fractures until additional data better elucidate both the benefits and harms associated with this approach.

Author

Dan Runde, MD

Published/Updated

Source

Hamilton HC, Foxcroft DR. Central venous access sites for the prevention of venous thrombosis, stenosis and infection in patients requiring long-term intravenous therapy. Cochrane Database Syst Rev. 2007 Jul 18;(3):CD004084

Merrer J et al. Complications of femoral and subclavian venous catheterizaiton in critically ill patients: a randomized controlled trial. JAMA. 2001 Aug 8;286(6):700-7

Efficacy Endpoints

Catheter-related infectious complications (colonization with or without sepsis), Catheter-related thromboses (fibrin sleeves, major and complete vessel thrombosis)

Harm Endpoints

Catheter-related mechanical complications (arterial puncture, minor bleeding, hematoma, misplaced catheter)

Narrative

Central venous access is often felt to be essential during management of the critically ill patient. Ideally, clinicians choosing anatomic sites for central venous access would be aware of the disadvantages and advantages associated with the most commonly used sites. These sites include the internal jugular vein, the subclavian vein and the femoral vein.

The Cochrane Review identified one randomized control trial that examined the difference between subclavian and femoral vein sites, concluding that the subclavian site is associated with a significantly decreased rate of both infectious and thrombotic complications. This conclusion however was based on disease-oriented (rather than patient-oriented) outcomes such as 'line colonization', and ultrasound detected femoral vein thrombosis. There was no difference between the two sites in terms of actual clinical infections. There was also no significant increase in mechanical complications during or secondary to line insertion.

Caveats

There is currently only one high quality randomized control trial that examines the differences among these sites. The study included a total of only 293 patients and was underpowered (too small) to be able to reliably detect differences in clinical infection rates. When combining clinical infection with 'line colonization', a surrogate marker measuring bacteria detected on the tip of a catheter that had been removed, the investigators were able to report a difference with statistical significance. There was no difference in clinical infection rates (a patient-oriented benefit) between groups, however. During the study, there were also four pneumothoraces that occurred in association with subclavian line placement, an important complication that is not associated with the femoral access site. Furthermore, the study only examined the difference between subclavian and femoral access sites. While there are multiple prospective and retrospective studies that include analysis of internal jugular central venous access, neither the Cochrane Review nor this author were able to find a randomized trial that examined complication rates among subclavian, internal jugular and femoral access sites.

Finally, the study from which the Cochrane review reports its data was performed nearly a decade ago, a time when the 'checklist' approach that includes new and diligent measures to prevent infection were not in widespread use. This has the potential to render these data irrelevant to the current routine practice of central line placement. There is therefore a lack of high quality, applicable evidence pertaining to the preferred site of central venous access.

Author

Dan Runde, MD and Jarone Lee, MD, MPH

Published/Updated

November 8, 2010

Source

Chlebowski RT, Anderson GL, Gass M. Estrogen Plus Progestin and Breast Cancer Incidence and Mortality in Postmenopausal Women JAMA. 2010; 304(15):1684-1692

Efficacy Endpoints

Mortality, breast cancer mortality, breast cancer diagnoses

Harm Endpoints

Death, due to breast cancer, diagnosis of breast cancer

Narrative

Hormone replacement therapy (HRT) remains a controversial intervention with complex effects. The study targeted here for appraisal and translation is a follow-up analysis examining participants in the Women"s Health Initiative study originally published in 2002.1 The original study was the largest and best of its kind and surprised many with results suggesting that coronary heart disease was not reduced among those taking HRT, but indeed may have been slightly increased. This was a stark contrast to both the conventional wisdom and the results of many observational (cohort) studies examining the issue. In addition, the study was stopped early due to a secondary finding of increased breast cancer diagnoses among those taking HRT. In this analysis the authors extend the analysis of breast cancer outcomes by an additional 7 years and report on not just breast cancers, but also deaths due to these breast cancers among women who had been enrolled in the original study.

The result suggests an increase in the chances of death due to breast cancer, although the differences are small for individuals. Taken at face value the report indicates that more than 99.9% of women were unaffected in terms of breast cancer by HRT for each year of therapy, and 99% were unaffected after 11 years cumulatively. Perhaps most importantly, one in every 5000 women taking HRT per year lost her life to breast cancer because of the HRT. These numbers may be contextualized for women and physicians attempting to make a decision: First, based on this study, roughly 1 in every 7 additional breast cancers due to HRT would result in death. Second, based on the earlier report, it is also true that fractures related to osteoporosis and colorectal cancers are favorably affected. For every 8 cases of breast cancer caused by HRT there would also be 6 colorectal cancers prevented, 5 hip fractures prevented, 6 back fractures prevented, and hot flashes would be reduced by approximately 75%.

Caveats

The topic is complex and difficult to study even in randomized trials, and an individual"s baseline risk of cancer (due to genetics, lifestyle, etc.) are clearly much more important determinants than any pill. Therefore the numbers reported here and from any other study on the topic, should be tailored whenever possible to an individual"s baseline risk. This would offer a more individualized estimate of true risk.

An important additional note is that this analysis is a secondary analysis of a trial for which breast cancer was a secondary endpoint. Analyzing subjects years after they have left their original study groups means that in most cases they are no longer taking the original treatment or placebo and are now subject to risks and confounders that make cause-effect conclusions unstable and prone to error. In addition, as the authors point out, the original study was designed to detect coronary events (heart attacks, cardiovascular deaths, etc.) and therefore the study was designed to answer a different question than the question asked here. This limits, though does not invalidate, the importance of these findings. An ideal study would be designed to seek and report breast cancer outcomes primarily rather than secondarily. Despite these caveats this is certainly the best available data on the topic and should be considered a robust and reasonable best estimate of the impact of HRT on breast cancer.

Author

David Newman, MD

Published/Updated

October 20, 2010

Source

Gøtzsche PC, Nielsen M. Screening for breast cancer with mammography. Cochrane Database Syst Rev. 2006 Oct 18;(4):CD001877. Review. Update in: Cochrane Database Syst Rev. 2009;(4):CD001877. PubMed PMID: 17054145.

Efficacy Endpoints

Mortality, breast cancer mortality

Harm Endpoints

Unnecessary surgical procedure, false positives, overdiagnosis leading to treatment

Narrative

Screening mammography (mammography in women without any signs or symptoms of possible breast cancer) has been studied in large randomized trials of nearly a half million women. The theoretic basis for the intervention is sound. It is presumed that therapeutic intervention at a point when cancer is visible on a mammogram but not yet palpable in the breast will, for a small number, result in earlier, ultimately life-saving, therapy. Overall mortality rate, however, was not improved in the groups in these studies assigned to receive regular mammograms. When aggregating data from those trials in which randomization was appropriate (resulting in balanced groups), there was also no identifiable reduction in deaths due to breast cancer.

The statistical result is slightly different when one accepts all trial data rather than restricting data to appropriately randomized studies. While overall mortality remains unchanged, in this analysis breast cancer mortality is reduced by approximately 15%. If accurate, this would represent a NNT of approximately 2000. However, breast cancer mortality is not as important as overall mortality, because individuals deciding whether to undergo screening mammography will typically want to avoid death, rather than simply avoiding death from one possible cause. The lack of overall mortality benefit with a potentially small benefit noted in breast cancer mortality suggests that mammography is likely to be taking as many lives as it is saving.

This is an important, though uncommonly discussed, issue in the translation of evidence from cancer screening trials.1 It is known that overdiagnosis (treatment of cancers that would have been no threat), and high false positive rates (misdiagnosis) lead to medical harms and unnecessary surgeries, chemotherapy, and radiation. It is presumed that a small number of these individuals also die due to aggressive therapies such as chemotherapy and major surgery, and given the small, though disputed, margin of benefit suggested by the analysis above it seems likely that if there is a benefit to screening mammography it is balanced out by mortal harms from overdiagnosis and false-positives.1 If screening mammography reduces the rate of one type of death, but the overall rate of death is not affected, it is mathematically implied that there is an increase in some cause of death that is offsetting the reduction in breast cancer deaths. Unfortunately, because most massive trials of mammography are simply too large to reliably track every cause of death, it is often difficult to tell with precisions where this increase occurs.

Harms due to breast cancer screening are discussed at length in the Cochrane review, and in more recent analyses.2,3 These include false positives that result in surgical procedures, and emotional distress due to false positives. In addition, despite the lack of identifiable benefit to screening mammography, women in groups assigned to receive mammograms were 20% more likely to undergo mastectomy and 30% more likely to undergo surgery. Finally, if it is true that breast cancer deaths are reduced it has been estimated that for every one patient who avoids death from breast cancer approximately 10 to 20 women are treated unnecessarily as cancer patients, typically receiving surgery, radiation, and chemotherapy.3,4 This additional estimate of harm is not represented in the NNH statistics listed above.

Caveats

This topic is complex and difficult to study even in randomized trials, where some patients assigned to mammography do not comply with plans for screening and others assigned not to receive mammograms undergo screening (i.e. 'contamination'). Moreover the studies are dated, and there have been significant improvements in breast cancer therapy since these trials. Therefore these data are highly imperfect. This has led many researchers to examine observational data, from which inferences regarding cause and effect are less valid, but which may be useful in explaining epidemiologic trends showing no significant reduction in breast cancer deaths since the advent of screening mammography. Most recently, an observational study of the effects of screening in the United States over the past thirty years suggests that 50% of screen-detected cancers represent overdiagnosis, and that while early stage cancer diagnoses have doubled, advanced stage breast cancer is about as common today as it was before mammography was in widespread use.5

Importantly, overall mortality may not be affected by mammography because breast cancer deaths are only a small fraction of overall deaths. This would make it very difficult to affect overall mortality by targeting an uncommon cause of death like breast cancer. If this is the reason for trial data demonstrating no overall mortality benefit then it means that it would take millions of women in trials before an overall mortality difference was apparent, a number far higher than the current number of women enrolled in such trials. If this is the proper explanation then any important impact on mortality exists, it is small enough that it would take millions of women in trials to identify it. This belies the public perception of mammography.

It is important to note that screening mammography saves some number of individual lives in the sense that there are almost certain to be women whose breast cancer is amenable to curative treatment at the time it is identified by mammography, but not amenable to cure at the time it is identified by physician examination or self-examination. However the number of cancers for which the window of curative potential exists exclusively during this time is likely to be small, and however many lives are saved in this way an equal number appears to be lost due to radiation from mammograms, or overdiagnosis, or false positives, or some combination of the three. The NNT that we calculate here is designed to be a tool for those attempting to make decisions about health care interventions, and in aggregate the best available evidence suggests there is no overall benefit to screening mammography. This should not be taken to mean that some individuals are not saved—it is likely that they are—however an equal number of individuals appear to lose their lives due to mammography, and there is no way to know which group one will eventually fall into.

Two recent data reviews deserve further mention. The United Kingdom commissioned an independent review after dissenting voices swelled, for the purpose of better informing shared decision making and educational materials about the harms and benefits of screening.6 Unfortunately the review concluded that screening mammography trials were inadequately powered to detect an impact on all-cause mortality, and therefore used breast cancer mortality as a primary outcome. They concluded a 20% reduction and used this in their discussion of harms and benefits. As noted above, cause-specific mortality is both scientifically unstable (conclusion reversal is common when all-cause mortality is considered)7 and disease-centered rather than patient centered (patients would prefer to avoid death altogether). Thus, either screening mammography does not save lives or else we have inadequate data to say whether it does or does not. In neither case can a benefit be scientifically claimed.

A circumspect, and patient-centered, response to the UK review provides a brief, smart data analysis suggesting that screening mammography is taking significantly more lives than it is saving.8 This latter review takes into account improvement in cancer treatments and likely overdiagnosis (using the US data, which were published after the UK-commissioned review), and finds that the short and long-term sequelae of radiation therapy in cancer treatment regimens is leading to two to three lives taken by mammography for every one life saved. We have nonetheless chosen to retain the 'Red' (data suggest no benefit) rating of screening mammography rather than change to 'Black' (data suggest harms>benefits), as this review presents a best-guess projection about modern day impact.

As a final note, it is important to be aware that the primary reason for the failure of screening mammography is the ongoing inability to determine which cancer cells in which humans represent a true threat to the host. If this could be determined successfully then any screening test like mammography would be a success. Should technology and science advance to the point where prognosticating based on clinical material (i.e. biopsy cells or mammogram characteristics or individual's biologic characteristics) was overwhelmingly accurate and reliable, early detection programs of all sorts would all become far more likely to result in life-saving benefits.

Author

David Newman, MD

Published/Updated

October 6, 2010

Source

Gillespie LD, Robertson MC, Gillespie WJ, Lamb SE, Gates S, Cumming RG, Rowe BH. Interventions for preventing falls in older people living in the community. Cochrane Database Syst Rev. 2009 Apr 15;(2):CD007146. Review. PubMed PMID: 19370674.

Efficacy Endpoints

Fall prevention

Harm Endpoints

None known

Narrative

As we experience the grey tsunami that is the evolving demographic of our era, the number of fragile elderly who suffer falls and who subsequently experience a downward spiral in their health is also rising. A great deal of research activity has been invested in fall prevention both for identifying high risk individuals and in developing both simple and complex interventions to mitigate this risk.

Benefits: The evidence as assessed in this Cochrane review which did not revise its search since 2003 reports on 62 trials having enrolled nearly 22 000 patients. The most robust evidence on fall prevention pertains not to the simple intervention of strength and balance training reported here but rather multidisciplinary, multifactorial, health and environmental risk factor screening and interventions programmes i.e. a well established fall prevention outreach program. This latter approach demonstrates efficacy in both high risk groups and in terms of a broader outreach program to all community-dwelling elderly. This NNT summary focuses on the single intervention of strength and balance training as it is a more feasible recommendation for healthcare providers to make and have followed. Also of note is that these findings extend to not only simple falls but falls resulting in injury.

Harms: With the exception of a single trial within the systematic review that reported adverse events with a program of brisk walking in women with recent osteoporotic fractures, there are no reported downsides to this intervention aside from cost and the invested effort that it demands of both subjects and their caregivers. 1

Caveats

While the total population sample in the few trials included for the intervention of strength and balance training in high risk community-dwelling elders, a recent "From Evidence to Action" review in JAMA2 conducted a more up to date search which confirmed these benefits and suggested that they are even more robust. Another systematic review looking at a broad range of exercise-based interventions across a wider spectrum of patients noted the same effect size with fair precision around the degree of benefit.3

There are some methodological weaknesses in the component studies of the Cochrane review namely the way in which a fall is defined and potential problems with recall bias influencing the manner in which the outcomes were measured. However, until such time that we are conducting fall prevention trials with closed circuit TV monitoring and fall sensors attached to study subject wrists the benefits of this simple intervention are likely to outweigh the harms.

Author

Eddy Lang, MD

Published/Updated

September 23, 2010

Source

de Lorgeril M, Salen P, Martin, J, et.al. Mediterranean Dietary Pattern in a Randomized Trial. Archives of Internal Medicine 1998, 158: 1181-7.
de Lorgeril M, Renaud S, Mamelle N, et al. Mediterranean alpha-linolenic acid-rich diet in secondary prevention of coronary heart disease. The Lancet; Jun 11, 1994; 343:1454-9.

Efficacy Endpoints

Overall survival, new cancers, heart attack

Harm Endpoints

None

Narrative

Background: The Mediterranean population is known to have a lower incidence of coronary heart disease compared to other groups.1 It is unclear what the reason for this is, speculation is mostly dietary and includes the protective effects of olive oil (linoleic acid) or flavonoids in red wine, the heavy reliance on fruits and vegetables, and the relative paucity of dairy and red meat. The “typical” Mediterranean diet:

• abundant fruits and vegetables
• olive oil as the principal source of fat
• fish and poultry consumed in low to moderate amounts
• dairy consumption is mostly of cheese and yogurt (less butter/cream)
• red meat consumed in low amounts
• red wine in low to moderate amounts

Mediterranean diets are low in saturated and omega-6 fats, but high in omega-3 fats, oleic acids, fiber, antioxidants, vegetable proteins, and B vitamins.

Study Descriptions: In this single blinded secondary prevention trial 605 survivors (10% female) of a first acute myocardial infarction were randomized to the American Heart Association Step 1 diet or to a Mediterranean-type diet for 5 years. Patients randomized to the Mediterranean-type diet were given free quantities of a rapeseed oil-based margarine. Rapeseed oil is similar in composition to olive oil, though it is composed of slightly higher amounts of linolenic acids.

The study was terminated early due to significant benefit found at intermediate analysis at ~4 years. Death occurred in 34/303 (8.0%) in the control group versus 14/302 (4.6%) in the Mediterranean diet group (p=0.03). There were new cancers discovered in 17/302 (5.6%) of the control group versus 7/302 (2.3%) in the Mediterranean diet group (p=0.05). There were 33 (8.2%) non-fatal MI’s in the control group versus 8 (2.6%) in the Mediterranean diet group.

Patients who developed cancer in both groups had much higher rates of smoking. There were no significant differences in the 2 groups in terms of tobacco use, medications used (including anti-lipids), exercise, weight, blood pressure, and psychosocial factors that were addressed in a separate paper.2 The only significant difference found between the control and experimental groups was in nutrient intake. Interestingly, there was no significant difference between the groups in terms of serum cholesterol, triglycerides, or HDL at the end of the study.

Caveats

It is not certain to what degree each diet was followed in each group, though surveys and serum samples of the experimental group point to strong Mediterranean diet adherence. The authors specifically tried to evaluate for a placebo effect from the diet counseling via surveys and concluded there was none. The fact that the patients were 90% male makes it harder to draw conclusions about the possible impact of diet in women.

The magnitude of this study’s results is astonishing. To compare saving a life post-heart attack with this diet (NNT= 30) and with statins (NNT=83) suggests that diet is nearly three times more powerful as a life-saving tool. A few factors make this particularly remarkable. Cancers were also reduced, while some authors have raised concerns about statins increasing cancer risk (without supporting evidence in the industry-sponsored trials to date). Imagine that the control group had been following a typical dietary pattern rather than the AHA recommended diet; the size of the effect could be even greater. Finally, the study suggests that cholesterol, which was not reduced by the Mediterranean diet, may not be as important a dietary consideration for heart disease prevention as currently thought and practiced.

It is difficult to make sweeping statements from a single study, but given the existing data and lack of harms the Mediterranean diet seems beneficial and should be strongly recommended at this time.

Author

Joshua Quaas, MD

Published/Updated

September 17, 2010

Source

Jarvik JG, Hollingworth W, Martin, B, Emerson SS, et.al. Rapid Magnetic Resonance Imaging vs Radiographs for patients with low back pain. JAMA 2003, 289(21): 2810-18.

Efficacy Endpoints

Back-related disability measured by the modified Roland questionnaire, Medical Outcomes Study 36-item Short From Health Survey (SF-36), pain preference scores, satisfaction, cost

Harm Endpoints

None

Narrative

380 patients with low back pain for > 1 year were randomized to receive either lumbar spine radiographs versus a lumbar spine MRI. The patient groups appear to be well randomized. The primary hypothesis was that a validated back pain disability scale would demonstrate better scores in the MRI patients because of accurate diagnoses and better management.

Scores were statistically unchanged at 12 months. Both groups improved (had decreases in scores) that were most substantial in the first 3 months. There were many more diagnoses in the MRI group, however none were consequential (i.e. no diagnosis of metastasis or abscess or spinal cord compression). There were no adverse events in either group. 10 patients in the MRI group versus 4 patients in the radiograph group underwent lumbar spine operations (p = 0.09).

There were no differences between the groups in “pain bothersomeness”, pain frequency, SF-36 score (functional health and well-being), and physical functioning. The MRI group had a mean cost of $2380 versus $2059 for the radiograph strategy, which includes consults, physical therapy, etc. Patients were equally satisfied in both groups; however they and their physicians were slightly more reassured if an MRI was performed.

This randomized controlled trial suggests that patients with low back pain without pathologic features are not helped by having an MRI. This strategy results in increased cost, possibly more procedures, and no change in patient long-term pain, disability, or functional status. The phenomenon of reassurance via diagnostic testing seems apparent in these data.

Caveats

Larger studies (non-randomized) looking at MRI for subacute/chronic back pain have found higher rates of malignancy: one study of primary care patients had ~0.5% incidence, another large study that included hospitalized patients found a rate of 1.6%.

Author

Joshua Quaas, MD

Published/Updated

August 25, 2010

Source

Ram, FS, Rodriguez-Roisin, R, Granados-Navarrete, A, et al Antibiotics for exacerbations of chronic obstructive pulmonary disease. Cochrane Database Syst Rev 2006.

Efficacy Endpoints

Mortality, Treatment Failure (Lack of resolution, worsening, or death)

Harm Endpoints

Diarrhea

Narrative

Chronic obstructive pulmonary disease (COPD), a term that encompasses both patients diagnosed with chronic bronchitis and emphysema, is an obstructive lung disease, in many cases caused by tobacco smoking. It is thought that patients with COPD ‘exacerbation’ (increased shortness of breath or change in their chronic cough and sputum) may benefit from antibiotics, though the reasons for this are not well elucidated.

Benefits: Benefits were robust. 11 randomized trials are included from this review, totaling 817 subjects. The data suggest that overall COPD exacerbations benefit from antibiotics – both by reducing subjects’ short term (1-2 week) mortality and by reducing the chance of treatment failure (not getting better or getting worse). Mortality was reduced by 11.6%, a NNT of 8—a number that held consistently across subgroups. Treatment failure was reduced by 30.7% (NNT of 3) but this seemed most applicable to hospitalized subjects. The best effects on primary outcomes likely apply to the sickest patients: those admitted to the hospital and to the intensive care unit.

Harms: Only two studies collected data on diarrhea (a common side effect of antibiotics): antibiotics increased the risk of developing diarrhea by 5.0%, for a NNH of 20.

Caveats

It isn’t clear if this translates well, or to what degree, for outpatients (patients who don’t require hospitalization). In addition, many of these studies were conducted in the mid- to late-20th century, possibly with a sicker-than-average group. Also, rates of smoking and packs-per-day have since declined. It is possible that different bacteria may be involved today. These studies also had a varying definition of COPD and COPD exacerbation. Finally, some studies used clinical criteria to rule out pneumonia, instead of using x-ray evaluation.

Author

Graham Walker, MD

Published/Updated

August 19, 2010

Source

Ranji SR, et al. Do Opiates Affect the Clinical Evaluation of Patients With Acute Abdominal Pain? JAMA. 2006;296(14): 1764-1774

Efficacy Endpoints

Pain Control

Harm Endpoints

Missed or mistaken diagnoses

Narrative

This review examined a question that has stirred controversy for decades. Based on a concern that treatment with opiates may obscure the signs and symptoms that lead to accurate diagnosis of abdominal pain, opiates were often withheld as initial treatment.

There are now 12 randomized trials of varying quality (9 adult and 3 pediatric), aggregated into one systematic review. Based on both the findings of the individual trials and also the pooled results of these trials, it is safe to say that there is no evidence of an increase in misdiagnosis or worse outcomes following the administration of opiates for pain when compared to the administration of placebo or nothing for pain.

Caveats

In the aggregated data, particularly in studies where pain was demonstrably less in the treatment group, the administration of opiates did alter physical examination findings. There were, however, no increases in patient harms, diagnostic errors, or management errors.

In the early 20th century, when diagnostic error caused by opiates was posed as a concern in a surgical textbook by Dr. Zachary Cope, morphine tartrate often came in either 30mg (1/2 grain) or 60mg (1 grain) doses. These doses may well have been adequate to obscure any diagnostic yield from abdominal examination. In the studies covered in this review the doses are smaller, typically 10mg of morphine, its equivalent, or less.

In addition, while the studies on this topic were not designed to precisely measure the degree of pain relief that the medications afforded, there was a detectable decrease in pain in the treatment arm in 8 of the 12 studies. Although we cannot list the NNT’s for this benefit (see methods for the distinction between continuous and dichotomous variables and the NNT), and the benefit depends on adequate doses it is apparent that there was a considerable benefit in pain relief for those in the opiate groups—without any increase in patient harms.

Author

David Newman, MD

Published/Updated

Source

Olasveengen, TM, Sunde K, Brunborg C, Thowsen J, et al. Intravenous Drug Administration During Out-of-Hospital Cardiac Arrest: A Randomized Trial. JAMA. 2009; 302(20):2222-2229.
IG Stiell, GA Wells, B Field, Advanced cardiac life support in out-of-hospital cardiac arrest..N Engl J Med 2004;351:647-656

Efficacy Endpoints

Mortality

Harm Endpoints

Mortality

Narrative

Hundreds of thousands of individuals suffer sudden cardiac arrest (SCA) each year in the United States and abroad. Advanced Cardiac Life Support (ACLS) is an algorithm-based set of recommendations and instructions assembled by the American Heart Association for the management of this condition. While rapid defibrillation appears highly effective for SCA, the role of intravenous agents is unproven.

This review examined the utility of the ACLS algorithms in the only two high quality trials to test them. The first used historical controls during the ‘phase-in’ of an ACLS system to an existing rapid-defibrillation-and-CPR-only system. Despite the pitfalls and biases inherent to this design, which tend to favor a treatment effect, ACLS failed to produce a change in survival though it did lead to more intensive care unit admissions.

The second study randomized 851 nontraumatic cardiac arrest patients to receive treatment with or without intravenous access using ACLS algorithms. There was no difference in mortality, the primary outcome. Some have argued that the 1.3% nonsignificant difference may represent a potential benefit, which assumes that a trial of 851 subjects was inadequately powered. This remains unproved and, if correct, suggests a very small hypothetical effect.

The harms and costs of ACLS are not typically visible in such studies. It is important to note that first, because patients are pulseless (dead) at the moment of enrollment, any intervention that does not produce an improvement should be seen as an opportunity cost if other approaches are available. For instance non-ACLS based approaches including those based on physician judgment rather than undifferentiated algorithms, are commonly eschewed in favor of ACLS, despite the apparently established futility of ACLS. Second, ACLS tends to increase intensive care unit admissions and generate substantial resource utilization (both during resuscitation and in the implementation of post-resuscitation care), in the absence of a detectable benefit. In the two studies above there was an additional 5 to 10% absolute increase in intensive care admissions in the group receiving intravenous, ACLS-based medications.

Caveats

*One of the above studies uses a before/after historical controls design in a Canadian milieu, and the second is in a single Norwegian emergency medical services setting. Despite their being the best available evidence, the data are limited and the external validity of these data is therefore unclear.

While all of these trials refer to out-of-hospital cardiac arrest, there have not been any in-hopsital cardiac arrest trials showing benefit, and there have been a handful of observational studies suggesting possible harm from ACLS medications.

Author

David Newman, MD

Published/Updated

Source

Perez MI, Musini VM, Wright JM. Effect of early treatment with anti-hypertensive drugs on short and long-term mortality in patients with an acute cardiovascular event. Cochrane Database Syst Rev. 2009;(4):CD006743.
Al-Reesi A, MD; Al-Zadjali N, MD; Perry J, Fergusson D, et al. Do β-blockers reduce short-term mortality following acute myocardial infarction? A systematic review and meta-analysis. CJEM. 2008;10(3):215-23
Sinert R, NewmanDH, Paladino L, Brandler E. Immediate Beta-blockade in Patients with Myocardial Infarctions: Is There Evidence of Benefit? An Evidence-Based Review. Annals of Emergency Medicine. 2010; online,doi:10.1016/j.annemergmed.2010.03.036

Efficacy Endpoints

Mortality

Harm Endpoints

Mortality, Congestive Heart Failure

Narrative

This review examined the effects of beta-blocking agents during major heart attacks ('STEMI'). There are 27 trials and at least 3 systematic reviews addressing this issue. A Cochrane review and a CJEM review from 2008 ask this question: do trials randomizing patients to receive immediate beta-blockers or a control (typically placebo) in the initial phases of a heart attack demonstrate improved outcomes? Both reviews find that they do not. One review (Al-Reesi et al.) suggests a potential benefit in one subgroup (patients with little or no risk of heart failure). However this subgroup analysis is of unclear importance given the clinical heterogeneity and varying quality of the included trials.

One trial has driven a longstanding perception of immediate beta-blockers as beneficial. This is ISIS-1, a large, unblinded trial that found a 0.7% reduction in mortality in the beta-blocker group.1 There are 26 trials prior to ISIS-1, however, that suggest no aggregate mortality benefit to beta-blockade. The lack of blinding in ISIS-1 may therefore explain the 0.7% difference, and a more recent, and larger, trial that utilized double-blind techniques (COMMIT, 2004) found no benefit.2

Harm due to beta-blockers is largely in the form of cardiogenic shock. Because this harm appears to occur in the first day of treatment, while the modest benefits of beta-blockers (0.5% less re-infarction and 0.5% less ventricular arrhythmia)2 occur days later, it seems that immediate treatment is, in aggregate, harmful, while delayed treatment may be modestly beneficial.

Caveats

The third review (Sinert et al), while reaching the same conclusion as the others, asks a more nuanced question: do beta-blockers during STEMI result in improved outcomes? Other reviews have examined studies that randomize to two groups: Early beta-blockade followed by continuing beta-blockade after discharge, versus no early AND no continuing beta-blockade. Because beta-blockers are shown to reduce mortality when given after heart attacks, these reviews have conflated immediate treatment with later treatment. Only one trial (identified by the third review) separated these two elements. However, all other trials are biased in favor of immediate beta-blockers by attributing benefits of delayed treatment to immediate treatment. The lack of a benefit even in these biased analyses suggests strongly that the one large study that found no benefit to immediate treatment was indeed correct.

Virtually all of the high quality studies of this topic have utilized intravenous beta-blocker agents, which may increase cardiogenic shock episodes more than orally ingested versions. These shock episodes may offset potential benefits. Some have argued therefore that oral administration would cause less cardiogenic shock and allow the benefits of the drug to be realized. Though plausible this remains hypothetical and untested, while the overall failure of beta-blockers in the one format studied remains apparent and robust.

Note: We have chosen to list beta-blockers as "No benefits" rather than "Caution" based on the apparently offsetting nature of some small benefits (decreased ventricular arrhythmias and decreased repeat heart attacks) with the established harms. We believe that it is likely that beta-blockers, in aggregate offer no benefit, but we cannot state with confidence that harms overwhelm benefits.

Author

David Newman, MD

Published/Updated

Source

Bucher et al. Adjunctive corticosteroids for Pneumocystis jiroveci pneumonia in patients with HIV-infection. Cochrane Database Syst Rev. 2006;(3):CD006150.
Bozzette et al. A controlled trial of early adjunctive treatment with corticosteroids for pneumocystis carinii pneumonia in the acquired immunodeficiency syndrome. NEJM. 1990; 323(21):1451-57.
Gallant e tal. The effect of adjunctive corticosteroids for the treatment of pneumocystis carinii pneumonia on mortality and subsequent complications. Chest. 1998;114(5):1259-63.

Efficacy Endpoints

Mortality

Harm Endpoints

Increase in AIDS-related Complications

Narrative

Pneumocystits jiroveci pneumonia (PCP), formerly know as Pneumocystis carinii pneumonia, is the most common opportunistic infection among people with human immunodeficiency virus (HIV), and mortality remains high at 10 to 20% during the initial infection.

This review examined the role of corticosteroids among HIV patients with PCP and hypoxemia defined as arterial oxygen partial pressure <70mmHg or an alveolar-arterial gradient >35 mmHg on room air. The authors included six trials with 489 patients that randomized subjects to receive either corticosteroids or placebo as an addition to antimicrobials. The review reported outcomes for undeveloped nations with minimal access to antiretroviral therapy and for developed nations with access. Control group mortality rates were 25% and 10% in these two groups, respectively.

Note that despite the greater number of patients that will experience a harmful effect of steroids, the beneficial effect is of greater impact to most patients. Based on this value judgment we classified the therapy as beneficial overall (a green light).

Most of the studies gave intravenous methylprednisolone with a starting dose of either 40mg or 60mg every 6 hours, then tapered over 18 to 24 days. Two studies gave a burst of methylprednisolone for 10 days with no taper at either 40mg every 12 hours or based on weight (2mg/kg every 6 hours). Two studies gave oral prednisone at either 80mg per day for five days or 60mg per day for 7 days and then tapered to complete a course of 21 days.

Caveats

The pooled study population was small with only 489 subjects, and only three of the six studies were completed and fully published. One study was published only in abstract form and the other two were stopped early because of benefit. Overall, there was moderate heterogeneity in the pooled results for mortality at 1 month (P=0.12; I2=43%), but the results became homogeneous at 3-4 months (P=0.46; I2=0%). In terms of harm, the data is scant. Only two studies reported harms.

Author

Jarone Lee, MD, MPH

Published/Updated

Source

Valenzuela TD Roe DJ, Nichol G, Clark LL, et al. Outcomes of rapid defibrillation by security officers after cardiac arrest in casinos. NEJM. 2000; 343(17): 1206-9.

Efficacy Endpoints

Mortality

Harm Endpoints

Mortality

Narrative

Hundreds of thousands of individuals suffer sudden cardiac arrest (SCA) each year in the United States and abroad. Rapid defibrillation has been thought highly effective for SCA, but data have been sparse. This review examined the utility of rapid defibrillation under the best of possible circumstances: in a public location, with otherwise relatively healthy patients who suffer witnessed, sudden cardiac arrest, and in whom defibrillation is typically available within 3-4 minutes.

Under these unusual circumstances 90 of 148 (61%) arrests were associated with an initial cardiac rhythm of ventricular fibrillation when the defibrillator was placed on the patient. Survival to hospital discharge was 38% overall and 59% among those patients found to be in ventricular fibrillation.

Caveats

We are including this study in our NNT review database because it represents a remarkable biologic description of SCA and is likely to remain the highest form of research available on this topic. There will never be a randomized trial of rapid defibrillation versus standard care without rapid defibrillation, nor should there be. However, this type of data makes bold presumptions. It presumes that all of those untreated would die. While this is likely true, it is less clear whether or not CPR without electrical countershock would provide some degree of life saving benefit. It seems probable that it would, though it is not possible to quantify this. Furthermore, the maximized conditions (witnessed arrests, healthy subjects, immediate response) do not represent a typical circumstance for the current use of defibrillation or general resuscitation in virtually any setting.

Finally, there may be harms associated with attempts to perform resuscitation. Intensive care unit admissions, comatose or vegetative state survivors, and substantial resource utilization (both during resuscitation and in post-resuscitation care) are all important potential harms that should be considered both for society at large and for those individuals who survive bur do not regain neurologic function, and for their families.

Author

David Newman, MD

Published/Updated

Source

Magee K et al. Heparin versus placebo for acute coronary syndromes. Cochrane Database Syst Rev. 2008 Apr 16;(2):CD003462.

Efficacy Endpoints

Mortality, Primary or Secondary myocardial infarction

Harm Endpoints

Major bleeding, Minor bleeding

Narrative

'Acute coronary syndrome' in this review includes cardiac conditions believed to be caused by arteries partially or completely occluded due to a combination of longstanding plaques in artery walls and new tears in the innermost lining of the artery (the tears are caused by plaques poking through). The blood typically misinterprets these tears as bleeding and forms a clot where the tears are, clogging the artery. Heparin inhibits clotting which seems intuitively like it would help but by the time the patient has symptoms like chest pain a clot has already formed, so it's unclear how much this can help. In addition, anything that inhibits clotting can lead to bleeding.

Eight randomized control trials of >3100 patients are included. Heparin did not reduce deaths or non-fatal heart attacks. There was no difference in major bleeding in the included studies, though minor bleeding events increased by 7.5%, or 1 in every 17 subjects.

Caveats

Heparin did not measurably prevent deaths but because so few people died in either group the review did not have the ability to detect very small differences. Notably, the Cochrane review concludes a 3% absolute benefit for reducing nonfatal heart attacks in the first week after the initial acute coronary syndrome, however this benefit is gone at 30 days of follow-up and at 3-6 months. After the first week, any reduction in heart attacks is neutralized as subjects in the heparin groups rapidly 'catch up' by suffering more heart attacks than those in the groups receiving placebo.

The Cochrane review reports early studies of heparin versus placebo for ACS. Later studies are more diligent and transparent in their reporting of major bleeding events. These later studies appear to represent an experience that is much closer to the community experience and suggest a roughly 4-5% major bleeding rate (above and beyond aspirin) with any heparin.1 Therefore it is likely that the true NNH for heparins in ACS is roughly 20 to 25 for major bleeding events. We added this as a quantifiable harm because we find the proposition that heparin has no effect on major bleeding events to be implausible, and the later study results are the highest quality that we could identify. Actual bleeding event rates may be higher, though they seem unlikely to be lower.

Author

Koustav Mukherjee, MD, Ashley Shreves, MD and David Newman, MD

Published/Updated

January 19, 2010

Source

Indications for fibrinolytic therapy in suspected acute myocardial infarction: collaborative overview of early mortality and major morbidity results from all randomized trials of more than 1000 patients. Lancet 1994; 343:311-22.

Efficacy Endpoints

Death at 1 month following acute heart attack (STEMI)

Harm Endpoints

Major bleeding (brain bleeding or bleeding requiring transfusion).

Narrative

This systematic review includes 9 trials and 58600 patients randomized to receive a fibrinolytic drug or placebo for suspected heart attack. Patients were enrolled based on strong suspicion of heart attack by the treating doctor. Most (76%) were men, and most had ST-elevations on their EKG (68%). There was an overall mortality benefit of 1.9% (9.6 vs. 11.5) in favor of fibrinolytics. There was also a 0.4% increase in hemorrhagic stroke (1.2 vs. 0.8). Benefit was demonstrably greater with earlier treatment, with the most benefit apparent for treatment given within a few hours of symptom onset. Benefits were smaller and less statistically robust in the 12 to 24 hour period. Patients with ST-depressions were harmed rather than helped.

Caveats

There was no gold standard to prove STEMI (e.g. catheterization or biomarkers) and some patients had normal EKG’s (5%). There is heterogeneity between trials, including inconsistency in the use of aspirin. However, groups appear to have been randomized well and treated equally with respect to other interventions.

Author

Joshua Quaas, MD

Published/Updated

Source

Rowe BH et al. Early emergency department treatment of acute asthma with systemic corticosteroids. Cochrane Database Syst Rev. 2001;(1):CD002178.

Efficacy Endpoints

Hospital admission, asthma relapse, subsequent hospitalization for asthma

Harm Endpoints

Steroid adverse drug effects

Narrative

Admissions: In the included 12 studies (N = 863), steroids were always given within an hour. The impact seen in these trials might be surprising to many given the conventional wisdom that effects of steroids are delayed. The effects were consistent across the pediatric and adult populations. Of note, all steroids were administered intravenously or intramuscularly in the adults but given orally in pediatrics.

Relapse Rates/Subsequent Hospitalizations: There were 6 studies (N = 374), both pediatric and adults. Steroids were administered intramuscularly in one, but given orally in the remaining five, for a duration of 3-10 days. The baseline relapse rates for asthmatics has been noted to be 12-16% at 2 weeks. In this review, most studies followed patients up for a maximum of 7-10 days; however, one extended the period to 21 days.

Caveats

Admissions: The average admission rate for asthmatics is 10-20%; however, in the included studies, the admission rate was generally greater than 40%, indicating a relatively sick asthmatic population. In addition, the authors noted significant heterogeneity between the studies and performed a pre-planned subgroup analysis to assess the impact of steroids based on severity of disease. As expected, the impact of steroids was larger (NNT = 5) in the more severe asthmatic populations. For harm, the only aggregated data we were able to find appears to suggest that there may be a small, clinically difficult-to-detect negative impact of systemic steroids on growth in children due to intermittent steroids in asthma, while there was no detectable impact of inhaled corticosteroids. These effects were from small, heterogeneous studies and should be considered preliminary.1

Relapse Rates/Subsequent Hospitalizations: The actual steroid, dosage and duration of course varied between studies so judgment still informs these areas of management. For harm, the only aggregated data we were able to find appears to suggest that there may be a small, clinically difficult-to-detect negative impact of systemic steroids on growth in children due to intermittent steroids in asthma, while there was no detectable impact of inhaled corticosteroids. These effects were from small, heterogeneous studies and should be considered preliminary.1

Author

Ashley Shreves, MD

Published/Updated

January 10, 2010

Source

Camargo CA et al. Continuous versus intermittent beta-agonists for acute asthma. Cochrane Database Syst Rev. 2003;(4):CD001115.

Efficacy Endpoints

Hospital Admission

Harm Endpoints

Tachycardia, tremor

Narrative

There were 8 studies (N=461) included, 3 of patients with mild-moderate attacks and 5 of patients with severe attacks. There was 1 pediatric study, 1 pediatric-adult and the remaining 6 were adult only. There was variability between the studies in terms of dosing, duration of therapy, and co-interventions. Generally in the continuous groups, 10-30 mg of albuterol was administered in normal saline over a period of 2-4 hours, with 1 study group receiving only 5mg and a separate study lasting 6 hours. The control groups received 2.5-5.0 mg of albuterol in nebulized form over the same period of time compared to the respective experimental groups. Specifically, 3 groups received q20 minute nebulizers, 3 q30 minute nebulizers and 2 q1 hour nebulizers.

Caveats

The patients from the “severe” studies (baseline admission rate 40%) contributed 91% of the weight of this review, so it is not clear whether these results should be generalized to all asthma patients presenting with acute exacerbations. Also, a significant potential benefit to continuous nebulizers is decreased patient time in the ED and utilization of staff resources. Unfortunately, most studies did not report this outcome so it is not clear whether there is an important impact in this regard.

Author

Ashley Shreves, MD

Published/Updated

Source

ISIS­2: Randomized trial of intravenous streptokinase, oral aspirin, both, or neither among 17187 cases of suspected acute myocardial infarction. Lancet. 1988 Aug 13;2(8607):349-60.

Efficacy Endpoints

Death at 1 month following major heart attack (STEMI)

Harm Endpoints

Major bleeding (brain bleeding or bleeding requiring transfusion) and minor bleeding.

Narrative

The lone high quality trial addressing this question was multi-institutional across the U.S and Western Europe and was a randomized controlled trial of over 17000 subjects. Subjects were believed to be suffering acute MI by treating physicians and were randomized to aspirin, a thrombolytic (‘clot-busting’ medication), both, or neither. The groups were similar. There was increased death in the first month among those assigned to placebo 1016/8600 (11.8%) versus those assigned to aspirin 811/8587 (9.4%) with no significant increase in transfusion or intracerebral hemorrhage. There was a 0.6% increase in minor bleeding.

Aspirin provides among the best mortality benefits of any intervention for MI, with minimal downside. The anti-platelet mechanism seems sensible and the data is strong. Aspirin should be given to all cases of confirmed or suspected STEMI unless a significant contraindication exists.

Caveats

We don’t know for certain that the patients entered into this trial were having heart attacks, as the entry criterion was an EKG showing ST elevation and MI was not confirmed with any other testing (i.e. cardiac biomarkers). There has been no attempt to confirm these results (there probably never were).

Author

Joshua Quaas, MD

Published/Updated

November 28, 2009

Source

Cundiff DK et al. Anticoagulants vs non-steroidal anti-inflammatories or placebo for treatment of venous thromboembolism. Cochrane Database Syst Rev. 2006 Jan 25;(1):CD003746.

Efficacy Endpoints

Death, Recurrent PE, Recurrent/Extension of DVT

Harm Endpoints

Major hemorrhage (intracerebral hemorrhage, GI bleed). Risk varies depending on population. In one Cochrane review of trial data major bleeding occurred in 0.9% to 2% of subjects, a NNH of 50 to 111 (Hutten BA), and there were no major hemorrhages in the Nielsen study of 90 ambulatory outpatients. However subjects eligible for randomized trials tend to be at lower risk for bleeding than typical coumadin patients in the community. In one population-based cohort of coumadin patients 2% experienced a fatal bleeding event (NNH=50) and the overall major bleeding rate was 12% (NNH=8). (Landefeld CS et al. Am J Med 1989). It is estimated that 15000-52000 patients die annually in the U.S. of brain bleeding caused by coumadin.

Narrative

There has been one adequately designed study addressing whether anticoagulation improves important clinical outcomes in patients with VTE (Nielson et al., 1994). Subjects were enrolled if they were ambulatory with symptomatic DVT. Half of the 90 subjects also had silent pulmonary emboli. Subjects were randomized to anticoagulation or NSAIDs and there was 1 fatal PE during the study period (anticoagulation group).

In 1960, Barritt and Jordan performed a randomized trial evaluating anticoagulation for presumed PE. While the study is often cited as evidence of the benefit of anticoagulation, methodologic weaknesses render it unreliable (the Cochrane group excludes the study). Patients with suspected PE were included and no confirmatory testing was performed; randomization was not described; no blinding and no placebo were used; all subjects were inpatients for other reasons and all were critically ill at enrollment; the study was stopped inappropriately early. The results of this study appear to have led the medical community to erroneously conclude that anticoagulation is proven beneficial for VTE.

Caveats

While the Nielson study does not prove that anticoagulation is unhelpful for VTE it is our best and only methodologically reliable source. Large, properly designed trials are needed.

Author

Ashley Shreves, MD

Published/Updated

Source

Bosch X et al. Platelet glycoprotein IIb/IIIa blockers for percutaneous coronary revascularization, and unstable angina and non-ST-segment elevation myocardial infarction. Cochrane Database Syst Rev. 2001;(4):CD002130.

Efficacy Endpoints

Death at 6 months, heart attack at 6 months

Harm Endpoints

Major bleeding events (brain bleeding or bleeding requiring transfusion)

Narrative

There are four major trials that address this question, including just over 3000 subjects. Studies were of relatively high quality and the use of glycoprotein inhibitors in this situation represents a best-case-scenario: The subjects are low risk for bleeding, high risk for death or recurrent heart attack, and are undergoing invasive management. If the mechanism of action and conventional wisdom are to be believed glycoprotein inhibitors should be most helpful for high risk patients undergoing invasive procedures. The lack of benefit in this population, and an increase in harmful events, is remarkable.

Caveats

There are only four trials, all industry sponsored. While neither death nor MI was affected, some may note that 'revascularizations' (repeat angioplasties) were reduced. However, at TheNNT.com we see this as neither a patient-oriented nor a clinically important endpoint. In these studies 'revascularization' was separate from 'MI' and death, which were both unaffected. This means that the condition that led to the repeat angioplasty did not result in death was not an MI, suggesting that the ‘need’ for the procedure was highly questionable. This is particularly true when one considers that only STEMI patients have been shown to benefit from angioplasty (see COURAGE trial and related data).

Author

David Newman

Published/Updated

October 23, 2009