Early Decrease in Blood Pressure after Acute Intracerebral Hemorrhage

Aggressive blood pressure lowering did not improve patient outcomes compared to less aggressive management after spontaneous intracerebral hemorrhage

Benefits in NNT

No one was helped (no improvement in functional recovery) No one was helped (no death prevented)
No one was helped

Harms in NNT

No one was harmed
No one was harmed
View As:


Moullaali TJ, Wang X, Sandset EC, et al. Blood pressure in acute stroke (BASC) investigators. Early lowering of blood pressure after acute intracerebral haemorrhage: a systematic review and meta-analysis of individual patient data. J Neurol Neurosurg Psychiatry. 2022;93(1):6-13.

Study Population: 16 studies with individual participant data from 6221 patients receiving a blood pressure lowering strategy within 7 days of acute spontaneous intracerebral hemorrhage

Efficacy Endpoints

Functional status based on distribution of modified Rankin scale scores, death or dependency, death or severe dependency, death

Harm Endpoints

Early deterioration, symptomatic hypotension, and other serious adverse events


Acute spontaneous intracerebral hemorrhage (ICH) is associated with poor clinical outcomes, particularly in the setting of elevated blood pressure.1, 2, 3 Some trial data have suggested early aggressive lowering of blood pressure (BP) may improve patient outcomes,4 theoretically by reducing hematoma growth. However, rigorous randomized controlled trials (RCTs) have in many cases failed to show a benefit with a target systolic BP <140 mm Hg compared with a less stringent target.5, 6 While American Heart Association (AHA) guidelines suggest it is safe to decrease systolic BP to 140 mm Hg during ICH,7 there is uncertainty about the effects of BP lowering in this setting.8, 9, 10

The systematic review summarized here included individual participant data (IPD) from trials evaluating BP management after acute ICH.11 In an IPD, the data points from each enrolled subject are collected from the study investigators and are used for performing the meta-analysis. This is different from a trial level data meta-analysis where only the final findings of trials are included in the analysis.

The authors of the review included trials of patients >18 years with acute primary spontaneous ICH <7 days from onset, and which randomized patients to more intensive or less intensive blood pressure lowering. In some cases, this meant drug versus placebo, while in others it meant lower BP targets, such as a systolic BP of 140 mm Hg versus 180 mm Hg. Primary outcomes included functional status defined by the ordinal distribution of modified Rankin Scale (mRS) scores ranging from 0 (no symptoms) to 6 (death) at the end of follow-up. Secondary outcomes included death or dependency (mRS, 3–6), death or severe dependency (mRS, 4–6), and death. Safety outcomes included early neurological deterioration, symptomatic hypotension, and other serious adverse events, with all outcomes defined by their respective trial.

The authors contacted investigators of 50 studies and obtained IPD from 16 (n = 6221 patients). Mean patient age from available studies was 64 years, and 36% of patients were female. Patients had a median National Institutes of Health Stroke Scale score of 11 on presentation. Mean systolic and diastolic BP at randomization were 177 and 100 mm Hg, respectively. Median hematoma volume was 11 ml. Median time to randomization from symptom onset was 4 h. One trial (n = 274) evaluated renin-angiotensin system blockers, three evaluated α- and β-blockers (n = 3310), four evaluated calcium channel blockers (n = 1319), five evaluated nitrates (n = 789), and one evaluated magnesium (n = 387), while the remainder evaluated multi-drug regimens.

Blood pressure was significantly lower in patients receiving more intensive management: mean difference in the systolic BP 8 mm Hg (95% CI: 6–9) at 1 h, 12 mm Hg (95% CI: 11–13) at 1–24 h, and 7 mm Hg (95% CI: 6–8) at 2–7 days. However, intensive interventions were not associated with improved recovery or death. Intensive BP-lowering interventions also had equivocal effects on hematoma growth, with a nonsignificant mean difference of roughly 1 ml at 24 h, while no effect was seen on rates of neurologic deterioration, severe hypotension, and cardiac or renal events.


The most obvious weakness in these data is that they represent only half (54%) of patients randomized into the 50 relevant studies identified by the authors. Additionally, only 36% of patients were female, which may reflect underrepresentation. It is unclear how these issues might have introduced bias and impacted the results. However, just as publication bias tends to favor intervention groups, missing data (particularly those based on subjective decisions by investigators about whether to share data) are likely to have a similar effect.

In this sense, it is reassuring that two other systematic reviews arrived at the same conclusion.8, 12

The systematic review did not report baseline mRS scores. Therefore, it is not clear how baseline imbalances between patients were addressed. The degree of clinical heterogeneity in the 16 included studies is a major limitation, raising concerns about combining data for a meta-analysis. The clinical heterogeneity could be attributed to various settings; differences in subject inclusion/exclusion criteria; and variations in agents, classes, and routes of administration for BP lowering. The meta-analysis also included multiple lower-quality RCTs but did not conduct a subgroup analysis of high-quality studies only.

The latest AHA ICH guideline takes a favorable stand on BP lowering, concluding: “surviving patients show modestly better functional recovery, with a favorable trend seen toward a reduction in the conventional clinical endpoint of death and major disability.”7 This recommendation is primarily based on INTERACT-2, but additional trials including ATACH-2 have failed to find evidence of benefit with aggressive BP lowering.5, 6 Given the lack of statistical benefit in the data we reviewed (odds ratios at roughly 1.0 indicating equal odds of outcomes in control and study groups), the current data do not support a “favorable trend.” Moreover, bias from studies cited in the guideline7 (e.g., selective data inclusion, unblinded observational designs) favoring BP lowering, suggests that these results may actually overestimate potential benefits and underestimate harms.8, 10, 11

In summary, the existing evidence currently offers no support for aggressive attempts to lower BP in acute ICH patients. ICH remains a heterogenous condition, and the absolute difference in BP achieved in this meta-analysis was modest. Because of the heterogeneity of this evidence and its limited ability to represent published studies, we have assigned a color recommendation of yellow (unclear if benefit) for active/intensive BP-lowering interventions in the setting of acute ICH. Further high-quality data are needed.

The original manuscript was published in Academic Emergency Medicine as part of the partnership between TheNNT.com and AEM.


Brit Long, MD; Michael Gottlieb, MD
Supervising Editors: Shahriar Zehtabchi, MD


March 15, 2022