Intracerebral hemorrhage (ICH) has a poor prognosis. Almost half of patients are dead by one month and many survivors are dependent on others for activities of daily living. Many patients have haematoma expansion and deteriorate in the first hours after symptom onset. Higher blood pressure (BP) and haematoma growth are both predictors of adverse outcome. The American Stroke Association guidelines for the management of spontaneous ICH have few Class 1 treatment options . These include surgical removal of cerebellar hemorrhage in patients who are neurologically deteriorating or who have brainstem compression or hydrocephalus, and correction of coagulation factor and platelet deficiencies. These therapies are not relevant to most patients presenting with ICH. Intermittent pneumatic compression in addition to elastic stockings to prevent venous thromboembolism, initial intensive monitoring, maintenance of normoglycemia, and treatment of clinical and electrographic seizures, are also recommended. These recommendations are also important but don't directly address the issue of damage caused by the haematoma.
The first Intensive Blood Pressure Reduction in Acute Intracerebral Hemorrhage Trial (INTERACT1) study gave cause for hope when early intensive BP lowering seemed to attenuate haematoma growth when compared with a more conservative guideline based policy . Clinicians were therefore waiting with anticipation for the results of INTERACT2, in which 2839 patients with spontaneous ICH and a systolic BP between 150 and 220 mmHg were randomly assigned to receive intensive anti-hypertensive therapy with a systolic target of <140 mmHg within one hour, or a standard guideline recommended treatment of <180 mmHg . Investigators were free to choose oral and intravenous BP lowering agents according to local availability.
INTERACT2 failed to show a significant reduction in the rate of the primary outcome of death or major disability, defined as a modified Rankin scale (mRS) score of 3–6, with early intensive BP lowering at 90 days (odds ratio with intensive therapy of 0·87: 95% confidence interval 0·75–1·01; P = 0·06) . However, in the key secondary endpoint of an ordinal analysis of the distribution of mRS scores, there was a significant favorable shift in those patients with aggressive therapy. There were also more patients who were normal or near normal (mRS of 0–1) at 90 days. The effects of early intensive BP lowering on the primary outcome were also consistent across all pre-specified subgroups. Participants receiving aggressive therapy reported better overall health-related quality of life. Most importantly, there were no differences in the rate of death or numbers of serious adverse events between the two groups.
If we assume that the positive endpoints of the trial reflect a true treatment benefit, rather than chance effects or confounding from the lack of physician and patient blinding, then INTERACT2 raises several intriguing questions. Most haematoma growth occurs early, and patients were randomized at around 4 h. Would earlier treatment have proven more beneficial? Only a third of the aggressive treatment group achieved thetarget systolic BP of 140mmHg within the first hour. Would a positive result have been obtained if more patients achieved this target? Did the frequent use of additional therapies in the trial such as mannitol or haemostatic agents confound the results? Last, and most perplexingly, what is the mechanism of benefit? The trial was based on the premise that early intensive BP lowering improves outcome by reducing haematoma expansion . An alternative explanation is not yet immediately obvious, although reduced intracranial pressure, cerebral oedema or extravasation of harmful plasma components, and amelioration of the harmful effects of systemic hypertension, are all possible mechanisms .
How are clinicians to interpret INTERACT2 and should the results lead to a change in practice? We can be confident that a strategy of early and aggressive BP lowering is safe in a wide range of clinical settings and is probably effective. This study gives us the best evidence to date regarding the optimal BP target in ICH. Stroke units could therefore consider adapting or developing protocols, based on locally available oral and intravenous agents, to lower BP to a target of <140 mmHg in spontaneous ICH patients. Treatment should be given after discussion of the limitations of the INTERACT2 study with patients and their families. However, clinicians should not disregard the fact that the magnitude of benefit is modest at best with a number needed to treat of 29 to prevent one poor outcome. For this reason, a policy of routinely transferring patients to an intensive care unit to intensively lower BP cannot be recommended without further evidence. And we should wait with anticipation for the results of the Antihypertensive Treatment of Acute Cerebral Haemorrhage (ATACH) II trial, in which ICH patients will be randomized to intensive BP lowering with nicardipine alone, compared with standard care, using similar BP targets to INTERACT2.