In last week’s journal club, we turned our gaze back to neuro critical care, with a paper that revisited the question of blood pressure control in acute intracerebral haemorrhage. It’s a scenario that is familiar to many of us; the systolic blood pressures well into the 200s in a patient with a new bleed in their head. Qureshi and colleagues have tried to revisit the question of what is the best way to respond to this through the ATACH2 trial (Antihypertensive Treatment in Acute Cerebral Haemorrhage II). The paper is available to read through the link below:
The ATACH2 Trial
The ATACH2 Trial
What's it about?
Intracerebral haemorrhage is commonly met with a hypertensive response, but there is some suggestion that this elevated blood pressure may drive expansion of the haematoma if left unchecked. The authors particularly note the INTERACT2 study which suggested an improvement in outcomes when the blood pressure was quickly and aggressively controlled, though this didn’t reach statistical significance. As such, they have put together this study to try and finally answer this question.
What did they do?
The study was a multicentre, open label, randomised control trial looking at patients who displayed a hypertensive response following an intracerebral haemorrhage. Patients were recruitable if they were over 18, presented with a GCS higher than 5, a haematoma size less than 60cm3, at least one systolic blood pressure reading over 180mmHg, and were randomised within 4.5 hours (initially 3, put extended following their pilot data). Patients were all treated with antihypertensive agents to a target of below 180mmHg, but with the ‘standard treatment’ group target being 140-179 mmHg, and the ‘intensive treatment’ being 110-139 mmHg for 24 hours. They were randomised to which group target they were aiming for, but the same drug titration protocol was used for each group.
Their primary outcome was the level of death or disability at 3 months, as defined by the modified Rankin scale. Secondary outcomes included the degree of haematoma expansion at 24 hours, quality of life (as assesses by the EQ-5D questionnaire) and the incidence of complications.
Their primary outcome was the level of death or disability at 3 months, as defined by the modified Rankin scale. Secondary outcomes included the degree of haematoma expansion at 24 hours, quality of life (as assesses by the EQ-5D questionnaire) and the incidence of complications.
What did they find?
To cut a long story short – no difference. They recruited a total of 1000 patients from 110 centres around the globe (ranging from USA to South Korea). There was pretty good differentiation between the groups in terms of the blood pressures achieved, with a mean of 128.9 +/- 16 mmHg compared to 141.1 +/- 14.8 mmHg.
However, analysis of the outcomes showed no difference in the primary outcome. The rate of death or disability was 38.7% in the intensive treatment arm compared with 37.7% in the standard treatment arm. Even with some statistical manipulation to ‘adjust’ for potential influencing factors (age, GCS, intraventricular extension) there wasn’t even a hint at any sort of significant difference in 3 month outcome.
However, two factors of note were different between the groups: the intensive treatment arm showed a tendency for reduced haematoma expansion (relative risk 0.78, CI 0.58-1.03), and also had a higher rate of adverse incidents over the 3 months (25.6 vs 20.0%).
However, analysis of the outcomes showed no difference in the primary outcome. The rate of death or disability was 38.7% in the intensive treatment arm compared with 37.7% in the standard treatment arm. Even with some statistical manipulation to ‘adjust’ for potential influencing factors (age, GCS, intraventricular extension) there wasn’t even a hint at any sort of significant difference in 3 month outcome.
However, two factors of note were different between the groups: the intensive treatment arm showed a tendency for reduced haematoma expansion (relative risk 0.78, CI 0.58-1.03), and also had a higher rate of adverse incidents over the 3 months (25.6 vs 20.0%).
Is it any good?
There are a number of interesting points that can be taken from this paper, with several good points to highlight first. This was a large study conducted across the globe with a fairly pragmatic approach. Their endpoint was clearly the specific systolic blood pressure of the patient, with a decent rationale behind it. The assessors of the study outcomes were blinded, even if the treated clinicians weren’t.
Perhaps the first interesting point to discuss was the discrepancy in the treatment failure rate. This was basically the number of patients that were allocated to on BP target but were unable to achieve it within the initial time target (primary failure), or to maintain it over the 24 hours (secondary failure). Unsurprisingly, both of these were higher in the intensive treatment group, with primary and secondary failure occurring in 12.2% and 15.6% of these patients respectively, compared to just 0.8% and 1.4% in the standard treatment arm. This has the potential to confound the results and conclusions to some degree, as it meant that there was a not insignificant part of the ‘intervention’ arm that didn’t adequately receive the intervention.
A further observation provides some pause for thought when analysing the results. Though the groups appear to be well differentiated from a statistical perspective, the difference between the mean systolic blood pressures is only just over 12 mmHg. Now this is still a difference, but it makes you wonder (especially when combined with the treatment failure rate) as to whether the overall difference between the 2 groups was less than appears at first glance.
However, perhaps the most interesting point arises when we look at the nature of the patient characteristics. The majority of these patients (55%) were GCS 15 at presentation. Similarly, the incidence of ‘poor outcome’ was notably less than the 60% the authors anticipated. This raises questions about whether this is the correct patient population to be looking at for this, and whether the aggressive lowering of blood pressure is not appropriate for these patients.
Perhaps the first interesting point to discuss was the discrepancy in the treatment failure rate. This was basically the number of patients that were allocated to on BP target but were unable to achieve it within the initial time target (primary failure), or to maintain it over the 24 hours (secondary failure). Unsurprisingly, both of these were higher in the intensive treatment group, with primary and secondary failure occurring in 12.2% and 15.6% of these patients respectively, compared to just 0.8% and 1.4% in the standard treatment arm. This has the potential to confound the results and conclusions to some degree, as it meant that there was a not insignificant part of the ‘intervention’ arm that didn’t adequately receive the intervention.
A further observation provides some pause for thought when analysing the results. Though the groups appear to be well differentiated from a statistical perspective, the difference between the mean systolic blood pressures is only just over 12 mmHg. Now this is still a difference, but it makes you wonder (especially when combined with the treatment failure rate) as to whether the overall difference between the 2 groups was less than appears at first glance.
However, perhaps the most interesting point arises when we look at the nature of the patient characteristics. The majority of these patients (55%) were GCS 15 at presentation. Similarly, the incidence of ‘poor outcome’ was notably less than the 60% the authors anticipated. This raises questions about whether this is the correct patient population to be looking at for this, and whether the aggressive lowering of blood pressure is not appropriate for these patients.
Final thought
Overall I think the study provides quite a good deal of useful information about the management of patients who have a hypertensive response to an intracerebral bleed. Aggressive lowering of systolic blood pressure to below 140 mmHg is unlikely to provide any benefit to a target of less than 180 mmHg. It may be that there is some reduction in haematoma expansion, but it suggests that harm is done elsewhere via other mechanisms that mean this isn’t beneficial. There is perhaps a lingering doubt as to whether this question need revisiting but looking specifically at those with a worse presentation (lower GCS for example) but I’m not convinced there is sufficient justification to do this, despite the INTERACT2 study. There is no evidence of benefit, there is some evidence of harm, and it’s quite hard to do. Let’s leave it there.
Thanks again for reading, and I must once again thank everyone at our journal club for all of their contributions. As always please leave your thoughts on the topic. I’d also really recommend checking out The Bottom Line website through the link below for another great review of this paper (as always).
BW
Tom Heaton
Thanks again for reading, and I must once again thank everyone at our journal club for all of their contributions. As always please leave your thoughts on the topic. I’d also really recommend checking out The Bottom Line website through the link below for another great review of this paper (as always).
BW
Tom Heaton
References & Links
- Qureshi A Et al. Intensive blood-pressure lowering in patients with acute cerebral haemorrhage (ATACH2). NEJM. 2016.
- Anderson C Et al. Rapid blood-pressure lowering in patients with acute intracerebral haemorrhage. NEJM. 2013. 368:2355-2365.
- Wong A, Chambler D. Summaries: ATACH2. Thebottomline.org.uk. 2016.
Image courtesy of www.freedigitalphotos.net and Stockdevil
RSS Feed
