Several studies have recently shown that even in the setting of better BP control, blood pressure variability may still have a negative effect on outcomes. In this article, authors Tanaka et al utilized a prospective, multicenter, observational study to reinforce some of the previous findings.
205 patients were included, with inclusion criteria being age ≥ 20; initial SBP > 180, total GCS ≥ 5, supratentorial IPH and initial volume measurement < 60ml. BP was measured every 15min for the first 2 hours after starting anti-HTN therapy, then every hour after that for the next 22 hours. Outcome measures included hematoma expansion, neurologic deterioration, and unfavorable outcomes of an MRS of 4-6 at 3 months. On multivariate regression analyses, authors found standard deviation (SD) and successive variation (SV) of SBP to be associated with neurologic deterioration, and SV of SBP to be associated with unfavorable MRS outcomes at 3 months. Neither SD or SV of SBP appeared to have significant associations with hematoma expansion. There did not seem to be a significant link between SV/SD of DBP and any of the outcomes.
This study raises interesting questions about what the true causative agent is of the worse outcomes—is it a direct effect of variations in SBP, or are the negative effects somehow due to whatever caused that variation? The authors touch on this, noting that “autonomic dysfunction, including sympathetic overactivity and diminished baroreflex sensitivity may be one of the mechanisms,” the thought being that these may contribute to altered perfusion, edema and secondary injury.
This study, much like post-hoc analysis of INTERACT2, seems to imply that minimizing blood pressure variability in addition to aggressively lowering SBP may diminish rates of neurologic deterioration and improve long-term outcomes. It will be interesting to see how this continues to play out in further studies, and if deterioration is found to be a direct effect of that variability or rather is just the effect of another, more damaging process. If it is simply directly due to variability and we utilize effective methods to minimize that, what will we find “acceptable” levels of variability to be?