QUICK HIT ARTICLE #10: How Low Should you Go? BP lowering without hypertensive emergency

Miller, JB. Cerebrovascular risks with rapid blood pressure lowering in the absence of hypertensive emergency. Am J Emerg Med. 2018 Aug 21. pii: S0735-6757(18)30690-9. doi: 10.1016/j.ajem.2018.08.052.

There are not many studies of affects of blood pressure lowering on ED patients in the absence of end organ disease. This one is an interesting one and should give cause before lowering blood pressure for the sake of a number.


This was a non-randomized trial of 39 patients with out of control BP who had their blood pressure lowered not for end organ disease but for the sake of the number being too high. The primary outcome was measure of cerebral blood flow. One of the secondary outcomes was the number of adverse events in the groups with their blood pressure lowered. There was no difference in the primary outcome, however 5% of patients developed adverse events including stroke and altered mental status when their SBP was lowered by >25%. The practice of lowering BP for the sake of the number being too high is not recommended and can cause harm as seen in this study.


This was a prospective quasi-experimental(1) study occurring in a single academic emergency department. A convenience sample of patients presenting to the ED with severely elevated BP (SBP >180 mmHg) were included if the treating clinician was going to give them medications to lower blood pressure DESPITE not having end organ damage. In other words these were patient with “hypertensive urgency” and not hypertensive EMERGENCY. To evaluate these patients the investigators decided to use transcranial doppler (TCD) to measure cerebral blood flow and determine if lowering the bp acute resulted if decreases in cerebral blood flow. In other words does lowering blood pressure in those who have chronically elevated blood pressure cause problems. The primary outcome was the change in middle cerebral artery (MCA) mean cerebral blood flow (CBF) before and after blood pressure lowering medications (paired samples). In order to detect a 10% relative change in CBF a sample size of 38 patients was obtained for an 80% power. The inclusion and exclusion criteria are listed below in the table. Secondary outcomes were the rate of adverse neurological effects and comparison of patients that had intensive (≥25%) drop in MAP versus those with modest BP reductions. The study recruited 48 patients but 9 were excluded leaving 39 patients. There was an average drop in SBP of -38 mmHg ((95% CI −49 to −27) and MAP of −27 (95% CI−36 to −20) mmHg representing an overall reduction of BP by 18%. For the primary outcome there was no difference in the 2 groups in reference to the CBF (a difference of -5 cm/s). There were no differences in any of the secondary outcomes involving TCD measurements. However, interestingly there were 2 patients with adverse events; one with a reversible stroke like syndrome after a reduction in SBP by54% and one with reversible altered mental status after an 29% reduction in SBP. Both of these adverse events occurred after TCD measurements were completed. This gives the incidence of adverse events as 5% in this group. Also both patients had their adverse events after second attempt at lowering BP. Unfortunately, the authors do not give mention to what drugs were administered. I think this study is actually pretty telling. If you try to lower BP too aggressively in patients without end organ damage you are going to adversely affect some of them. I think this really gets back to my point of what makes a great physician: knowing when to not just do something, but stand there. Maybe we need to change the term “conventional wisdom” to “conventional ignorance”.

A SIDE OF STATISTICAL SWOL: The Quasi-experimental study.

Remember there are 2 basic types of studies observation and experimental. The prototypical experimental study is the randomized controlled trial. A quasi-experimental study is one where you have a controlled trial without the randomization. Another name for this is the non-equivalent group controlled design (NECGD). I like this name better because it says in the name that without randomization the groups will be not be equivalent. Therefore, this is not a trial to determine cause-effect relationships. The typical statistics used in RCT’s may not be sufficient to determine significance here because there is bias between the groups. Usually analysis of co-variance with adjustments are used to determine if the groups are different.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.