Postoperative hypotension is a dose

Why it matters

Residents leave the operating room having learned to treat a low number. The PACU is where that number stops being watched continuously, and the ward is where it stops being watched at all. The sources for this day quantify hypotension as an exposure, by depth and duration, and find that the exposure keeps accruing after surgery ends. The POQI postoperative statement puts it plainly: ward vital signs are taken only every 4 to 6 h, so prolonged hypotension on postoperative days 1 to 4 can go unseen, and the odds of a poor outcome in that period are described as almost three times as high mcevoy-2019.

The consequence of getting this wrong is not one missed reading. It is treating hypotension as a moment rather than as an accumulating dose, and handing the largest part of that dose to the setting least able to see it.

Mechanism

Perfusion of the myocardium and the kidney depends on the pressure gradient across each organ’s vascular bed. When mean pressure falls below the range those beds can autoregulate, flow falls with pressure, and oxygen delivery falls with it. In a heart that is already working harder after surgery, from pain, catecholamines and inflammation, supply falls while demand rises. Injury is a function of how far pressure falls and for how long, because the deficit accumulates over time rather than occurring at a threshold. That is why every source on this day expresses the exposure in minutes below a level, not as a single reading.

Evidence

The POQI consensus statement on intraoperative hypotension records that a mean arterial pressure below 60 to 70 mmHg is associated with myocardial injury, acute kidney injury and death, and that injury is a function of both severity and duration sessler-2019. The same statement notes that a 2007 systematic review found 140 different definitions of intraoperative hypotension across 130 studies, which is why the day insists on an absolute threshold and a duration rather than a word sessler-2019.

Sessler 2018 analysed 9765 patients from the POISE-2 trial, in which clinically important hypotension was a systolic pressure below 90 mmHg requiring treatment sessler-2018. Hypotension occurred in 42% of patients; 590 (6.0%) had a myocardial infarction and 116 (1.2%) died within 30 days sessler-2018. For the composite of infarction and death, the intraoperative odds ratio was 1.08 (98.3% CI 1.03 to 1.12, P less than 0.001) per 10-minute increase in hypotension duration, and for the remainder of the day of surgery it was 1.03 (98.3% CI 1.01 to 1.05) per 10-minute increase sessler-2018. On postoperative days 1 to 4 the odds ratio was 2.83 (98.3% CI 1.26 to 6.35, P=0.002), but that exposure was binary, hypotension present or absent, so it cannot be quoted in the same per-time units as the other two sessler-2018.

Liem 2020 followed 1710 patients aged 60 or older after intermediate-to-high-risk noncardiac surgery, with frequent haemodynamic sampling on a high-dependency ward for the first 24 h, and characterised multiple absolute MAP thresholds from 50 to 75 mmHg by cumulative minutes, duration and area under threshold liem-2020. Two cumulative hours below 60 mmHg occurred in 144 patients (8%), and four hours below 75 mmHg in 824 (48%) liem-2020. The primary outcome was a peak high-sensitivity troponin T of 50 ng/L or more within three postoperative days, and adjusted odds ratios across the thresholds ranged from 2.18 to 3.26 liem-2020.

Salmasi 2017 supplies the comparison that makes a protocol possible: absolute MAP thresholds perform about as well as percentage reductions from each patient’s baseline salmasi-2017. A PACU can therefore use one fixed number without pretending to know every patient’s baseline.

The POQI postoperative statement gathers these into a bedside form: a systolic pressure below 90 mmHg, or more than 30% below baseline, is likely to put most patients at risk of end-organ injury, and risk rises with each 10-minute epoch of hypotension intraoperatively and on the day of surgery mcevoy-2019.

What this does not show

None of these sources randomised anyone to a blood pressure. Sessler 2018 sits inside a randomised trial, but hypotension there was an observed exposure, not an allocated one, so the design is a cohort analysis within a trial population sessler-2018. Liem 2020, Salmasi 2017 and both POQI statements are cohort and consensus liem-2020 salmasi-2017 mcevoy-2019 sessler-2019. The dose-and-duration structure is real and quantified, and it is associational. Whether treating the exposure changes any outcome is not established by anything on this page.

Liem 2020 also carries a result that cuts against the intuition the day builds on: in that cohort, intraoperative hypotension was independently not associated with myocardial injury liem-2020. The exposure that carried the association was the postoperative one. That is the reason this day is about the PACU and the ward rather than the operating room, and it is also a warning against assuming that the intraoperative and postoperative exposures are the same thing measured twice.

The 98.3% confidence intervals in Sessler 2018 are not a misprint: they reflect the trial’s correction for multiple comparisons, and they should be quoted as printed sessler-2018.

At the bedside

Record hypotension the way the sources measure it: how far below which number, and for how long. A single reading is a sample of an exposure, not the exposure.

When you set a PACU threshold, use an absolute number and say what it is. Salmasi 2017 is the reason you may do that without a baseline for every patient salmasi-2017. [PRACTICE VARIES: the absolute MAP or systolic threshold at which a PACU protocol triggers treatment, and the first-line treatment, differ between institutions; follow local protocol.]

Before you sign a patient out to a ward where vital signs are taken every 4 to 6 h, ask how much of their hypotension dose has already accrued and whether the next four hours unobserved are acceptable mcevoy-2019. The decision to discharge from the PACU is a decision about who watches the pressure next.

Treat the association as a reason to look and to act on a protocol, not as proof that the protocol changes outcome. Say so when you teach it.

Check yourself

Four items. Every option carries an explanation. Progress is not saved.