Perioperative Outcomes—The Limits of Blood Pressure–Centered Strategies

Intraoperative hypotension is very common, affecting 30% to 80% of surgical patients.^1–3 Results of observational studies suggest that even brief episodes of intraoperative hypotension are associated with an increased risk of postoperative complications, but the definition of hypotension varies across studies.^3–5 Prior observational studies suggest that exposure to a mean arterial pressure (MAP) lower than 60 mm Hg to 65 mm Hg increases the risk of postoperative myocardial and kidney injury.⁶ Based on these data, current guidelines recommend maintaining MAP at 60 mm Hg or higher during surgery.^6,7 However, several questions remain unsettled: Is intraoperative hypotension is a marker of underlying illness or a modifiable contributor to harm? Does the optimal MAP target vary across individuals, suggesting that a more personalized approach is warranted?

In this issue of JAMA, the IMPROVE-multi⁸ and PRETREAT⁹ trials address some of those uncertainties using different personalized approaches for the management of intraoperative hypotension.

In IMPROVE-multi, Saugel and colleagues⁸ hypothesize that tailoring the intraoperative MAP target to an individual’s preoperative values might reduce the risk of postoperative organ injury. Across 15 centers in Germany, the investigators randomized 1142 patients undergoing major abdominal surgery to the mean nighttime MAP recorded before surgery (individualized care) or to a MAP target of 65 mm Hg or higher (control). Patients in the individualized care group had higher intraoperative blood pressure values and received more vasoactive medications. However, the investigators found no difference in the composite primary outcome of acute kidney injury, myocardial injury, nonfatal cardiac arrest, or death at 7 days (33.5% of patients in the individualized care vs 30.5% in the control group; relative risk, 1.10 [95% CI, 0.93–1.30]; P = .31). None of the secondary outcomes differed significantly, including infectious complications, 90-day mortality, or hospital readmissions.

In PRETREAT,⁹ 3247 patients undergoing elective noncardiac surgery at a single institution in the Netherlands were randomized to a fixed MAP target of 65 mm Hg (control) or an individualized strategy in which higher intraoperative MAP targets were selected for all patients but varied according to the risk of intraoperative hypotension, evaluated based on a score calculated before surgery (≥70, ≥80, or ≥90 mm Hg for low, middle, and high intraoperative hypotension risks, respectively). The primary outcome was functional disability at 6 months, as measured by the World Health Organization Disability Assessment Schedule (WHODAS). As expected, patients randomized to the individualized care group received more vasoactive medications than patients in the control group and therefore experienced fewer episodes with MAP lower than 65 mm Hg (48% vs 71%). However, the functional disability score was not different at 6 months (WHODAS score, 17.7 vs 18.2; mean difference, −0.5% [95% credible interval, −1.9% to 0.9%]). Secondary outcomes—including quality of life, complications, or mortality—also did not differ significantly, and the trial was stopped for futility.

While both trials evaluated treatment protocols for intraoperative hypotension, some differences are noteworthy. In IMPROVE-multi,⁸ the intraoperative MAP target was guided by preoperative nighttime blood pressure recordings, while in PRETREAT,⁹ it was guided by an intraoperative hypotension risk score calculated preoperatively. These distinctions result in subtle differences in the study hypotheses across these 2 trials. In IMPROVE-multi, the underlying premise is that the nocturnal blood pressure is the lowest tolerable blood pressure, while in PRETREAT, the assumption is that the risk of intraoperative hypotension can be predicted, and the hypothesis is that a patient with a higher risk would benefit from higher intraoperative blood pressure. There were also important differences in clinical characteristics: in IMPROVE-multi, patients were 45 years or older undergoing elective major abdominal surgery and had 1 or more additional high-risk criterion, whereas patients in PRETREAT were older than 18 years and undergoing elective noncardiac surgery; 21% were at low risk of intraoperative hypotension, 56% were at intermediate risk, and 23% were at high risk.

There are also many similarities between IMPROVE-multi⁸ and PRETREAT.⁹ Both trials defined intraoperative hypotension as an episode of MAP lower than 65 mm Hg and assumed that targeting higher MAP in patients with higher preoperative MAP (IMPROVE-multi) or patients with higher baseline risk of hypotension (PRETREAT) would improve postoperative outcomes. Hence, although they defined personalized care differently, these 2 trials ultimately both tested a different version of the same hypothesis—ie, that maintaining higher MAP values in higher-risk patients could improve outcomes. The results in these 2 trials are strikingly similar and do not support the view that more aggressive management of intraoperative hypotension improves outcomes compared with a strategy based on the avoidance of MAP values below 65 mm Hg. These findings also add to a growing body of evidence from randomized trials in which higher perioperative blood pressure targets did not improve outcomes.¹⁰

What New Information Do These New Trials Contribute?

A frequent criticism of past trials comparing higher vs lower intraoperative MAP targets is that they compared fixed rather than personalized protocols. For example, the POISE-3 trial compared a combination of interventions delivered before, during, and after surgery to avoid hypotension in one group or hypertension in the other group.¹¹ In that 7490-patient trial, no significant difference was observed in the risk of major cardiovascular complications or postoperative acute kidney injury. In the single-center BBB trial,¹² an intraoperative MAP target of 75 mm Hg or higher did not affect the risk of postoperative acute myocardial injury or acute kidney injury compared with a MAP target of 60 mm Hg or higher. Similar findings were observed in the recent BP-CARES trial.¹³ The 292-patient INPRESS trial¹⁴ that examined individualized blood pressure management reported a lower risk of organ dysfunctions, a composite outcome based mostly on biomarker measures, in patients randomized to individualized blood pressure management using a continuous norepinephrine infusion vs standard of care using ephedrine boluses. The results of the larger IMPROVE-multi and PRETREAT trials provide compelling new evidence that, even when individualized, aiming for higher MAP targets primarily through the use of vasoactive medications does not improve important patient outcomes.

Future Directions

Both IMPROVE-multi⁸ and PRETREAT⁹ rely on an asymmetric equipoise, ie, there is clinical uncertainty regarding the benefit of maintaining higher intraoperative blood pressure but not regarding the potential harms of lower blood pressures. As clinicians, we share a deeply entrenched belief that hypotension is harmful and that correcting it will improve outcomes. Observational studies that have consistently reported an association between low MAP and adverse postoperative outcomes provide the rationale for all clinical trials to date testing the hypothesis that preventing hypotension may improve outcomes. However, this hypothesis has important limitations. First, hypotension may only be a surrogate for underlying physiological disturbances, which may not be corrected by targeting higher blood pressures. There are examples of interventions that increase the probability of hypotension without affecting important patient outcomes. For example, in the Stop-or-Not trial, the higher incidence of hypotension in the group in which renin-angiotensin system inhibitors were not held before surgery was not associated with more postoperative adverse events.¹⁵ Second, in most studies, higher MAPs are obtained with higher vasopressor use. The risk-benefit balance of increased exposure to vasopressorsmay not support their expanded use.^16,17 Third, while the prespecified subgroup analyses of the IMPROVE-multi and PRETREAT trials did not show substantial differences in the response to the blood pressure interventions, these trials did not explore all possible sources of heterogeneous treatment effects. Fourth, until a blood pressure threshold under which vasopressors should be started is identified, it is premature to draw conclusions about the benefits of vasoactive therapy for intraoperative hypotension. Meanwhile, precedents of supportive care interventions, such as tight glucose control in critically ill patients or excessive sedation, that have caused harm when administered too liberally warrants caution and a reassessment of current practices.¹⁸ If targeting MAPs above 65 mm Hg does not improve outcomes, it may be time to explore the possibility of tolerating MAPs less than 65 mm Hg in selected patients, because reducing exposure to vasopressors could offer additional benefits such as decreasing organ injury or the risk of cardiac arrhythmia. Moreover, the long-standing research focus on correcting hypotension may have diverted attention from investigation of many other factors—eg, bleeding risk, cardiovascular comorbidities, anesthetic drug dosing, surgical techniques—that may directly affect outcomes and also cause hypotension.

In conclusion, the results of the IMPROVE-multi and PRETREAT trials advance the field of perioperative medicine by challenging the hypothesis that a personalized strategy targeting higher MAP improves postoperative outcomes. The long-standing assumption that “higher is better” in patients undergoing high-risk surgery deserves reexamination, and future efforts should focus on the true drivers of perioperative risk rather than higher blood pressure targets.