Hypotension Is Associated With Perioperative Myocardial Infarction: Individualized Blood Pressure Is Important

See Article, p 6 In this issue of the Journal, Hallqvist et al1 report on a nested case–control study, based on several Swedish medical registries, that showed a strong association between reductions in systolic blood pressure and acute perioperative myocardial infarction (MI), drawing attention to the danger of hypotension. For decades, much attention was paid to perioperative arterial hypertension. In the early 1970s, the treatment of arterial hypertension was relatively ineffective. As a result, hypertensive episodes in response to laryngoscopy and intubation were frequent and associated with electrocardiographic signs of myocardial ischemia and dysrhythmias.2,3 Thus, the clinical focus was on the perioperative risk of hypertensive crises, often made more severe by repeated attempts at intubation.4 Concern about hypertensive episodes continues. Indeed, in 2020, a study reported on a variety of agents being used to blunt the hypertensive response to laryngoscopy, highlighting a continued focus on the risks associated with hypertension.5 However, for more than a decade, perioperative hypotension has been shown to be detrimental. In the Perioperative Ischemic Evaluation (POISE) study,6 perioperative hypotension was the major population-based risk factor for all-cause mortality. Indeed, over the past 10 years, many studies have demonstrated an association between hypotension and adverse perioperative outcome, whether hypotension occurred before, during, or after the surgery.7 Indeed, in vascular surgical patients, a 40% reduction from the baseline mean arterial pressure was associated with myocardial injury.8 In the study by Hallqvist et al,1 326 high-risk patients suffered acute MI (according to the third universal definition9) within 30 days of surgery; they were undergoing noncardiac surgery at 3 University Hospitals in Sweden between 2007 and 2014. For each patient developing MI, a control was selected and matched by age, sex, American Society of Anesthesiologists (ASA) physical status, cardiovascular disease, surgical year, hospital, surgical code, acute/elective surgery, and duration of surgery (classified as less or more than 3 hours). While many studies have used 1 or several thresholds to define hypertension or hypotension, Hallqvist et al1 adopted a more individualized approach: reductions in systolic blood pressure were defined as absolute or relative changes from the baseline blood pressure obtained before surgery, lasting ≥5 minutes, and classified according to their magnitude, that is, 20 to 40, 41 to 50, >50 mm Hg, or similar percentage reductions. The authors have also defined “quartiled” thresholds of blood pressure. In other studies, only 1 or 2 thresholds were used.4,10 A fairly similar approach was used by Monk et al11 in respect of 30-day mortality. It showed that hypotension not hypertension was a significant determinant of adverse outcome. To ensure the validity of the preoperative blood pressure value, the patient’s baseline blood pressure was estimated from all blood pressures—5 on average—documented within 2 months before surgery and obtained from the surgical ward, preoperative anesthetic consultation, or documentation from the primary health care physician. The results showed that an intraoperative hypotensive event >41 to 50 mm Hg was associated with more than tripled MI risk (odds ratio [OR] = 3.42; 95% confidence interval [CI], 1.13-10.3) and a hypotensive event >50 mm Hg with a considerable risk increase (OR = 22.6; 95% CI, 7.69-66.2) after adjustment for preoperative covariates such as high blood pressure (systolic arterial pressure [SAP], ≥140 mm Hg), diabetes mellitus, ischemic heart disease, intraoperative blood loss (>1800 mL), anemia (Hb <85 g/L), hypoxia (arterial oxygen saturation [Sao2], <90%), and positive fluid balance (>2000 mL). Whether absolute, percentile reductions, or “quartiled” thresholds of blood pressure, all showed, unsurprisingly, that the lower the systolic blood pressure, the worse the risk of MI. Of course, the study by Hallqvist et al1 suffers limitations. The data were gathered from registries, and only systolic, not mean, arterial blood pressure data were available. However, for many clinicians, it is systolic blood pressure that catches the eye on monitors rather than the mean arterial pressure. One of the advantages of registry-based studies is that the data reflect the current practice at the time. By contrast, randomized controlled trials and prospective observational studies generally include protocols that may differ from the standard management and, therefore, the data may not be fully representative of general clinical practice. This has been recently highlighted in the study by Titz et al.12 The authors found that protocols for 36 randomized controlled trials of beta-blockade were overly restrictive: the proportion of patients who could be enrolled was as low as 6.5% in some studies. Thus, conclusions of such studies may not be representative of routine practice. In most surgical populations, a substantial proportion of patients present with undiagnosed or poorly controlled arterial hypertension. In this study, the range of preoperative systolic blood pressures extended from 140 to 160 mm Hg, indicating that many patients can be considered as hypertensives. Intraoperatively, the range of systolic blood pressures extended from 65 to 90 mm Hg in patients with MI and between 80 and 110 mm Hg in the controls. In this study, a reduction in systolic blood pressure of 41 to 50 mm Hg was associated with a 3-fold increase in the risk of MI. It follows, therefore, that a fall in the systolic blood pressure to about 120 mm Hg from 160 mm Hg preoperatively would potentially be associated with an acute MI; yet, little attention would have been paid in the real world to such a change. However, most hypotensive episodes exceeded 50 mm Hg (260 vs 66 cases with smaller reductions). This in itself suggests that blood pressure reductions of this magnitude were probably not considered a significant risk factor for postoperative cardiovascular complications, including MI. As MI patients had a significantly higher systolic pressure (mean, 150 mm Hg) than controls (mean, 133 mm Hg), a 41 mm Hg reduction in systolic blood pressure would have resulted in a mean systolic blood pressure of 109 mm Hg, not such a worrying mean level of systolic hypotension. Yet such a reduction was associated with an increase in the risk of MI (OR = 3.42; 95% CI, 1.13-10.3) that is hard to ignore. It is interesting to note that the very high-risk patients suffered fewer episodes of hypotension (>50 mm Hg) than lower risk patients. Those identified as high risk by clinicians were probably monitored more closely resulting in better control of the circulation during surgery. Note that there was no difference in the occurrence of intraoperative hypotension among patients with fatal (<30 days) and nonfatal MI. The authors have chosen hard outcome, acute type 1 and type 2 MI, according to the third universal definition,9 for their study. The diagnosis of perioperative MI is challenging because it is very frequently silent. The authors have chosen MI rather than myocardial injury; though, both are associated with increased mortality.13 As cardiac biomarkers were not evaluated, it is possible that a number of totally silent MIs may have been unrecorded, thus reducing the apparent burden of MI and/or myocardial injury associated with hypotension. In a previous study, Hallqvist et al1 showed that intraoperative hypotension, demonstrated as a 50% reduction in the systolic blood pressure lasting >5 minutes, was associated with organ damage including myocardial injury and MI.14 Note that, as opposed to a mortality of 27% for MI in the present study, mortality of myocardial injury was 9.8% in the VISION study15 and 4% in the previous study by Hallqvist et al.14 While MI is a relatively rare postoperative condition, it is important as it has a much higher mortality (27% in their study) than in nonsurgical settings, where it was <4% in a study including >300,000 patients with ST-elevation myocardial infarction (STEMI) and non-ST elevation myocardial infarction (NSTEMI).16 In other studies, the mortality of perioperative MI ranged between 3.5% and 25%.17,18 A mortality rate of 75.6% has also recently been reported in patients who had undergone vascular surgery.19 Why is there such a difference between the surgical and nonsurgical settings? Advances in the treatment of STEMI and non-STEMI in nonsurgical patients have considerably reduced the in-hospital mortality of such patients. This is not the case with perioperative MI after vascular surgery as, even with recommended therapies, a very high proportion (37.4%) were dead at 1 year as opposed to 5.5% for those without perioperative MI.20 The high mortality of postoperative MI is not surprising as the standard treatment, outside of surgical settings, is immediate revascularization (primary percutaneous coronary intervention [PPCI]). PPCI necessitates administration of antiplatelet agents and anticoagulants. PPCI is not frequently used within the first days after surgery because of the perceived high risk of surgical site bleeding yet protocols to minimize the risk of surgical site bleeding exist,21 and revascularization reduces mortality.19,22 The large difference in MI mortality indicates that every effort should be made to prevent its occurrence in the surgical settings. From the data obtained in the study by Hallqvist et al,1 it is easy to determine, for each individual patient, the level of hypotension that could be associated with such a severe adverse outcome. Whether preventing hypotension can reduce the risk of MI will have to be demonstrated in further studies. Though, it is suggested in the study by Futier et al23 that, among patients predominantly undergoing abdominal surgery who were at an increased postoperative risk, management targeting an individualized systolic blood pressure (ie, blood pressure maintained within 10% of baseline) compared with standard management reduced the risk of postoperative organ dysfunction.23 A study of myocardial injury, not MI, by Salmasi et al24 is at variance with the data of Hallqvist et al1: the association based on the relative blood pressure thresholds was no stronger than that based on absolute thresholds. Furthermore, there was no clinically important interaction with preoperative arterial pressure. The authors concluded that anesthetic management, in respect of myocardial injury, can thus be based on intraoperative pressures without regard to preoperative pressure. However, this may not apply to MI as seen in the study by Hallqvist et al.1 If hypotension is an important factor of perioperative MI, then, in the POISE study,6 there is an apparent paradox: β-blockade was associated with more hypotension and an increase in all-cause mortality. Yet the risk of MI was reduced. This may be regarded as an inconsistency in respect of the study by Hallqvist et al,1 but beta-blockers are known to be cardioprotective, and in Hallqvist et al,1 study the proportion of patients on β-blockers is not given. In POISE, the next dose of slow-release metoprolol was given if the systolic blood pressure >100 mm Hg, irrespective of the preoperative blood pressure. This shows that clinicians accepted this particular threshold as they did not consider it in the context of the patients’ preoperative blood pressure. With a mean systolic blood pressure of 138·7 (standard deviation [SD], 19.9), a significant proportion (≤16%) of patients’ preoperative systolic blood pressure would have been close to 160 mm Hg; yet, the β-blockers would have been given in the face of a 50 to 60 mm Hg reduction in the systolic pressure. It easier to consider blood pressure reductions in respect of a threshold—and the threshold is often in the region of 100 mm Hg, but it makes more physiological sense to consider reductions as a proportion of the preoperative blood pressure. The great merit of the study by Hallqvist et al1 is to demonstrate the association with absolute and relative reductions in the blood pressure in the case of MI—a hard outcome that matters in the short and long term. The results suggest that patients could benefit from a more individualized approach to assessing blood pressure. DISCLOSURES Name: Pierre Foëx, DM, DPhil. Contribution: This author helped write the original draft. Name: Helen Higham, MD, DPhil. Contribution: This author helped by reviewing the draft and suggesting modifications. This manuscript was handled by: Stefan G. De Hert, MD.