Relationship Between First Systolic Blood Pressure in the Operating Room and Perioperative Ischemic Stroke in Non‐Brain Non‐Cardiac Surgical Patients

Authors: Yan Zhou, Liqing Xu, Lin Liu, Hongzhou Duan

Abstract

Hypertension is considered a potential risk factor for perioperative ischemic stroke (PIS). However, the association between elevated first systolic blood pressure measured in the operating room (first‐OR‐SBP) and the incidence of PIS has not been well documented. We conducted a single‐center retrospective cohort study including patients who underwent elective non‐brain, non‐cardiac surgery at Peking University First Hospital between January 1, 2018, and December 31, 2024. Data were extracted from a perioperative database, and patient demographics, intraoperative and perioperative variables—particularly the relationship between first‐OR‐SBP and PIS—were analyzed. Multivariate logistic regression was performed before and after propensity score matching to adjust for perioperative confounders. The minimum p value approach was used to identify a potential threshold of first‐OR‐SBP independently associated with PIS risk. We found that, among 105 059 surgeries, 195 patients (0.19%) experienced PIS. The threshold for first‐OR‐SBP associated with PIS was identified as 186 mm Hg. The adjusted odds ratios for PIS were 1.69 (95% CI, 1.12–2.55; p = 0.013) before matching and 1.62 (95% CI, 1.03–2.54; p = 0.036) after propensity score matching. We conclude that a first‐OR‐SBP≥186 mm Hg was significantly associated with an increased risk of perioperative ischemic stroke in patients undergoing elective non‐brain, non‐cardiac surgery.

Introduction

Stroke is a major and growing global health burden, accounting for approximately 6.2 million deaths annually and ranking as the leading cause of premature mortality and disability worldwide [1]. Perioperative stroke, a devastating complication for surgical patients, is defined as a cerebrovascular event—manifesting as motor, sensory, or cognitive dysfunction—that occurs intraoperatively or within 30 days after surgery, resulting from ischemia or hemorrhage [2]. The majority of these events are ischemic in nature [3]. Reported incidence rates vary widely, from 0.08% to 10% [4, 5, 6, 7, 8], with associated mortality being alarmingly high, ranging between 20% and 60% depending on patient population and surgical context [9].

Established risk factors for perioperative stroke include advanced age, hypertension, renal disease, history of transient ischemic attack or stroke, myocardial infarction, atrial fibrillation, and diabetes mellitus [2]. Among these, hypertension is a well‐established independent risk factor for both stroke in general and perioperative stroke specifically. However, the specific relationship between blood pressure measured in the operating room and the risk of perioperative ischemic stroke (PIS) remains poorly quantified [10].

We hypothesize that a threshold exists for the first systolic blood pressure recorded in the operating room (first‐OR‐SBP), beyond which the risk of PIS rises significantly. Therefore, this study aims to investigate the association between first‐OR‐SBP and the occurrence of PIS in patients undergoing elective non‐brain, non‐cardiac surgery, and to determine the optimal threshold of first‐OR‐SBP for predicting PIS risk.

Materials and Methods

Study Design, Setting, Population, and Data Collection

This retrospective cohort study was conducted at Peking University First Hospital. Ethical approval was granted by the Institutional Review Board (IRB), which waived the need for informed consent as the study used existing, de‐identified data, and required no patient follow‐up.

This study utilized data from the perioperative database of Peking University First Hospital, an institutional registry of inpatient surgical cases from 2012 onward. The analysis was restricted to adult patients (age ≥ 18 years) undergoing elective non‐brain, non‐cardiac surgery between January 1, 2018, and December 31, 2024. Eligible surgeries were ascertained based on the International Classification of Diseases and Procedures, Ninth Clinical Revision, volume 3 (ICD‐10).

The included surgical specialties comprised otolaryngology, general surgery, urology, gynecology, orthopedics, and vascular and thoracic surgery. Exclusion criteria were brain or cardiac procedures, obstetrics, emergency surgeries, and operations performed under local infiltration anesthesia. For patients with multiple surgeries, only the first procedure within a calendar year was analyzed; subsequent operations occurring within a 3‐month interval were not included in the registry.

Variables in the Study

The study variables encompassed patient demographics, preoperative comorbidities, and intraoperative characteristics. Demographic data included age, gender, body mass index, smoking history, and alcohol use. Preoperative comorbidities—hypertension, diabetes mellitus, hyperlipidemia, hyperhomocysteinemia, coronary artery disease, and prior myocardial infarction—along with the CHA2DS2‐VASc score, were ascertained from ICD codes and medical records using plain text analysis (Table S1). Intraoperative variables consisted of anesthesia type, incidence of hypotension, blood transfusion, mean heart rate, and risk assessment by the modified John Hopkins Hospital criteria (MJHSC) [11, 12, 13, 14].

First‐OR‐SBP

The ‘first systolic blood pressure in the operating room’ (first‐OR‐SBP) was defined as the initial non‐invasive blood pressure reading recorded after the patient's arrival in the operating room. This measurement was obtained with the patient in the supine position using an upper‐arm cuff.

Definitions Endpoints

The study endpoint was in‐hospital PIS, comprising cerebral infarction or TIA, which was initially identified via ICD‐10 codes. Case confirmation was based on a triad of neurological consultation reports in the Electronic Medical Record, discharge diagnoses, and perioperative head CT or MRI reports. All text was screened for neurologist‐defined keywords (Table S1). Finally, each potential ischemic stroke event was validated by two independent neurologists (L.‐Q. X. and H.‐Z. D.).

Non‐Linear Relationship Detection Between First‐OR‐SBP and PIS

The unadjusted relationship between first‐OR‐SBP and PIS risk was initially examined using a cubic spline function within a Generalized Additive Model. This preliminary analysis aimed to identify a potential inflection point that could serve as a clinically meaningful threshold for categorizing first‐OR‐SBP. Subsequently, if an interval of increased risk was observed, the optimal threshold was precisely determined using the minimum p value approach. This method systematically tests every possible 1 mm Hg increment of first‐OR‐SBP as a cutoff point in multivariate logistic regression models. The value associated with the smallest statistically significant p value was selected as the final threshold for dichotomizing first‐OR‐SBP.

Multivariate Logistic Regression to Detect Any Association Between the First‐OR‐SBP and PIS

A fully adjusted multivariate logistic regression model was constructed to assess the independent association between the first‐OR‐SBP and PIS. The model included the following pre‐specified age, CHA2DS2‐VASc score, anemia, hypertension, cancer surgery, surgical duration, intraoperative blood transfusion, and intraoperative hypotension. Multicollinearity was assessed using the variance inflation factor, with a threshold of 3. Model discrimination was evaluated using the c‐statistic, and calibration was assessed with the Hosmer–Lemeshow test.

To determine whether the inclusion of the first‐OR‐SBP improved predictive performance, we calculated the category‐free net reclassification improvement (NRI) and the integrated discrimination improvement (IDI).

Analysis of the Propensity Score Matching

Given the observed systematic differences in baseline characteristics (e.g., age, comorbidities, surgery duration) between patients with higher and lower first‐OR‐SBP, we employed propensity score matching (PSM) to minimize confounding. The propensity score was modeled using a comprehensive set of covariates encompassing patient demographics, comorbidities (e.g., hypertension, diabetes, CHA2DS2‐VASc score), and operative details (e.g., cancer surgery, duration, anesthesia type). A 4 nearest‐neighbor matching algorithm was applied. All statistical analyses were performed with R (v.4.2.2).

Statistical Analysis

All statistical analyses were performed using R programming language (version 4.2.2). Patients were categorized into two groups based on the predefined first‐OR‐SBP cutoff. Continuous variables were summarized as mean ± standard deviation or median with interquartile range depending on their distribution. Normality was assessed using the Kolmogorov–Smirnov test. Between‐group comparisons were conducted using the Student's t‐test for normally distributed variables and the Mann–Whitney U test for non‐normally distributed variables. Categorical variables were expressed as frequencies and percentages and compared using the Chi‐square test or continuity‐corrected Chi‐square test, as appropriate. Ordinal variables were analyzed using the Kruskal–Wallis H‐test. A two‐tailed p value < 0.05 was considered statistically significant.

Results

Baseline Characteristics

Between January 2018 and December 2024, a total of 126 462 elective surgical cases were screened, of which 105 059 non‐brain, non‐cardiac surgeries were included in the final analysis (Figure 1).

The median age of the study population was 56 years, 52.8% were male, and the mean body mass index was 24.4 ± 3.7 kg/m^2^. Most patients were classified as ASA physical status I or II. The most common surgical categories were genital/urinary (40.9%), digestive tract (14.6%), and musculoskeletal (10.4%). The median operative duration was 137 min (IQR: 93‐209), and the median first systolic blood pressure recorded in the operating room (first‐OR‐SBP) was 141.05 ± 24.4 mm Hg (Table 1).

Among the 105 059 surgeries, 195 (0.19%) were complicated by perioperative ischemic stroke (PIS). Patients with a first‐OR‐SBP≥186 mm Hg had a longer median hospital stay (4 days vs. 3 days; p < 0.001). Logistic regression identified older age, higher CHA2DS2‐VASc score, hypertension, anemia, surgery time, blood transfusion, and first‐OR‐SBP≥186 mm Hg as significant risk factors for PIS (Table 2).

Non‐Linear Relationship Between First‐OR‐SBP and PIS

A generalized additive model with cubic spline revealed a J‐shaped relationship between first‐OR‐SBP and the probability of PIS. The risk curve showed an inflection interval approximately between 160 and 210 mm Hg, beyond which the probability of PIS increased sharply (Figure 2).

Association Between First‐OR‐SBP and PIS in Multivariable Analysis

Multivariable logistic regression confirmed a significant association between first‐OR‐SBP and PIS. Using a threshold of 186 mm Hg, the adjusted odds ratio for PIS in the high SBP group (≥186 mm Hg) was 1.69 (95% CI: 1.12–2.55; p = 0.013), with the low SBP group (<186 mm Hg) as reference (Figure 3).

No evidence of multicollinearity was observed among variables in the model, with all variance inflation factors below 3. Models with and without first‐OR‐SBP both demonstrated good discriminative ability, with c‐statistics of 0.828 and 0.833, respectively (Table S2).

Subgroup and Sensitivity Analyses

Subgroup analyses performed before (Figure 3) and after propensity score matching (Figure 4) showed no significant interaction effects across patient characteristics. Sensitivity analysis yielded an E‐value of 3.27, indicating that an unmeasured confounder would need to be moderately strongly associated with both exposure and outcome to fully explain the observed association. The relationship between first‐OR‐SBP and PIS remained consistent across all sensitivity analyses (Table 2).

Propensity Score‐Matched Analysis

After propensity score matching, patients were stratified into two groups based on the first‐OR‐SBP threshold of 186 mm Hg. Baseline characteristics were well‐balanced between the matched cohorts (Table 3). The adjusted odds ratio for PIS in the high versus low SBP group was 1.62 (95% CI: 1.03–2.54; p = 0.036) (Table 4).

Discussion

Our study found that PIS occurred in 0.19% of adult patients undergoing elective non‐brain, non‐cardiac surgery. Both multivariable analysis and propensity score‐matched models consistently demonstrated that a first‐OR‐SBP ≥ 186 mm Hg was significantly associated with an increased risk of PIS.

PIS remains a serious complication in non‐cardiac, non‐brain surgical patients, particularly among those with pre‐existing conditions such as atrial fibrillation, valvular disease, renal impairment, or prior stroke [1]. The incidence of PIS in our cohort (0.19%) was slightly lower than rates reported in other contemporary 0.23% in a Chinese multicenter study [2], 0.25% in an analysis of the ACS‐NSQIP database [17], and 0.3% in the multinational VISION study [18]. This modest difference may be partly explained by our exclusion of neurosurgical procedures, which are known to carry a substantially higher risk of stroke—particularly vascular operations such as carotid endarterectomy, aneurysm clipping, and bypass surgery for moyamoya disease [19, 20, 21].

Identifying the mechanisms and risk factors for PIS is essential for developing effective prevention strategies, which may ultimately improve clinical outcomes and reduce healthcare costs. Established risk factors include advanced age, hypertension, renal disease, prior transient ischemic attack or stroke, myocardial infarction, atrial fibrillation, and diabetes mellitus [1]. Among these, hypertension is a well‐established independent predictor of PIS. Previous studies indicate that 4%–28% of patients experience elevated blood pressure upon arrival in the operating room [22]. In some cases, this elevation meets the criteria for hypertensive crisis, leading to postponement or cancellation of surgery. Our prior research suggests that factors such as inadequate preoperative blood pressure control, preoperative stress, poor sleep quality, and disease‐related symptoms (e.g., pain) may contribute to elevated in‐operating‐room blood pressure [10].

Numerous studies have examined the relationship between perioperative blood pressure and PIS in patients undergoing elective surgery [4, 5, 6, 7, 8]. One notable focus has been blood pressure variability (BPV), defined as the degree of fluctuation in a patient's ambulatory blood pressure over time [23]. The etiology of BPV is likely multifactorial, involving intrinsic factors such as autonomic dysfunction and vascular wall remodeling, as well as extrinsic influences like anxiety and medication adherence [24, 25, 26]. Regardless of its cause, elevated preoperative BPV has been associated with an increased risk of myocardial infarction, stroke, renal failure, hospital readmission, and mortality following elective surgery [27]. However, strategies to mitigate BPV remain limited; one large study found that pharmacologic reduction of BPV did not significantly lower mortality risk [28], underscoring the need for further research into the relationship between perioperative blood pressure dynamics and postoperative outcomes.

Our findings confirm that preoperative hypertension—particularly elevated first‐OR‐SBP—is associated with PIS. Notably, this relationship was not simply linear. We identified an inflection point in the association between first‐OR‐SBP and PIS, beyond which the risk increased sharply.

Currently, there is no consensus regarding the specific cerebral risk associated with elevated first‐OR‐SBP, particularly among patients with otherwise normal blood pressure [29]. Current US and ESC/ESH clinical guidelines recommend postponing elective surgery in patients with systolic blood pressure >180 mm Hg or diastolic pressure >110 mm Hg, except in emergencies [30]. Previous studies have yielded inconsistent Colombo et al. reported an increased risk of perioperative hypertension and neurological deficits in patients undergoing carotid endarterectomy [19], whereas Crowther et al. found no correlation between preoperative hypertension and perioperative hemodynamic changes, despite the known association of such changes with morbidity and mortality [31]. Many of these studies were limited by small sample sizes, contributing to uncertainty regarding the relationship between SBP and surgical risk.

In contrast, our study—comprising over 100 000 non‐brain, non‐cardiac surgeries—revealed a clear association between first‐OR‐SBP and PIS. Multivariable analyses, both before and after propensity score weighting, consistently demonstrated that elevated first‐OR‐SBP significantly increases PIS risk. It is important to note that PIS is a multifactorial outcome influenced by age, smoking, alcohol use, glucose and lipid metabolism, and cardiovascular health. To enhance the validity of our findings, we rigorously controlled for these confounders using propensity score matching. Through this process, a first‐OR‐SBP threshold of 186 mm Hg was robustly validated as a significant predictor of PIS.

The mechanisms underlying the association between elevated first‐OR‐SBP and PIS remain incompletely understood. However, insights may be drawn from studies linking intraoperative hypertension to myocardial ischemia, where inflammatory activation, reperfusion injury, and enhanced platelet aggregation have been implicated in adverse perioperative outcomes [32]. Furthermore, intraoperative blood pressure variability has been associated with 30‐day mortality in patients undergoing cardiac surgery [33].

We speculate that poorly controlled preoperative blood pressure—reflected by high first‐OR‐SBP—may contribute to cumulative target organ damage, including cerebrovascular impairment. The brain is particularly vulnerable to perfusion changes, and the hemodynamic stresses of surgery and anesthesia may unmask this susceptibility, increasing PIS risk [34]. Additionally, the use of dexmedetomidine for preoperative anxiolysis may induce peripheral vasoconstriction and transient hypertension [35]. Aggressive treatment of such hypertension could precipitate intraoperative hypotension or significant BP fluctuations. Since general anesthesia blunts cerebral oxygen consumption, early ischemic symptoms may remain clinically silent [34]. Postoperative fluid mismanagement in this context may further predispose to cerebral ischemia. Thus, enhanced hemodynamic and volume monitoring is warranted in patients with elevated first‐OR‐SBP.

Several limitations should be considered when interpreting our results. First, as a single‐center retrospective study, the generalizability of our findings may be limited, and the observational design precludes causal inference. Although we identified a significant association between first‐OR‐SBP and PIS, future prospective studies are needed to confirm this relationship and elucidate underlying mechanisms. Second, the observed PIS incidence of 0.19% may be an underestimate. As PIS is defined as a devastating ischemic complication for surgical patients that occurs intraoperatively or within 30 days after surgery; however, the mean duration of hospital stay is a few days in our study and systematic follow‐up information after hospital discharge was not available, this will almost certainly led to an underestimation of the true incidence of PIS in our cohort. What's more, transient ischemic attacks or mild deficits may go unreported by patients or unrecognized by clinicians, and stroke‐related neurobehavioral changes can be misattributed to postoperative delirium or anesthetic effects. Such under‐ascertainment could influence the precision of the 186 mm Hg threshold identified. Third, non‐invasive cuff measurements of first‐OR‐SBP may not fully reflect intra‐arterial values. However, given the large sample size, the resulting measurement error is unlikely to substantially alter our conclusions. Finally, the relatively small number of patients with intraoperative hypotension in our cohort limited the ability to detect its potential effect on PIS. Larger studies are needed to clarify its role.

Conclusions

A first systolic blood pressure in the operating room ≥186 mm Hg is independently associated with an increased risk of perioperative ischemic stroke in patients undergoing elective non‐brain, non‐cardiac surgery.

Author Contributions

Yan Zhou performed data collection and analysis, drafted the manuscript. Liqing Xu helped acquisition of data, analysis and interpretation of data. Hongzhou Duan conceived and designed the study and revised the manuscript. Lin Liu helped in collecting laboratory test results in the database and helped with statistical analysis.

Funding

This research was supported by Peking University First Hospital Cross disciplinary Research Project (No. 2023IR15) and Peking University Clinical Scientist Training Program (BMU2025PYJH003).

Conflicts of Interest

The authors declare that they have no competing interests.

Ethics Statement

This study was approved by the Peking University First Hospital Ethics Committee (2022yan487). This study was conducted in accordance with the 1964 Declaration of Helsinki and its amendments. All methods were performed in accordance with the relevant guidelines and regulations.

Consent

Due to the retrospective feature and none patient follow‐up, the requirement for written informed consent was waived by the IRB. IRB Contact Information: Peking University First Hospital Ethics Committee. Tel: 010‐82805563.

Supporting information