Background: Preoperative blood pressure (BP) thresholds associated with increased postoperative mortality remain unclear. We investigated the relationship between preoperative BP and 30-day mortality after elective non-cardiac surgery.
Methods: We performed a cohort study of primary care data from the UK Clinical Practice Research Datalink (2004-13). Parsimonious and fully adjusted multivariable logistic regression models, including restricted cubic splines for numerical systolic and diastolic BP, for 30-day mortality were constructed. The full model included 29 perioperative risk factors, including age, sex, comorbidities, medications, and surgical risk scale. Sensitivity analyses were conducted for age (>65 vs <65 years old) and the timing of BP measurement.
Results: A total of 251 567 adults were included, with 589 (0.23%) deaths within 30 days of surgery. After adjustment for all risk factors, preoperative low BP was consistently associated with statistically significant increases in the odds ratio (OR) of postoperative mortality. Statistically significant risk thresholds started at a preoperative systolic pressure of 119 mm Hg (adjusted OR 1.02 [95% confidence interval (CI) 1.01-1.02]) compared with the reference (120 mm Hg) and diastolic pressure of 63 mm Hg [OR 1.24 (95% CI 1.03-1.49)] compared with the reference (80 mm Hg). As BP decreased, the OR of mortality risk increased. Subgroup analysis demonstrated that the risk associated with low BP was confined to the elderly. Adjusted analyses identified that diastolic hypertension was associated with increased postoperative mortality in the whole cohort.
Conclusions: In this large observational study we identified a significant dose-dependent association between low preoperative BP values and increased postoperative mortality in the elderly. In the whole population, elevated diastolic, not systolic, BP was associated with increased mortality.
Keywords: blood pressure; mortality; surgery.
© The Author 2017. Published by Oxford University Press on behalf of the British Journal of Anaesthesia. All rights reserved. For Permissions, please email: firstname.lastname@example.org