Age-Based Preoperative Transthoracic Echocardiography, Intraoperative Hypotension, and Postoperative Outcomes in Older Adults Undergoing Noncardiac Surgery: A Regression Discontinuity Analysis.

Abstract

Background: The clinical value of routine preoperative transthoracic echocardiography (TTE) in older adults remains controversial. Existing evidence is largely observational and vulnerable to confounding by indication, limiting causal interpretation of whether preoperative TTE improves intraoperative management and perioperative outcomes. We aimed to evaluate the local effect of routine age-based preoperative TTE using a quasi-experimental design.

Methods: This single-center retrospective cohort study used a fuzzy regression discontinuity design based on an institutional policy of routine preoperative TTE in patients aged ≥75 years. Eligible patients were aged 60 to 89 years and underwent elective noncardiac surgery under general anesthesia from July 2017 to August 2024. To minimize confounding by indication, patients with guideline-based indications or other clinical conditions likely to prompt preoperative TTE independent of age were excluded. The primary outcome was intraoperative hypotension (IOH), defined as the duration of mean arterial pressure <65 mm Hg. Exploratory secondary outcomes included 30- and 90-day mortality, acute kidney injury (AKI), planned and unplanned intensive care unit (ICU) or cardiac care unit (CCU) admission, and postoperative length of hospital stay. Continuity of baseline characteristics and density of the running variable were assessed.

Results: Among 7270 patients, the probability of undergoing preoperative TTE increased by 43.1 percentage points (95% confidence interval [CI], 33.3-52.9) at the 75-year-old cutoff, confirming a strong first-stage discontinuity. The estimated local effect of preoperative TTE on IOH was a 24.8-minute reduction (95% CI, -71.1 to 21.5). Sensitivity analyses using alternative bandwidths and covariate adjustment yielded similar findings. Point estimates for 30-day mortality (-2.1%; 95% CI, -6.0 to 1.8), 90-day mortality (-3.2%; 95% CI, -10.4 to 3.9), and AKI (-3.6%; 95% CI, -15.9 to 8.6) were below zero but imprecise. Point estimates were also below zero for unplanned ICU or CCU admission (-6.0 percentage points; 95% CI, -12.2 to 0.2) and postoperative length of hospital stay (-7.0 days; 95% CI, -14.6 to 0.5). Baseline characteristics and the density of the running variable (age) showed no clear discontinuity at the cutoff.

Conclusions: Using a quasi-experimental design that reduces susceptibility to confounding by indication, we found no clear evidence that routine age-based preoperative TTE reduced IOH or improved postoperative outcomes near the threshold. The estimates were imprecise, and clinically meaningful benefit or harm could not be excluded. These findings suggest that age alone may be insufficient to justify routine preoperative TTE and support further evaluation of selective risk-based strategies.