Perspective:

The authors performed a retrospective analysis of 801,396 patients from the IBM MarketScan Research Database who underwent arthroplasty (72.1% knee, 27.1% hip) between 2004-2017, examining frequency trends in preoperative cardiovascular stress testing performed within 60 days of surgery, and guideline adherence, specifically the extent to which absence versus presence of Revised Cardiac Risk Index (RCRI) conditions influenced the probability of preoperative stress testing. The association between use of preoperative stress testing and adverse postoperative outcome (myocardial infarction [MI] or cardiac arrest during the inpatient claim) was explored.

The rate of preoperative stress testing increased between 2004-2006, but subsequently declined progressively after 2006, at a time coinciding with publication of simplified preoperative cardiovascular care guidelines and introduction of measures intended to limit the use of routine cardiovascular testing. This decline in preoperative stress testing coincided with declining rates of stress testing among the nonoperative population.

Logistic regression demonstrated that advanced age, male sex, and ≥1 RCRI factors were associated with greater chance of preoperative stress testing. Although membership in a capitated insurance plan was associated with lesser chance of preoperative stress testing, the temporally related decrease in testing rate over time after 2006 was observed across patients regardless of health plan coverage.

The rate of postoperative MI and cardiac arrest among patients with ≥1 RCRI conditions did not differ significantly among patients who did or did not undergo preoperative stress testing. Conversely, in patients with 0 RCRI factors, rates of postoperative MI and cardiac arrest were twice as frequent in patients who underwent preoperative stress testing compared to patients who did not undergo preoperative stress testing. Among all patients undergoing preoperative stress testing, 49% had 0 RCRI factors, suggesting poor overall guideline adherence and/or presence of comorbidity predictors not reflected in the RCRI assessment.