Frequency and Outcomes of Preoperative Stress Testing in Total Hip and Knee Arthroplasty from 2004 to 2017

“The American College of Cardiology and American Heart Association (ACC/AHA) practice guidelines for perioperative cardiovascular evaluation for noncardiac surgery aim to improve the effectiveness of perioperative care, optimize patient outcomes, and improve resource use. The guidelines have consistently deemphasized preoperative cardiac testing prompted solely by the upcoming surgery in the absence of signs or symptoms that would warrant testing outside of the preoperative setting.

[..] Using MarketScan inpatient claims data from 2003 to 2017, we selected patients according to International Classification of Diseases, Ninth Revision, Clinical Modification and International Statistical Classification of Diseases, Tenth Revision, Clinical Modification (ICD-9-CM and ICD-10-CM, respectively) who underwent total knee arthroplasty and total hip arthroplasty surgeries . After the initial cohort selection (n = 1,369,098), we applied the following exclusion criteria to patients: (1) undergoing elective surgery only, based on the principal diagnosis of osteoarthrosis, (2) 18 years and older, (3) first surgery observed only, and (4) active enrollment in the insurance plan 12 months prior to surgery. This selection process yielded 801,396 elective total hip and knee replacement surgeries performed between 2004 and 2017.

The primary outcome measure was the presence of any cardiac stress test within the 60 days prior to elective lower extremity joint replacement surgery. A 60-day period was chosen owing to previous analyses of the frequency of preoperative stress testing. [..] Secondary outcomes included a diagnosis of myocardial infarction or cardiac arrest during the inpatient claim associated with the total joint arthroplasty.

[..] The study cohort consisted of 801 396 elective total hip arthroplasty procedures (27.9%) and total knee arthroplasty procedures (72.1%) in the MarketScan database from 2004 to 2017. Overall, 10.4% of patients (n = 83,307 of 801,396) underwent a preoperative cardiac stress test in the 60 days prior to surgery.

[..] Joinpoint analysis identified 2 joinpoints in the yearly percentage of preoperative cardiac stress tests in the third quarter (Q) of 2006 (95% CI, 2005 Q4 to 2007 Q4) and Q4 of 2013 (95% CI, 2011 Q3 to 2015 Q3). Prior to the 2006 inflection point, preoperative stress testing increased 0.65% (95% CI, 0.09-1.21) per year. From 2006 Q4 to 2013 Q4, preoperative stress testing decreased −0.71% (95% CI, −0.79 to −0.63) per year. For the remainder of the study period, 2013 Q4 to 2017 Q4, the percentage change was slower (−0.40%; 95% CI, −0.57% to −0.24%).

Trends in Frequency of Preoperative Cardiac Stress Testing

[..] Patient characteristics that were associated with increased odds of a cardiac stress test include age, male sex, and RCRI [Revised Cardiac Risk Index] greater than zero. As expected, recent cardiac stress test was associated with decreased odds of another stress test within the 60 days prior to the surgical procedure. Patients enrolled in a capitated insurance plan had decreased odds of a stress test. Each increase in the Elixhauser score increased the odds of a preoperative cardiac stress test because the Elixhauser score represents comorbidities not accounted for by the RCRI.

[..] Rates of complications for patients with 0 RCRI conditions were 0.14% (n = 762 of 527,047) overall and decreased throughout the study period (P < .001). The rate of myocardial infarction and cardiac arrest was 0.58% (n = 915 of 159,020) for patients with 1 or more RCRI conditions, and also decreased over time (P = .005). Among patients with 1 or more RCRI conditions, stress testing was not associated with lower rates of complications. Patients with 0 RCRI conditions who received a stress test had a mean complication rate of 0.27% (n = 95 of 35,351) that decreased from 2004 to 2017 (P < .001) but was double the complication rate of patients without RCRI conditions who did not receive a stress test (0.14%; n = 667 of 491,696; P < .001).

[..] Our study identified a sustained decline in the frequency of preoperative cardiac stress tests in patients who underwent a total hip or knee arthroplasty beginning in Q4 of 2006. We did not observe any difference in myocardial infarction or cardiac arrest in patients with at least 1 RCRI condition who underwent stress testing, which suggests stress testing may not contribute to risk stratification in this patient population. Further, the proportion of stress tests conducted for patients with zero RCRI conditions remains high, which suggests an opportunity for interventions to further reduce preoperative cardiac stress testing. Additional investigation is needed to evaluate the optimal patient conditions that would warrant stress testing and whether our results are generalizable to other surgical procedures.”

Full article, Rubin DS, Hughey R and Gerlach RM. JAMA Cardiology 2020.9.30