Risk of Acquiring Perioperative COVID-19 During the Initial Pandemic Peak

A Retrospective Cohort Study

Lucas G. Axiotakis Jr., BS; Brett E. Youngerman, MD, MS; Randy K. Casals, MD; Tyler S. Cooke, MD; Graham M. Winston, MD; Cory L. Chang, AB; Deborah M. Boyett, MD, MS; Anil K. Lalwani, MD; Guy M. McKhann, MD


Annals of Surgery. 2021;273(1):41-48. 

In This Article


Setting and Participants

In this retrospective cohort study, we reviewed the medical records of all adult patients who underwent surgery at New York-Presbyterian Hospital—Columbia University Irving Medical Center, a quaternary, acute care hospital in northern Manhattan, between March 15 and May 15, 2020. During this period, surgeries were limited to emergent or urgent cases that were triaged by a multi-disciplinary hospital committee or the operating room medical director (AKL). Three to 5 operating rooms and an adjacent combined preoperative and recovery room were designated for surgical care. Postoperatively, COVID-19 negative patients were segregated from positive patients in the hospital when possible. Suspected or confirmed preoperative COVID-19 patients were isolated. Hospital census data (Figure 1) of confirmed inpatient COVID-19 hospitalizations showed a rise in cases throughout the study period until a peak around the week of April 13 before gradually declining through May 15, when there were still over 200 inpatient cases. This study was approved by the Columbia University Institutional Review Board (#AAAT0618).

Figure 1.

Daily counts of confirmed COVID-19 cases at NYP/CUIMC between March 15 and May 15, 2020. X-axis represents data, Y-axis represents raw count. All cases were confirmed by SARS-CoV-2PCR testing. Squares = Total inpatients; Circles = ICU inpatients; Diamonds = Patients requiring mechanical ventilation. The periods before and after routine preoperative testing took place are indicated. COVID-19 indicates coronavirus disease 2019; CUIMC, Columbia University Irving Medical; ICU, intensive care unit; NYP, New York-Presbyterian.

Preoperative testing for the SARS-COV-2 virus was performed on nasopharyngeal swab specimens with an in-house reverse-transcriptase–polymerase-chain-reaction (PCR) assay. We defined preoperative test as any test within 3 days of the index surgery and excluded patients with earlier test-confirmed COVID-19 infection (Figure 2). Between March 15th and April 6th, preoperative testing was performed based on clinical indication and availability; after April 6, preoperative testing was mandated for all patients undergoing surgery.

Figure 2.

Eligibility for preoperative SARS-CoV-2 testing. * = Preoperative negative patients (n = 501), forming the cohort for determination and analysis of perioperative COVID-19. COVID-19 indicates coronavirus disease 2019.

Preoperatively, patients were presumed positive without a test if they had (i) documentation of 2 or more of 7 core Centers for Disease Control and Prevention defined symptoms[11] in the 2 weeks preceding surgery, or (ii) symptom onset and a positive test by postoperative day 1; otherwise they were presumed negative. Postoperatively, patients were determined to have acquired symptomatic COVID-19 if: (i) they had a negative test within 3 days of surgery or were presumed negative preoperatively; and, (ii) they had documentation of 2 or more Centers for Disease Control and Prevention defined symptoms and clinical suspicion for COVID infection occurring in the 2 to 14 days after surgery, with or without a test. Patients without any documented clinical events or encounters after surgery allowing for adequate determination of COVID status based on symptoms were excluded.

Variables and Data Collection

Patient data were manually abstracted from the electronic medical record, including patient characteristics (age, sex, select comorbidities, and medications), surgery characteristics (surgical service, case length, case urgency, anesthesia type), and clinical course (pre- and postoperative length of stay (LOS), intensive care unit admission, adverse events, readmission, and mortality) based on American College of Surgeons National Surgical Quality Improvement Project (NSQIP) definitions[12] for a follow-up period of 30 days postoperatively. Pre- and postoperative COVID-19 testing and clinical symptom documentation were also abstracted from the record.

Statistical Methods

Data are presented as mean ± standard deviation or median [interquartile range (IQR)] for continuous variables and frequency (percentage) for categorical variables. For infected patients, the NSQIP surgical risk calculator was used to calculate expected mortality and LOS for comparison to actual outcomes. Univariate analyses were conducted via Firth logistic regression with Bonferroni correction for multiple comparisons to determine patient and surgery characteristics and postoperative outcomes associated with perioperative COVID-19. All analyses were performed in Stata/IC.[13]