Browse the corpus

Editor—During the early coronavirus disease 2019 (COVID-19) pandemic, there was nationwide guidance from institutions and policymakers in the USA to postpone elective surgeries,1 , 2 leading to a significant decrease surgical volume in March 2020.3 However, the global impact of the early pandemic on postoperative mortality remains unexplored.

coronavirus disease 2019 (COVID-19) pandemic, there was nationwide guidance from institutions and policymakers in the USA to postpone elective surgeries,1 , 2 leading to a significant decrease surgical volume in March 2020.3 However, the global impact of the early pandemic on postoperative mortality remains unexplored. We performed this retrospective cohort study among adult patients undergoing surgery between January 6, 2020 and June 28, 2020 across US hospitals in the Multicenter Perioperative Outcomes Group (MPOG) database, to explore the association between the early phase of the pandemic with postoperative mortality (full methods in Supplemental Methods). The primary endpoint was 30 day mortality. A mixed-effect Cox proportional hazard model was used to analyse the risk of 30 day mortality after surgery. Patients discharged alive before day 30 or still alive at 30 days after surgery were censored. The Cox model was adjusted for age, sex, ASA physical status, emergency status, hospital admission status, comorbidities derived from the preoperative score to predict postoperative mortality4 (cardiac arrhythmia, congestive heart failure, peripheral vascular disease, paralysis, chronic pulmonary disease, alcohol abuse, cancer, diabetes mellitus, renal failure), BMI, surgical subspecialty, and pandemic period (period of a three-phase pandemic model as described below). A random effect for the institution ID and week number and an interaction term between the pandemic period and the COVID-19 regional severity quartile was included in the model. Surgical case volume trends across all included MPOG institutions were identified using joinpoint regression. The identified joinpoints were used to split the study into three periods of surgical volumes: surgeries performed from January 6, 2020 through March 15, 2020 were defined as ‘pre-pandemic’; surgeries performed during the nadir surgical volume period from March 16, 2020 through April 5, 2020 were defined as ‘nadir pandemic’; and surgeries performed from April 6, 2020 through June 28, 2020 were defined as ‘restart pandemic’ (Supplementary Fig. S1). States in the USA were distributed into quartiles based on cumulative COVID-19 mortality per capita on June 30, 2020 to assess regional COVID-19 severity (Supplementary Fig. S2).

nadir pandemic’; and surgeries performed from April 6, 2020 through June 28, 2020 were defined as ‘restart pandemic’ (Supplementary Fig. S1). States in the USA were distributed into quartiles based on cumulative COVID-19 mortality per capita on June 30, 2020 to assess regional COVID-19 severity (Supplementary Fig. S2). A total of 618 251 surgeries across 36 MPOG institutions were analysed. The average weekly surgical case volume in the cohort was 35 876 cases per week at baseline during the week of January 6, 2020 and decreased the week of March 16, 2020, before increasing again the week of April 6, 2020. There was heterogeneity in weekly surgical volume decline from baseline to nadir between institutions, with the most severely affected institution having a 99% decrease from baseline surgical volume and the least severely affected institution having a 49% decrease from baseline surgical volume (Supplementary Fig. S3).

There was heterogeneity in weekly surgical volume decline from baseline to nadir between institutions, with the most severely affected institution having a 99% decrease from baseline surgical volume and the least severely affected institution having a 49% decrease from baseline surgical volume (Supplementary Fig. S3). When compared with pre-pandemic surgical patients, nadir pandemic surgical patients tended to be younger (53 [17] yr vs 56 [18] yr, P<0.001), less likely to be outpatient (39.1% vs 57.1%, P<0.001), and more likely to be undergoing emergency surgery (8.8% vs 4.2%, P<0.001) (Table 1 ). They were more likely to have comorbid conditions, including arrhythmia, coagulopathy, congestive heart failure, hypertension, and renal failure, and had higher 30 day mortality (hazard ratio, 1.75; 95% confidence interval [CI], 1.61–1.93). After adjustment, there was no difference in 30 day surgical mortality during the nadir pandemic period when compared with the pre-pandemic period (hazard ratio, 1.09; 95% CI, 0.77–1.41), nor was there a significant difference between pandemic nadir and restart periods. There was an increase in 30 day surgical mortality at hospitals from COVID-19 regional severity quartiles Q4, Q3, and Q1 during the nadir pandemic period when compared with the pre-pandemic period, with the greatest increase at hospitals in the highest COVID-19 regional severity quartile (30 day surgical mortality hazard ratio, 2.18 [95% CI, 1.94–2.45] in Q4; 1.43 [1.07–1.93] in Q3; 1.13 [0.90–1.43] in Q2; and 1.42 [1.13–1.80] in Q1). After risk adjustment, there was no significant difference in 30 day surgical mortality at hospitals within any of the COVID-19 severity quartiles during the nadir pandemic period when compared with the pre-pandemic period (Supplementary Table S1).Table 1Patient characteristics and outcomes across 36 US hospitals in the pre-pandemic, nadir pandemic, and restart pandemic periods.

30 day surgical mortality at hospitals within any of the COVID-19 severity quartiles during the nadir pandemic period when compared with the pre-pandemic period (Supplementary Table S1).Table 1Patient characteristics and outcomes across 36 US hospitals in the pre-pandemic, nadir pandemic, and restart pandemic periods. CI, confidence interval; HR, hazard ratio; sd, standard deviation.Table 1CharacteristicsPre-pandemic (340 505)Nadir pandemic (38 833)Nadir vs pre-pandemic P valueRestart pandemic (238 913)Restart vs pre-pandemic P valueAge, yr (sd)55.8 (17.3)53.3 (18.3)<0.00154.2 (18.0)<0.001Race (%) African-American48 758 (14.3%)5778 (14.9%)0.00335 321 (14.8%)<0.001 White247 183 (72.6%)27 967 (72.0%)0.02173 246 (72.5%)0.51 Other/unknown44 564 (13.1%)5088 (13.1%)0.9430 346 (12.7)<0.001Sex, female (%)192 157 (56.4%)22 384 (57.6%)<0.001135 193 (56.6%)0.27Outpatient (%)194 308 (57.1%)15 195 (39.1%)<0.001113 921 (47.7%)<0.001Emergency case (%)14 134 (4.2%)3399 (8.8%)<0.00116 104 (6.7%)<0.001ASA physical status 1–2161 643 (47.5%)13 853 (39.6%)<0.001104 691 (43.9%)<0.001 3–5177 381 (52.1%)23 166 (59.7%)<0.001132 865 (55.6%)<0.001Patient comorbidities, number (%) Arrhythmia49 778 (14.6%)7507 (19.3%)<0.00140 620 (17.0%)<0.001 Coagulopathy15 521 (4.6%)2944 (7.6%)<0.00115 597 (6.5%)<0.001 Diabetes mellitus, complicated13 595 (4.0%)2181 (5.6%)<0.00111 479 (4.8%)<0.001 Heart failure23 472 (6.9%)3640 (9.4%)<0.00119 763 (8.3%)<0.001 Hypertension, complicated32 370 (9.5%)5067 (13.0%)<0.00126 834 (11.2%)<0.001 Metastatic cancer16 723 (4.9%)3425 (8.8%)<0.00114 129 (5.9%)<0.001 Peripheral vascular disease20 076 (5.9%)3211 (8.3%)<0.00117 072 (7.1%)<0.001 Renal failure28 128 (8.3%)4461 (11.5%)<0.00123 661 (9.9%)<0.00130-day mortality (%)2817 (0.8%)561 (1.4%)<0.0012559 (1.1%)<0.001HR unadjusted 30-day mortality (95% CI)11.75 (1.61–1.93)1.30 (1.23–1.37)HR risk-adjusted 30-day mortality (95% CI)11.09 (0.77–1.41)0.95 (0.74–1.15)

76 (5.9%)3211 (8.3%)<0.00117 072 (7.1%)<0.001 Renal failure28 128 (8.3%)4461 (11.5%)<0.00123 661 (9.9%)<0.00130-day mortality (%)2817 (0.8%)561 (1.4%)<0.0012559 (1.1%)<0.001HR unadjusted 30-day mortality (95% CI)11.75 (1.61–1.93)1.30 (1.23–1.37)HR risk-adjusted 30-day mortality (95% CI)11.09 (0.77–1.41)0.95 (0.74–1.15) Patient characteristics and outcomes across 36 US hospitals in the pre-pandemic, nadir pandemic, and restart pandemic periods. CI, confidence interval; HR, hazard ratio; sd, standard deviation. Previous studies have shown worse postoperative outcomes in individual patients with COVID-19 infection,5 , 6 even in settings with low overall incidence of perioperative COVID-19 infection.7 , 8 However, there has been little evidence on the outcome of patients undergoing surgery during the early phase of the pandemic regardless of the infective status. The lack of association of early pandemic COVID-19 regional severity with adjusted surgical mortality in our study would suggest that despite any health system limitations in perioperative staff, intensive care unit beds, ventilators, medications, testing, or risk of viral transmission, outcomes of patients undergoing surgery were not compromised even in states with the highest COVID-19 severity.

ted surgical mortality in our study would suggest that despite any health system limitations in perioperative staff, intensive care unit beds, ventilators, medications, testing, or risk of viral transmission, outcomes of patients undergoing surgery were not compromised even in states with the highest COVID-19 severity. Our study has limitations. Owing to the need to preserve anonymity of the MPOG institutions meeting inclusion criteria of this study, it was not possible to ascertain the COVID-19 mortality rate within individual institutions being studied, or their precise geography; as such, some granularity as to COVID-19 regional severity was compromised by only associating institutional surgical volumes and patient outcomes with national quartiles of COVID-19 mortality rates. COVID-19 severity in this study was evaluated based on June 30, 2020 COVID-19 statistics. It is possible that this assessment of COVID-19 severity does not correlate completely with early COVID-19 regional severity. Furthermore, only mortality was explored as an outcome, whereas other potentially meaningful outcomes (e.g. postoperative infections) were not explored. Although the MPOG consortium includes a mix of private and public hospitals from across the USA, we had an over-representation of centres from states most severely impacted. Furthermore, county hospitals might be underrepresented.

other potentially meaningful outcomes (e.g. postoperative infections) were not explored. Although the MPOG consortium includes a mix of private and public hospitals from across the USA, we had an over-representation of centres from states most severely impacted. Furthermore, county hospitals might be underrepresented. To conclude, patients undergoing surgery during the early phase of the pandemic had a greater burden of comorbidities, were more frequently undergoing emergency surgeries and had higher 30 day mortality, with the highest surgical mortality noted at hospitals with highest COVID-19 regional severity. However, after adjusting for patient risk factors, surgical mortality was not different across regions of varying COVID-19 regional severity, suggesting that this finding was related to an inherently sicker patient population undergoing higher-risk surgeries, rather than impact from regional COVID-19 severity per se. Taken together, these findings support that the quality of care and safety of anaesthesia, surgery and perioperative medicine were maintained during the early phase of the pandemic.