key: cord-0854522-cgs20s1b
authors: Jonker, Pascal KC.; van der Plas, Willemijn Y.; Steinkamp, Pieter J.; Poelstra, Ralph; Emous, Marloes; van der Meij, Wout; Thunnissen, Floris; Bierman, Wouter FW.; Struys, Michel MRF.; de Reuver, Philip R.; de Vries, Jean-Paul PM.; Kruijff, Schelto
title: Perioperative SARS-CoV-2 infections increase mortality, pulmonary complications and thromboembolic events: a Dutch multicenter matched-cohort clinical study
date: 2020-09-24
journal: Surgery
DOI: 10.1016/j.surg.2020.09.022
sha: ddd133c1c1197294eecd4c6ff0ce53090dc7770b
doc_id: 854522
cord_uid: cgs20s1b

BACKGROUND: A direct comparison of SARS-CoV-2 positive patients with a SARS-CoV-2 negative control group undergoing an operative intervention during the current pandemic is lacking, and a reliable estimate of the assumed difference in morbidity and mortality between both patient categories remains unknown. METHODS: We included all consecutive patients with a confirmed pre- or postoperative SARS-CoV-2 positive status (operated in 27 hospitals) and negative controls (operated in 4 hospitals) undergoing emergency or elective operations. A propensity score-matched comparison of clinical outcomes was performed between SARS-CoV-2 positive and negative tested patients (control group). Primary outcome was overall 30-day mortality rate between both groups. Main secondary outcomes were overall, pulmonary, and thromboembolic complications. RESULTS: In total, 161 SARS-CoV-2 positive and 342 control SARS-CoV-2 negative patients were included in this study. The 30-day overall postoperative mortality rate was greater in the SARS-CoV-2 positive cohort compared to the negative control group (16% vs. 4% respectively; p=0.007). After propensity score-matching, the SARS-CoV-2 positive group consisted of 123 patients (median 70 years of age [IQR 59-77] and 55% male) were compared to 196 patients in the matched control group (median 69 years (IQR 58 – 75] and 53% male). The 30-day mortality rate and risk were greater in the SARS-CoV-2 positive group compared to the matched control group (12% versus 4%, p=0.009 and OR 3.4 [95%CI 1.5 – 8.5], p=0.005, respectively) . Overall, pulmonary and thromboembolic complications occurred more often in SARS-CoV-2 positive patients (p<0.01). CONCLUSIONS: Patients diagnosed with perioperative SARS-CoV-2 have an increased risk of 30-day mortality, pulmonary complications, and thromboembolic events. These findings serve as an evidence-based argument to postpone elective surgery and selected emergency cases.

from Italy reported a 20% 30-day postoperative overall-mortality rate (OR 9.5, 95% CI 1.8 -96.5) and a greater 72 risk of pulmonary complications (OR 35.6, ) and thromboembolic complications (OR 13.2, 73 95% CI 1.5 -∞) in 41 SARS-CoV-2 positive patients compared to a mainly historic control cohort. 7 Although of 74 relevance, this single-center study has limitations due to the small sample of SARS-CoV-2 positive patients and 75 the overall design resulting in inaccurate estimates. It is therefore, important to provide clinicians with more 76 accurate data to improve perioperative clinical decision-making for this patient category during the foreseen new 77 waves of SARS-CoV-2 infections. This current multicenter, nationwide, matched-cohort study compares the 30-78 day postoperative morbidity and mortality rates between SARS-CoV-2 positive and negative patients undergoing 79 elective or emergency operations in hospitals within a similar health care system and standardized surgical 80 guidelines. Results of this study will provide a more reliable insight into the actual difference in overall mortality University, Nashville, United States). 9 Pseudo anonymized data was entered manually from the electronic patient 118 file in the electronic case report form by local researchers of the respective participating centers and cross-119 checked for inconsistencies and missing data by the coordinating researchers prior to data locking. Per site, a de-120 identification key was stored in an on-site, secured, digital data storage area. Center-specific data were accessible 121 by local researchers and the coordinating researchers from the UMCG (SK, PKCJ, WYvdP, PJS, JPPMdV).

Prior to analysis, parameters were checked for completion per case, and data were curated by coordinating analyses. The χ2 test was used for categorical data, and logistic modeling was used for calculating odds ratios 145 with 95% confidence intervals (CIs). Propensity score-matching was used to help control for differences at 146 baseline between those patients undergoing an operation who had a preoperative SARS-CoV-2 infection or who 

Assessment of the covariate balance was done by comparing standardized differences before and after matching.

The analysis of the primary and secondary outcomes was done by using the matched sample. Multiple Before propensity matching, the 30-day overall postoperative mortality rate in SARS-CoV-2 positive and 211 negative patients was 16% and 4%, respectively (p=0.007; Table 3 ). In the propensity score-matched cohort, 30-212 day overall-mortality was associated with an OR of 3.4 (95% confidence interval 1.5 -8.5) for patients with a 

An overview of the complications for matched and unmatched cohorts is provided in Table 3. Supplemental   227   Table 3 gives a detailed description of the diagnosed pulmonary complications and thromboembolic events. 

Comparing outcomes of surgery between hospitals within a uniform health care system allows for accurate 272 assessment of differences in morbidity and mortality between SARS-CoV-2 positive and negative patients. We 273 used propensity-score-matching in an attempt to account for a wide variety of baseline differences. Our study 274 still, however, has limitations we need to address. Despite the propensity matching, it is still possible that some 

COVID-19 Map -Johns Hopkins Coronavirus Resource Center Available

Recommendations for general surgery activities in a pandemic 342 scenario (SARS-CoV-2)

Elective surgery cancellations due to the COVID-19 pandemic: global 344 predictive modelling to inform surgical recovery plans

Mortality and pulmonary complications in patients undergoing 12

Classification of Surgical Complications A New Proposal 369

With Evaluation in a Cohort of 6336 Patients and Results of a Survey

Little RJA, Rubin DB. Statistical analysis with missing data. 375 16. van Buuren S, Groothuis-Oudshoorn K. mice: Multivariate imputation by chained equations in R

Nonparametric preprocessing for parametric causal inference

Sex, Demographics (COVID-19) -Worldometer Available from

Confirmed SARS-CoV-2 positive cases per 1,000 people