Effects of pre-operative isolation on postoperative pulmonary complications after elective surgery: an international prospective cohort study

Original Article Effects of pre-operative isolation on postoperative pulmonary complications after elective surgery: an international prospective cohort study COVIDSurg Collaborative* and GlobalSurg Collaborative* NIHR Global Health Research Unit on Global Surgery, Birmingham, UK Summary We aimed to determine the impact of pre-operative isolation on postoperative pulmonary complications after elective surgery during the global SARS-CoV-2 pandemic. We performed an international prospective cohort study including patients undergoing elective surgery in October 2020. Isolation was defined as the period before surgery during which patients did not leave their house or receive visitors from outside their household. The primary outcome was postoperative pulmonary complications, adjusted in multivariable models for measured confounders. Pre-defined sub-group analyses were performed for the primary outcome. A total of 96,454 patients from 114 countries were included and overall, 26,948 (27.9%) patients isolated before surgery. Postoperative pulmonary complications were recorded in 1947 (2.0%) patients of which 227 (11.7%) were associated with SARS-CoV-2 infection. Patients who isolated pre-operatively were older, had more respiratory comorbidities and were more commonly from areas of high SARS-CoV-2 incidence and high-income countries. Although the overall rates of postoperative pulmonary complications were similar in those that isolated and those that did not (2.1% vs 2.0%, respectively), isolation was associated with higher rates of postoperative pulmonary complications after adjustment (adjusted OR 1.20, 95%CI 1.05–1.36, p =0.005). Sensitivity analyses revealed no further differences when patients were categorised by: pre-operative testing; use of COVID-19-free pathways; or community SARS-CoV-2 prevalence. The rate of postoperative pulmonary complications increased with periods of isolation longer than 3 days, with an OR (95%CI) at 4–7 days or ≥8 days of 1.25 (1.04– 1.48), p =0.015 and 1.31 (1.11–1.55), p =0.001, respectively. Isolation before elective surgery might be associated with a small but clinically important increased risk of postoperative pulmonary complications. Longer periods of isolation showed no reduction in the risk of postoperative pulmonary complications. These findings have significant implications for global provision of elective surgical care. ................................................................................................................................................................. Correspondence to: J. Simoes Email: [email protected] Accepted: 14 July 2021 Keywords: SARS-Cov-2; COVID-19; surgery; pathways; pre-operative isolation *Collaborating authors are listed in online Supporting Information Appendix S1. This article is accompanied by an editorial by Charlesworth & Grossman. Anaesthesia 2021; 76: 1439–41. Twitter: @CovidSurg; @GlobalSurg Introduction Several strategies have been explored to mitigate against the risk of peri-operative SARS-CoV-2 infection, given the high associated rate of postoperative pulmonary complications and mortality [1]. It has become clear that a range of measures are needed to ensure safe surgery, including: COVID-19-free surgical pathways; patient testing for SARS-CoV-2; and delaying surgery in patients with SARS-CoV-2 infection [2, 3]. These measures will still be needed despite the roll-out of vaccination 1454 ©2021 Association of Anaesthetists Anaesthesia 2021, 76, 1454–1464 doi:10.1111/anae.15560 13652044, 2021, 11, Downloaded from https://associationofanaesthetists-publications.onlinelibrary.wiley.com/doi/10.1111/anae.15560 by Universidad De Sevilla, Wiley Online Library on [16/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License programmes, which may take years to achieve globally, be less effective against SARS-CoV-2 variants and not achieve universal implementation [4, 5]. Isolation before elective surgery has been recommended by several national surgical associations [6– 8]. This attempts to reduce the risk of asymptomatic carriers undergoing surgery, thereby protecting individual patients and reducing in-hospital transmission to other patients and staff. It presents potential problems for patients, including: logistical considerations; reducing patient mobility before major surgery; and social isolation. It also means that lastminute additions to operating theatre lists are less likely, thereby representing an additional potential burden to surgical recovery plans. These limitations would be acceptable if there was clear evidence of benefit in regard to reduction of postoperative complications, both related to SARS-CoV-2 and otherwise. Demonstrating benefit and optimum duration will support wider rollout of global best practice in elective surgery. Demonstrating no benefit will allow units to tailor clinical guidance and consider reducing the burden of isolation. We aimed to determine the impact of pre-operative isolation on postoperative pulmonary complications after elective surgery. Methods This was a planned sub-study of an international prospective cohort study of patients undergoing surgery in hospitals, regardless of local SARS-CoV-2 infection prevalence or isolation policies during at least one 7-day study period in October 2020. Full methodology has been previously reported [9]. Hospitals providing surgical care for patients under any surgical specialty were eligible to participate. Local and national approvals were obtained according to local regulations. In the UK, this study was registered as clinical audit or service evaluation at each participating site. Patient consent was obtained when demanded by local governance requirements, which was not required in the UK. The study was registered prospectively and we adhered to the strengthening of the reporting of observational studies in epidemiology (STROBE) statement for observational studies [10]. All consecutive patients undergoing elective surgery for any indication were included. Elective surgery was defined as any procedure routinely performed in an operating theatre by a surgeon during a planned admission. A list of excluded procedures (such as central line insertion) was provided to all collaborators (see online Supporting Information Table S1). Patients undergoing emergency surgery or who had a pre-operative SARS-CoV-2 diagnosis were not included. Pre-operative isolation was defined as limitation of social contacts before surgery during which time patients stayed at home, avoided public spaces and transport and did not receive visitors from outside of their household. Patients who were not required to follow any social distancing or isolation measures before surgery were considered to have not been exposed to pre-operative isolation. The number of days that each patient was required to follow the isolation measures was recorded as the duration of isolation. Patients were categorised according to the duration of pre-operative isolation into three groups: up to 3 days; 4 to 7 days; or ≥8 days before their planned surgery. Multiple variables were collected for each patient: age; sex; ASA physical status; revised cardiac risk index (RCRI); presence of respiratory comorbidities; indication for surgery; and grade of surgery. Consistent with previous analyses, age was categorised as <70 or ≥70 y. Countrylevel income was defined according to the World Bank index classification (updated in 2019) based on gross national income per capita, which determines three income groups: high income; upper-middle income; lower-middle income (including patients from both low- and lowermiddle income countries). Pre-operative testing was defined as a reverse transcription-polymerase chain reaction (RT-PCR) or antigen swab test performed within 3 days before surgery. A patient was considered to have undergone surgery in a COVID-19-free surgical pathway if a completely segregated pathway was provided (including ward, operating theatre and critical care areas) or the hospital was not admitting patients with SARS-CoV-2 infection. If otherwise, the surgical pathway was recorded as non-segregated. Community SARS-CoV-2 prevalence was determined for each patient based on the 14-day case notification rate at the time of surgery in each participating country. These rates were extracted from the World Health Organization, European Centre for Disease Control, US Centre for Disease Control and specific national registries via the Our World in Data platform [11]. Hospitals were classified as being in communities with either a low (<100 cases per 100,000 population) or high (≥100 cases per 100,000 population) SARS-CoV-2 prevalence. The primary outcome of this study was postoperative pulmonary complications, defined as pneumonia, acute respiratory distress syndrome or unexpected mechanical ventilation within 30 days after surgery. The main secondary outcome was 30-day mortality. Other secondary outcomes ©2021 Association of Anaesthetists 1455 COVIDSurg Collaborative and GlobalSurg Collaborative | Isolation and postoperative pulmonary complications Anaesthesia 2021, 76, 1454–1464 13652044, 2021, 11, Downloaded from https://associationofanaesthetists-publications.onlinelibrary.wiley.com/doi/10.1111/anae.15560 by Universidad De Sevilla, Wiley Online Library on [16/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License were postoperative SARS-CoV-2 infection and SARS-CoV-2 postoperative pulmonary complications within 30 days for surgery. Postoperative SARS-CoV-2 diagnosis was based on any one of the following criteria: positive RT-PCR nasopharyngeal swab; positive rapid antigen test; chest computed tomography (CT) scan showing changes consistent with pneumonitis secondary to SARS-CoV-2 infection; positive immunoglobulin-G or immunoglobulin- M antibody test showing active infection; or clinical diagnosis (in the absence of negative RT-PCR swab results). A SARS-CoV-2 postoperative pulmonary complication was defined as a pulmonary complication within 30 days of surgery in a patient with a postoperative SARS-CoV-2 infection (see online Supporting Information Appendix S2). Descriptive results are reported with absolute and proportional frequencies. Chi-square tests were used to assess unadjusted differences between groups. Multivariable logistic regression models were performed to test for an independent effect of pre-operative isolation on postoperative pulmonary complications and mortality, adjusting for patient, surgical and local setting factors. The results of the adjusted models were summarised using univariable and multivariable OR (95%CI). The statistical significance threshold used was p <0.05. Missing data were reported in all tables and figures. Statistical analyses were performed using R studio V 3.6.1 (RStudio, Boston, USA) and the packages dplyr, tidyverse, gmodels, finalfit, ggplot2 and forestplot. The adjusted model for the primary outcome was repeated in key sub-groups to determine the effect of preoperative isolation in patients: with high vs low ASA physical status; in settings where pre-operative testing was or was not available; where COVID-19-free surgical pathways were or were not established; and where the community SARS- CoV-2 prevalence was high or low. A sensitivity analysis was performed to explore the impact of duration of isolation on postoperative pulmonary complications, comparing patients who were isolated for: up to 3 days; 4 to 7 days; and ≥8 days before surgery. An interaction term was used in adjusted models to account for the fact that isolation was likely to be less common in low- and middle-income countries where there is a proven higher rate of postoperative complications and mortality [12]. Results A total of 96,454 patients were included from 1634 hospitals in 114 countries. There were 62,839 (65.1%) patients who underwent surgery in areas with high SARS-CoV-2 prevalence and 33,615 (34.9%) patients in areas of low prevalence. The study included 65,228 (67.6%) patients from high-income countries and 31,226 (32.4%) patients from low- and middle-income countries. There were 74,347 (77.1%) patients who underwent surgery for a benign condition and 57,079 (59.2%) who underwent major surgery (see online Supporting Information Table S2). Overall, 26,948 (27.9%) patients isolated before surgery, 80,200 (83.1%) underwent surgery in a COVID-19-free pathway and 67,612 (70.1%) had a pre-operative SARS-CoV-2 test. Patients who isolated before surgery were older, had more respiratory comorbidities and higher ASA physical status. Pre-operative isolation was more common in areas of high SARS-CoV-2 prevalence and in high-income countries (Table 1). Patients who isolated pre-operatively were also more frequently tested for SARS-CoV-2 before surgery and underwent surgery in a COVID-19-free surgical pathway more often (see online Supporting Information Figure S1). The overall rate of postoperative pulmonary complications was reported in 1947 (2.0%) and postoperative mortality was reported in 648 (0.7%). Of all the postoperative pulmonary complications, 227 (11.7%) occurred in patients with a concomitant postoperative SARS-CoV-2 infection. Compared with patients who had isolated before surgery, those who did not isolate had similar postoperative pulmonary complication rates (2.1% vs 2.0%), lower mortality (0.4% vs 0.8%) and similar SARS-CoV- 2 infection rates (0.8% vs 0.8%) (Fig. 1). After adjustment for measured confounders, isolation was associated with higher postoperative pulmonary complication rates (adjusted OR 1.20, 95%CI 1.05–1.36, p =0.005) (Fig. 2) but with no significant difference in postoperative mortality (adjusted OR 0.80, 95%CI 0.62–1.02, p =0.081) (see online Supporting Information Table S3). Multiple factors were independently associated with increased postoperative pulmonary complication rates, including: age >70 y; male sex; high ASA physical status; previous respiratory comorbidities; cancer surgery; and major surgical procedures (Fig. 2). Pre-operative testing was associated with reduced postoperative pulmonary complication rates (OR 0.81,95%CI 0.73–0.89, p <0.001) (online Supporting Information Table S4). Isolation was not associated with reduced postoperative pulmonary complications across any of the predefined sub-groups (Fig. 3). Pre-operative isolation was associated with higher postoperative pulmonary complication rates in: patients with ASA physical status 1–2; those undergoing surgery in a high SARS-CoV-2 prevalence area; a COVID-19-free pathway; or those who did not undergo pre-operative testing (see online Supporting Information Tables S5-S12). 1456 ©2021 Association of Anaesthetists Anaesthesia 2021, 76, 1454–1464 COVIDSurg Collaborative and GlobalSurg Collaborative | Isolation and postoperative pulmonary complications 13652044, 2021, 11, Downloaded from https://associationofanaesthetists-publications.onlinelibrary.wiley.com/doi/10.1111/anae.15560 by Universidad De Sevilla, Wiley Online Library on [16/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Table 1 Baseline characteristics of patients who isolated and those who did not isolate before surgery. Values are number (proportion). Proportions are expressed as per column total. No isolation n=69,436 Isolation n=26,948 p value Age <70 y 56,593 (81.5%) 20,129 (74.7%) <0.001 ≥70 y 12842 (18.5%) 6819 (25.3%) Missing 1 0 Sex Female 36,609 (52.7%) 14,370 (53.3%) 0.096 Male 32,824 (47.3%) 12,578 (46.7%) Missing 3 0 ASA physical status 1–2 53,452 (77.0%) 20,275 (75.3%) <0.001 3–5 15,973 (23.0%) 6654 (24.7%) Missing 11 19 Revised cardiac risk index 0 32,045 (46.2%) 12,401 (46.0%) <0.001 1–2 35,484 (51.1%) 13,950 (51.8%) ≥3 1881 (2.7%) 583 (2.2%) Missing 26 14 Respiratory comorbidities No 63,651 (91.7%) 23,676 (87.9%) <0.001 Yes 5740 (8.3%) 3250 (12.1%) Missing 45 22 Surgical indication Benign 54,148 (78.0%) 20,143 (74.8%) <0.001 Cancer 15,287 (22.0%) 6803 (25.2%) Missing 1 2 Surgery grade Minor/intermediate 27,716 (39.9%) 11,611 (43.1%) <0.001 Major 41,703 (60.1%) 15,336 (56.9%) Missing 17 1 Surgery area Abdominopelvic 34,202 (50.2%) 12,717 (48.2%) <0.001 Thoracic 3188 (4.7%) 1317 (5.0%) Head and neck 7276 (10.7%) 2727 (10.3%) Limb 9498 (13.9%) 4533 (17.2%) Other 13,930 (20.5%) 5111 (19.4%) Missing 18 1 Anaesthesia Local 6525 (9.4%) 2789 (10.4%) <0.001 Regional 11,172 (16.1%) 3492 (13.0%) General 51,712 (74.5%) 20,655 (76.7%) Missing 27 12 SARS-CoV-2 community prevalence Low 30,282 (43.6%) 3326 (12.3%) <0.001 High 39,154 (56.4%) 23,622 (87.7%) (continued) ©2021 Association of Anaesthetists 1457 COVIDSurg Collaborative and GlobalSurg Collaborative | Isolation and postoperative pulmonary complications Anaesthesia 2021, 76, 1454–1464 13652044, 2021, 11, Downloaded from https://associationofanaesthetists-publications.onlinelibrary.wiley.com/doi/10.1111/anae.15560 by Universidad De Sevilla, Wiley Online Library on [16/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Of patients who isolated before surgery, 6971 (26.0%) isolated for up to 3 days, 10,691 (39.9%) for 4 to 7 days, and 9164 (34.2%) for ≥8 days (see online Supporting Information Table S13). Patients isolating for longer durations were progressively older and had a greater comorbid burden, including respiratory comorbidities. A duration of isolation of up to 3 days before surgery did not show a difference in postoperative pulmonary complication rates compared with no isolation (OR 0.90, 95%CI 0.70–1.13, p =0.377). Patients isolating for longer durations had higher rates of postoperative pulmonary complications when isolating for 4 to 7 days (OR 1.25,95% CI 1.04–1.48, p =0.015) and for ≥8 days (OR 1.31, 95%CI 1.11–1.55, p =0.001) (Table 2). Discussion Pre-operative isolation was associated with a small but clinically important increase in postoperative pulmonary complications. A sensitivity analysis looking at the duration of pre-operative isolation found that patients who isolated for longer periods had higher rates of postoperative pulmonary complications. Although there was a difference in mortality rates, the numbers were small and the difference was not statistically significant following adjustment. There were no clinically relevant differences in SARS-CoV-2 positivity or SARS-CoV-2-related postoperative pulmonary complication rates. We also showed no benefit with pre-operative isolation: in areas of different SARS-CoV- 2 community prevalence; when pre-operative testing was Table 1 (continued) No isolation n=69,436 Isolation n=26,948 p value Country income High 42,025 (60.5%) 23,166 (86.0%) <0.001 Upper middle 13,192 (19.0%) 2979 (11.1%) Low/lowmiddle 14,219 (20.5%) 803 (3.0%) Figure 1 Flowchart of patient inclusion, with postoperative outcomes by isolation group. SARS-CoV-2 associated postoperative pulmonary complications were defined as the presence of postoperative pulmonary complications (PPC) in patients with a postoperative diagnosis of SARS-CoV-2 infection. Missing data for the presented variables: pre-operative isolation n =70; postoperative pulmonary complications n =64; postoperative mortality n =55. 1458 ©2021 Association of Anaesthetists Anaesthesia 2021, 76, 1454–1464 COVIDSurg Collaborative and GlobalSurg Collaborative | Isolation and postoperative pulmonary complications 13652044, 2021, 11, Downloaded from https://associationofanaesthetists-publications.onlinelibrary.wiley.com/doi/10.1111/anae.15560 by Universidad De Sevilla, Wiley Online Library on [16/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Figure 2 Multivariable logistic regression model exploring the association between pre-operative isolation and postoperative pulmonary complications, adjusting for patient, surgery and surgical setting factors. Number in dataframe =96,454; number in model =96,067; missing =387; AIC =16,680.6; C-statistic =0.784. Full model presented in online Supporting Information Table S4, including an interaction term of isolation and country income. Community prevalence of SARS-CoV-2 was defined as the median 14-day cumulative country case notification rate per 100,000 population during October 2020. Country income groups defined as per the World Bank classification. Figure 3 Forest plot of the adjusted odds ratio (95%CI) for the effect of isolation in postoperative pulmonary complications across patient sub-groups for ASA physical status, pre-operative testing, COVID-19-free pathways and community prevalence. Odds ratios are adjusted for the same variables used in the main model, including an interaction term of isolation and country income. ©2021 Association of Anaesthetists 1459 COVIDSurg Collaborative and GlobalSurg Collaborative | Isolation and postoperative pulmonary complications Anaesthesia 2021, 76, 1454–1464 13652044, 2021, 11, Downloaded from https://associationofanaesthetists-publications.onlinelibrary.wiley.com/doi/10.1111/anae.15560 by Universidad De Sevilla, Wiley Online Library on [16/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Table 2 Multivariable logistic regression model exploring the association between pre-operative isolation length and postoperative pulmonary complications, adjusting for patient, surgery and surgical setting factors. Community incidence of SARS-CoV- 2 was defined as the median 14-day cumulative country case notification rate per 100,000 population during October 2020. Values are number (proportion). No postoperative pulmonary complications Postoperative pulmonary complication(s) Univariable OR 95%CI p value Multivariable OR 95%CI p value Duration of isolation No isolation 67,993 (98.0%) 1386 (2.0%) Reference –Reference – ≤3 days 6877 (98.7%) 91 (1.3%) 0.69 (0.54–0.87) 0.002 0.90 (0.70–1.13) 0.377 4–7 days 10,483 (98.1%) 206 (1.9%) 0.99 (0.83–1.17) 0.887 1.25 (1.04–1.48) 0.015 ≥8 days 8910 (97.2%) 252 (2.8%) 1.51 (1.29–1.77) <0.001 1.31 (1.11–1.55) 0.001 Age <70 y 75,397 (98.3%) 1328 (1.7%) Reference –Reference – ≥70 y 19,045 (96.9%) 619 (3.1%) 1.85 (1.67–2.03) <0.001 1.22 (1.09–1.36) <0.001 Sex Female 50,218 (98.5%) 783 (1.5%) Reference –Reference – Male 44,222 (97.4%) 1164 (2.6%) 1.69 (1.54–1.85) <0.001 1.53 (1.39–1.68) <0.001 ASA physical status 1–2 72,837 (98.8%) 884 (1.2%) Reference –Reference – 3–5 21,566 (95.3%) 1063 (4.7%) 4.06 (3.71–4.45) <0.001 2.33 (2.09–2.60) <0.001 Revised cardiac risk index 0 44,029 (99.0%) 425 (1.0%) Reference –Reference – 1–2 48,109 (97.3%) 1314 (2.7%) 2.83 (2.54–3.16) <0.001 1.52 (1.30–1.77) <0.001 ≥3 2258 (91.7%) 205 (8.3%) 9.41 (7.91–11.15) <0.001 2.44 (1.95–3.03) <0.001 Respiratory comorbidities No 85,760 (98.2%) 1560 (1.8%) Reference –Reference – Yes 8609 (95.7%) 384 (4.3%) 2.45 (2.19–2.74) <0.001 1.67 (1.48–1.89) <0.001 Surgical indication Benign 73,141 (98.5%) 1151 (1.5%) Reference –Reference – Cancer 21,299 (96.4%) 796 (3.6%) 2.37 (2.17–2.60) <0.001 1.70 (1.54–1.88) <0.001 Surgery grade Minor/intermediate 39,019 (99.2%) 306 (0.8%) Reference –Reference – Major 55,409 (97.1%) 1638 (2.9%) 3.77 (3.34–4.27) <0.001 2.15 (1.88–2.46) <0.001 Anaesthesia Local 9261 (99.5%) 48 (0.5%) Reference –Reference – Regional 14,524 (99.0%) 154 (1.0%) 2.05 (1.49–2.86) <0.001 1.68 (1.20–2.40) 0.003 General 70,620 (97.6%) 1745 (2.4%) 4.77 (3.62–6.44) <0.001 2.49 (1.86–3.43) <0.001 Surgical area Abdominopelvic 47,310 (98.1%) 921 (1.9%) Reference –Reference – Thoracic 10,196 (98.4%) 168 (1.6%) 0.85 (0.71–1.00) 0.049 1.20 (0.98–1.45) 0.069 Head and neck 13,883 (98.8%) 167 (1.2%) 0.62 (0.52–0.73) <0.001 1.08 (0.89–1.30) 0.448 Limb 18,909 (98.8%) 226 (1.2%) 0.61 (0.53–0.71) <0.001 1.04 (0.87–1.24) 0.671 Other 4129 (89.9%) 462 (10.1%) 5.75 (5.11–6.45) <0.001 2.73 (2.40–3.10) <0.001 Pre-operative SARS–CoV–2 testing No test 39,375 (97.8%) 871 (2.2%) Reference –Reference – Test 54,957 (98.1%) 1066 (1.9%) 0.88 (0.80–0.96) 0.004 0.82 (0.74–0.91) <0.001 (continued) 1460 ©2021 Association of Anaesthetists Anaesthesia 2021, 76, 1454–1464 COVIDSurg Collaborative and GlobalSurg Collaborative | Isolation and postoperative pulmonary complications 13652044, 2021, 11, Downloaded from https://associationofanaesthetists-publications.onlinelibrary.wiley.com/doi/10.1111/anae.15560 by Universidad De Sevilla, Wiley Online Library on [16/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License implemented; or when COVID-19-free surgical pathways were in place. Although this study did not directly identify causes for an association between isolation and increased postoperative pulmonary complication rates, it provides an opportunity to generate hypotheses. Isolation is associated with decreased physical activity, worse nutritional habits and higher levels of anxiety and depression [13–16]. These effects in already vulnerable patients may have contributed to an increased risk of pulmonary complications. Further, there is increasing evidence demonstrating that prehabilitation before surgery improves patient recovery and outcomes [17–19]. It is possible that isolation may have, therefore, conversely led to patient deconditioning and functional decline [20], adversely influencing patient outcomes. Although patients who isolated were slightly more comorbid than those who did not, these findings remained present after adjustment. Our evidence suggests that removing pre-operative isolation strategies is unlikely to lead to worse postoperative outcomes for patients, but institutions should monitor their postoperative pulmonary complication rates as strategies evolve. Whereas unadjusted mortality rates appeared lower for patients who isolated, this did not remain significant after adjustment. Combined with small numbers, a clear benefit of isolation on 30-day mortality cannot be drawn from our data. There is a known interaction between mortality, failure to rescue patients from surgical complications and lowerincome surgical settings [12, 21]. Although such an interaction is likely to have contributed to this finding, it was beyond the scope of this predefined analysis to explore further, especially in the context of small event numbers and the likelihood of further unmeasured confounders. However, benefits of pre-operative isolation are not only for the individual patient but also to other patients and staff in hospitals who are at risk from asymptomatic carriers of SARS-CoV-2. Although this study included all patients operated upon during the same week in each specialty and hospital, it was not designed to capture cross infection in surgical wards, as only individual level outcomes were collected. Social isolation can have a system-wide benefitin preventing the admission of patients incubating SARS-CoV- 2 that was not captured by our study. Further, a systems risk of undiagnosed SARS-CoV-2 infection in institutions that operate COVID-19-free surgical pathways is the risk of individual patients affecting these pathways and other patients within them. For example, a COVID-19-free ward could become a COVID-19-positive ward if one patient tests positive, thereby undermining the entire pathway. Finally, symptomatic screening and pre-operative testing are likely Table 2 (continued) No postoperative pulmonary complications Postoperative pulmonary complication(s) Univariable OR 95%CI p value Multivariable OR 95%CI p value Surgical pathway Non-segregated 15,830 (97.8%) 354 (2.2%) Reference –Reference – COVID-19-free 78,559 (98.0%) 1587 (2.0%) 0.90 (0.81–1.02) 0.087 0.99 (0.87–1.12) 0.828 SARS-CoV-2 community prevalence Low 32,876 (97.9%) 698 (2.1%) Reference –Reference – High 61,567 (98.0%) 1249 (2.0%) 0.96 (0.87–1.05) 0.341 1.08 (0.95–1.22) 0.235 Country income* High 64,024 (98.2%) 1190 (1.8%) Reference –Reference – Upper middle 15,740 (97.3%) 438 (2.7%) 1.47 (1.29–1.67) <0.001 1.99 (1.72–2.29) <0.001 Lower middle/low 14,679 (97.9%) 319 (2.1%) 1.19 (1.04–1.37) 0.010 2.03 (1.71–2.40) <0.001 Interaction term Isolated ≤3 days, LMIC ––1.71 (0.59–3.91) 0.256 1.54 (0.53–3.57) 0.366 Isolated ≤3 days, UMIC ––0.93 (0.41–1.84) 0.847 0.77 (0.34–1.56) 0.505 Isolated 4–7 days, LMIC ––0.62 (0.28–1.20) 0.197 0.52 (0.23–1.03) 0.087 Isolated 4–7 days, UMIC ––0.89 (0.63–1.26) 0.531 0.80 (0.55–1.13) 0.212 Isolated ≥8 days, LMIC ––1.23 (0.43–2.76) 0.660 0.82 (0.28–1.89) 0.682 Isolated ≥8 days, UMIC ––1.28 (0.86–1.86) 0.213 0.82 (0.55–1.21) 0.340 *Country income groups defined as per the World Bank classification. LMIC, low-to-middle income country; UMIC, upper-to-middle income country. ©2021 Association of Anaesthetists 1461 COVIDSurg Collaborative and GlobalSurg Collaborative | Isolation and postoperative pulmonary complications Anaesthesia 2021, 76, 1454–1464 13652044, 2021, 11, Downloaded from https://associationofanaesthetists-publications.onlinelibrary.wiley.com/doi/10.1111/anae.15560 by Universidad De Sevilla, Wiley Online Library on [16/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License to remain key components of elective surgery admissions processes to prevent nosocomial infection by SARS-CoV-2 [22]. Our study has limitations. Firstly, postoperative SARS- CoV-2 rates were similar in both groups, suggesting that pre-operative isolation is not effective in reducing nosocomial SARS-CoV-2 infection. However, we did not include patients who isolated and then tested positive who may have had their surgery delayed or cancelled. This might have underestimated the SARS-CoV-2 incidence and postoperative pulmonary complication rates in patients who did not isolate. Secondly, although a definition of preoperative isolation was stated in the study protocol, slightly different strategies could have been reported as preoperative isolation. Patient compliance with isolation recommendations was not measured, which could have contributed to the underestimation of the benefits of isolation. However, the large numbers and heterogeneous sites contribute to a pragmatic study design and generalisability of our conclusions. Thirdly, although adjustment was performed for all the available variables, there might be residual confounding that affected results. We addressed this through multiple sensitivity analyses, in which the findings were consistent. Finally, community SARS-CoV-2 prevalence was collected from the most reliable sources available, but we acknowledge that they might be inaccurate in some settings, influencing the adjusted analysis [23, 24]. These were assessed at a national level, possibly lacking the granularity of regional variation within countries. Healthcare providers may wish to take these findings into consideration when reviewing local and national guidance. Relaxation of pre-operative isolation policies appears to be safe for individual patients, especially in the presence of pre-operative testing, which this and previous studies showed to be beneficial [2]. Selected isolation practices may remain in place in certain conditions (such as high-risk patients and periods of high community prevalence). Further research is needed to explore the most effective method of maintaining patient fitness and conditioning for patients that are isolating, which may include home or remote prehabilitation [25, 26]. Postoperative pulmonary complications related to SARS- CoV-2 accounted for only a small proportion of observed postoperative pulmonary complications. Going forward, research to target other causes of postoperative pneumonias is needed, since endemic pathogens may be the more common organisms. This study demonstrates that patient isolation before elective surgery might be associated with a small increased risk in postoperative pulmonary complications. Longer periods of isolation did not reduce the risk of postoperative pulmonary complications. These findings have implications for the global provision of elective surgery. Acknowledgements This trial was registered at clinicaltrials.gov (NCT04509986). Funding was provided by the: National Institute for Health Research (NIHR) Global Health Research Unit; Association of Coloproctology of Great Britain and Ireland; Bowel and Cancer Research; Bowel Disease Research Foundation; Association of Upper Gastrointestinal Surgeons; British Association of Surgical Oncology; British Gynaecological Cancer Society; European Society of Coloproctology; Medtronic, NIHR Academy; Sarcoma UK; Urology Foundation; Vascular Society for Great Britain and Ireland; and Yorkshire Cancer Research. The views expressed are those of the authors and not necessarily those of the funding partners. No other competing interests declared. References 1. COVIDSurg Collaborative. Mortality and pulmonary complications in patients undergoing surgery with perioperative SARS-CoV-2 infection: an international cohort study. Lancet 2020; 396:27–38. 2. COVIDSurg Collaborative. Preoperative nasopharyngeal swab testing and postoperative pulmonary complications in patients undergoing elective surgery during the SARS-CoV-2 pandemic. Bristish Journal of Surgery 2021; 108:88–96. 3. COVIDSurg Collaborative. Elective cancer surgery in covid-19- free surgical pathways during the SARS-CoV-2 pandemic: an international, multicenter, comparative cohort study. Journal of Clinical Oncology 2021; 39:66–78. 4. Sah P, Vilches TN, Moghadas SM, et al. Accelerated vaccine rollout is imperative to mitigate highly transmissible COVID-19 variants. EClinicalMedicine 2021; 35: 100865. 5. Burki TK. Challenges in the rollout of COVID-19 vaccines worldwide. Lancet Respiratory Medicine 2021; 9: e42–3. 6. National Institute for Health and Care Excellence. COVID-19 rapid guideline: arranging planned care in hospitals and diagnostic services. 27/07/2020. www.nice.org.uk/guidance/ ng179 (accessed 29/06/2021). 7. Associacion Espanola de Cirurgia Mayor Ambulatoria. Recomendaciones para la reapertura de las unidades de CMA durante el periodo de transicion de la pandemia por el COVID- 19 (SARS-CoV-2). 02/09/2020. https://www.asecma.org/doc umentos/Recomendaciones-apertura-CMA-tras-Covid.pdf (accessed 29/06/2021). 8. Amaral C, Orfao R, Lanca F, Azenha M. Recomendacoes para retoma da atividade cirurgica electiva apos condicionamento pela pandemia COVID-19. http://www.spanestesiologia.pt/ webstspa/wp-content/uploads/2020/05/recomendacoes-de- anestesiologia-obstetrica_final_04052020.pdf (accessed 29/ 06/2021). 9. COVIDSurg Collaborative, GlobalSurg Collaborative. Timing of surgery following SARS-CoV-2 infection: an international prospective cohort study. Anaesthesia 2021; 76: 748–58. 10. Elm EV, Altman DG, Egger M, et al. Strengthening the reporting of observational studies in epidemiology (STROBE) statement: 1462 ©2021 Association of Anaesthetists Anaesthesia 2021, 76, 1454–1464 COVIDSurg Collaborative and GlobalSurg Collaborative | Isolation and postoperative pulmonary complications 13652044, 2021, 11, Downloaded from https://associationofanaesthetists-publications.onlinelibrary.wiley.com/doi/10.1111/anae.15560 by Universidad De Sevilla, Wiley Online Library on [16/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License