Outcomes of gynecologic cancer surgery during the COVID-19 pandemic: an international, multicenter, prospective CovidSurg-Gynecologic Oncology Cancer study

Background The CovidSurg-Cancer Consortium aimed to explore the impact of COVID-19 in surgical patients and services for solid cancers at the start of the pandemic. The CovidSurg-Gynecologic Oncology Cancer subgroup was particularly concerned about the magnitude of adverse outcomes caused by the disrupted surgical gynecologic cancer care during the COVID-19 pandemic, which are currently unclear. Objective This study aimed to evaluate the changes in care and short-term outcomes of surgical patients with gynecologic cancers during the COVID-19 pandemic. We hypothesized that the COVID-19 pandemic had led to a delay in surgical cancer care, especially in patients who required more extensive surgery, and such delay had an impact on cancer outcomes. Study Design This was a multicenter, international, prospective cohort study. Consecutive patients with gynecologic cancers who were initially planned for nonpalliative surgery, were recruited from the date of first COVID-19-related admission in each participating center for 3 months. The follow-up period was 3 months from the time of the multidisciplinary tumor board decision to operate. The primary outcome of this analysis is the incidence of pandemic-related changes in care. The secondary outcomes included 30-day perioperative mortality and morbidity and a composite outcome of unresectable disease or disease progression, emergency surgery, and death. Results We included 3973 patients (3784 operated and 189 nonoperated) from 227 centers in 52 countries and 7 world regions who were initially planned to have cancer surgery. In 20.7% (823/3973) of the patients, the standard of care was adjusted. A significant delay (>8 weeks) was observed in 11.2% (424/3784) of patients, particularly in those with ovarian cancer (213/1355; 15.7%; P<.0001). This delay was associated with a composite of adverse outcomes, including disease progression and death (95/424; 22.4% vs 601/3360; 17.9%; P=.024) compared with those who had operations within 8 weeks of tumor board decisions. One in 13 (189/2430; 7.9%) did not receive their planned operations, in whom 1 in 20 (5/189; 2.7%) died and 1 in 5 (34/189; 18%) experienced disease progression or death within 3 months of multidisciplinary team board decision for surgery. Only 22 of the 3778 surgical patients (0.6%) acquired perioperative SARS-CoV-2 infections; they had a longer postoperative stay (median 8.5 vs 4 days; P<.0001), higher predefined surgical morbidity (14/22; 63.6% vs 717/3762; 19.1%; P<.0001) and mortality (4/22; 18.2% vs 26/3762; 0.7%; P<.0001) rates than the uninfected cohort. Conclusion One in 5 surgical patients with gynecologic cancer worldwide experienced management modifications during the COVID-19 pandemic. Significant adverse outcomes were observed in those with delayed or cancelled operations, and coordinated mitigating strategies are urgently needed.


Introduction
The COVID-19 pandemic took the entire world off guard and led numerous healthcare systems to redesign their clinical services for the reallocation of available resources and accommodate the changes in treatment priorities. 1e8 This study was conducted when little was known about the true magnitude of the virus and when effective therapeutic strategies were lacking. 2e6 Women with gynecologic cancers were among the most affected populations during the pandemic, and patients appealed for more robust provision of high standard care, even in times of crisis. 9 The pandemic has particularly challenged surgical care delivery for those with more advanced or relapsed disease, where surgery can be life-prolonging but not curative. Surgeries were being delayed or replaced by systemic or palliative care options that had previously been associated with poorer and less favorable outcomes. 10 Moreover, other COVID-19-related events such as delayed diagnosis through the repeatedly imposed lockdowns, increased perioperative complications and mortality through the active COVID-19, and exhausted healthcare resources such as theater space and intensive care capacity were also observed. 1e7,11 This study aimed to assess the early impact of the pandemic-related modifications to clinical management on the outcomes in surgical gynecologic cancer patients. We hypothesized that the COVID-19 pandemic had led to a delay in surgical cancer care, especially in patients who required more extensive surgery, and such delay had an impact on cancer outcomes. This study is an international prospective multicenter analysis of patients with gynecologic cancers treated during the pandemic.

Study design
This was a multicenter, international, prospective cohort study analyzing the clinical and surgical outcomes of patients with gynecologic cancer who had or were planned to have treatment as first line during the COVID-19 pandemic era. The study was part of the wider GlobalSurg-CovidSurg Consortium (https://globalsurg.org/ CovidSurg/)-an umbrella study encompassing all solid cancer types and aiming to explore the impact of COVID-19 in surgical patients and services across multiple specialties. 12,13 The study was prospectively registered (NCT04384926) and designed to inform clinical care as the pandemic evolves; the short study period encouraged wider participation at the time that resources were severely restricted owing to the competing needs to the pandemic.
Before the start of the study, each cancer group had the opportunity to add additional questions on the Case Report Form (CRF) to support the need to explore the impact of the COVID-19 pandemic particularly relevant to them. The gynecologic cancer group was particularly concerned about the impact of the pandemic on treatment adjustments and the subsequent clinical outcomes. The related outcomes are stated as secondary outcomes in this report to be consistent with the main CovidSurg-Cancer study.
Equivalent results of other cancer types and generic methods applying to all cancers (eg, inclusion criteria and patient identification) have been published previously. 14 This study included hospitals that included all patients discussed for surgery regardless of whether they were operated on or not.
Any hospital that performed elective cancer surgery and was affected by the COVID-19 pandemic, was eligible to participate. The participating hospitals were identified by local principal investigators. Study approvals for participating hospitals were secured by local principal investigators before entry into the study and data collection. The study protocol was either registered as a clinical audit with institutional review or a research study obtaining ethical committee approval, depending on local and national requirements. Investigators were invited to identify a start date, representing the start of the emergence of COVID-19 in their respective hospitals.
All consecutive adult (age18 years) surgical patients with multidisciplinary team (MDT) decisions supporting surgery were captured by the participating centers (local principal investigators) from that point for the next 3 months (representing the first peak period of the COVID-19 pandemic). Only centers confirming the complete inclusion and follow-up of all nonoperated patients (labeled as "Group 1 hospitals" by the CovidSurg study team) were included in analyses comparing operated and nonoperated patients. The definition of lockdown stringency and the level of COVID-19 burden areas (classified as a median of at least 25 cases per 100,000 per 14 days, representing the World Health Organization [WHO] recommendations at the time of the study) were previously described. 13,14 Patients and procedures Consecutive eligible patients were identified from MDT meetings, operating lists, and outpatient or virtual clinics. The day of surgery was defined as day zero, with patients followed up for 30 days postoperatively using routine follow-up pathways. COVID-19 diagnosis was made through nasopharyngeal swab and polymerase chain reaction, computed tomography thorax, or clinical symptoms consistent with COVID-19.
Patients who had a therapeutic operation for suspected cancer that was subsequently shown to be a preinvasive or benign lesion after histologic examination and full organ resection (eg, highgrade dysplasia or carcinoma in situ) were still included in this study. Elective surgery was defined as any surgery booked in advance of a planned admission to hospital. 15 The primary outcomes of the cross-specialty COVIDSurg study were the 30-day postoperative pulmonary complications (COVID-19 infection, pneumonia, acute respiratory distress syndrome, and unexpected ventilation), and they were previously reported. 13 Key study outcomes of the gynecooncological cohort The primary outcome of this analysis was as follows: The incidence of pandemic-related changes in care The secondary outcomes were as follows:

Statistical analysis
The study was conducted according to guidelines set by the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement for observational studies. 17 Chi-square tests were used to compare differences in categorical data apart from when cell sizes were small, when the Fisher exact tests were used. Continuous nonparametric data are presented as medians and interquartile ranges, and median differences between groups were compared using the ManneWhitney U-test. Missing data are included in summary tables when applicable. Analysis was performed used Stata SE version 16

Patients and tumor-related characteristics
We recruited a total of 3973 patients from 227 centers across 52 countries with the following geographic distribution: 2402 patients (60%) from Europe and Central Asia, 484 (12%) from Latin America & Caribbean, 376 (10%) from East Asia & Pacific, 277 (7%) from South Asia, 244 (6%) from North America, 160 (4%) from the Middle East and North Africa, and 24 (1%) from Sub-Saharan Africa (Table 1). Only 189 patients did not undergo surgery (4.8%); these were excluded from all analyses pertaining to surgical data alone.
Examining separately the surgical morbidity profile of the patients who acquired postoperative COVID-19 infections (Supplemental Table 4  We also compared the patient characteristics and outcomes for those who had a prolonged hospital stay of 14 days. The cutoff is used for the continual National Health Service England audit that required mandatory data submission for every hospital in the United Kingdom. Those patients with poorer performance status and higher ASA score, higher disease stage, higher comorbidities index, having had an open surgery, respiratory complications, bowel surgery, and low COVID-19 burden had a higher risk of staying in the hospital longer than 14 days postoperatively (Supplemental Table 5).
When evaluating the impact of the COVID-19 pandemic on the type of treatment the patients received, we demonstrated that 20.7% (823/3973) of all patients (both operated and nonoperated) had their standard of care adjusted (Table 4 and Supplementary  Table 3 (Table 5).
In those who had an operation more than 8 weeks after the initial decision, the operations were more likely to be open surgery (P<.0001), involve midabdominal surgery (P<.0001), and require bowel resections (P¼.011).
There was no significant difference in 30-day mortality (    Although treatment alterations were not associated with a significant difference in 30-day postoperative mortality, significant delay (>8 weeks) in the time to surgery was associated with a composite of adverse outcomes (Table 5), including disease progression

Results in the context of what is known
Our results suggest that the surgical morbidity profile appeared to be equivalent to the historic surgical morbidity data outside of the COVID-19 pandemic 18e20 in contrast to previous single-center data, suggesting higher morbidity rates. 11 It is plausible that the modifications of surgical approaches have influenced the overall morbidity profile during the pandemic. For example, in patients undergoing colorectal cancer surgery, 14 there was a significant increase in the overall rate of stoma formation (34.2% vs 27.2% in the prepandemic era), especially end stoma formation (70% vs 43.6% prepandemic).
Despite this study being conducted during the first wave of the COVID-19 pandemic, a low perioperative COVID-19 infection rate (0.6%) was observed. Our data appeared more favorable than the previously reported experience, 13,14,21 with a lower incidence of postoperative COVID-19 infection (0.6% vs 3.8%) and lower 30-day mortality in the non-COVID19-infected patients' cohort (0.7% vs 1.8%). Consistent with previous reports, 13,14,21 patients infected with COVID-19 had a significantly longer postoperative stay and higher surgical morbidity and mortality (63.6% and 18.2%, respectively) than uninfected patients. The cause of the more favorable outcomes in patients with gynecologic cancer than previously reported experiences is likely to be multifactorial and could be related to the overall less radical surgical procedures. Only 3.5% of our operated patients underwent bowel resections, which is a surrogate marker of surgical radicality. Nevertheless, similar to previous reports, 14 our data also confirmed that bowel resection was associated with higher risks of perioperative COVID-19 infection, other surgical morbidity, and In addition to the well-established risk factor of prolonged hospital stay, we showed that low COVID-19 burden areas were associated with significantly higher rates of increased hospitalization (>14 days). It is possible that the healthcare systems that were least affected by the pandemic continued to operate for complex patients with a high disease burden, who in turn required a longer hospital stay.
Although published recommendations highlighted expected delays in gynecologic cancer delivery, the published evidence demonstrated some variations of the impact of COVID-19. For example, Bruce et al 22 suggested that the number of referrals to gynecologic oncology decreased during the early stages of the pandemic, but the time to evaluation and treatment initiation were unaffected. In contrast, a number of surveys and retrospective cohort studies have described delays in consultations and treatments. 23e26 This study specifically evaluated the impact of the significant delay in the time to surgery on oncologic outcomes in patients with gynecologic cancers. One in 10 patients experienced significant delay (>8 weeks) to surgery ( . The results highlighted this delay disproportionately affect those who required more complex surgery, at higher anesthetic risks and in areas where resources were scarce (LMICs and areas in full lockdown with subsequently low COVID-19 burden). Consistent with a previous systematic review, 10 treatment delay was associated with poorer outcomes. There are very limited existing multicentered international studies with similar coverage of LMICs reporting detailed outcomes and complication rates of gynecologic cancer surgery. 18e20 However, they had different classifications and recording methods compared with this current study. Therefore, direct comparison was inappropriate, which limited our ability to identify any change in the morbidity and mortality rates because of the COVID-19 pandemic.
The study and follow-up periods for CovidSurg-Cancer were deliberately short to encourage wider participations and a rapid turnover of data to inform practice. Although the comparison of a retrospective cohort could aid direct comparison of management adjustments and outcomes and it would have been performed in a single-center setting, 11 the resources required to do so globally at the first peak of the pandemic could not be justified. We also acknowledge that post-MDT patient selections have contributed to the associations between surgical delay or cancellation and poorer outcomes, but the results highlight the need to target resources for those with complex care needs to ensure that they receive their planned urgent life-prolonging operations.

Clinical implications
Gynecologic cancer surgery during the early stages of the COVID-19 pandemic and before the availability of vaccination and effective treatments appeared to be safe, with a low risk of COVID-19 infection. 14 With significant contribution from LMICs (26.9%; 1067/3973), these data also provide a global snapshot of the morbidity and mortality rates experienced by patients undergoing gynecologic cancer surgery. This study identified that a significant proportion of patients had their surgical plans delayed or cancelled, which was associated with poorer outcomes. The results represent an early and clear signal that robust mechanisms and pathways are urgently needed to ensure adequate cancer care even in times of crisis without detrimental oncologic compromise. With this large data set estimating the number of treatment plan alternations globally, the results could be used to coordinate plans and allocate resources as required to rescue and salvage the detrimental effects. Yet, few health systems have established consensus agreements to date. 1,8 These results will also add to the evidence for future pandemics and the planning of relevant studies during a worldwide crisis in the future.

Research implications
The data were collected during the initial phase of the pandemic when the impingement of the virus was completely new, when none of the patients were vaccinated, and when protective measures were immature. Continued evaluation of the impact of the COVID-19 pandemic on cancer care delivery in the medium-and long-term is required. Population-based studies will be required to evaluate which patient subgroups (eg, cancer types or stage) were the most adversely affected. Comparative studies are also required to identify effective ways to mitigate the detrimental impact of the pandemic on cancer care.

Conclusions
The new rising incidence of COVID-19 worldwide owing to emerging variants combined with incomplete vaccination coverage has continued to disrupt healthcare delivery, and our results remain highly relevant. Despite the very low risk of perioperative COVID-19 infection, those infected had substantially poorer outcomes. The high rates of pandemic-related treatment plan modifications were associated with early negative impact on oncological outcomes. Robust strategies for the safe provision of surgical cancer care when the pandemic is transitioning into an endemic state are urgently needed. n