key: cord-0926113-56u8bb5o authors: Johnson, Brett A.; Waddimba, Anthony C.; Ogola, Gerald O.; Fleshman, James W.; Preskitt, John T. title: A systematic review and meta-analysis of surgery delays and survival in breast, lung and colon cancers: Implication for surgical triage during the COVID-19 pandemic date: 2020-12-08 journal: Am J Surg DOI: 10.1016/j.amjsurg.2020.12.015 sha: cb01cfcd09660f59aaa50a1e4a6652edc0af23b4 doc_id: 926113 cord_uid: 56u8bb5o BACKGROUND: Thousands of cancer surgeries were delayed during the peak of the COVID-19 pandemic. This study examines if surgical delays impact survival for breast, lung and colon cancers. METHODS: PubMed/MEDLINE, EMBASE, Cochrane Library and Web of Science were searched. Articles evaluating the relationship between delays in surgery and overall survival (OS), disease-free survival (DFS) or cancer-specific survival (CSS) were included. RESULTS: Of the 14,422 articles screened, 25 were included in the review and 18 (totaling 2,533,355 patients) were pooled for meta-analyses. Delaying surgery for 12 weeks may decrease OS in breast (HR 1.46, 95%CI 1.28–1.65), lung (HR 1.04, 95%CI 1.02–1.06) and colon (HR 1.24, 95%CI 1.12–1.38) cancers. When breast cancers were analyzed by stage, OS was decreased in stages I (HR 1.27, 95%CI 1.16–1.40) and II (HR 1.13, 95%CI 1.02–1.24) but not in stage III (HR 1.20, 95%CI 0.94–1.53). CONCLUSION: Delaying breast, lung and colon cancer surgeries during the COVID-19 pandemic may decrease survival. A systematic review and meta-analysis of surgery delays and survival in breast, lung and colon cancers: Implication for surgical triage during the COVID-19 pandemic The Coronavirus Disease 2019 (COVID-19) pandemic has indirectly threatened the health of thousands of cancer patients by disrupting their treatment schedules. 1 To conserve resources and limit the spread of the virus, many hospitals were forced to delay elective surgeries. 2, 3 Nearly 38% of cancer surgeries are estimated to have been canceled worldwide during the 12-week peak of the pandemic. 4 During this time numerous professional associations 5e7 published guidelines for triaging cancer cases. The speed at which these were created provided rapidly needed guidance to individual healthcare institutions, given that triaging of cancer surgeries in the United States (U.S.) is not a common practice. 8 However, the need for immediate guidance also limited the amount of anticipatory research that could occur before issuance of these recommendations. The majority of guidelines were based on the opinions of a small panel of experts, which resulted in discordance between recommendations. 9 Therefore, it is imperative to adduce sufficient medical evidence to update and strengthen guidelines in preparation for future waves of COVID-19. Few reviews and meta-analyses have examined the relationship between delays in surgery and survival for breast, lung and colon cancers. A meta-analysis 10 published in 1999 reported a 3 month delay in treatment for breast cancer was associated with a 12% lower rate of survival. That review did not examine delays in surgery alone but instead included all initial treatment modalities. A 2007 meta-analysis 11 found delays in surgery for colorectal cancers did not worsen survival, however, the interpretation of these results is limited since colon and rectal cancers were not evaluated separate. A systematic review 12 from 2018 examined the relationship between time-to-surgery (TTS) for colon cancer and outcomes, however no meta-analysis was performed. Two reviews 13,14 evaluated treatment delays in lung cancer, but neither examined surgery-specific delays. Although numerous studies have examined the impact that increasing TTS has on survival in cancer patients, there is no evidence-based standard that can serve as an empirical benchmark for when a cancer surgery should be considered delayed. This leads to inconsistent results between studies and potential confusion in efforts to determine the true impact of delaying surgery. We addressed this gap in the literature by conducting a systematic review and meta-analysis, whose primary objective was to evaluate if delaying surgery by 12 weeks impacts survival for breast, lung, and colon cancers. This study was conducted following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 15 and the Meta-analysis of Observational Studies in Epidemiology (MOOSE) 16 guidelines. Structured searches of the PubMed/MED-LINE, EMBASE, Cochrane Library and Web of Science databases were conducted (April 2020) to identify relevant articles. The defined search strategy, available in Appendix A, yielded a preliminary collection of articles. Filters were used to limit results to English language and full-text papers published prior to April 1, 2020. Abstracts and full-texts of articles with relevant titles were reviewed. Additionally, references of included publications were examined to identify studies not captured in the preliminary search. Assessment of eligibility was conducted independently by the first author (B.A.J.) and studies in question were further evaluated by a second reviewer (A.C.W). Eligible studies fulfilled the following inclusion criteria: (1) the majority of the sample population diagnosed with invasive ductal or lobular breast cancer, non-small cell lung cancer or adenocarcinoma of the colon; (2) surgery was the initial treatment; (3) surgical delays analyzed separate from all other treatment delays; (4) TTS interval clearly defined as diagnosis to surgery; and (5) appropriate reporting of outcomes (discussed below). To minimize the effect of confounding variables between comparison groups, only studies which fulfilled the following additional criteria were included in the meta-analyses: (1) relevant prognostic factors identified; and (2) published findings incorporated multivariate risk adjustment for relevant prognostic factors or balanced such covariates between comparison groups. 17 Studies were excluded for any of the following: (1) abstracts only and review articles; (2) sample sizes <100; (3) non-invasive tumors included in analysis; (4) the majority of the patient population analyzed had metastatic (stage IV) disease; (5) multiple cancer types included in analysis; or (6) duplicate study populations. Included studies were evaluated using the Newcastle-Ottawa Scale (NOS) which identifies the quality of nonrandomized studies. The scale ranges from 0 to 9 with a score of 7 indicating a high quality study. 18 Descriptions of the categories used to evaluate each study are in Appendix A. The primary endpoint was overall survival (OS). Secondary endpoints included cancer-specific survival (CSS) and disease-free survival (DFS). For inclusion into this paper, outcomes must be reported as hazard ratios (HRs), odds ratios (ORs), risk ratios (RRs) or as p-values. Studies that reported HRs with 95% CIs for OS were considered for inclusion in the meta-analyses. Additional variables extracted from individual studies included: years of diagnosis of sample population; country study took place; source of data; inclusion criteria; sample sizes; mean/median TTS; mean/median follow-up, ages of sample population; any adjuvant therapy received; and prognostic factors controlled for. HRs were converted to represent a 12-week delay in surgery and then pooled using the method described in prior meta-analyses. 19e22 A random-effects model was fitted given the prospect of moderate to high heterogeneity. 23 Heterogeneity was quantified by calculating I 2 and a sensitivity analysis further evaluated the impact each study had on the overall heterogeneity. 24e26 Publication bias was assessed via visual inspection of funnel plots and the Egger's regression test for analyses with 10 or more studies. 27 Statistical significance was defined as a p-value <0.05 and all tests were 2sided. Statistical analyses were performed using SAS® software version 9.4 (SAS, Inc., Cary, NC) and R® software version 3.6.0 (R Development Core Team, Vienna, Austria). After 14,422 publications were screened, a total of 186 abstracts/ full texts were evaluated. Ultimately, 25 observational studies met inclusion criteria for the review, of which 18 studies, consisting of 2,533,355 patients, met inclusion criteria for meta-analysis. Of the seven studies excluded from the meta-analyses, six 28e33 did not report HRs and/or 95% CIs for OS and one 34 contributed an overall weight of <0.0%. The inclusion process is illustrated by the flow diagram in Appendix A. Years of publication of included studies ranged from 2002 to 2020 and sample sizes ranged from 398 to 683,604 patients. A total of 23 studies examined OS, 5 examined DFS and 5 examined CSS. A summary of findings from each study is listed in Table 1 and further study characteristics are available in Appendix A. The mean score on the Newcastle-Ottawa Scale was 7.8 for all included studies and 7.9 for only the studies included in the metaanalyses. Among the studies included in the meta-analyses, 10 did not have an average follow-up of at least 5 years, 3 did not control for tumor stage and 3 had sample populations not representative of the total population at risk. Little to no asymmetry was observed in the funnel plots for lung and colon cancers, indicating a low risk of publication bias. Although minor asymmetry existed on the breast cancer funnel plot, the value for Egger's test was 0.11, indicating non-significant asymmetry. Detailed results of the Newcastle-Ottawa Scale and the funnel plots are available in Appendix A. OS was reported in twelve breast cancer studies; including eleven analyses of non-stage specific (multiple stage) disease 29,35e44 ; four of stage I only disease 39 Results of the meta-analysis for non-stage specific breast cancers are illustrated as a forest plot in Fig. 1 . The results suggest a delay in surgery of 12 weeks is associated with decreased OS (HR 1.46, 95%CI 1.28e1.65, I 2 86%). Results of the stage-specific breast cancer meta-analyses are illustrated as forest plots in Fig. 2 . The results suggest a delay in surgery of 12 weeks is associated with decreased OS for stages I (HR 1.27, 95%CI 1.16e1.40, I 2 97%) and II (HR 1.13, 95%CI 1.02e1.24, I 2 89%) diseases but not for stage III (HR 1.20, 95%CI 0.94e1.53, I 2 63%) disease. The sensitivity analysis for studies of non-stage specific breast cancers indicates excluding Polverini et al. 42 39 decreases overall heterogeneity from I 2 ¼ 63% to 0% and the overall HR remains insignificant. OS was reported in seven lung cancer studies; including three analyses of non-stage specific (multiple stages) disease 32, 35, 36 ; four of stage I only disease 31, 34, 45, 46 ; and one of stage II only disease. 45 Sample sizes ranged from 398 to 242,444 with a mean of 40,798 patients. Increasing TTS was associated with decreased OS in 1/3 non-stage specific studies, 4/4 stage I only studies, and 1/1 stage II only studies. Results of the meta-analysis for lung cancer are illustrated as a forest plot in Fig. 3 . The results suggest a delay in surgery of 12 weeks is associated with decreased OS (HR 1.04, 95%CI 1.02e1.06, I 2 84%). Too few HRs were reported to perform meaningful stagespecific analyses. In the sensitivity analysis for lung cancer all studies were found to contribute equally to overall heterogeneity. OS was reported in 8 colon cancer studies 33,36,47e52 and all analyses were of non-stage specific (multiple stages) disease. Sample sizes ranged from 458 to 118,504 with a mean of 19,874 patients. Increasing TTS was associated with decreased OS in 3/8 studies. Results of the meta-analysis for colon cancer are illustrated as a forest plot in Fig. 4 . The results suggest a delay in surgery of 12 weeks is associated with decreased OS (HR 1.24, 95%CI, 1.12e1.38, I 2 71% ). The sensitivity analysis for studies of colon cancer indicates excluding Grass et al. 47 decreases overall heterogeneity from The lack of evidence-based standards for what is considered a delay in cancer surgery has led to inconsistent study designs and few attempts at synthesizing data between studies. By converting HRs to a common 12-week delay in surgery, for the first time we were able to pool findings from studies evaluating surgical delays and survival in breast, lung and colon cancers. The results of the meta-analyses indicate a surgical delay of 12 weeks is associated with decreased overall survival in these three cancer types. Results from the stage-specific analyses for breast cancer suggest delaying surgery by 12 weeks decreases overall survival for stage I and II diseases but not stage III disease. Findings from this study should be utilized to empirically strengthen and modify existing triage guidelines in preparation for future waves of COVID-19. The pooled HR for breast cancer was the largest among the three cancer types examined, which may indicate survival in these patients is especially sensitive to surgical delays. The stage-specific analyses performed for breast cancer are of particular interest since COVID-19 triage guidelines differ by tumor stage. Most guidelines recommend delaying early-stage breast cancers and to instead consider neoadjuvant treatment with chemotherapy or endocrine therapy. However, our results suggest these cancers are most impacted by delays in surgery. Although alternative treatments exist, it is important to consider that under normal circumstances chemotherapy is not given for many early-stage (especially stage I) breast cancers. 53 While the current literature does suggest endocrine therapy can safely be utilized in the neoadjuvant setting, the appropriate patient population and the exact treatment regimen are not yet well defined. 54 For stage III, the impact of surgical delays trended toward worse survival, however, significance was not reached. Compared to stages I/II, it is possible delaying surgery for stage III disease has a negligible impact on survival since these patients already experience significantly poorer outcomes from their 'delay' in diagnosis. Therefore, our results suggest surgeries for stage III breast cancers should be delayed prior to stages I/II during future waves of COVID-19. If delays for stages I/II become necessary, surgeries should first be delayed for patients who likely would receive adjuvant chemotherapy under normal circumstances, given the well-established efficacy of neoadjuvant chemotherapy. The pooled HR for lung cancer suggests a delay in surgery of 12 weeks is associated with a slight decrease in overall survival. It is noteworthy that 4/5 of the studies in our analysis included only stage I and/or II diseases. While this limits the comparability of the Fig. 3 . Meta-analysis of the estimated hazard ratio for overall survival for a delay in surgery of 12 Weeks among patients with lung cancer. Fig. 4 . Meta-analysis of the estimated hazard ratio for overall survival for a delay in surgery of 12 Weeks among patients with colon cancer. Legend (Figs. 1, 2, 3, 4) The size of the boxes represents the weight the study has in the overall analysis; the black lines coming off of each box represent the 95% confidence intervals of the study; confidence interval lines that cross the line of no effect indicate results are not significant; the center of the diamond(s) represents the overall HR and the edges represent the 95% confidence interval; Abbreviations: OS, overall survival; HR, hazard ration; CI, confidence interval. results to the triage recommendations for stage III disease, our analysis suggest early-stage lung cancers are sensitive to delays in surgery. Many of the guidelines recommend delaying small and early stage lung cancers and to consider alternative non-surgical treatments such as stereotactic ablative radiotherapy, cryotherapy or radiofrequency ablation. 5, 7 However, curative surgery is often the single treatment for early-stage lung cancer 53 and the efficacy of alternative treatments has been shown to be inconsistent. 55e57 Our results suggest if delaying surgery for stages I/II lung cancers becomes necessary during future waves of COVID-19 it should be done with caution and rescheduled for the earliest possible date. The pooled HR for colon cancer suggests a delay in surgery of 12 weeks is associated with decreased overall survival. Most triage guidelines discourage delaying curative-intent surgery for colon cancer, which our findings support. 5, 6 If it becomes necessary to delay surgery for colon cancer during future waves of COVID-19 our results suggest that it should be done with extreme caution and rescheduled for the earliest possible date. The most substantial limitation to this paper is the heterogeneity found between pooled studies, most notably for breast and lung cancers. However, results of the meta-analyses remained similar following the sensitivity analyses. Second, since individual studies did not compare consistent TTS lengths, it was necessary to standardize each time interval in order to compare findings. To accomplish this, we relied on the assumption a log-linear relationship exists between the effect of TTS and OS. Third, even though the overall assessment of quality indicates the inclusion of highquality studies, there were numerous studies which did not control for stage and/or did not have an appropriate length of followup. Fourth, only full-text articles in English were included which may cause publication bias, although the funnel plots/Egger's test indicate the risk for bias was low. The decision to only include fulltext articles was made after a cursory reading of seemingly relevant abstracts revealed that upon their full-text reviews, studies often fell short of our strict inclusion criteria. 58 Lastly, our analysis is limited to non-randomized observational studies, however, the prospect of randomizing cancer patients to different categories of surgical delays is unlikely due to the ethical implications. The results of the meta-analyses suggest a delay in surgery of 12 weeks is associated with decreased overall survival in breast (especially for stage I and II), lung and colon cancers. Triage guidelines for cancer surgeries during the COVID-19 pandemic should take into consideration the emerging medical evidence adduced by this and other similar studies. Future research would benefit from the synchronization of study designs, implementation of a constant TTS interval and use of similar TTS comparison groups in order to improve comparability of findings. To help better guide surgical triage recommendations during future waves of COVID-19, studies should further evaluate which stages and subtypes are most impacted by delays in surgery in these and other cancer types. Covid-19 d implications for the health care system COVID-19: Recommendations for Management of Elective Surgical Procedures CMS Releases Recommendations on Adult Elective Surgeries Elective surgery cancellations due to the COVID-19 pandemic: global predictive modelling to inform surgical recovery plans Elective Case Triage Guidelines for Surgical Care Management of cancer surgery cases during the COVID-19 pandemic: considerations COVID-19 guidance for triage of operations for thoracic malignancies: a consensus statement from thoracic surgery outcomes research network How to risk-stratify elective surgery during the COVID-19 pandemic? Discordance of COVID-19 guidelines for patients with cancer: a systematic review Influence of delay on survival in patients with breast cancer: a systematic review Relationship of diagnostic and therapeutic delay with survival in colorectal cancer: a review G€ ogenur I. The effect of time from diagnosis to surgery on oncological outcomes in patients undergoing surgery for colon cancer: a systematic review How long is too long? A scoping review of health system delays in lung cancer Is increased time to diagnosis and treatment in symptomatic cancer associated with poorer outcomes? Systematic review Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement Meta-analysis of observational studies in epidemiology: a proposal for reporting Methodological issues in systematic reviews and meta-analyses of observational studies in orthopaedic research Critical evaluation of the Newcastle-Ottawa scale for the assessment of the quality of nonrandomized studies in meta-analyses Association between time to initiation of adjuvant chemotherapy and survival in colorectal cancer: a systematic review and meta-analysis Association between delayed initiation of adjuvant CMF or anthracycline-based chemotherapy and survival in breast cancer: a systematic review and meta-analysis Survival and time to initiation of adjuvant chemotherapy among breast cancer patients: a systematic review and metaanalysis Association between adjuvant chemotherapy duration and survival among patients with stage II and III colon cancer: a systematic review and meta-analysis Meta-analysis in clinical trials revisited Modern Epidemiology Measuring inconsistency in meta-analyses Sensitivity of between-study heterogeneity in meta-analysis: proposed metrics and empirical evaluation Bias in meta-analysis detected by a simple, graphical test Timing of breast cancer surgery-how much does it matter? Delay of treatment initiation does not adversely affect survival outcome in breast cancer Effect of interval to definitive breast surgery on clinical presentation and survival in early-stage invasive breast cancer Effects of delayed surgical resection on shortterm and long-term outcomes in clinical stage I non-small cell lung cancer Influence of delays on survival in the surgical treatment of bronchogenic carcinoma. Lung Canc Treatment delay in surgically-treated colon cancer: does it affect outcomes? Delayed surgery after histologic or radiologic-diagnosed clinical stage I lung adenocarcinoma Delay to curative surgery greater than 12 weeks is associated with increased mortality in patients with colorectal and breast cancer but not lung or thyroid cancer The influence of hospital volume and surgical treatment delay on long-term survival after cancer surgery Time to surgery and the impact of delay in the non-neoadjuvant setting on triple-negative breast cancers and other phenotypes Impact of delayed treatment in women diagnosed with breast cancer: a population-based study Time-to-surgery and overall survival after breast cancer diagnosis in a universal health system Time from breast cancer diagnosis to therapeutic surgery and breast cancer prognosis: a population-based cohort study Time to surgery and breast cancer survival in the United States Time to treatment: measuring quality breast cancer care Delay in surgical treatment and survival after breast cancer diagnosis in young women by race/ethnicity The association of waiting times from diagnosis to surgery with survival in women with localised breast cancer in England Time to initial cancer treatment in the United States and association with survival over time: an observational study Impact of timing of lobectomy on survival for clinical stage IA lung squamous cell carcinoma Impact of delay to surgery on survival in stage I-III colon cancer Delay to colectomy and survival for patients diagnosed with colon cancer Association between the time to surgery and survival among patients with colon cancer: a population-based study Do moderate surgical treatment delays influence survival in colon cancer? Influence of delays to nonemergent colon cancer surgery on operative mortality, disease-specific survival and overall survival Therapeutic delay reduces survival of rectal cancer but not of colonic cancer National Cancer Data Base Neoadjuvant endocrine therapy for estrogen receptor-positive breast cancer: a systematic review and meta-analysis A systematic review and meta-analysis of stereotactic body radiation therapy versus surgery for patients with nonesmall cell lung cancer Survival of primary sbrt compared to surgery for operable stage i/ii non-small cell lung cancer Non-small cell lung cancer: current treatment and future advances Issues relating to study design and risk of bias when including non-randomized studies in systematic reviews on the effects of interventions None. Supplementary data related to this article can be found at https://doi.org/10.1016/j.amjsurg.2020.12.015. Study concept and design e Johnson, Preskitt, Fleshman, Waddimba. Acquisition of data e Johnson, Waddimba, Ogola. Analysis and interpretation of data e All authors. Drafting of manuscript e Johnson, Waddimba, Preskitt, Fleshman. Critical revision of the manuscript for intellectual content e All authors. Final approval of version to be published e All authors. Accountable for all aspects of the work e Johnson. This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Mr. Johnson, Dr. Waddimba, Dr. Ogola, Dr. Fleshman and Dr. Preskitt do not have proprietary or commercial interest in any product mentioned or concept discussed in this article to report.