Similarities and differences between study designs in short‐ and long‐term outcomes of laparoscopic versus open low anterior resection for rectal cancer: A systematic review and meta‐analysis of randomized, case‐matched, and cohort studies

Abstract Aim Randomized controlled trials (RCT) are the gold standard in surgical research, and case‐matched studies, such as studies with propensity score matching, are expected to serve as an alternative to RCT. Both study designs have been used to investigate the potential superiority of laparoscopic surgery to open surgery for rectal cancer, but it remains unclear whether there are any differences in the findings obtained using these study designs. We aimed to examine similarities and differences between findings from different study designs regarding laparoscopic surgery for rectal cancer. Methods Systematic review and meta‐analyses. A comprehensive literature search was conducted using PubMed, Scopus, and Cochrane. RCT, case‐matched studies, and cohort studies comparing laparoscopic low anterior resection and open low anterior resection for rectal cancer were included. In total, 8 short‐term outcomes and 3 long‐term outcomes were assessed. Meta‐analysis was conducted stratified by study design using a random‐effects model. Results Thirty‐five studies were included in this review. Findings did not differ between RCT and case‐matched studies for most outcomes. However, the estimated treatment effect was largest in cohort studies, intermediate in case‐matched studies, and smallest in RCT for overall postoperative complications and 3‐year local recurrence. Conclusion Findings from case‐matched studies were similar to those from RCT in laparoscopic low anterior resection for rectal cancer. However, findings from case‐matched studies were sometimes intermediate between those of RCT and unadjusted cohort studies, and case‐matched studies and cohort studies have a potential to overestimate the treatment effect compared with RCT.


| INTRODUC TI ON
Observational studies inherently include confounding factors that can affect their results, and several differences have been noted between the findings of randomized controlled trials (RCT) and those of observational studies 1 . For adequate comparison of interventions in observational studies, matching methods are frequently used in surgical research and the most common method is propensity score matching 2 . First introduced by Rosenbaum et al, propensity score matching is expected to be an alternative to RCT 3,4 , which are currently considered the gold standard in surgical research investigating the treatment effect of an intervention. Although it may be favorable to conduct RCT, this is often difficult to do for practical and ethical reasons in surgical research 2 . In addition, RCT generally require substantial resources including time, money, and collaboration among diverse specialists in order to ensure patient safety, data correctness, and standardization of interventions.
Recently, nationwide databases such as the National Clinical Database (NCD) and the National Database (NDB) have become available, despite several restrictions on their use 5,6 . If findings from case-matched studies are similar to those from RCT, case-matched studies using a nationwide database may supersede RCT. There are methodological differences between RCT and case-matched studies such as patient selection and adjustment for confounding factors 7,8 .
However, it remains unclear whether there are differences in findings between RCT and case-matched studies.
We aimed to investigate similarities and differences in findings between RCT, case-matched studies, and cohort studies regarding laparoscopic low anterior resection (LAR) for rectal cancer. LAR is defined as a procedure representing the performance of surgery in NCD 5 . These study designs have been used to investigate the potential superiority of laparoscopic LAR to open LAR for rectal cancer, which is of major interest to surgeons.

| ME THODS
We conducted a systematic review and meta-analyses.

| Eligibility
Studies in which laparoscopic LAR was compared with open LAR for rectal cancer were eligible. When multiple surgical procedures were included in a study, studies in which over 70% of patients underwent LAR were included. Small studies that included less than 50 patients for each intervention group were excluded. Study design was restricted to RCT, case-matched study, or cohort study. Both prospective and retrospective studies were included, and the method of randomization or matching was not restricted. The language was restricted to English.

| Outcome measures
Short-term outcomes were the incidence of postoperative overall complications, the incidence of anastomotic leakage, mortality, reoperation rate, length of stay, operative time, estimated blood loss, and rate of positive circumferential resection margins. Long-term outcomes were 3-year overall survival (OS), 3-year disease-free survival (DFS), and 3-year local recurrence rate (LRR).

| Literature search and study selection
A comprehensive literature search was conducted on June 12, 2019, using PubMed, Scopus, and Cochrane Central Register of Controlled Trials. The search terms used were "rectal cancer", "anterior resection", "laparoscopy", "open", and related terms (Appendix S1).
Duplications were excluded by checking the names of study authors, publication year, and study characteristics such as study design, setting, and period. Two review authors (NH and YF) independently screened the titles and abstracts of studies identified by literature search, and then assessed the full texts of potential eligible articles. Disagreement was resolved by discussion.

| Data extraction
The same authors (NH and YF) also independently extracted data from the included studies; the data included study design and setting, number and characteristics of patients, surgical procedure, and short-and long-term outcomes. Each double-checked the extracted data for the other, and any discordance between them was resolved by discussion. For cohort studies, unadjusted data were extracted to assess the results without adjusting for confounding factors.

| Statistical analysis
Data synthesis was carried out using Review Manager 5.3 (Cochrane Collaboration Software, Nordic Cochrane Centre). A random-effects model was used for all meta-analyses because of presumed heterogeneity in the surgical quality of LAR across the included studies.
The Mantel-Haenszel method was used for dichotomous variables and inverse-variance weighting was applied for continuous variables. Risk ratio (RR) with 95% confidence interval (CI) was used for dichotomous variables in a meta-analysis. Risk difference (RD) was applied instead of RR when a rare outcome was assessed. Mean difference (MD) with 95% CI was used for continuous variables when a single measure was included in a meta-analysis. The median with range was converted to mean with standard difference (SD) by the method of Hozo et al. 9 A two-sided P-value <.05 was considered statistically significant.

| Characteristics of included studies
The comprehensive literature search identified 3229 articles. Of these, 1368 duplications were removed. Screening was conducted for 1861 articles by checking the titles and abstracts for the potential to be included in this review. After screening, the full text of 106 articles was checked to assess whether they met the inclusion criteria. Finally, 35 studies (40 articles) were included in this review ( Figure 1)  . Nine articles were reported from 4 RCT and were treated as 4 studies [10][11][12][13][14][15][16][17][18] .
Included studies comprised 5 RCT, 10 case-matched studies, and 20 cohort studies. Four studies were conducted internationally and the remaining 31 were reported from 11 countries. One case-matched study was prospective and the remaining 9 were retrospective. Among cohort studies, 1 was prospective and 19 studies were retrospective (Table 1).   Figure S1).

| Three-year OS
A total of 20 studies with 9157 patients including 3 RCT with 1834 patients, 6 case-matched studies with 2956 patients, and 11 cohort studies with 4367 patients reported on the 3-year OS and were included in a meta-analysis stratified by study design.   Figure S8).  Figure S9).  Figure 3).

This review was supported by a grant from Kondou Kinen Medical
Foundation.

D I SCLOS U R E
Funding: This review was supported by a grant from Kondou Kinen Medical Foundation.
Conflicts of interest: The authors declare no conflicts of interest for this study.
Author Contributions: All authors contributed to the study concept and design. Literature search and data collection were performed by NH and YF. Statistical analysis was conducted by NH and checked by the other authors. The first draft of manuscript was written by NH and all authors commented on previous version of the manuscript. All authors read and approved the final.