Page 904 - Read Online

Park et al. Mini-invasive Surg 2020;4:87 I http://dx.doi.org/10.20517/2574-1225.2020.87 Page 3 of 16

Inclusion and exclusion criteria
The following inclusion criteria were prerequisite to be included in the meta-analysis: (1) direct comparison
of the pre-determined outcomes of IA with EA involving right-sided and/or left-sided colectomies; and
(2) reported data concerning at least the primary endpoint (i.e., anastomotic leakage). If two studies were
reported by the same institution and/or authors, the one with more comprehensive data was included,
unless the studies were of different design and encompassed distinctive study population.

Non-comparative studies such as case series, description of particular techniques, along with animal
studies, conference abstracts, review articles, opinions and editorials were excluded from the analysis.
Furthermore, studies with inadequate data or that described other types of resections (e.g., single-incision
approach, purely robotic, sub-total colectomy, primary rectosigmoid resection, and palliative resection)
were excluded as well. The natural orifice extraction studies were excluded as it is currently not a widely
[9]
practiced method and its validity is still to be confirmed .
Outcome measures
The primary endpoint was anastomotic leakage since the safety of a surgical technique is considered the
most vital. An anastomotic leak was defined as a defect in the intestinal wall integrity at the anastomotic
site leading to a communication between the intraluminal and extraluminal compartments either clinically
[10]
or radiologically .
With regard to the secondary outcomes, we chose the following clinical endpoints to best reflect crucial
clinical consequences of colonic resection:

Intraoperative:
(1) Operative time
(2) Number of lymph nodes harvested

Post-operative:
(1) Mortality, defined as any deaths occurred during hospitalisation or within 30 days post-operatively
(2) Need for re-intervention
(3) Time to first flatus
(4) Surgical site infections
(5) Incidence of post-operative incisional hernia

Data analysis
Statistical analysis
The meta-analysis was performed using Review Manager 5.3 (Cochrane Community) and was conducted in
accordance with recommendations from the Cochrane Collaboration and Meta-Analysis of Observational
Studies in Epidemiology Guidelines.

The statistical analysis for dichotomous variables was summarised by calculating odds ratios (OR) with a
confidence interval (CI) of 95%. Mantel-Haenszel method was used to calculate the effect size by combining
the odds ratios of the outcomes using a random-effects model. Odds ratio < 1 favoured the IA group while
odds ratio > 1 favoured the EA group. This was considered statistically significant if P < 0.05 and if the
confidence interval did not include 1. Continuous variables were statistically analysed by calculating the
weighted mean difference (WMD) with a 95% confidence interval. A positive WMD indicated that the
pooled mean value of the outcome was higher in the IA group and was considered statistically significant
2
if P < 0.05. Study heterogeneity was evaluated using I statistics. I > 50% was considered substantial (i.e.,
2
2
serious heterogeneity) while I < 50% was considered low-moderate risk of heterogeneity. In studies which