Sources

We conducted a systematic review of robotic gynecologic surgery by searching PubMed Central, OVID Medline, EMBASE, the Cochrane database, and ClinicalTrials.gov databases. Searches were conducted using the terms “robotic” or “robotics,” “intraoperative complications” or “postoperative complications” or “intestinal perforation” or “bowel injury” or “colon injury,” and “gynecologic surgical procedures.” Searches were limited to English-language articles, without restriction on initial publication date and published from November 2001 through December 31, 2014. There were no initial limitations in regards to study design. Additional studies were deemed relevant by reviewing the references of included articles.

Study selection

The Guidelines for Meta-Analyses and Systematic Reviews of Observational Studies (MOOSE) was used to complete the systematic review with the exception of scoring study quality and a single primary reviewer. The identified abstracts and the additional references were reviewed by one of the authors (T.P., a gynecology resident) for inclusion. Another author (N.L.S., a senior epidemiologist) also reviewed a subset (n = 82) of abstracts and articles that the data abstractor (T.P.) found questionable; full agreement on the decision regarding inclusion occurred. N.L.S. also reviewed all included articles for a subset of analyzed indices. Manuscript authors were not contacted.

Abstracts were evaluated for relevance after identification utilizing the search criteria listed above. Articles were included if they were published in refereed journals and mentioned bowel or colon injury or gynecologic surgical complications in the title and/or abstract. The full text of relevant abstracts was then reviewed in detail. For inclusion in the review, articles had to report the incidence of injury to the gastrointestinal tract. If articles commented on multiple other complications but did not include bowel injuries, it was assumed no bowel injuries occurred. Articles only including isolated complications (ie, cystotomy) that were not bowel injuries were excluded.

In the articles included in the review, bowel injury definitions varied, together with but not limited to: serosal abrasion, bowel perforation or tear, enterotomy, or bowel injuries for which the details were not further specified. We report on all of the cases as best possible, although details such as surgical indication, etiology, and management were not specified for most of the study subjects experiencing injury. The authors were not contacted to more precisely categorize these indices because the lack of original manuscript specificity reflects information that was either unavailable to them or that was not abstracted or recorded by them. Articles using national databases, case reports, data presented in review articles, and duplicate data presented in multiple manuscripts were also excluded to avoid double-counting.

Fisher exact test and odds ratios (OR) with 95% confidence intervals (CIs) are presented for all incidence rate comparisons given the varied study designs and uncommon occurrence of bowel injury. The Wilson method for proportions was used to calculate 2-tailed 95% CIs and P values for incidence. As surgical procedure and anatomic location play such potentially important roles in the incidence of bowel injury, exact OR are presented for analyses stratified by these variables, using the most precise classifications possible based on the reviewed articles, comparing the incidence rates to the most common procedure (which was procedures not otherwise specified), with P ≤ .05 considered statistically significant. The data were also analyzed by study design. All analyses were conducted using Stata (Version 11.0; StataCorp, College Station, TX) except for basic incidence rates that were calculated using the Excel (Microsoft Corp, Redmond, WA)data abstraction database for which 2 authors (T.P. and N.L.S.) developed a variable-specific coding scheme.

Sources

We conducted a systematic review of robotic gynecologic surgery by searching PubMed Central, OVID Medline, EMBASE, the Cochrane database, and ClinicalTrials.gov databases. Searches were conducted using the terms “robotic” or “robotics,” “intraoperative complications” or “postoperative complications” or “intestinal perforation” or “bowel injury” or “colon injury,” and “gynecologic surgical procedures.” Searches were limited to English-language articles, without restriction on initial publication date and published from November 2001 through December 31, 2014. There were no initial limitations in regards to study design. Additional studies were deemed relevant by reviewing the references of included articles.

Study selection

The Guidelines for Meta-Analyses and Systematic Reviews of Observational Studies (MOOSE) was used to complete the systematic review with the exception of scoring study quality and a single primary reviewer. The identified abstracts and the additional references were reviewed by one of the authors (T.P., a gynecology resident) for inclusion. Another author (N.L.S., a senior epidemiologist) also reviewed a subset (n = 82) of abstracts and articles that the data abstractor (T.P.) found questionable; full agreement on the decision regarding inclusion occurred. N.L.S. also reviewed all included articles for a subset of analyzed indices. Manuscript authors were not contacted.

Abstracts were evaluated for relevance after identification utilizing the search criteria listed above. Articles were included if they were published in refereed journals and mentioned bowel or colon injury or gynecologic surgical complications in the title and/or abstract. The full text of relevant abstracts was then reviewed in detail. For inclusion in the review, articles had to report the incidence of injury to the gastrointestinal tract. If articles commented on multiple other complications but did not include bowel injuries, it was assumed no bowel injuries occurred. Articles only including isolated complications (ie, cystotomy) that were not bowel injuries were excluded.

In the articles included in the review, bowel injury definitions varied, together with but not limited to: serosal abrasion, bowel perforation or tear, enterotomy, or bowel injuries for which the details were not further specified. We report on all of the cases as best possible, although details such as surgical indication, etiology, and management were not specified for most of the study subjects experiencing injury. The authors were not contacted to more precisely categorize these indices because the lack of original manuscript specificity reflects information that was either unavailable to them or that was not abstracted or recorded by them. Articles using national databases, case reports, data presented in review articles, and duplicate data presented in multiple manuscripts were also excluded to avoid double-counting.

Fisher exact test and odds ratios (OR) with 95% confidence intervals (CIs) are presented for all incidence rate comparisons given the varied study designs and uncommon occurrence of bowel injury. The Wilson method for proportions was used to calculate 2-tailed 95% CIs and P values for incidence. As surgical procedure and anatomic location play such potentially important roles in the incidence of bowel injury, exact OR are presented for analyses stratified by these variables, using the most precise classifications possible based on the reviewed articles, comparing the incidence rates to the most common procedure (which was procedures not otherwise specified), with P ≤ .05 considered statistically significant. The data were also analyzed by study design. All analyses were conducted using Stata (Version 11.0; StataCorp, College Station, TX) except for basic incidence rates that were calculated using the Excel (Microsoft Corp, Redmond, WA)data abstraction database for which 2 authors (T.P. and N.L.S.) developed a variable-specific coding scheme.