|
Laparoscopic vs. open surgery for diverticular disease: a meta-analysis of nonrandomized studies |
Purkayastha S, Constantinides V A, Tekkis P P, Athanasiou T, Aziz O, Tilney H, Darzi A W, Heriot A G |
|
|
CRD summary This review compared outcomes for patients undergoing laparoscopic versus open surgical resections for diverticular disease. Evidence is limited to a small number of comparative non-randomised trials that suggested outcomes for laparoscopic and open surgeries were in general equivalent. However, the authors failed to find any clinically applicable conclusions, owing to the absence of randomised controlled trials.
Authors' objectives To compare operative, functional and adverse event outcomes in patients undergoing laparoscopic versus open surgery for the resection of diverticular disease.
Searching MEDLINE, EMBASE, Ovid (databases not specified) and the Cochrane Library were searched between 1996 and 27 June 2005; the search terms were reported. In addition, the reference lists of retrieved articles were checked for additional studies. No language restrictions were applied.
Study selection Study designs of evaluations included in the reviewAll comparative study designs were eligible for inclusion. Retrospective and prospective non-randomised designs were included in the review.
Specific interventions included in the reviewThe review compared techniques that were clearly documented as either laparoscopic or open techniques for the surgical resection of diverticular disease.
Participants included in the reviewEligible studies included patients undergoing surgical resection for diverticular disease, where the reasons for surgery were clearly reported. The included studies varied in the patients they recruited, both in terms of their age and gender. Studies including elective cases (9 studies) or a combination of elective and emergency cases (3 studies) were included. All of the studies included patients with uncomplicated diverticular disease, while three also included patients with complicated diverticular disease. The intervention groups were matched for one or more of the following patient characteristics: age, gender, weight or body mass index (BMI), American Society of Anesthesiologists (ASA) grade, number of admissions, history of abdominal surgery, diagnosis, procedure and proportion of anterior resection cases.
Outcomes assessed in the reviewEligible studies had to report one of the following outcomes:
operative outcomes (i.e. blood loss, specimen length, operative time, whether the splenic flexure was mobilised and the cost of the procedure in $US);
adverse event rates grouped according to the International Classification of Diseases (ICD-9) criteria (e.g. mechanical wound complications, infections, urinary complications, pulmonary complications, gastrointestinal tract complications, cardiovascular complications, systemic complications, intra-operative complications and overall mortality);
functional outcomes (i.e. time to normal diet, time to liquid diet, time to first bowel movement and length of hospital stay).
Studies were excluded if the necessary outcomes could not be calculated or extrapolated from the reported data.
How were decisions on the relevance of primary studies made?The authors did not state how the papers were selected for the review, or how many reviewers performed the selection.
Assessment of study quality The validity of the included studies was assessed using a modified Newcastle-Ottawa Scale consisting of 7 criteria. Details of the criteria and scoring system were reported and divided into three domains: patient selection, comparability of the study groups and outcome assessment. Each study was awarded a score of between 0 and 10 stars; studies scoring 6 or more stars were considered to be of a high quality.
Two reviewers assessed the validity of the studies. The authors did not report how any disagreements were resolved.
Data extraction Two reviewers independently extracted the data from the included studies. In addition to standard data concerning study design, population, outcome and intervention, the study inclusion and exclusion criteria, matching criteria, number of left and/or right sided resections and the conversion rate were also recorded. Where possible 2x2 tables were constructed for the dichotomous outcome data and used to calculate odds ratios (ORs) with 95% confidence intervals (CIs). Where 'zero' cells were present in the 2x2 tables, '0.5' was added to all cells. If there were zero event rates for both intervention groups, the data were discarded from the meta-analysis. Weighted mean differences (WMDs) with 95% CIs were calculated for continuous outcomes.
Methods of synthesis How were the studies combined?The studies were combined to provide a pooled OR for dichotomous outcomes and a pooled WMD for continuous variables. A random-effects model was used according to the Mantel-Haenszel method.
How were differences between studies investigated?Differences between the studies were investigated using the chi-squared test. In addition, where heterogeneity was suspected, the data were analysed using both fixed-effect and random-effects models. Sensitivity analyses were also performed according to the following groups: studies with more or less than 50 patients; high-quality (6 or more stars) versus low-quality studies; studies with publication dates before or after 2001; studies of elective surgeries versus elective and emergency surgeries; and studies that included only left sided surgeries versus right or right and left sided surgeries. One large study of 18,444 patients accounted for 94.5% of the total study sample and so dominated the analysis. The effects of this study were assessed in subgroup analyses.
Results of the review Twelve non-randomised studies assessing a total of 19,608 patients (1,192 laparoscopic surgeries and 18,416 open surgeries) were included in the review. Six of the studies were retrospective and six were prospective.
The quality scores for the 12 studies ranged from 5 to 8. Seven studies were considered to be of a high quality. Only 5 studies contained intervention groups that were fully matched for the age and gender of the participants; 3 studies were matched for BMI and two for ASA score. One large study of 18,444 patients accounted for 94.5% of the total study sample and so dominated the analysis. The effects of this study were considered in the sensitivity and subgroup analyses.
Significant differences favouring laparoscopic surgery over open surgery were found in terms of reduced infections (OR 0.61, 95% CI: 0.41, 0.90; 10 studies), wound infections (OR 0.35, 95% CI: 0.15, 0.81; 8 studies), pulmonary complications (OR 0.4, 95% CI: 0.26, 0.62; 7 studies), gastrointestinal tract complications (OR 0.75, 95% CI: 0.58, 0.98; 8 studies), specimen length (WMD -2.19, 95% CI: -3.65, -0.72; 2 studies) and cardiovascular complications (OR 0.28, 95% CI: 0.13, 0.59; 5 studies). No significant heterogeneity was detected for any of these findings.
Operation time was longer for laparoscopic surgery (WMD 67.59, 95% CI: 3.79, 131.39; 7 studies), but the length of time in hospital was significantly shorter (WMD -3.81 days, 95% CI: -4.63, -2.98; 8 studies). However, significant heterogeneity (P<0.001) was detected for both of these outcomes.
In terms of functional outcomes, significant differences in favour of laparoscopic surgery were found for time to normal diet (WMD -2.4 days, 95% CI: -2.99, -1.81; 2 studies), time to liquid diet (WMD -2.77 days, 95% CI: -4.68, -0.86; 2 studies) and time to first bowel movement (WMD -1.01 days, 95% CI: -1.55, -0.47; 2 studies). However, significant heterogeneity was found for time to first liquid and all three functional outcomes were based on only 2 small studies.
The main results generally remained significant in the sensitivity analyses, although some were shown to demonstrate significant heterogeneity which was not observed in the main analyses. However, one exception to this was the analysis of high-quality studies (which excluded the large but poorer quality study of 18,444 patients). Only blood loss and length of stay were found to be significantly in favour of laparoscopic surgery when only high-quality studies were considered.
The funnel plot showed some evidence of publication bias when all studies were included; this suggested an absence of smaller studies favouring open surgery. However, when only studies of high quality were considered, the 5 remaining studies showed no evidence of publication bias.
Cost information Operating costs (2 studies) were more for laparoscopic surgery than for open surgery (the authors reported WMD $699.56, 95% CI: -2,361, 376, but this appears to have been incorrectly reported as the CI does not contain the estimate of effect size). However, this finding was not statistically significant and significant heterogeneity was detected (P<0.001).
Authors' conclusions The authors stated that the extraction of clinically applicable conclusions was limited by the absence of randomised controlled trials (RCTs). Only a small number of comparative non-randomised trials were available, and these suggested that the results for laparoscopic and open surgical resections were in, general, equivalent. There may, however, be a potential reduction in complications and hospital stay for patients undergoing laparoscopic surgery for diverticular disease.
CRD commentary This was a clearly reported review based, in general, on well-defined inclusion criteria and methods. It was difficult, however, to assess the potential for selection bias as the authors failed to state how the studies were selected for inclusion in the review. The literature searches appeared quite extensive but were mainly focused on electronic databases. The limited search for unpublished material and the authors' funnel plot suggest that the findings may be subject to publication bias. The reliability of the funnel plot is questionable though, given the relatively small number of included studies.
Details of the studies and the main outcomes were tabulated clearly. The authors also carried out a number of assessments to determine the influence of study heterogeneity on their findings, including the dominating influence of one very large but poorer quality study. The syntheses of data appeared appropriate, although survival analyses would have been preferable to WMDs for the time to event outcomes such as hospital stay and length of surgery. By not calculating hazard ratios the authors are likely to have lost potentially useful data, but this type of analysis is very much dependent on the relevant data being available in the original studies. Overall, the findings of the review are limited given the absence of RCTs and the authors are right to be cautious in their findings.
Implications of the review for practice and research Practice: The authors did not state any implications for practice.
Research: The authors stated the need for RCTs that are structured and consider both long- and short-term outcomes, including quality of life and costs.
Bibliographic details Purkayastha S, Constantinides V A, Tekkis P P, Athanasiou T, Aziz O, Tilney H, Darzi A W, Heriot A G. Laparoscopic vs. open surgery for diverticular disease: a meta-analysis of nonrandomized studies. Diseases of the Colon and Rectum 2006; 49(4): 446-463 Indexing Status Subject indexing assigned by NLM MeSH Adult; Aged; Aged, 80 and over; Digestive System Surgical Procedures; Diverticulum, Colon /surgery; Female; Humans; Laparoscopy; Male; Middle Aged; Recovery of Function; Treatment Outcome AccessionNumber 12006001472 Date bibliographic record published 28/02/2007 Date abstract record published 28/02/2007 Record Status This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn. |
|
|
|