A systematic methodological review of reported perioperative variables, postoperative outcomes and hernia recurrence from randomised controlled trials of elective ventral hernia repair: clear definitions and standardised datasets are needed.

Background: This systematic review assesses the perioperative variables and post-operative outcomes reported by randomised controlled trials (RCTs) of VH repair. This review focuses particularly on definitions of hernia recurrence and techniques used for detection.

Objective: Our aim is to identify and quantify the inconsistencies in perioperative variable and postoperative outcome reporting, so as to justify future development of clear definitions of hernia recurrence and a standardised dataset of such variables.

Methods: The PubMed database was searched for elective VH repair RCTs reported January 1995 to March 2016 inclusive. Three independent reviewers performed article screening, and two reviewers independently extracted data. Hernia recurrence, recurrence rate, timing and definitions of recurrence, and techniques used to detect recurrence were extracted. We also assessed reported post-operative complications, standardised operative outcomes, patient reported outcomes, pre-operative CT scan hernia dimensions, intra-operative variables, patient co-morbidity, and hernia morphology.

Results: 31 RCTs (3367 patients) were identified. Only 6 (19.3%) defined hernia recurrence and methods to detect recurrence were inconsistent. Sixty-four different clinical outcomes were reported across the RCTs, with wound infection (30 trials, 96.7%), hernia recurrence (30, 96.7%), seroma (29, 93.5%), length of hospital stay (22, 71%) and haematoma (21, 67.7%) reported most frequently. Fourteen (45%), 11 (35%) and 0 trials reported CT measurements of hernia defect area, width and loss of domain, respectively. No trial graded hernias using generally accepted scales.

Conclusion: VH RCTs report peri- and post-operative variables inconsistently, and with poor definitions. A standardised minimum dataset, including definitions of recurrence, is required.