Open abdomen

Rencüzoğulları 2015 — Turkey, single centre, 40 randomised, NPWT versus Bogota bag, no sample-size calculation, funding and interests not reported.

Auer 2021 — Austria, single centre, August 2010 to September 2012, 34 randomised, Suprasorb CNP versus ABThera, sample-size calculation present, funded by a medical university scientific budget, no declared interests.

There is no trial comparing NPWT with any temporary abdominal closure other than the Bogota bag. The review says so in terms.

CAT: OCEBM 1 / The review's OWN GRADE rating is reproduced - VERY LOW certainty for every outcome · The review's ratings are reproduced rather than re-derived. The evidence base is two trials and 74 adults. Against a Bogota bag (Rencuzogullari 2015, Turkey, 40 patients, no sample-size calculation, funding and interests not reported): time to fascial closure MD -3.60 days (95% CI -8.16 to 0.96), P = 0.12; fistula formation 2/20 against 1/20, RR 2.00 (0.20 to 20.33); mortality 5/20 against 7/20, RR 0.71 (0.27 to 1.88); length of stay MD 1.10 days (-13.39 to 15.59). Comparing two commercial devices (Auer 2021, Austria, 34 patients): fascial closure 15/17 against 12/17, RR 0.80 (0.56 to 1.14); FISTULA 0/17 AGAINST 4/17, RR 0.11 (0.01 to 1.92); mortality 1/17 against 3/17; reoperation 17/17 against 17/17. Not one estimate excludes the null. The review's own risk-of-bias output: both trials at high risk overall - Auer low risk for sequence generation and allocation concealment, high risk for performance and detection blinding, and high risk on 'other' for baseline BMI imbalance; Rencuzogullari unclear on sequence generation and allocation concealment and high risk for blinding. A further limitation the review flags itself, and which the wiki should carry, is that the time-to-closure estimate is unusable: 'these may not be reliable data, because this study did not report whether all participants reached primary fascial closure' - and with 12 of 40 deaths in that trial, a mean computed over 20 per arm is not a closure time. Indirectness is also conceded: 10 of 74 participants (13.5%) are trauma patients in a review whose title and objective exclude them, handled by downgrading rather than exclusion. The value of this review is not its effect estimates, which are uninformative. It is that it constitutes a formal, grade-rated statement that the fistula-causation question remains unresolved by randomised evidence - a documented uncertainty the safety panel can cite instead of an opinion - and that it establishes that no trial compares NPWT with any temporary abdominal closure other than the Bogota bag.

Figures: Figures checked

• No significant difference in fascial closure (63.5% NPWT vs 69.5%; OR 0.74).
• No significant difference in morbidity, enteroatmospheric fistula, bleeding, or abscess.
• Significantly lower postoperative mortality with NPWT: 28.5% vs 41.4% (OR 0.46, 95% CI 0.23–0.91, p = 0.03).
• Significantly shorter ICU stay with NPWT (MD −4.53, p < 0.00001).
• Level of evidence III.

CAT: OCEBM 3 · Start Low (cohort); −1 risk of bias; −1 inconsistency. AMSTAR-2 confidence CRITICALLY LOW — three critical flaws (no a priori protocol or registration; inappropriate statistical combination of two poor RCTs with four retrospective and two historically controlled cohorts; no investigation of publication bias) plus six non-critical weaknesses. Inconsistency serious: I²=83% for the primary outcome and 72% for mortality, with the review stating that the included studies show 'effects in opposite directions'. NOTE — the +1 large-effect upgrade carried on the 2026-07-28 version of this CAT is WITHDRAWN on reading the full text: the mortality signal (OR 0.46) disappears in the review's own sensitivity analysis restricted to the two RCTs (OR 0.74, 95% CI 0.29-1.91, p=0.54), two of the eight studies compared a prospective NPWT group against a retrospective control, and the review itself names 'a risk of selection bias favoring NPWT' — so confounding by indication and era is a plausible explanation and a GRADE upgrade for magnitude is not available. This is not a borderline call, so the benefit-of-the-doubt tie-breaker does not apply to the certainty rating; it is applied only to hold the OCEBM level at 3, matching the paper's own self-designated level III.

Figures: Figures checked

• Defines the open abdomen and proposes an open-abdomen classification system.
• Recommends decompressive laparotomy for overt ACS, and — relevant here — negative pressure wound therapy together with efforts to achieve same-hospital-stay fascial closure in patients with an open abdomen.
• Recommends protocolised IAP measurement and management, and avoidance of sustained IAH.
• Suggests avoiding routine early biologic-mesh use in open abdominal wounds.
• Overall quality of evidence guiding the recommendations was generally low; appropriately designed trials are needed.

AGREE II: Recommended with modifications — Scope 89% · Stakeholders 44% · Rigour 54% · Clarity 94% · Applicability 33% · Independence 58%

Figures: Figures checked

• Ruptured AAA accounted for 60–100% of cases across studies.
• Primary fascial closure was achieved in 79–100%.
• Median time to closure was 10.5 and 17 days in two prospective studies (closure 100% and 96%) using VACM for anticipated open abdomen ≥5 days.
• Graft infection rate was 0% in three studies.
• No VACM patient was left with a planned ventral hernia.
• In-hospital survival ranged 46–80%.

CAT: OCEBM 4 · No pooled estimate exists anywhere in this paper. The authors state that because no randomised trial comparing temporary-closure devices after AAA repair was identified, 'there was no basis to perform a meta-analysis'. Every number is a single-series proportion from between 4 and 30 patients. The headline 79 to 100 per cent fascial closure is a per-protocol rate among survivors, not a closure rate. Table 1 labels that column 'primary fascial closure rate per protocol (survival permitting closure)'. Recomputed on every patient enrolled, the same series give Mayer 11/20 (55%), Kimball 13/21 (62%), Sörelius 25/30 (83%), Seternes 8/9 (89%), Ross 18/20 (90%) and Pettersson 4/4 (100%) — A true range of 55 to 100 per cent. In-hospital survival was 46 to 80 per cent, so the excluded denominator is large. The two prospective VACM series carrying the 100% and 96% figures are 9 and 30 patients, 39 in total, with median closure at 10.5 and 17 days. Graft infection '0% in three studies' rests on 4, 16 and 6 patients — 26 PATIENTS — while Sörelius reported 2/30 (6.7%) and Kimball 1/14 (7.1%), and two studies did not study it at all. Barker 2007 (n = 13) reported no closure rate. GRADE begins Low for a body of observational evidence and falls to Very low for risk of bias — no instrument was applied and the authors concede selection, performance and information bias — and for imprecision on 117 patients. Because there is no comparator arm anywhere in the review, no relative effect of NPWT over any alternative temporary closure can be graded from it at all.

Figures: Figures checked

CAT: OCEBM 4 (retrospective cohort, descriptive with a prognostic regression). GRADE not applicable to any NPWT effect — the closure technique is not compared. QUIPS applied to the prognostic analysis: moderate risk · This page carries a baseline, not an effect. In-hospital mortality 62.4% in 101 non-trauma open abdomens; among survivors to discharge, fascial closure in 75.7%; age over 60 and APACHE II over 18.5 independently predicting death. Those are the numbers a reader needs when a motivation for negative-pressure open-abdomen management in a septic or post-haemorrhage patient is being written, because they say what the expected outcome is regardless of the closure device. The series is 88% vacuum-based, so it is also a description of vacuum-based temporary closure in routine use, but the technique was not compared with anything and the paper makes no claim about it. Prognostic quality (Quips): participation adequate (all laparostomies in the period); attrition not an issue for in-hospital death; prognostic factors measured routinely (APACHE II, SOFA, age); outcome unambiguous; confounding handled by a multivariable model on 63 deaths with few covariates; reporting adequate. The vascular reader should note that three ruptured aortic aneurysms and ten iatrogenic or postoperative haemorrhages are in the denominator, which makes the series more relevant to this practice than a pure sepsis cohort would be, though the subgroup is too small to analyse.

Figures: Figures checked