Abdominal wound dehiscence, the partial or complete disruption of a surgically closed abdominal wound with or without protrusion of abdominal contents, remains one of the most feared complications of laparotomy. In a retrospective cohort study conducted at Wolaita Sodo University Comprehensive Specialized Hospital in Southern Ethiopia, 47 of 450 adult patients who underwent midline laparotomy between June 1, 2020 and June 30, 2023 developed the complication during their index admission, a cumulative incidence of 10.4 percent with a 95 percent confidence interval of 7.9 to 13.6 percent. Nearly all of these cases, 45 of 47, followed emergency operations, and every one of the ten in-hospital deaths recorded in the cohort occurred in a patient who had developed the condition. The findings, published in Discover Medicine, quantify a burden that is far higher than the 0.4 to 3.5 percent incidence typically reported in developed countries, yet sits within the 10 to 30 percent range described for some centres in India and Pakistan and close to the 12.4 percent reported in Egypt.
The study population comprised a complete enumeration of eligible patients rather than a probability sample. Every adult aged 18 or older who underwent midline laparotomy at the hospital during the three-year window and had a complete medical record was included, yielding 450 patients with no missing values on any analyzed variable. Records were identified from operation theatre logbooks, and data were extracted from admission registrations, operating room logs, postoperative monitoring sheets, and intensive care and discharge records using a pretested structured checklist. The authors report that the number of logbook entries screened and set aside under each exclusion criterion was not tallied at the time of abstraction, so exact exclusion counts cannot be given, a limitation they state transparently. Patients were excluded if they had a prior history of wound dehiscence, underwent the index laparotomy as a re-operation, had a gynecological indication, lacked a complete record, or died on the operating table during the initial surgery.
Dehiscence was defined strictly as full-thickness fascial separation occurring before complete healing, with or without evisceration, during the index admission; isolated superficial skin or subcutaneous separation did not count. Diagnosis relied on the operating surgeon’s documentation of fascial disruption, corroborated by the management received, which in every confirmed case was relaparotomy with retention sutures or open-abdomen Bogota bag application. The complication was recognized between postoperative days 5 and 10 in 39 patients, or 83.0 percent of cases, and before day 5 in the remaining eight, with no case recorded after day 10. This timing coincides with the period in which the tensile strength of the healing fascia is at its lowest, and the authors note that it supports close wound inspection during the second postoperative week.
The cohort was predominantly male, with 297 patients, or 66.0 percent, and a mean age of 44.9 years. Most patients, 385, or 85.6 percent, lived in rural areas. Emergency midline laparotomy accounted for 352 of the 450 operations, or 78.2 percent, while 98 were elective. Preoperative sepsis was documented in 72 patients, 16.0 percent, and gastrointestinal contamination of at least 500 milliliters was present in 74 patients, 16.4 percent. Trauma was the most common surgical indication at 24.2 percent of cases, followed by perforated appendicitis at 20.9 percent and large bowel obstruction at 16.7 percent. Yet the distribution of dehiscence across indications was strikingly uneven: the proportion of patients affected was highest after surgery for typhoid perforation, 13 of 18 cases, or 72.2 percent, and small bowel obstruction, 25 of 36, or 69.4 percent, while no case occurred after surgery for perforated appendicitis, perforated peptic ulcer, large bowel obstruction, or gastric cancer. In absolute terms, small bowel obstruction and typhoid perforation together accounted for four-fifths of all dehiscence cases.
The statistical challenge at the heart of the analysis was quasi-complete separation, a situation in which combinations of predictors almost perfectly determine the outcome. All 34 patients who had both a surgical site infection and an operative duration of at least three hours developed dehiscence, and all 335 patients with neither a surgical site infection nor preoperative sepsis remained free of it. Under such conditions, standard maximum-likelihood logistic regression produces inflated and unstable coefficient estimates. The authors therefore fitted their multivariable model using Firth’s penalized, bias-reduced logistic regression, the recommended approach for separated binary outcome data, and confirmed that pairwise multicollinearity was not the underlying issue, with all variance inflation factors below 2. Because surgical site infection is a postoperative event whose timing relative to fascial dehiscence could not be established from the records, they also fitted a pre-specified sensitivity model that excluded it.
In the full adjusted model, operative duration shorter than three hours was associated with lower odds of dehiscence, with an adjusted odds ratio of 0.02 and a 95 percent confidence interval of 0.00 to 0.50, as was gastrointestinal contamination below 500 milliliters, with an adjusted odds ratio of 0.03 and a confidence interval of 0.00 to 0.57. Postoperative cough or vomiting showed an adjusted odds ratio of 111.3 and emergency surgery an adjusted odds ratio of 25.7, but the authors emphasize that the confidence intervals for these two estimates span more than two orders of magnitude, indicating the direction but not the magnitude of the associations. When surgical site infection was removed from the model, the estimates for cough or vomiting fell to an adjusted odds ratio of 3.01 and emergency surgery to 1.39, with the latter no longer statistically significant, while the associations for shorter operative duration and lower contamination persisted with considerably narrower confidence intervals. Preoperative sepsis, which could not be estimated stably in the full model because of its correlation with surgical site infection, showed a strong and precise association in the sensitivity model, with an adjusted odds ratio of 7.46 and a confidence interval of 3.41 to 16.30.
Surgical site infection itself showed by far the strongest crude association with the outcome: 42 of 47 patients with dehiscence, or 89.4 percent, had a surgical site infection, compared with 42 of 403 patients without dehiscence, or 10.4 percent, yielding a crude odds ratio of 72.2. Preoperative sepsis was documented in 66.0 percent of dehiscence patients versus 10.2 percent of others, a crude odds ratio of 17.1. The authors are careful to interpret these as associations rather than causal effects, noting that the retrospective design could not establish whether infection preceded or followed fascial breakdown, and that infection and dehiscence may be manifestations of the same underlying process. They nonetheless point to a plausible biological mechanism involving bacterial collagenase activity that disrupts collagen deposition, together with increased matrix metalloproteinase activity and insufficient inhibition. The practical implication they draw is that a patient with a surgical site infection after midline laparotomy, particularly after a prolonged operation, should be regarded as being at very high risk of fascial dehiscence and monitored accordingly.
The clinical burden of the complication was heavy. Patients who developed dehiscence stayed in hospital a median of 26 days, compared with a median of 8 days among those who did not, a difference tested with the Mann–Whitney U test and significant at p less than 0.001. Some 87.2 percent of dehiscence patients were hospitalized for more than 12 days, versus 6.0 percent of the rest of the cohort. All ten in-hospital deaths in the study occurred among patients with dehiscence, giving an in-hospital mortality of 21.3 percent in that group and 0 percent among patients without it, a difference confirmed by Fisher’s exact test. Management required relaparotomy with retention sutures in 45 of the 47 affected patients and Bogota bag application in the remaining two.
The study has limitations that the authors address directly. It relied on retrospective document review, so diagnoses of dehiscence, surgical site infection, and sepsis were based on clinical documentation rather than independent standardized re-adjudication, and no formal inter-rater reliability statistic was computed. Several clinically important covariates could not be modeled: body mass index, ASA physical status class, formal nutritional status, and quantitative serum albumin were not documented, while diabetes mellitus, hypertension, obesity, pulmonary disease, renal disease, steroid use, and smoking were recorded as absent in every patient, which the authors treat as not documented rather than as evidence the conditions were truly absent. With only 47 events and seven predictors, roughly 6.7 events per parameter, the multivariable model is at risk of overfitting, and follow-up was limited to the index admission, meaning dehiscence occurring after discharge and any post-discharge deaths would not have been captured, so the reported incidence and mortality may be underestimates.
The authors conclude that about one in ten patients undergoing midline laparotomy at the hospital developed abdominal wound dehiscence, almost exclusively after emergency surgery, and that shorter operative duration and lower-volume gastrointestinal contamination were consistently associated with lower odds in both adjusted models. They call for prospective studies with systematic documentation of comorbidity, nutritional status, and the timing of surgical site infection to quantify these associations precisely. In the meantime, they suggest that preoperative sepsis, heavy contamination, and prolonged surgery identify patients who warrant particular attention to fascial closure and postoperative wound surveillance in resource-limited settings where emergency, contaminated procedures dominate the surgical case mix.
Subject of Research: Cumulative incidence and associated factors of abdominal wound dehiscence after midline laparotomy in Southern Ethiopia
Article Title: Abdominal wound dehiscence and associated factors among patients undergone midline laparotomy at a comprehensive specialized hospital in Southern Ethiopia: a retrospective cohort study
Article References: Dana, H. T., Shuke, S., Meskele, M., Olani, M., Abiye, A. A., Tila, M., & Kuche, A. D. (2026). Abdominal wound dehiscence and associated factors among patients undergone midline laparotomy at a comprehensive specialized hospital in Southern Ethiopia: a retrospective cohort study. Discover Medicine, 3(1), Article 148. https://doi.org/10.1007/s44337-026-00720-7
Image Credits: AI Generated
DOI: 10.1007/s44337-026-00720-7
Keywords: abdominal wound dehiscence, midline laparotomy, retrospective cohort, Ethiopia, surgical site infection, preoperative sepsis, emergency surgery, Firth logistic regression, hospital stay, in-hospital mortality, fascial closure, resource-limited surgery
News Source: Ophelia Keating. (November 5, 2026). One in Ten Laparotomy Patients Developed Wound Dehiscence in Southern Ethiopia Cohort. Scienmag.
