Use of Prophylactic Mesh When Creating a Colostomy Does not Prevent Parastomal Hernia

A Randomized Controlled Trial-STOMAMESH

Christoffer Odensten, MD; Karin Strigård, MD, PhD; Jörgen Rutegård, MD, PhD; Michael Dahlberg, MD, PhD; Ulrika Ståhle, MD; Ulf Gunnarsson, MD, PhD; Pia Näsvall, MD, PhD


Annals of Surgery. 2019;269(3):427-431. 

In This Article


Study Design

The study was designed as a double-blinded multicenter RCT. Patients were randomized to 2 groups, those with and those without prophylactic mesh around the stoma. The study protocol adhered to the Helsinki Declaration and was approved by the Regional Ethics Committee at Umeå University, Sweden (DNR 07–081 M). The study was registered at (Identifier: NCT00917995) and adhered to the CONSORT 2010 criteria for RCTs.


All patients scheduled for permanent colostomy at one of the participating hospitals and who met the inclusion criteria (no previous stoma; older than 18 years; and with informed consent), were eligible to participate in the study. Exclusion criteria were: expected survival less than 3 years; fecal peritonitis; previous stoma; and no informed consent.


Colostomy without mesh was taken through the rectus muscle, at a site marked by a stoma therapist before surgery. The height of the stomal nipple was at least 1 cm, and neither intraabdominal lateral closure nor fixation to the fascia was performed. Mucocutaneous interrupted monofilament sutures were used to fix the stoma to the skin.

Colostomy with mesh incorporated a lightweight polypropylene mesh (density 25–40 g/m2) in the sublay position. A 10 × 10 cm space for the mesh was created dorsal to the rectus abdominis muscle but anterior to the posterior rectus sheath and peritoneum. The bowel was passed through the rectus muscle via a cruciform incision in the middle of the mesh. Single nonabsorbable monofilament sutures anchored the mesh laterally in the pocket. Medially the mesh was incorporated in the running suture closing the fascia. A potential risk is that the mesh would not reach the midline; however, no such effect was described in the original article.[26] The surgical technique used had been described in an earlier study comparing stoma with and without prophylactic mesh.[26] An instruction video, made by Israelsson et al at the Department of Surgery, Sundsvall, was distributed to all participating centers to ensure that a uniform surgical technique was used. Fixation to the skin was performed with the same technique as when creating a colostomy without mesh. Surgery was performed by an experienced colorectal surgeon with an annual volume of at least 100 major surgical procedures.

Postoperative mobilization was according to each hospital's routine.

Study Outcome

Early complications were evaluated at 1 month, and late complications and possible recurrence of parastomal hernia were assessed at a 1-year follow up. The examiner was blinded to the group the patient was randomized to. Primary endpoint was the rate of parastomal hernia in both groups, judged clinically and by computerized tomography (CT) 1 year after surgery. Clinical examination was performed in both upright and supine positions by digital palpation within the stoma at rest and during the Valsalva maneuver. Findings were classified as a bulge or PSH. Definition of a bulge was a protrusion around the stoma noticed by the surgeon and/or the patient but deemed not to be a hernia clinically. CT was performed in the supine position, with or without intravenous contrast; the Valsalva maneuver was not performed. All CT examinations were evaluated by a radiologist with a special interest in parastomal hernia. The findings were classified according to the Moreno-Matias scale, classified as validated by the European Hernia Society, based on the content of the hernia: no hernia = 0; bowel forming the stoma with hernia sac < 5 cm = IA; bowel forming the stoma with hernia sac > 5 cm = IB; hernia sac containing omentum = II; and hernia sac containing loop other than bowel forming the stoma = III.[27,28]

Sample Size

As the vast majority of patients receiving a colostomy are operated for colorectal cancer, a 5-year survival rate around 50% is likely. Assuming a parastomal hernia incidence of 20%[7] without and 5% with mesh, a 5-year survival rate of just over 50%, and a significance level of 95%, it would require 220 patients to achieve a power of 80%.

Randomization and Blinding

Randomization was performed using sealed envelopes, stratified per hospital in blocks of four, to ensure balance between the 2 arms. The envelopes were prepared by the Regional Cancer Center North, Umeå University Hospital, in bundles of 100 per hospital, without the help of the investigators. Neither the patient nor the surgeon assessing the patient postoperatively was informed on which arm the patient had been randomized to. Postoperative assessment was made by a surgeon not involved in the primary procedure. If randomization of a patient was performed incorrectly, the patient was replaced by 3 new patients to ensure maintenance of power.

Data Collection

Data were collected prospectively and patients were followed by protocol with individual case report forms (CRF). All data were entered into an Excel-database.

After completion of inclusion, a separate analysis was performed to describe the cohort of patients who met the inclusion criteria but were not included in the study. To obtain a representative group of "not included" patients, all other colostomy operations performed during the inclusion/trial period at Umeå University Hospital, Sunderby Regional Hospital and Mora County Hospital were assessed for demographic data and reasons for not being included.

Statistical Analysis

The IBM SPSS Statistics 24 software program was used. All analyses of group variables were performed with the χ 2 test and all analyses of continuous variables were performed with the independent Student t test; a P< 0.05 was considered significant. The population was tested for skewness and kurtosis (measure describing outliers in a data set) using SPSS Statistics. As normal distribution was shown, parametric statistics were used.