All adult orthopaedic patients undergoing a non-emergent procedure between December 2010 and November 2011 were screened for eligibility either in the preadmission clinics for elective cases or after admission for trauma-related procedures. Patients were consented for participation prior to screening. Patients were included if they were to undergo an orthopedic procedure requiring a wound greater than two centimeters in length. Exclusion criteria included, open fractures, procedures of the foot or hand, arthroscopic procedures and chemotherapy or radiation therapy treatment. The study was approved by the local Human Research Ethics Board. A schematic of the trial design can be found in Figure
Baseline parameters including age, weight, height were collected. Medical comorbidities affecting wound healing including: diabetes, peripheral vascular disease, renal failure requiring dialysis, liver failure and inflammatory arthritis were identified. Medications including corticosteroids, transplant anti-rejection medications, disease-modifying anti-rheumatic drugs warfarin, asprin, clopidogrel and heparin were noted by patient report. A smoking history was also obtained. A trained research assistant or orthopaedic surgery resident collected patient data on case report forms which were later entered into an electronic database (MS Excel, Microsoft, Redmond, WA).
Allocation of patients to treatment groups occurred at the completion of deep wound closure using sequentially-numbered opaque envelopes with randomization sequence generated by the principle investigator using Microsoft Excel (Redmond, WA). Any participant felt to be endangered by the use of one closure method was withdrawn from the study prior to allocation.
After completion of the procedure, deep tissues were closed with absorbable braided suture (Polysorb, Covidien, Mansfield, MA). In all patients the subcutaneous tissue was also closed with an absorbable braided suture. Patients allocated to the sutures intervention had their wounds closed using the suture material chosen by the primary surgeon. The primary surgeon also decided on the most appropriate technique of closure. Those allocated to the staples group were closed using a commercially-available stapler (Weck Visistat 35W, Limerick, PA). Closure material was removed, when necessary, during a wound check two weeks after surgery by medical staff or orthopedic technologists.
Patients were blinded to treatment allocation by use of an adhesive bandage or plaster which remained in place until the first planned post-operative visit when the closure material was removed or the wound was checked, including the removal of steristrips in the case of absorbable subcuticular closure. During closure material removal the wound was hidden from the view of the patient when necessary to maintain blinding. Immediately prior to unblinding patients completed an outcome assessment survey including a questionnaire determining complication occurance and a 100 mm VAS pain scale. Pain measurements were normalized to worst pain imaginable reported by the patient.
Wound complications were identified by hospital staff during the post-operative inpatient stay and were reported by patients at two- and six-week follow-up appointments in a survey conducted by the study coordinator. Patients were asked whether they had taken antibiotics, required dressing changes or a re-operation at two- and six-week follow-up appointments, with affirmative answers resulting in further questioning to confirm the type of complication that had occurred. Suspected infections were defined as: reoperation, use of intravenous antibiotics or patient report of use of oral antibiotics related directly to the surgical procedure. Wound drainage, necrosis, dehiscence and suture abscess were suspected when dressing changes or additional medical treatment was reported by the patient. The hospital and clinic charts of all suspected infections and five negative controls were reviewed by an orthopaedic surgeon to confirm the diagnosis of surgical site infection as defined by the CDC. Secondary outcome measures of time taken to close wounds and pain with wound closure removal were also collected.
The primary outcome was a composite measure of all-cause, patient-reported wound complications. Primary outcome data for the entire cohort of patients encompassing four anatomic sites was analyzed using the Fisher’s exact test (Stata, College Station, TX) and are presented as relative risk and 95% confidence intervals. Strata-specific relative risks were calculated for each anatomic site using the primary outcome measure. Continuous secondary outcome measures were analyzed using a 2-tailed Student’s t-test. The results were reported as per the guidelines of the CONSORT group