This protocol was approved by the local institutional review board in May 2010. Participation is purely voluntary and withdrawal from the study after randomization will be allowed. The study was also registered at clinicaltrials.gov after institutional review board approval. Registration identification NCT01146236 (registered June 14, 2010)
The study will be executed in five medical centers affiliated with two accredited universities. Patients will be drawn from a catchment of 2.7 million inhabitants. The study population will be a mix of rural and urban patients from diverse ethnic backgrounds. A variety of procedures will be included in the study encompassing both urgent and elective operations.
Parallel Group Randomized Controlled Trial with Economic Analysis
Recruitment of patients under the care of subspecialists in Orthopaedic Trauma, Spine, Adult Reconstruction and Upper Extremity Reconstruction will take place concurrently with the same trial infrastructure. Patients will be recruited by the office assistant, research assistant or delegate of the treating surgeon administering the surgical consent form. Verbal consent will be obtained for trauma patients unable to sign a consent form. Written consent will be obtained within 24 hours of treatment allocation. Patients eligible but not enrolled will be documented. Participant flow through the study can be seen visually in Figure
Participant flow diagram representing the planned design of the study.
• Adult patients (> 18
• All open orthopaedic procedures
• Any wound
cm in length
• Open fracture
• Known nickel allergy
• Active infection (any site)
• Foot surgery (any site)
• Hand surgery (including carpal surgery)
• Arthroscopic procedures
• Skin closure with metallic staples
• Skin closure with suture material – type of suture (non-absorbable or absorbable) and technique (simple, horizontal mattress, vertical mattress, subcuticular) at the discretion of the surgeon.
Method of allocation to groups
Randomization will be concealed and will be allocated at the time of skin closure. Randomization will be completed using an online randomization program. Block randomization with randomly sized blocks between eight and 12 participants will be employed to ensure an equal number of participants in each group for each subspecialty. Group assignments will be revealed by the online randomization program and the time of the randomization will be recorded electronically. Times of randomization will be compared with operative notes retrospectively to ensure concealment is being maintained.
Methods of blinding
Blinding of patients and providers was attempted in a pilot study. It was found that it was not feasible to blind these groups to the treatment allocation after post-operative dressings were changed, generally at postoperative day two. Additionally, post-operative radiographs revealed stapled patients in the pilot, further unblinding both providers and participants.
Outcome assessors will be blinded to the closure method through the use of long sleeved shirts, pants or a sleeve provided for the patient to wear. Blinding will be maintained by trained clinic staff aware of the purpose and methods of blinding. Participants will be asked not to reveal randomization when they become aware of treatment allocation during the course of the follow-up.
The outcomes adjudication committee will be blinded to treatment group during the determination of complications of individual patients. The data analysis team will also be blinded to the treatment groups during the synthesis of the results which will be presented as Group A and Group B in the draft manuscript.
Primary outcome measure
The primary outcome measure will be a composite outcome encompassing all causes of wound complication. The clinical relevance of the primary outcome measure stems from the fact that the components of the composite outcome all represent occurrences that are patient important. The components consist of the following events:
1. Surgical site infection as defined by:
• Use of intravenous antibiotics
• Use of oral antibiotics
• Re-operation at same site
2. Wound drainage occurring after post-operative day two requiring a dressing change.
3. Wound Necrosis defined as blackening of the skin edges at the incision site or skin slough observed by providers or the participant.
4. Suture Abscess defined as the expulsion of deep suture material and purulent material without surrounding erythema.
5. Peri-incision Blistering defined as blistering at the edge of the incision along the entire length. Blistering at dressing tape site will be excluded from this definition, however, blistering due to wound tape application which is contiguous with the wound edge will be considered an event.
6. Wound Dehiscence as defined by the loss of apposition of the skin edges visible to the eye along the length of the incision.
7. Material Sensitivity as defined by a local reaction to metal or suture material resulting in skin changes along the entire incision length.
The primary outcome measure will be assessed during admission by review of the patient chart by site coordinators who will be blinded to the treatment allocation. Outcome assessors will be uniformly trained in the definitions of the components of the composite outcome score. Outpatient complications will be recorded through self-report questionnaires administered by site coordinators. All deceased patients will also be reviewed for occurrence wound complications. Final determination of events will be made by an independent outcomes adjudication committee based on blinded clinical data identified by the committee as necessary for the determination of complications. A sample of non-infected participants will also be reviewed by the committee as a negative control.
Secondary outcome measures
Treatment preference reported by the participant
Additional unscheduled episodes of care as defined by:
• Dressing changes by homecare/patient at home or self-reported visits to other healthcare professionals.
• Length-of-stay – Based on admission and discharge dates
• Visual analogue pain score for suture/staple removal
Justification of the length of follow-up
The occurrence of SSIs happens primarily in the first three months post-operatively. The Centers for Disease Control (CDC) defines the timing of SSI as within the first 30
days unless a foreign body is implanted. In the presence of an implanted foreign body the surveillance period is extended to one year. Since the majority of orthopaedic surgeries result in the implantation of a metallic body we have extended the follow-up period to one year. Given the fact that the primary outcome measure is a composite measure related to the healing of a wound we will analyze the data after all participants have completed six-week follow-up. A phone survey will be performed at one year in order to conform with the CDC definition of SSI.
Sample size determination
Estimates of sample size necessary to definitively test the hypothesis in this study were calculated using rates of wound complications found in a pilot study which included 148 participants. A relative risk reduction of 25% in wound complications was chosen as a minimal clinically important difference in wound complications. Sample sizes are presented graphically for varying powers and with a constant α level of 0.05 (Figure
). Microsoft Excel (Redmond, WA) was utilized to perform calculations. The details of the sample size calculation are presented in Figure
Projected sample size based on a relative risk reduction of 25% for wound complications and using an alpha of 0.05.
Anticipated recruitment rate
In a pilot site it was found that 60% of patients approached consented to participate. A recruitment model was created taking into account expected site fatigue and differences in consenting efficiency between the United States and Canada. Based on the combined clinical volume at the five sites it is expected that the study will be completed with enrolment after 24
months of recruitment. Additional international sites are currently being developed to ensure that this target is met and to improve the generalizability of the results.
All anatomic sites will be analyzed together and separately using Stata (College Station, TX). Dichotomous primary and secondary outcome measures will be analyzed using Fisher’s exact test. Stratification based on potential confounding variables will be performed. A multivariable analysis employing logistic regression will be performed to assess the possible impact of empirical confounding by putative risk factors of the binary outcomes. Length of stay, visual analogue scale pain measurements and patient satisfaction data will be compared using a two-way ANOVA. Linear regression will be performed on continuous variables to assess the possible impact of empirical confounding by putative risk factors. A 4.7% level of significance will be considered significant based on the adjustment of significance level for the interim analysis.
Subgroup analyses will be performed comparing the use of sutures and staples in trauma versus elective procedures and primary versus revision procedures as appropriate.
Interim analysis will be performed by a Data Safety and Monitoring Committee blinded to the treatment groups. Analyses will take place at the half-way point of patient recruitment. The alpha spending function approach to setting significance levels will be used. The trial will be stopped if a difference is found in the complication rates using a significance level of 0.15% based on the O’Brien-Flemming boundaries for a group sequential trial.