Fourteen subjects were retrospectively recruited for this case-controlled study. In test group (seven patients), primary closure was not achieved and membrane was left exposed at the initial surgery or it became exposed during the first week of healing. In the control group (seven patients), primary wound closure was achieved and no exposure of the membrane occurred until the placement of a dental implant 4 months after augmentation. Each patient received a particulate cancellous allograft (500 to 800 μm, RegenerOss, BioMet 3i), and then, the grafted defect area was covered with a bioresorbable matrix membrane (Sunstar, Suisse SA, Etoy, Switzerland). Longer span edentulous spaces were divided into individual sites based on a 10–12-mm width per site, and each site was bordered by at least one tooth. The subject inclusion criteria included a treatment that was planned to receive a dental implant in the future. At least 18-year-old males and females were included in this study. All subjects signed an informed consent approved by the University of Louisville Institutional Review Board in July 2010. Exclusion criteria excluded patients with uncontrolled diabetes, who are smokers, and with immune diseases or other systemic diseases that significantly affect the periodontium; patients with an allergy to any material or medication used in this study; and patients who need prophylactic antibiotics, previous head and neck radiation therapy, and chemotherapy in the previous 12 months and with severe psychological problems.

All surgical procedures were done by one surgeon (ME). The surgical procedure consisted of the reflection of a full thickness flap to expose the residual alveolar ridge. Following complete exposure of the defect area, horizontal ridge width was measured with a digital caliper at the midridge crestal level. Horizontal ridge measurements (at the crestal level) included initial ridge, initial augmented ridge (residual ridge plus graft), and the ridge at the re-entry after the 4-month healing time. All measurements were performed by a masked examiner (MEG), who was unaware of the treatments. Cortical perforations were performed with a ½ round bur to increase vascularization in the defect area [30]. The bone screws (Salvin Dental Specialties, Charlotte, NC) and cancellous particulate allograft were placed to allow for augmentation to achieve 8 mm horizontal width available for implant placement. Horizontal measurements were again taken with the same caliper. Then, the grafted area was covered with a bioresorbable matrix barrier. In the buccal surface, the graft material was completely covered by the membrane. However, in the lingual or palatal surface, the membrane was tucked least 5 mm between the alveolar bone and gingival tissue. Periosteal releasing incisions were performed when needed, and all wound was tried to close primarily and flap re-positioned coronally without compromising MGL. In the test group, the membrane could not be closed primarily (Fig. 1a) or became exposed to the oral environment in 1 week after primary closure was achieved during surgery (Fig. 1b). In the control group, primary closure was achieved with a monofilament suture (Cytoplast, PTFE Suture, Osteogenics Biomedical, Lubbock, TX) (Fig. 1c). The patients were prescribed doxycycline hyclate 50 mg once a day for 2 weeks and hydrocodone when needed. Sutures were removed in 10 days. The subjects were seen every other week to clean the area with hydrogen peroxide.