PURPOSE: The aim was (1) to evaluate the soft-tissue reaction of a synthetic polyethylene glycol (PEG) hydrogel used as a barrier membrane for guided bone regeneration (GBR) compared with a collagen membrane and (2) to test whether or not the application of this in situ formed membrane will result in a similar amount of bone regeneration as the use of a collagen membrane. MATERIAL AND METHODS: Tooth extraction and preparation of osseous defects were performed in the mandibles of 11 beagle dogs. After 3 months, 44 cylindrical implants were placed within healed dehiscence-type bone defects resulting in approximately 6 mm exposed implant surface. The following four treatment modalities were randomly allocated: PEG+autogenous bone chips, PEG+hydroxyapatite (HA)/tricalcium phosphate (TCP) granules, bioresorbable collagen membrane+autogenous bone chips and autogenous bone chips without a membrane. After 2 and 6 months, six and five dogs were sacrificed, respectively. A semi-quantitative evaluation of the local tolerance and a histomorphometric analysis were performed. For statistical analysis, repeated measures analysis of variance (ANOVA) and subsequent pairwise Student's t-test were applied (P<0.05). RESULTS: No local adverse effects in association with the PEG compared with the collagen membrane was observed clinically and histologically at any time-point. Healing was uneventful and all implants were histologically integrated. Four out of 22 PEG membrane sites revealed a soft-tissue dehiscence after 1-2 weeks that subsequently healed uneventful. Histomorphometric measurement of the vertical bone gain showed after 2 months values between 31% and 45% and after 6 months between 31% and 38%. Bone-to-implant contact (BIC) within the former defect area was similarly high in all groups ranging from 71% to 82% after 2 months and 49% to 91% after 6 months. However, with regard to all evaluated parameters, the PEG and the collagen membranes did not show any statistically significant difference compared with sites treated with autogenous bone without a membrane. CONCLUSION: The in situ forming synthetic membrane made of PEG was safely used in the present study, revealing no biologically significant abnormal soft-tissue reaction and demonstrated similar amounts of newly formed bone for defects treated with the PEG membrane compared with defects treated with a standard collagen membrane.