BACKGROUND The loss in the retentive capability of implant-supported overdenture attachments has been related to the wear of the retentive inserts. Wear of the abutment coating material when following the replacement period for the retentive inserts requires investigation. OBJECTIVE The purpose of this in vitro study was to compare the changes in retentive force of 3 polyamide and a polyetheretherketone denture attachments under repeated insertion and removal cycles in wet conditions while following their manufacturers' replacement time recommendations. METHODS Four different denture attachments (LOCKiT, OT-Equator, Ball attachment, and Novaloc) with their retentive inserts were tested. Four implants were embedded into individual acrylic resin blocks, and 10 abutments for each attachment were used. Forty metal housings with their retentive inserts were attached to polyamide screws with autopolymerizing acrylic resin. A customized universal testing machine was used to simulate insertion and removal cycles. The specimens were mounted on a second universal testing machine at 0, 540, 2700, and 5400 cycles, and the maximum retentive force was recorded. The retentive inserts for LOCKiT (light retention), OT-Equator (soft retention), and Ball attachment (soft retention) were replaced every 540 cycles, while the Novaloc (medium retention) attachments were never replaced. All the abutments were weighed with a precision scale at 0, 2700, and 5400 cycles. The surface of every abutment was examined under a stereomicroscope at ×10 magnification. Data were analyzed with descriptive statistics. Two-way repeated measures ANOVA was used to compare the mean retentive force and mean abutment mass of all groups and time evaluation points. Bonferroni corrections were made to adjust for multiple tests (α=.05). RESULTS The mean retention loss for LOCKiT was 12.6% after 6 months and 45.0% after 5 years of simulated use. The mean retention loss for OT-Equator was 16.0% after 6 months and 50.1% after 5 years of simulated use. The mean retention loss for Ball attachment was 15.3% after 6 months and 39.1% after 5 years of simulated use. The mean retention loss for Novaloc was 31.0% after 6 months and 59.1% after 5 years of simulated use. The mean abutment mass difference was statistically significant (P<.05) for LOCKiT and Ball attachment but not statistically significant (P>.05) for OT-Equator and Novaloc at baseline, 2.5 years, and 5 years. CONCLUSIONS All tested attachments recorded retention loss under the experimental conditions, even when their manufacturers' recommendations for replacement time for the retentive inserts were followed. Patients should be aware that implant abutments should be replaced after a recommended period since their surfaces also change over time.