Studies suggest that long-term use of proton pump inhibitors (PPIs) may be associated with an increased risk of fracture. However, the role of medication adherence in this association is not fully understood. A retrospective cohort study was conducted to examine the relationship between PPI use/adherence and fracture risk among elderly subjects by combining administrative pharmacy claims data, survey data, and Medicare data. The study cohort included 1,604 PPI users and 23,672 nonusers who were enrolled in Pennsylvania’s Pharmaceutical Assistance Contract for the Elderly program. PPI adherence was measured by the proportion of days covered (PDC). Time-dependent Cox proportional hazards models were used to estimate adjusted hazard ratios (HRs) of PPI use/adherence for fracture risk while controlling for demographics, comorbidity, body mass index, smoking, and non-PPI medication use. The overall incidence of any fracture per 100 person-years was 8.7 for PPI users and 5.0 for nonusers. A gradient in fracture risk according to PPI adherence was observed. Relative to nonusers, fracture HRs associated with the highest (PDC ≥ 0.80), intermediate (PDC 0.40–0.79), and lowest (PDC <0.40) adherence levels were 1.46 (p < 0.0001), 1.30 (p = 0.02), and 0.95 (p = 0.75), respectively. In addition, the fracture risk of PPI use was significant for hip (HR = 1.32, p = 0.04) and vertebral (HR = 1.69, p = 0.0005) fractures, and risk was similar between major osteoporotic and other fractures. These results provide further evidence that PPI use may increase fracture risk in the elderly and highlight the need for clinicians to periodically reassess elderly patients’ individualized needs for ongoing PPI therapy, while weighing potential risks and benefits.