BACKGROUND: Intravenous (IV) bisphosphonate (BPN) use is associated with claims for osteonecrosis of the jaw (ONJ), but the applied claims codes have yet to be validated. Little is known about risk factors for ONJ incidence in patients unexposed to BPN.
OBJECTIVES: We validated potential ONJ cases by chart review to assess risk factors for ONJ and measure ONJ incidence rate (IR) in cancer patients.
METHODS: Using US health insurance claims, we created a cohort of patients with breast cancer (BRCA), prostate cancer (PRCA), or multiple myeloma (MM) and used 4 diagnosis and 11 procedure codes to identify potential ONJ cases after cohort entry. We abstracted medical records for identified potential ONJ cases, Blinded to exposure, experts reviewed chart information to classify probable and possible cases. We estimated the range for age- and sex-adjusted ONJ IR from probable cases (lower estimate) and from a combination of probable and possible cases (upper estimate), adjusting IRs upwards to account for cases missed owing to lack of chart availability. To adjust for changes in covariables, we used logistic regression to estimate odds ratios (ORs). We used months of follow-up as separate observations and reclassified exposure and covariable status in each month, using GEE to correct for multiple observations per patient.
RESULTS: Lower and upper IR per 1000 person-yrs were 0.045 and 0.22 in unexposed, 4.8 and 6.6 in IV-BPN exposed, and 0.31 and 0.31 in oral-BPN exposed. After controlling for sex, corticosteroid use, and age, OR (95% CI) relative to no BPN for IV BPN was 8.8 (2.0–38) and for oral BPN was 0.6 (0.08–4.9); for MM relative to BRCA was 4.5 (0.7–28), and PRCA relative to BRCA was 0.37 (0.06–2.3).
CONCLUSIONS: MM was a strong risk factor for ONJ, even in BPN-unexposed patients, but this finding is based on few cases. IV BPN was also a strong risk factor.
