Anti-osteoporosis medication (AOM) does not abolish fracture risk, and some individuals experience multiple fractures while on treatment. Therefore, criteria for treatment failure have recently been defined. Using data from the Global Longitudinal study of Osteoporosis in Women (GLOW), we analyzed risk factors for treatment failure, defined as sustaining ≥2 fractures while on AOM. GLOW is a prospective, observational cohort study of women aged ≥55 years sampled from primary care practices in 10 countries. Self-administered questionnaires collected data on patient characteristics, fracture risk factors, previous fractures, AOM use, and health status. Data were analyzed from women who used the same class of AOM continuously over 3 survey-years and had data available on fracture occurrence. Multivariable logistic regression was used to identify independent predictors of treatment failure. Data from 26,918 women were available, of whom 5550 were on AOM. During follow-up, 73/5550 women in the AOM group (1.3%) and 123/21,368 in the non-AOM group (0.6%) reported occurrence of ≥2 fractures. The following variables were associated with treatment failure: lower SF-36 score (physical function and vitality) at baseline, higher FRAX score, falls in the past 12 months, selected comorbid conditions, prior fracture, current use of glucocorticoids, need of arms to assist to standing, and unexplained weight loss ≥10 lb (≥4.5 kg). Three variables remained predictive of treatment failure after multivariable analysis: worse SF-36 vitality score (odds ratio [OR] per 10-point increase 0.85; 95% confidence interval [CI] 0.76-0.95; p = 0.004), ≥2 falls in the past year (OR 2.40; 95% CI 1.34-4.29; p = 0.011), and prior fracture (OR 2.93; 95% CI 1.81-4.75; p < 0.0001). The C statistic for the model was 0.712. Specific strategies for fracture prevention should therefore be developed for this subgroup of patients.