This population-based study aimed to compare direct health care costs related to the pharmacological treatment of osteoporosis and to the management of osteoporotic fractures among compliant and noncompliant users of alendronate and risedronate. During a 2-year follow-up period, compared to those with medication possession ratio (MPR) > or = 80%, women with MPR < 80% incurred significantly higher physician care costs and hospital care costs. This study aimed to compare direct health care costs related to the treatment of osteoporosis and osteoporotic fractures among compliant and noncompliant users of alendronate and risedronate. A cohort of 15,027 women having initiated alendronate or risedronate was identified. MPR and direct health care costs (physician care, hospital care, drugs) were assessed during a 2-year period. Regression models were used to estimate mean predicted cost for compliant (MPR > or = 80%) and noncompliant (MPR < 80%) women. Mean predicted physician care cost (in Canadian dollars) was $51 among women with MPR < 80% and $34 among those with MPR > or = 80%: mean difference $17, 95% confidence interval (CI) $2-22. Mean predicted hospital care cost was $568 among women with MPR < 80% and $379 among those with MPR > or = 80%: mean difference $189, 95% CI $56-320. Mean predicted drug cost was $439 among women with MPR < 80% and $1,068 among those with MPR > or = 80%: mean difference $-639, 95% CI $-649 to -629. Compared to compliant women, noncompliant women incurred significantly higher physician care and hospital care costs. Due to lower drug costs, total direct health care costs were lower among noncompliant women.