To describe daily practices regarding safety monitoring of methotrexate prescribed at low-(i.e.≤30mg/week). To identify determinants of these practices. To assess association between monitoring and early methotrexate discontinuation.
Population-based cohort study using the French claims database échantillon généraliste de bénéficiaires (EGB) over the period 2009–2015. Incident methotrexate users were included. The pre-treatment and post-treatment monitoring prescribed to these patients was analyzed. Determinants of monitoring were identified using a logistic regression model. Association between monitoring and early methotrexate discontinuation was assessed using Cox proportional-hazards model.
During the study period, 615924 individuals had data in the EGB and 2472 (0.40%) were incident methotrexate users (63.3% women; mean age: 54.7±17.8 years; mean weekly dosage: 13.0±5.3mg). Among these incident users, only 50–70% had an albumin testing (67.0%); HIV (49.7%), hepatitis B (54.8%) or C (55.0%) serology; or chest X-ray (57.4%) within the year before initiating methotrexate. Only 65.7% had a least one CBC, transaminase and urea-creatinine testing combined within the three months before initiation. During the first three months of exposure, the median number of CBC, transaminase and urea-creatinine testing was 2 [1–4], 2 [1–4], and 2 [1–3], respectively. The monitoring modalities depend more on prescriber characteristics than on patient or treatment characteristics. There was a significant positive association between frequency of monitoring during exposure and early methotrexate discontinuation.
Monitoring of patients prescribed low-dose methotrexate is much less frequent than recommended. Frequent monitoring is associated with early methotrexate discontinuation.Le texte complet de cet article est disponible en PDF.
Keywords : Methotrexate, Safety, Monitoring, Population-based