Very few epidemiological data focused on the impact of non-adherence on survival and stroke incidence in France.

The objective of this study was to describe time-trends adherence to NOAC from 2012 to 2015, and to estimate the impact of non-adherence to NOAC on stroke incidence and mortality as a competing cause.

All AF patients>20 years old, residing in France, and newly treated by NOAC between 2012 and 2015 were included using the nationwide SNDS database. Patients were followed for 2 calendar years or till the first occurrence of one of the following events: death, switch to a VKA, hospitalisation for stroke, any systemic embolism, or hemorrhages. Adherence to NOAC treatment was estimated by the proportion of days covered (PDC) by the treatment. We used a Fine and Gray regression model to estimate sub-distribution hazard ratio (SHR) for stroke and Cox regression models to compute cause-specific hazard ratios (CSHR) death.

In 2015, 123,540 patients were estimated newly treated by NOAC for AF indication versus 54,129 in 2012. The median follow-up was 2 years. Proportion of patients with PDC>80% increased between 2012 and 2015, from 59.1% to 65.0%. Among treated patients, the proportion of days covered by standard dosage increased from 33.2% to 59.2%. Adherent patients were older, were more often women, had more frequently history of stroke, heart failure, treated hypertension and diabetes, compared with non-adherent patients. In multivariate analysis, non-adherence to NOAC treatment was associated with 14% increase in stroke occurrence and 23% increase in death.

Adherence to NOAC increased between 2012 and 2015 with a higher proportion of patients treated with standard doses, and less patients treated with reduced or out-of-recommendation dosage. Non-adherence to NOAC was associated with important increased risk of stroke and death highlighting the crucial need improve patients’ adherence to NOAC.