On-treatment opioid agonists reduce mortality
Periods on opioid agonist treatment were linked to substantially fewer deaths than periods off treatment in French primary care patients.
*Nationwide retrospective cohort study; Level 2b (OCEBM).

Citation

Dupouy J, Druel V, Verges Y, Pariente A, Lapeyre-Mestre M. Opioid agonist treatment and risk of mortality in French primary care: a nationwide, retrospective cohort study. Lancet Public Health. 2026;11:e449–e456.

Background

Opioid agonist treatment is widely used in France, especially buprenorphine, but long-term mortality data in primary care have been limited. This study compared death risk during time on versus off treatment.

Patients

175,191 people aged ≥15 years in France starting opioid agonist treatment (2010–2022) with at least two dispensings within 35 days; excluded non-incident users (no opioid agonist treatment-free year before starting) and those outside the main national insurance scheme.

Intervention

Time on opioid agonist treatment (buprenorphine, buprenorphine–naloxone, or methadone), based on pharmacy dispensing records.

Control

Time off opioid agonist treatment.

Outcome

(Primary) All-cause mortality at 1 year. (Secondary) All-cause mortality at 2, 5, and 7 years; selected causes of death.

Follow-up Period

Median 3,320 days (about 9.1 years).

Results

Comparison Time point Adjusted hazard ratio (95% confidence interval) Absolute difference (deaths per 1000 person-years)
On treatment vs off treatment 1 year (primary) 0.41 (0.37–0.44) 21.45 fewer
On treatment vs off treatment 2 years 0.36 (0.34–0.39) —
On treatment vs off treatment 5 years 0.36 (0.34–0.38) —
On treatment vs off treatment 7 years 0.40 (0.37–0.42) —
Buprenorphine vs off treatment 1 year 0.16 (0.14–0.18) —
Methadone vs off treatment 1 year 0.58 (0.53–0.64) —
Approximate number needed to treat: about 47 person-years on treatment to prevent 1 death (using the 1-year rate difference).

Limitations

Observational design limits causal claims; unmeasured differences between on- and off-treatment periods could partly explain results. Medication dispensing does not guarantee medication taken. The larger apparent benefit for buprenorphine could reflect patient selection (indication bias). Cause-of-death data were not available for the most recent years and may be incomplete for suspicious deaths.

Funding

French medicine-safety agency and French National Health Insurance (EPI-PHARE).

Clinical Application

Reinforces prioritizing rapid access and long-term retention in opioid agonist treatment within primary care; interpret the larger buprenorphine effect cautiously due to likely patient selection.