Abstract Background While data from USA and Canada demonstrate an opioid overdose epidemic, very little nation‐wide European studies have been published on this topical subject. Methods Using a nationally representative sample of the French Claims database (>700,000 patients), the exhaustive nationwide hospital discharge database, and national mortality registry, all patients dispensed at least one prescription opioid ( PO ) in 2004–2017 were identified, to describe trends in PO analgesic use, shopping behaviour, opioid‐related hospitalizations and deaths. Annual prevalence of PO use and shopping behaviour (≥1 day of overlapping prescriptions from ≥2 prescribers, dispensed by ≥3 pharmacies) was estimated. Results In 2004–2017, the annual prevalence of weak opioid use codeine, tramadol and opium rose by 150%, 123%, and 244%, respectively ( p < 0.05). Strong opioid use increased from 0.54% to 1.1% (+104%, p < 0.05), significantly for oxycodone (+1950%). Strong opioid use in chronic noncancer pain rose by 88% ( p < 0.05) and 1180% for oxycodone. Opioid shopping increased from 0.50% to 0.67% (+34%, p < 0.05), associated with higher mortality risk HR = 2.8 [95% confidence interval ( CI ): 1.2–6.4]. Opioid‐related hospitalizations increased from 15 to 40 per 1,000,000 population (+167%, 2000–2017), and opioid‐related deaths from 1.3 to 3.2 per 1,000,000 population (+146%, 2000–2015). Conclusions This study provided a first European approach to a nationwide estimation with complete access to several national registries. In 2004–2017 in France, PO use excluding dextropropoxyphene more than doubled. The increase in oxycodone and fentanyl use, and nontrivial increasing trend in opioid‐related morbidity–mortality should prompt authorities to closely monitor PO consumption in order to prevent alarming increases in opioid‐related morbidity–mortality. Significance In 2004–2017, prescription opioid use in France at least doubled and oxycodone use increased particularly, associated with a nontrivial increase in opioid‐related morbidity–mortality. Although giving no indication for an ‘opioid epidemic,’ these findings call for proper monitoring of opioid use.