AbstractPurposeIncreasing use of tramadol for chronic non‐cancer pain is concerning since tramadol users may be at risk of developing recurrent opioid use with increasing opioid consumption and co‐medication. Therefore, we investigated a complete national cohort of tramadol users.MethodsThe study population (154 042 adult individuals in Norway, who redeemed ≥ one tramadol prescription in 2012) was stratified into four groups according to their opioid use 2 years before their first tramadol prescription in 2012 and followed until 2016. Information on all dispensed opioid analgesics, benzodiazepines (BZDs), and BZD‐related Z‐hypnotics were retrieved from the Norwegian Prescription Database.ResultsSix percent of opioid naïve tramadol users (no opioid use 2 years before tramadol use in 2012) became recurrent users (received opioids annually during 4‐year follow‐up), almost doubled their mean opioid consumption (66 to 108 defined daily doses [DDD]). One‐quarter proceeded to strong opioids or was co‐medicated with BZDs, one‐third with Z‐hypnotics. Among former weak opioid users, 39.8% became recurrent users, 18.7% proceeded to strong opioids, mean opioid consumption increased slightly, one‐third used BZDs, or Z‐hypnotics concurrently. Among former strong opioid and users in palliative care; 61%, 70% became recurrent users and developed a similar prescription pattern (high and increasing mean opioid consumption, 301 to 318, 413 to 430 DDD); half of them proceeded to strong opioids and/or used BZDs or Z‐hypnotics concurrently.ConclusionsMany patients who developed recurrent opioid use received prescriptions which substantially conflicted with existing guidelines and might lead to problematic opioid use.