OBJECTIVE: Patients prescribed non-selective (ns) NSAIDs are at risk of gastrointestinal (GI) adverse events, hence the co-prescription of gastroprotective agents (GPAs) as prophylactic therapeutic options. Concerns over appropriate GPA co-prescribing, especially for high-risk patients, prompted this assessment of the level of GPA co-prescription for patients at GI risk. METHODS: A primary care database (DIN-LINK), representative of the UK, containing records of over 800,000 patients was used. Patients with osteoarthritis and/or rheumatoid arthritis who received nsNSAIDs and co-prescribed GPA between September 2003 and August 2005 were identified. Gastroprotection level (%) was defined by days on GPA within any three months of nsNSAID use. Patients were grouped by level of gastroprotection. For each month, the percentage of patients at different levels of gastroprotection was calculated and analysed according to GI risk factors (NSAID use frequency, serious co-morbidity, aspirin use, and previous GI adverse events). RESULTS: Approximately 25,000 patients with nsNSAID prescriptions were identified. Of these, over half were frequent nsNSAID users (prescribed nsNSAIDs for at least 75% of the observed period), 2.5% had previous GI adverse events, 12% were taking aspirin, and over two thirds had serious co-morbidity. Of total nsNSAID users, 18% used GPAs for the full period (100%); of frequent nsNSAID users, 19.5%. By analysing patients with risk factors and frequent nsNSAID use, the percentage of those using GPAs for the full period was less than 40% with previous GI adverse events, 30% for aspirin takers, and 22% with serious co-morbidities. High, but not full (80-99%), GPA use was achieved in similarly low numbers of patients with risk factors. CONCLUSION: Prescription data revealed frequent NSAID users receive little co-prescribed GPAs. Presence of risk factors of NSAID-induced GI risk did not increase GPA use, which may increase GI-related hospitalisation risk.