Background Nonsteroidal anti-inflammatory drugs are associated with gastrointestinal (GI) damage. The Celecoxib vs. Omeprazole and Diclofenac for At-Risk Osteoarthritis and Rheumatoid Arthritis Patients (CONDOR) trial showed that a haemoglobin drop ≥2 g/dL adjudicated as either of defined or presumed GI origin was the most frequent component/event for the composite GI primary end point. This adverse event is potentially clinically relevant in long-term NSAID treatment. Aim To define potential risk factors associated with a decrease in haemoglobin/haematocrit. Methods Post hoc analysis of the CONDOR trial was conducted in the intention-to-treat population. Clinically significant blood loss was defined as: (i) a haemoglobin drop ≥2 g/dL and/or a haematocrit drop ≥10%; and (ii) blood loss adjudicated as either of defined or presumed GI origin. Fifteen risk factors were evaluated by stepwise logistic regression. Each factor had to be significant at <0.20 α to be included in the model. Results A total of 64/3774 (1.7%) osteoarthritis (OA) patients had decreased haemoglobin/haematocrit and were adjudicated to the GI endpoint. Significant risk factors, at the 0.20 α level found to be associated with clinically significant blood loss in OA patients included [odds ratio (80% CI)] baseline C-reactive protein (CRP) levels [2.27 (1.46-3.53)], history of gastritis and history of GI intolerance [1.55 (1.06-2.28)], positive Helicobacter pylori at screening [1.54 (1.07-2.22)], increasing age [1.17 (1.04-1.32)] and body mass index [BMI; 1.03 (1.00-1.06)]. Conclusions Monitoring for decreases in haemoglobin should be considered for all OA patients and especially those with an increased age, BMI, history of gastritis and GI intolerance, CRP levels >1 mg/dL and/or positive H. pylori status, as this may affect their clinical management. © 2012 Blackwell Publishing Ltd.