Background: The diagnosis of hip osteoarthritis is often based on clinical symptoms, such as pain and stiffness, and radiographic features. However, the association between hip pain and radiographic hip osteoarthritis (ROA) remains uncertain. Aim: To examine the association between hip pain and hip ROA. Design and setting: Cross-sectional analysis of a Dutch cohort, the Cohort Hip and Cohort Knee (CHECK) study. Methods: Participants (45-65 years) experienced hip and/or knee pain for which they had no prior consultation or were within 6 months of their first consultation to a general practitioner. Using weight-bearing anterio-posterior pelvis radiographs, definite and early-stage hip ROA were defined as Kellgren and Lawrence grade ≥2 and ≥1, respectively. Presence of ROA and pain was assessed in both hips of all participants. We assessed the association between hip pain and ROA using generalized estimating equations. Results: The prevalence of definite ROA was 11.0% (218/1982 hips), while the prevalence in painful and pain-free hips was 13.3% (105/789) and 9.5% (113/1193), respectively. The early-stage prevalence was 35.3% (700/1982 hips), 41.2% (325/789) among painful and 31.4% (375/1193) among pain-free hips. Compared to pain-free hips, the odds ratio of painful hips was 1.51 (95% CI 1.16-1.98) for definite and 1.47 (95% CI 1.24-1.75) for early-stage ROA. Conclusion: Hip pain was associated with definite and early-stage hip ROA, yet the overall ROA prevalence was modest and the prevalence among pain-free hips was substantial. Therefore, radiographs hardly help to identify hip OA patients among patients who recently newly presented with hip or knee complaints.