AIM: Prevalence, optimal diagnostic approach and consequences of clinically unsuspected osteomyelitis in diabetic foot ulcers are unclear. Early diagnosis of this infection may be crucial to ensure correct management. METHODS: We conducted a prospective study in 20 diabetic patients with a chronic foot ulcer (>8 weeks) without antibiotic pretreatment and without clinical signs for osteomyelitis to assess the prevalence of clinically unsuspected osteomyelitis and to compare the value of magnetic resonance imaging (MRI), 18F-fluorodeoxyglucose positron emission tomography (18F-FDG PET) and 99mTc-labelled monoclonal antigranulocyte antibody scintigraphy (99mTc-MOAB). Those with suggestive scans underwent bone biopsy for histology (n = 7). RESULTS: Osteomyelitis was confirmed by biopsy in seven of the 20 clinically unsuspected foot ulcers. Presence of osteomyelitis was not related to age, ulcer size, ulcer duration, duration of diabetes or HbA1c. C-reactive protein was slightly elevated in patients with osteomyelitis (35.1 +/- 16.0 mg L(-1) vs. 12.2 +/- 2.6 mg L(-1) in patients with and without osteomyelitis respectively; P = 0.07). MRI was positive in six of the seven patients with proven osteomyelitis, whereas 18F-FDG PET and 99mTc-MOAB were positive only in (the same) two patients. Of the seven patients with osteomyelitis, five had lower limb amputation and in one patient the ulcer was persisting after 24 months of follow-up. In contrast, of the 13 patients without detectable signs of osteomyelitis on imaging modalities only two had lower limb amputation and two persisting ulcers. CONCLUSIONS: Clinically unsuspected osteomyelitis is frequent in persisting foot ulcers and is a high risk factor for adverse outcome. MRI appears superior to 18F-FDG PET and 99mTc-MOAB in detecting foot ulcer-associated osteomyelitis and might be the preferred imaging modality in patients with nonhealing diabetic foot ulcers.