Background <p>This prospective, long-term observational study, initiated in 2002, aimed to characterize clinical and laboratory data, whole body MRI detected lesions, and treatment responses in 37 juvenile patients with chronic non-bacterial osteomyelitis at a time when biological DMARDs were not yet standard therapy.</p> Methods <p>Patients were assessed at baseline and at 1 (without MRI), 3, 6, 12, 18, 24, 36, 48, 60 months. All patients received naproxen as first-line therapy. Clinical management allowed for escalation to sulfasalazine, pamidronate, and glucocorticoids as needed. Treatment response was evaluated using the pedCNO disease activity score (30/50/70/90% improvement). Further composite numeric disease activity (DA) scores– the CARRA CDAS and a new MRI DAS - were applied.</p> Results <p>The mean age at disease onset was 10.8 years, with a diagnostic delay of 5.8 months. Naproxen was the initial treatment in all patients. Second-line therapy was initiated in 10 patients due to inadequate improvement in physician global assessment of disease activity, patient-reported overall wellbeing or MRI lesions. Escalated therapies included sulfasalazine (<i>n</i> = 10), bisphosphonates (<i>n</i> = 1), methotrexate (<i>n</i> = 1), and short- (&lt; 4 wks) or long-term oral glucocorticoids (<i>n</i> = 5 and <i>n</i> = 3, respectively). The mean number of clinical lesions decreased from 2.1 to 0.4 at 12 months and reached 0.15 at 60 months. MRI-detected lesions declined from 5.0 to 2.25 at 12 months and to 1.1 at 60 months.</p> Conclusion <p>Most children experienced favourable long-term outcomes. Clinical improvement occurred more rapidly than radiologic resolution. Patients with insufficient response to NSAIDs should be considered for a treat-to-target approach, including the use of conventional and biologic DMARDs.</p> Trial registration <p>A trial registration EUDRA CT was not available at the time the study was started. Informed consent was given by all parents.</p>