To examine the 6-year course of nontraumatic knee symptoms in adults in general practice, to identify prognostic factors for unfavorable outcome, and to develop a clinical prediction rule.


Adults (ages >35 years) with incident nontraumatic knee symptoms (n = 549) were followed for 6 years. Multivariable logistic regression analysis was used to identify prognostic factors associated with an unfavorable outcome, the area under the receiver operating curve (AUC) was calculated to determine discriminative ability, and a clinical prediction rule was developed. Unfavorable outcome is defined as persistent knee symptoms at 6-year followup or having undergone knee replacement surgery during followup.


At 6-year followup, 42.1% of patients had an unfavorable outcome. Having persistent knee symptoms (odds ratio [OR] 5.31, 95% confidence interval [95% CI] 3.27–8.61) and fulfilling the clinical American College of Rheumatology (ACR) criteria for osteoarthritis (OA; OR 2.65, 95% CI 1.48–4.73) at 1-year followup were significantly associated with unfavorable outcome, while fulfilling the clinical ACR criteria for OA at baseline was not. Baseline factors independently associated with an unfavorable outcome were low/middle education level, comorbidity of the skeletal system, duration of knee symptoms of >3 months, bilateral knee symptoms, self-reported warm knee, history of nontraumatic knee symptoms, valgus alignment, pain at passive knee flexion/extension, and bony enlargement of the knee joint (AUC 0.80).


Nontraumatic knee symptoms in adults in general practice appear to become a chronic disorder in nearly half of the patients. The developed clinical prediction rule with 10 baseline prognostic factors can be used to select high-risk patients for an unfavorable outcome at long-term followup.