Karin Magnusson1, Aleksandra Turkiewicz2, Simon Timpka3, Martin Englund4. 1. Lund University, Lund, Sweden, and Diakonhjemmet Hospital, Oslo, Norway. 2. Lund University, Lund, Sweden. 3. Lund University, Malmö, Sweden. 4. Lund University, Lund, Sweden, and Boston University School of Medicine, Boston, Massachusetts.
Abstract
OBJECTIVE: To simplify the previously published Nottingham 12-year risk prediction model for knee osteoarthritis (OA) and examine whether it can be used to predict the 40-year risk of knee OA in young men. METHODS: Our cohort included 40,118 men who were 18 years of age and had undergone military conscription in Sweden from 1969 to 1970. Diagnostic OA codes were obtained from the Swedish National Patient Register for persons registered from 1987 to 2010. The original Nottingham model included as predictors age, sex, body mass index (BMI), knee injury, occupational risk, and family history of OA, with a receiver operating characteristic area under the curve (AUC) of 0.70 (95% confidence interval [95% CI] 0.61-0.79) in the model development sample, and AUC 0.60 (95% CI 0.58-0.63) in an external validation sample. In our sample, we used predictors that were available only in adolescence (age, BMI, and knee injury) and evaluated the discrimination of the simplified model using AUC. RESULTS: The AUC statistic of the modified knee OA model to predict 40-year risk was 0.60 (95% CI 0.59-0.61). Hence, using the reduced model, an 18-year-old man with a BMI of 30 and a knee injury would have 3 times the risk of developing knee OA within 40 years when compared to a man of similar age having a BMI of 25 and no knee injury (predicted risks 22% and 7%, respectively). CONCLUSION: The 40-year risk of knee OA on individual and population levels can be predicted in 18-year-olds from a few easily measured covariates with moderate discrimination. The discrimination of this simplified model based on data available in adolescents was comparable to that of the full Nottingham model in middle-aged individuals.
OBJECTIVE: To simplify the previously published Nottingham 12-year risk prediction model for knee osteoarthritis (OA) and examine whether it can be used to predict the 40-year risk of knee OA in young men. METHODS: Our cohort included 40,118 men who were 18 years of age and had undergone military conscription in Sweden from 1969 to 1970. Diagnostic OA codes were obtained from the Swedish National Patient Register for persons registered from 1987 to 2010. The original Nottingham model included as predictors age, sex, body mass index (BMI), knee injury, occupational risk, and family history of OA, with a receiver operating characteristic area under the curve (AUC) of 0.70 (95% confidence interval [95% CI] 0.61-0.79) in the model development sample, and AUC 0.60 (95% CI 0.58-0.63) in an external validation sample. In our sample, we used predictors that were available only in adolescence (age, BMI, and knee injury) and evaluated the discrimination of the simplified model using AUC. RESULTS: The AUC statistic of the modified knee OA model to predict 40-year risk was 0.60 (95% CI 0.59-0.61). Hence, using the reduced model, an 18-year-old man with a BMI of 30 and a knee injury would have 3 times the risk of developing knee OA within 40 years when compared to a man of similar age having a BMI of 25 and no knee injury (predicted risks 22% and 7%, respectively). CONCLUSION: The 40-year risk of knee OA on individual and population levels can be predicted in 18-year-olds from a few easily measured covariates with moderate discrimination. The discrimination of this simplified model based on data available in adolescents was comparable to that of the full Nottingham model in middle-aged individuals.