Investigating Risk Factors for Progression of Base of Thumb Osteoarthritis

Introduction: Risk factors for the development of basilar thumb arthritis are well described, but predictors of progression of the disease remain poorly understood. We hypothesize that clinical and biomechanical measures can help predict progression of early basilar thumb arthritis.
Methods: Ninety-one patients with basilar thumb pain and a modified radiographic Eaton-Littler Stage of 1 or less were followed prospectively. Patients received x-rays and computed tomography (CT) scans at baseline, 1.5 years, and 3 years. Baseline information on patients’ medical history, thumb and wrist range of motion, and grip and pinch strength were obtained. The primary outcome was progression at 3 years, defined as an increase in modified Eaton-Littler Stage. Following a univariate screen, a predictive model for progression was created with logistic regression using automated backward selection of clinical variables. Then, a least absolute squares shrinkage operator (LASSO) was employed to incorporate CT parameters into the model. The final model was selected based on the cross-validated area under the receiver operator curve (AUC). For secondary outcomes, the final Patient Rated Wrist and Hand Evaluation score at 3 years was modeled using linear regression; progression beyond the minimal clinically important difference in the Australian/Canadian Osteoarthritis Hand Index was modeled using logistic regression. In both models, clinical predictors were identified using automated backward variable selection, followed by selection using LASSO for incorporation of CT data.
Results: Fifty-seven patients (63%) had radiographic disease progression at 3 years; 1 was censored for having received basilar thumb arthroplasty surgery, and an additional 4 were lost to follow-up. Clinical predictors of radiographic progression included high body mass index (Odds Ratio (OR) = 1.22, 95% Confidence Interval (CI) = [1.08, 1.38]) and low range of motion in thumb abduction (OR = 0.90, 95% CI = [0.83, 0.96]). Assessment of CT parameters indicated that increased dorsal tilt of the base of the first metacarpal relative to the trapezium during grasping motions was also predictive of progression (OR = 1.26, 95% CI = [1.07, 1.48]). The cross-validated AUC for this model was 0.80. Baseline factors predictive of an increased Patient Rated Wrist and Hand Evaluation score at 3 years included the need for pain medications (p = 0.064), passive carpometacarpal extension ROM less than 20º (p = 0.008), and a higher numeric rating score for pain with pinching motions (p = 0.003). This model was associated with an r2 value of 0.41. Factors predictive of an increase in the Australian/Canadian Osteoarthritis Hand Index score beyond the minimal clinically important difference were the need for hand therapy (OR = 6.40, 95% CI = [1.25, 32.82]) and pain with grasping motions (OR = 4.32, 95% CI = [1.31, 14.27]) at baseline. The cross-validated AUC for this model was 0.55.
Conclusion: Increased body mass index and changes in carpometacarpal joint range of motion resulting in collapse of the first metacarpal into the palm, particularly during grasping motions, provided excellent predictive ability for radiographic progression of early basilar thumb arthritis by 3 years. Similar factors were associated with worsening scores on patient reported outcome metrics; however, the ability of these factors to predict clinical deterioration, as opposed to radiographic progression, was less robust. These findings may be helpful for educating patients with early basilar thumb arthritis about disease progression and for treatment planning.