Predicting Treatment Responses in Patients With Osteoarthritis: Results From Two Phase III Tanezumab Randomized Clinical Trials

Abstract

Prediction of treatment responses is essential to move forward translational science. Our question was to identify patient-based variables that predicted responses to treatments. We conducted secondary analyses on pooled data from two randomized phase III clinical trials (NCT02697773 and NCT02709486) conducted in participants with moderate to severe osteoarthritis randomized to subcutaneous placebo (n = 514) or tanezumab 2.5 mg (n = 514). We used gradient boosted regression trees to identify variables that predicted Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) Pain subscale scores at Week 16 and marginal plots to determine the directional relationship between each variable category and responses to placebo or tanezumab within the models. We also used Virtual Twins models to identify potential subgroups of response to the active treatment vs. placebo. We found that responses to placebo were predicted by baseline WOMAC Physical Function, baseline WOMAC Pain, the radiographic classification of the index joint, and the standard deviation of diary pain scores at baseline. In contrast, baseline WOMAC Pain along with failure of prior medications, duration of disease, and standard deviation of diary pain scores at baseline were predictive of tanezumab responses as expressed by the WOMAC Pain scores at Week 16. Those who responded to tanezumab vs. placebo were identified based on the radiographic classification of the index joint and either age or smoking status. These secondary-data analyses identified distinct and common patient-based variables to predict response to placebo or tanezumab. These findings will inform the design of future clinical trials, helping to move forward clinical pharmacology and translational science.

Figure 1

The black lines represent baseline values for each individual patient, sorted by severity. Individual colored lines represent the change from baseline at Week 16 for each individual patient (colored lines extending below the black line indicate lower score and, thus, improvement at Week 16). The pretreatment (baseline) and posttreatment (Week 16) boxplots show median (middle horizontal line in the box), and quartiles 1 and 3 (bottom and top lines of the box). Lines extend from the box to the smallest and largest observations no further than 1.5 times the interquartile range from quartile 1 and quartile 3, respectively, and any data beyond this range are plotted individually. WOMAC Pain subscale ranges from 0 = no pain to 10 = extreme pain. WOMAC, Western Ontario and McMaster Universities Osteoarthritis Index.

Figure 2

Gradient boosted regression trees for variables predictive of WOMAC Pain scores at Week 16 in the placebo arm. We used gradient boosted regression trees to identify the variables that contributed most to WOMAC Pain scores at Week 16 in patients who received placebo. These analyses identified baseline WOMAC Physical Function, baseline WOMAC Pain, the KL grade of the index joint, and the SD of diary pain scores at baseline as the most important predictors of WOMAC Pain scores at Week 16. The overall predictive ability of the model for the data provided was 22%, and the estimated predictive ability for new data was 9.8%. BMI, body mass index; KL, Kellgren-Lawrence; OA, osteoarthritis; SD, standard deviation; WOMAC, Western Ontario and McMaster Universities Osteoarthritis Index.

Figure 3

Gradient boosted regression trees for variables predictive of WOMAC Pain scores at Week 16 in the tanezumab 2.5-mg arm. We used gradient boosted regression trees to identify the variables that contributed most to WOMAC Pain scores at Week 16 in patients who received tanezumab 2.5 mg. These analyses identified baseline WOMAC Pain, failure of prior medications, duration of disease, and the SD of diary pain scores at baseline as the most important predictors of WOMAC Pain scores at Week 16. The overall predictive ability of the model for the data provided was 22.9%, and the estimated predictive ability for new data was 6.4%. BMI, body mass index; KL, Kellgren-Lawrence; OA, osteoarthritis; SD, standard deviation; WOMAC, Western Ontario and McMaster Universities Osteoarthritis Index.

Figure 4

Subgroup identification with enhanced response to tanezumab 2.5 mg vs. placebo. Virtual Twins analyses using a random forest model split the data by KL grade in the index joint and then age or tobacco class. Two subgroups of patients were predicted to have an increased response to tanezumab vs. placebo that was greater than that estimated by the linear model. The first subgroup contained patients aged >68 years with a KL grade of 2 or 3 in the index joint. The second subgroup contained patients who had never smoked with a KL grade of 4 in the index joint. Correction of these analyses indicated that optimism bias of the method was smaller than the enhanced treatment effect, suggesting that these findings may generalize to new data. For failed drug categories in Panel (a), Cat A contained patients with contraindication to NSAID, opioid, and/or tramadol, inadequate pain relief for all three classes, inadequate pain relief for only NSAID and/or tramadol, no contraindication but intolerability to NSAID, opioid, and/or tramadol. Cat B contained patients with inadequate pain relief for only NSAID and opioid, no contraindication, intolerability to NSAID and/or tramadol/opioid and unwilling to take tramadol/opioid. KL, Kellgren-Lawrence; Cat, Category; NSAID, nonsteroidal anti-inflammatory drug; WOMAC, Western Ontario and McMaster Universities Osteoarthritis Index.