Item Response Theory Quantifies the Relationship Between Improvements in Serum Phosphate and Patient-Reported Outcomes in Adults With X-Linked Hypophosphatemia

Abstract

Burosumab is indicated for treatment of a rare bone disease, X-linked hypophosphatemia (XLH). The aim of this analysis was to evaluate the relationship between a treatment response biomarker and patient-reported outcomes (PROs). Longitudinal data for PROs were obtained from adults with XLH from a phase III study. Individual rich time profiles of the biomarker, serum phosphate were simulated using a prior population pharmacokinetic-pharmacodynamic model to calculate serum phosphate exposure metrics for each 28-day treatment cycle, which were then merged with PROs data. Item response theory parameters were first estimated to map a latent variable, ψ, that is, disability score, relative to baseline. Next, the relationships between serum phosphate exposures and ψ were modeled using a nonlinear mixed-effect (NLME) modeling approach. A combined item response theory-NLME model with average serum phosphate as a predictor of ψ described PROs data well. The model estimates suggested 28%, 31%, and 25% reduction in Western Ontario and McMaster Universities Osteoarthritis Index, brief pain inventory, and brief fatigue inventory scores, respectively, with every unit increase in average serum phosphate from the lower limit of normal (2.5 mg/dL). Additionally, a time effect of ~ 0.08% improvements each week was estimated. The analysis suggested that burosumab treatment-induced improvements in serum phosphate levels are associated with improvements in PROs in adults with X-linked hypophosphatemia. The analyses confirmed the importance of prolonged serum phosphate level correction in adult patients with XLH. These results can be useful to guide the design of further studies and to design treatment optimization strategies.