Real-time prediction of rapid weight change in children with cystic fibrosis who have initiated modulator therapy.

BACKGROUND & AIMS: Broad uptake of elexacaftor/tezacaftor/ivacaftor (ETI) modulator therapy has substantially improved nutritional status among children with cystic fibrosis (CF), but post-ETI weight change may be suboptimal without personalized monitoring of body composition and physical function markers. We developed a prediction model including these markers and evaluated clinical thresholds assessing individualized risk of rapid weight change over time.

METHODS

A retrospective longitudinal cohort study was performed (n = 27 children with CF aged 6-11 years starting ETI) to predict age-related body mass index (BMI) percentile over time via random forest and Bayesian elastic net. Model-based clinical thresholds were identified for rapid weight change over time.

RESULTS

Median (Q1, Q3) rate of change in BMI-for-age percentile was 2.06 (-2.42, 5.34). Most (92.6%) met clinical thresholds for rapid weight change; 48.2% experienced weight cycling. Body composition and physical function predictors improved model deviance (-2 log-likelihood: 569.05 vs. 820.42) and forecasting performance (mean absolute percentage error [SD]: 0.12 [0.22] vs. 0.85 [1.81]). Greater hand grip strength corresponded to increased BMI-for-age percentile over time and was a criterion for clinical thresholds. Younger age and higher fat mass index (FMI) and fat-free mass index (FFMI) z-scores at enrollment were associated with lower BMI-for-age percentile improvement during treatment. Baseline body composition and physical function markers yielded similar predictive accuracy to inclusion as time-varying predictors. Model performance using BMI z-scores yielded minor differences.

CONCLUSION

Monitoring post-ETI body composition and hand grip strength may sufficiently predict rapid weight change in pediatric CF, thereby improving nutritional health and modulator benefit.