Mathematical modeling of the progression of Duchenne Muscular Dystrophy in San Diego in an era of allele specific oligonucleotide therapy

Abstract

Duchenne muscular dystrophy (DMD) is a relatively common, severe, progressive, X-linked muscle disorder caused by pathogenic variation in the dystrophin gene. Affected status can be determined at birth, 3-5 years before symptom onset. As effective corticosteroid and genetic therapies are becoming available for DMD, predictive models of DMD progression are needed to guide individualized care of patients and clinical trials of new treatments. We developed a multivariable, dynamical model of DMD progression, as measured by percent of predicted FVC (FVC%), through a longitudinal natural history study of 113 boys from a single center in San Diego County between 2008 – 2023. A mixed effect linear regression model of FVC% had marginal and conditional r2 values of 0.54 and y of 0.83 with five factors (Hispanic ethnicity, B= -10.0, 95% CI -16.2 to -3.89, p=.002; age at symptom onset, B= -2.4, 95% CI -3.6 to -1.3; p<.001; DMD exon 44 skip amenability, B= 13.14, 95% CI 4.0 to 22.31, p=.007; steroid therapy start <10 years, B=24.9, 95% CI 16.0 to 33.9, p<.001; years from symptom onset, B= -2.5, 95% CI -2.9 to -2.2, p<.001). Race, age when first seen at a tertiary referral center, age at molecular diagnosis and treatment with allele specific oligonucleotides were not significantly associated with respiratory progression of DMD as measured by longitudinal predicted forced vital capacity. This model of DMD respiratory progression suggests that care be individualized by earlier or more intensive treatment in Hispanic patients and those with early onset of symptoms. Clinical trials of new treatments for DMD that use FVC% as a primary or secondary end-point may benefit from designs that account for potentially confounding effects of ethnicity, DMD genotype, concurrent use of corticosteroids, and earlier age of onset of symptoms.