#MDA2021 – Factors Likely to Predict Wheelchair Use With FSHD1 Detailed

Editor’s note: The Muscular Dystrophy News Today team is providing in-depth coverage of the 2021 MDA Virtual Clinical and Scientific Conference, March 15–18. Go here to read the latest stories from the conference.

Factors related to age, medication use, and co-existing conditions predict the likelihood of a person with facioscapulohumeral muscular dystrophy type 1 (FSHD1) becoming dependent on a wheelchair, a machine learning analysis showed. 

Women and those with facial weakness are at higher risk of needing a wheelchair in the future, the study also found.

These findings were in the oral presentation “Predictors of functional outcomes in patients with facioscapulohumeral muscular dystrophy,” given at the 2021 MDA Virtual Clinical & Scientific Conference.

FSHD1 is caused by abnormal activity of the DUX4 gene due to missing repeat DNA segments — known as D4Z4 repeats. Normally, people carry more than 10 repeats, but those with FSHD1 have between one to 10 repeats.

The disorder is characterized by progressive weakening of muscles, starting in the face, shoulders, and upper arms. Generally, the progression is relatively slow, and symptoms can begin in infancy or adulthood. FSHD1 also is highly variable in clinical presentation, and an estimated 20% of individuals require a wheelchair by the age of 50. 

However, little is known about factors that may predict the loss of functional abilities over time. 

Researchers at the University of Kansas Medical Center, with colleagues at the local Children’s Mercy Hospital and computer specialists at AIbytes, analyzed data collected from the U.S. National Registry of FSHD to identify factors that may predict the transition to wheelchair use. Their study was financially supported by the FSHD Society. 

“For this study we specifically focused on wheelchair use given that it’s such a memorable moment in one’s lifetime when they go from being ambulatory [walking] to needing a wheelchair to get around,” said Natalie Katz, a child neurologist at Children’s Mercy who gave the presentation.

Anonymous data from 578 FSHD1 participants (52.1% men), with an average of nine years follow-up based on annual patient surveys (ranging up to 18 years), were collected and analyzed using traditional epidemiological approaches — which focus on the distribution and determinants of disease — as well as machine learning computer algorithms to predict wheelchair use.

Most of these patients (58.8%) had four to seven D4Z4 repeats, while 10.4% carried one to three repeats, and 20.6% had eight to 10 repeats. A significantly higher number of women had one to three repeats than men (37 vs. 23), whereas 76 men had eight to 10 repeats compared to 43 women. 

Epidemiological analysis found that those with one to three D4Z4 repeats were diagnosed on average at age 14, reported facial weakness as a first symptom (53.4%), and were significantly more likely to use a wheelchair at a younger age (average age of 13). 

Patients with four to seven D4Z4 repeats were diagnosed, on average, at age 30; experienced upper body weakness at symptom onset (37.3%); and progressed to wheelchair dependency at 45 years old, on average. Participants with eight to 10 repeats were diagnosed around age 40 and also first reported upper body weakness (38.1%), and, on average, required a wheelchair by age 54. 

Katz noted that women were more likely to progress to a wheelchair than men, which “may be due in part to the slightly higher proportion of women that we had in our study in that one to three repeat category.”

Machine learning was then used to investigate why a person with FSHD1 progresses to a wheelchair, evaluating 189 features and focusing on age, sex, number of D4Z4 repeats, age at symptom onset and diagnosis, education, body mass index (BMI, a measure of body fat), medication use, and co-existing medical conditions. 

Several age-related factors predicted wheelchair use, such as longer disease duration and older age at the time of the analysis. Younger age at symptom onset and disease diagnosis also predicted wheelchair use.

Patients using a higher number of medications were more likely to need a wheelchair in the future. In addition, co-existing conditions (comorbidities) also predicted wheelchair use. Conditions noted included breathing difficulties, pneumonia, or arthritis, high blood pressure, constipation, and heart problems, “but then not having them did not necessarily influence away from wheelchair use,” Katz said. 

A higher risk of wheelchair use was found in women, those with a lower BMI, and patients who experienced facial weakness. A lower risk of wheelchair dependency was seen in people with eight to 10 D4Z4 repeats.

Finally, an analysis to determine whether the use of certain types of medications influenced wheelchair use found all classes of therapies were associated with wheelchair use, except for amino acid supplements. 

“Smaller repeat lengths were associated with earlier age at diagnosis, facial weakness as the presenting symptom, and earlier progression to wheelchair use,” the team wrote. “Machine learning analysis found that age-related features, medication use, and medical co-morbidities had a stronger influence on wheelchair use than did genetics or presenting symptoms.”

Further studies are planned to verify machine learning data with larger groups of patients, further investigate the contribution of specific medications, and evaluate the impact of co-existing conditions on wheelchair use, Katz said. 

Their results, she added, may help to identify valid outcome measures for future clinical trials testing therapies for FSHD1.