Back

Fragile X Syndrome in Brazil: Development and Validation of a Clinical Checklist for Population Screening

Romero, L. M.; Santos, G. C. C.; Schubert, V.; Scalabrin, E. E.; Herai, R. H.

2025-10-23 epidemiology
10.1101/2025.10.21.25338500 medRxiv
Show abstract

Fragile X Syndrome (FXS) is the most common inherited cause of intellectual disability and syndromic autism, but diagnosis remains challenging due to phenotypic variability and limited access to molecular testing in low- and middle-income countries. We conducted a cohort study of 419 individuals with molecularly confirmed full mutation or mosaicism (364 males, 55 females) from the Buko Kaesemodel Institute database to characterize clinical features and develop a screening tool for referral to genetic testing. Twelve physical, cognitive, and behavioral symptoms were assessed, and statistical analyses (95% CI, Pearson correlation) combined with machine learning (Random Forest, K-means) identified the most discriminative traits. The most frequent symptoms were learning difficulties (86.9%), attention deficit (79.7%), delayed speech (73.3%), intellectual disability (72.8%), repetitive movements (70.2%), and characteristic facial features (65.4%). Macroorchidism was present in 41.8% of males. Symptom prevalence in females was similar but with fewer physical features. Moderate correlations were observed, including between intellectual disability and learning difficulties (r = 0.59). A scored checklist, incorporating sex-specific weights, achieved 95% sensitivity with ROC areas of 0.73 (males) and 0.76 (females). Key predictors differed by sex, with intellectual disability and facial features being most informative in males, and learning difficulties in females. This is the largest FXS cohort analyzed in Brazil and provides the most comprehensive symptom profile in an admixed population. The validated checklist represents a practical, low-cost tool to support early diagnosis and guide public health strategies, reducing unnecessary molecular testing while prioritizing high-risk cases.

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.