A Retrospective Longitudinal Study of 460 Patients with ABCA4-Associated Retinal Disease.

PURPOSE: To investigate the distribution of genotypes and natural history of ABCA4-associated retinal disease in a large cohort of patients seen at a single institution.

DESIGN: Retrospective, single institution, cohort review.

SUBJECTS: Patients seen at the University of Iowa between November 1986 and August 2022 clinically suspected to have disease caused by sequence variations in ABCA4.

METHODS: DNA samples from participants were subjected to a tiered testing strategy progressing from allele-specific screening to whole genome sequencing. Charts were reviewed and clinical data were tabulated. The pathogenic severity of the most common alleles was estimated by studying groups of patients that shared one allele. Groups of patients with shared genotypes were reviewed for evidence of modifying factor effects.

MAIN OUTCOME MEASURES: Age of first uncorrectable vision loss, best corrected visual acuity, and the area of the I2e isopter of the Goldmann visual field.

RESULTS: A total of 460 patients from 390 families had convincing clinical features of ABCA4-associated retinal disease. Complete genotypes were identified in 399 patients and partial genotypes in 61. The median age of first vision loss was 16 years (range 4-76 years). 265 of the families (68%) had a unique genotype and no more than 10 patients shared any single genotype. Review of the patients with shared genotypes revealed evidence of modifying factors that in several cases resulted in a greater than 15-year difference in age of first vision loss. 241 different alleles were identified among the members of this cohort and 161 of these (67%) were found in only a single individual.

CONCLUSIONS: ABCA4-associated retinal disease ranges from a very severe photoreceptor disease with an onset before age 5 years to a late-onset RPE-based condition resembling pattern dystrophy. Modifying factors frequently impact the ABCA4 disease phenotype to a degree that is similar in magnitude to the detectable ABCA4 alleles themselves. It is likely that the majority of patients in any cohort will have a unique genotype. The latter observations taken together suggest that patients' clinical findings will in most cases be more useful for predicting their clinical course than their genotype.