1. Introduction

Osteogenesis imperfecta (OI) is a connective tissue disorder primarily caused by heterozygous variants in Type I collagen and characterized by bone fragility, skeletal deformities, and short stature [1]. OI patients with deleterious collagen variants are classified phenotypically by the Sillence classification [2], in which OI Type II is perinatal lethal, OI Type III is severe progressive deforming, and OI Type IV is moderately‐severe, whereas Type I OI is mild. The dominant, recessive, and X‐linked OI types identified subsequently, primarily by genetic etiology, comprise Types V–XXII [1, 3]. In addition to fragility and deformity of the axial skeleton, patients with OI may display a variety of secondary connective tissue features, including blue sclerae, dentinogenesis imperfecta, impaired hearing, and cardiopulmonary impairment [3].

Hearing loss (HL) has been reported to impact people with OI with an overall prevalence ranging from 28% to 66% [4–9]. In the pediatric population, the prevalence is reportedly lower, 0%–25% [5, 10–13]. The degree of HL in OI is typically mild [6, 9], but can be moderate, severe, or profound. It usually begins as conductive hearing loss (CHL) that progresses to mixed hearing loss (MHL) or sensorineural hearing loss (SNHL) with aging [5, 12, 14, 15], although SNHL is reported to occur at all ages [4, 8, 11, 15, 16]. The majority of HL in OI is described as bilateral and symmetrical for type and degree [8, 14]. Studies of HL in OI are starting to emerge beyond Europe and North America. In 2015, Lin et al. [17] associated HL with helical mutations (17%) and haploinsufficiency (21%) in a Taiwanese OI population.

Machol et al. [9] recently reported on HL in OI in a large cross‐sectional population of North American participants in the Brittle Bone Research Consortium (BBRC). In contrast to studies predominantly reporting HL as beginning in the second or third decade of life [5, 6, 14, 15, 18], the BBRC [9] reported HL beginning in the first decade of life for approximately 20% of participants with OI Type III. This rate of childhood onset is higher than reported for a European OI population with known collagen‐based OI, which found 6.3% of total ears experienced HL onset before the age of 10 years [18]. The BBRC [9] also found a higher prevalence of SNHL in females at all ages, although this difference was not significant in individuals younger than 30 years of age. A 2019 literature review by Carre et al. [14], reported no gender differences in the prevalence of HL.

Heterozygous variants in the genes encoding Type I collagen, COL1A1 and COL1A2, are responsible for 80%–90% of OI cases reported in North America [19]. Studies on the relationship between gene variants and HL are limited. In 1990, Sykes et al. [20] reported “presenile hearing loss” restricted to individuals with a COL1A1 variant in the United Kingdom. In 2004, Hartikka et al. [7] genotyped a Finnish OI population, mean age 36.4 years, among which 17 (about one‐third) had null COL1A1 variants diagnostic for OI Type I. Among 21 patients with COL1A1 and 11 with COL1A2 heterozygous missense or exon skipping variants, they found no correlation between hearing status and mutated gene. In 2011, Swinnen et al. [18] examined the correlation between HL and the underlying variant in 114 OI patients (64 families from Belgium, Netherlands, and Italy) who had developed HL before the age of 40 or had normal hearing after the age of 40. In this group, two‐thirds of patients had null variants in one COL1A1 allele, indicative of OI Type I, whereas the remaining patients had heterozygous missense or exon skipping variants in COL1A1/COL1A2. Again, a correlation between HL and mutated gene or variant type was not found.

Here, we report on audiological findings among 73 participants in the Longitudinal OI Natural History Study of the National Institute of Child Health and Human Development (NICHD) at the NIH Clinical Center, Maryland, United States. Patients were enrolled as children and received repeat audiological evaluations and follow‐up as adults. All patients had heterozygous variants altering the primary protein sequence encoded by COL1A1 or COL1A2 as the etiology of OI Type III or IV. The goals of this study were to: (1) determine whether patients with missense/exon skipping variants in a specific collagen gene or combinations of causal gene and OI type experience HL; (2) leverage longitudinal data to identify patterns in progression of HL by type and degree in relation to gene variant and OI type; and (3) identify age of HL onset.