Reconsidering Individuals With Normal Hearing

Dementia, called the “greatest global challenge for health and social care in the 21st century,”^1(p2673) occurs in 47 million persons globally. This number is projected to triple by 2050. With no cure and no treatments to alter its natural history, public health prevention efforts are paramount.Hearing loss (HL) is a novel yet treatable risk factor for dementia.¹ In the article by Golub and colleagues² in this issue of JAMA Otolaryngology-Head&Neck Surgery, the importance of understanding the association between hearing and cognitive performance for older adults is highlighted. Novel to this study is the focus on adults with hearing in the normal range.

The association between HL and cognition has been consistently reported in observational studies. This association could be explained by a common cause between hearing and cognition that is associated with both conditions (eg, micro-vascular disease).³ Alternatively, the following mechanistic pathways have been hypothesized to link HL and cognition: (1) sensory deprivation from HL that places an individual at greater risk of cognitive decline; (2) information degradation from HL that leads to neuroplastic changes, deafferentation, or change in brain structure; and (3) HL requiring additional cognitive resources that detract from the brain’s ability to compensate for decreased cognition.³

Although an increasing number of studies have investigated the association between hearing and cognition in older adults, few have considered a targeted exploration of cognition for those with hearing levels within what has traditionally been deemed normal hearing. Golub and colleagues² examined the association of HL and cognition with a novel yet simple alteration by redefining and examining what is considered the normal range of hearing for adults. In creating a strict normal definition for HL (a pure-tone average ≤15 dB) compared with the more widely used 25 dB or less,⁴ Golub and colleagues² explored the association between divisions of what has been considered normal hearing and question how this relates to our understanding of the association between HL and cognitive performance. They found a 2.28-point decrease in the Digit Symbol Substitution Test cognitive score per 10-dB decrease in hearing ability among individuals classically defined as having normal hearing compared with a 0.97-point decrease among those with HL. Surprisingly, a greater association between hearing and cognitive impairment was found in individuals with normal hearing, especially within the strict normal range, compared with those with HL. Prior research on the association between HL and cognitive impairment has generally found a dose-dependent response, with a greater degree of HL associated with decreased cognitive scores and cognitive function.⁵ However, the results reported by Golub and colleagues² run counter to previous findings because they found that the greatest association with cognitive performance was seen in those with the lowest degree of HL and within the threshold of the traditional range of normal hearing (HL of 0-15 dB).

We caution the reader about the assertion by Golub and colleagues² that an association between HL and cognition begins in those with a pure-tone average greater than 25-dB HL. To our knowledge, no previous research has attempted to define the threshold at which hearing may influence cognition: this threshold is unknown. Likewise, the cut point for HL has not always been clearly defined at 25-dB HL and has a complicated history.⁶ Ultimately, the labeling of the normative values is determined via average responses per frequency within a sample population rather than a meaningful threshold above which we know that negative consequences may occur. However, Golub and colleagues² presented an important perspective that questioned the use of the classic delineation of normal hearing at 25-dB HL within the context of research on hearing and cognitive performance.

Before any formalized conclusions regarding this association can be drawn, more work is needed on how hearing that is classically defined as normal is associated with cognitive performance. Because of the surprising nature of these results, speculation for external factors that could lead to the present study’s findings is warranted. Golub and colleagues² are to be commended for their concerted effort to capture a diverse study population with inclusion of both a population of white race/ethnicity and a Hispanic population. Yet, the Hispanic population is at greater risk of impaired cognition owing to external factors that may remain as unmeasured confounders in this analysis,⁷ with cognitive impairment measured at one time in the study by Golub and colleagues² being more susceptible to confounding because of educational level and other social factors than are analyses of change. Study models by Golub and colleagues² adjust for age, sex, educational level, and cardiovascular disease, leaving many additional aspects of an individual’s circumstance (eg, socioeconomic status) unmeasured. It is particularly challenging in cross-sectional studies of cognition to adequately adjust for potential confounders; therefore, inferences cannot be made. Considering that more participants were Hispanic and had a hearing PTA within the strict normal range, we cannot assess whether the results demonstrating an equivalent or greater association in individuals with normal hearing vs those with HL are not instead attributable to external factors that are associated with Hispanic ethnicity, which remain unaccounted for in the study.

It is worth considering, as discussed by Golub and colleagues², that the use of pure-tone average may be an inappropriate way to define hearing in the association between hearing and cognition, and the surprising findings may instead be explained by speech-in-noise perception ability and not peripheral hearing. Further study is needed to answer these questions.

Although the research by Golub and colleagues² requires replication and further thought regarding the implications of the results, the findings bring awareness to the potential informative nature of hearing within a normal range. The perspective presented is important, and future work should address 2 key questions. What can we learn from those who do not have classic HL? How might this information, if replicated, inform our treatment recommendations for patients with hearing difficulties but normal hearing, especially when considering dementia prevention?