Many attempts have been made to identify possible environmental and genetic risk factors associated with ARHL (Fetoni et al., 2011). Epidemiological studies have indicated that many of these factors are indeed very significant in the trajectory of ARHL. Yamasoba et al. (2013) identified four categories of risk factors associated with ARHL in humans: genetic predisposition, environment, health co-morbidities and cochlear aging. The first three risk factors associated with ARHL are reviewed here and cochlear aging will be discussed in the section on the impact of aging on the peripheral hearing system.

The inner ear is more vulnerable to age-related changes than both the external and middle ear (Schmiedt, 2010). Hence, pathophysiological changes related to cochlea and afferent neurons are discussed here. Based on underlying case history information, audiometric configurations, and temporal bone analyses, Schuknecht (1974) divided ARHL into four distinct classes: sensory (loss of hair cells and supporting cells), neural (loss of afferent neurons), strial (atrophy of cochlear lateral wall and Stria vascularis), and cochlear conductive (cochlear atrophy mainly basilar membrane and organ of Corti). Rapid high-frequency sloping hearing loss, flat hearing loss, poor word discrimination and gradual hearing loss with no pathological evidence are characteristics of sensory, strial, neural, and cochlear conductive presbycusis, respectively (Schuknecht and Gacek, 1993). Later, Schuknecht and Gacek (1993) added two more categories: (i) mixed-consisting of a mixture of pathological characteristics and (ii) indeterminate-consisting of none of the aforementioned pathological characteristics. Contrary to the above-mentioned findings, Allen and Eddins (2010) reported that hearing phenotypes do not naturally form distinct classes of ARHL and rather form a continuum. However, subtypes of ARHL can still be identified if the distribution of data are in the extremes of either flat or sloping loss (Allen and Eddins, 2010).

Evidence for age-related pathological changes were reported in human temporal bone studies as well as in animal models. Histopathological studies have reported progressive loss and degeneration of spiral ganglion cells (Mills et al., 2006b; Hinojosa and Nelson, 2011), loss of nerve fibers in spiral lamina (Belal, 1975; Mills et al., 2006a) and hypertrophy of the internal elastic lamina of the internal auditory artery (Belal, 1975) in aged adults. Progressive outer hair cell loss with no degeneration of the cells have been reported in Stria vascularis (Sha et al., 2008), spiral ganglion cell degeneration (Keithley et al., 2004), and degeneration of organ of Corti and afferent neurons (Ohlemiller and Gagnon, 2004) in mouse models. Loss of outer hair cells and detachment of the stria vascularis from the spiral ligament was seen in aging Fisher rat F344 models (Buckiova et al., 2007). Loss of marginal and intermediate cells of Stria vascularis (Gates and Mills, 2005) and loss of Stria capillaries (Gratton and Schulte, 1995) were noted in aged gerbil models.

Two most prominent physiological changes have been observed in the aging cochlear: decline in cochlear endolymphatic potentials (EP) and increase in action potential (CAP) of the auditory nerve (Gates and Mills, 2005). Recent studies have concluded that metabolic/strial presbycusis or degenerative changes in the lateral wall and stria vascularis as the predominant cause of ARHL (Gates and Mills, 2005; Ohlemiller, 2009). The above mentioned pathological changes of the stria vascularis leads to loss of expression of key ion transport enzymes, such as Na+, K+-ATPase, and the Na+, K+, Cl− co-transporter, impairment in outer hair cell functions and decline in endolymphatic potential (EP) values (Salt et al., 1987; Wangemann, 2002). The EP provides voltage for the cochlear amplifier, thus, a decline in EP results in impairment in cochlear amplifier ultimately resulting in poor hearing thresholds (Schmiedt et al., 2002; Gates and Mills, 2005).

Another physiological impairment observed in aging cochlear is asynchronous firing of the auditory nerve fibers as indicated by increased thresholds of the compound action potential of the auditory nerve (Gates and Mills, 2005). A combination of pathological changes of the spiral ganglion cells and synaptic connections between the inner hair cells and afferents and decline in EP is thought to be one of the underlying causes of the auditory nerve malfunction (Gates and Mills, 2005).

Yamasoba et al. (2013) proposed a model to explain the development of ARHL. According to this model, a number factors (aging, genetic, environmental, health co-morbidity) and their interactions contribute to ARHL. Mitochondrial dysfunction associated with reactive oxygen species (ROS) plays a major role in the age-related cochlear cells degeneration (Cheng et al., 2005; Someya and Prolla, 2010). Noise trauma also generate excess of ROS in the cochlea (Henderson et al., 2006) triggering necrotic or apoptic cell death in cochlea (Henderson et al., 2006). Oxidative stress can be accelerated by hypoxic situations resulting from decline in cochlear blood supply due to noise trauma (Wen et al., 2017) or cardiovascular diseases (Nomiya et al., 2008) and genetic factors (Uchida et al., 2009), Ototoxic medication (Tabuchi et al., 2011), and health co-morbidity (Cruickshanks et al., 1998a) factors. Long-duration mitochondrial function leads to Bak-dependent apoptosis of the cochlear cells leading to increase in hearing thresholds (Someya and Prolla, 2010). Details of ARHL environmental and health-related co-morbid risk factors are described in Environmental Factors and Health Co-morbidities. Genetic investigations have identified a number genes associated with cochlear aging (see section Genetics). In addition, genes that are associated with antioxidant protection, mitochondrial dysfunction (see section Genetics), also contribute to the cochlear hair cell dysfunction.

The age-associated changes reported in cochlear nucleus and the auditory cortex have also been a focus of research. As Figure 1 shows afferent auditory fibers from the cochlea terminate at the cochlear nucleus (CN) complex which includes the anteroventral cochlear nucleus (AVCN), posteroventral cochlear nucleus (PVCN), and dorsal cochlear nucleus (DCN) (Brawer et al., 1974). First stage of centralized auditory processing occurs at the CN complex (Schmiedt, 2010). AVCN and PVCN carry out the spectral analysis of the auditory signal (Rhode and Greenberg, 1994) and AVCN and DCN process the localization cues (May, 2000). Both animal and human models have been used to investigate the underlying patho- physiological changes of the central auditory system associated with aging. The alterations in animal models from the CN to the auditory cortex are discussed first.

Difficulties expressed by older adults in understanding speech could be due to age-related deficits in peripheral and/or central auditory pathways (Gates and Mills, 2005). Among the peripheral factors that impact speech understanding, poor audibility resulting from cochlear pathology is considered to play a significant role (George et al., 2007). However, part of the problem in recognizing speech in older adults seem to be associated with temporal deficits resulting from aging and ARHL (Gordon-Salant and Fitzgibbons, 1993). Several studies have investigated the impact of aging on central auditory processing of speech stimuli (See Humes et al., 2010 for a review). These studies have used competing speech (Jerger et al., 1991; Gates et al., 2008), temporally-distorted speech (Gordon-Salant and Fitzgibbons, 2001), and/or binaural speech stimuli (Jerger et al., 1991; Gates et al., 2008). A review by Humes and Dubno (2010) concluded that aging significantly impairs all three speech test measures and hearing loss was the major factor on negatively impacting participants' performance on tasks that utilized competing and time compressed speech stimuli. The above mentioned temporal processing deficits could explain some of the difficulties faced by older adults, with or without hearing loss, in perceiving speech in challenging situations-i.e., accented speech (Gordon-Salant et al., 2013) and speech in noise or reverberation (Dubno et al., 1984; Gordon-Salant and Fitzgibbons, 1993). Further, speech comprehension in quiet and in noise also relies on cognitive functions (Pichora-Fuller et al., 1995). These include searched based attention (Pichora-Fuller and Singh, 2006), selective attention (Alain and Arnott, 2000), divided attention (Kemper et al., 2003), working memory (Pichora-Fuller et al., 1995), and processing speed (Pichora-Fuller and Souza, 2003).

There are two key components of the auditory system involved in the processing of incoming auditory stimuli; the peripheral hearing system and the central hearing system (Katz, 2015). Approximately 83% of adults aged 70 years and above suffer from a peripheral hearing loss (Cruickshanks et al., 1998b). Peripheral hearing loss not only affects the auditory processing of speech sounds but also the higher-level cognitive functions required to process linguistically demanding stimuli (Stewart and Wingfield, 2009; Peelle et al., 2011). For example, the hearing impaired individuals show difficulty in attending to working memory (Wingfield et al., 2006) and auditory and visual free recall tasks (Wingfield et al., 2006) and require longer latencies to make accurate perceptual judgments (Tun et al., 2010). Evidence from both cross-sectional (Valentijn et al., 2005; Tay et al., 2006; Jayakody et al., 2018a), and longitudinal (Valentijn et al., 2005; Lin et al., 2011a; Deal et al., 2016) studies found an association between peripheral hearing loss and cognitive impairment in older adults. A meta-analysis study reported an association between hearing loss and cognitive impairment, with the degree of cognitive impairment being significantly associated with the severity of both untreated and treated peripheral hearing loss (Taljaard et al., 2015). Further, an association between ARHL and risk of incident dementia (Lin et al., 2011a; Gallacher et al., 2012; Gurgel et al., 2014; Deal et al., 2016) and risk of incident Alzheimer's disease (AD) (Lin et al., 2011a) has also been reported. Most of the studies reported above have considered verbally loaded cognitive measures to assess the cognitive functions of older adults. Cumulative evidence strongly suggest that hearing loss could result in overestimation of the level of cognitive impairment in hearing impaired participants (Jorgensen, 2012; Dupuis et al., 2015). Therefore, non-verbal cognitive measures could be more appropriate for older adults with potential ARHL. A small number of recent studies that investigated the association between ARHL and cognitive impairment by removing auditory items of the cognitive screening tests (Dupuis et al., 2015) or using non-auditory tests of cognition (Lin et al., 2013, 2017; Wong et al., 2014; Jayakody et al., 2018a) have also reported an association between peripheral hearing loss and cognitive impairment.

In some individuals, difficulties experienced during comprehending speech in noisy backgrounds or competing speakers could be related to central auditory processing (CAP) difficulties rather than peripheral hearing deficits (Panza et al., 2015). Results from a number of longitudinal studies suggest that CAP in the absence of severe peripheral hearing loss is associated with high incidence of cognitive decline and dementia due to AD (Gates et al., 1996, 2002, 2008, 2011; Profant et al., 2015). These studies have demonstrated that individuals with central auditory dysfunction were at a significantly increased risk for incident dementia with hazard ratios ranging from 9.9 (95% confidence interval [CI], 3.6–26.7) to 23.3 (95% CI, 6.6–82.7) (Gates et al., 1996, 2011; Profant et al., 2015). Gates et al. (1996) reported that the relative risk of developing cognitive decline or dementia was twice higher for those with bilateral poor CAP scores as measured by Synthetic Identification-Ipsilateral Competing Message (SSI-ICM) scores compared to those with unilateral poor SSI-ICM scores. Furthermore, impaired central auditory functions have been observed five to ten years prior to an official AD diagnoses (Iliadou et al., 2003; Bateman et al., 2012).

Numerous attempts have been made to provide plausible explanations for the association between ARHL and cognitive decline. Cognitive load on perception hypothesis argues that the decline in cognitive capacity increases the cognitive load resulting in a sensory loss (CHABA, 1988; Lindenberger and Baltes, 1994). Sensory Deprivation hypothesis proposes that hearing loss leads to permanent deterioration in cognitive functions (CHABA, 1988; Lindenberger and Baltes, 1994; Schneider and Pichora-Fuller, 2000; Pichora-fuller, 2003; Lin et al., 2011b, 2013; Humes et al., 2013). Cortical reorganization following ARHL provides substantial evidence to support the sensory deprivation hypothesis (Husain et al., 2011; Peelle et al., 2011; Eckert et al., 2012; Boyen et al., 2013). Findings demonstrating that severe to profound hearing impaired children and young adults with a long duration of hearing impairment performed similar to their hearing peers on tests of intelligence seem to contradict the sensory deprivation hypothesis (Vernon, 2005). Hence, one could argue that hearing is not the sole contributor to cognitive impairment, rather one of the factors. Information degradation hypothesis postulates that hearing loss results in reversible cognitive decline (CHABA, 1988; Schneider and Pichora-Fuller, 2000; Pichora-fuller, 2003). Supportive evidence for the information degradation hypothesis can be found in a large number of existing research studies. In sum, these studies suggest that older adults rely more on cognitive resources to interpret degraded speech signals, resulting from hearing loss thereby placing more demand on executive functions and working memory (Pichora-Fuller et al., 1995; Pichora-fuller, 2003; Pichora-Fuller and Singh, 2006; Amichetti et al., 2013).

Common cause hypothesis argues that a common mechanism underlies age-associated changes observed in cognition, hearing, and other sensory modalities (CHABA, 1988; Lindenberger and Baltes, 1994; Baltes and Lindenberger, 1997). It has been suggested that general age-related neuropathological changes in the brain may actually explain this relationship, but current data do not support this general domain and function specific underlying mechanism theory (Lindenberger and Ghisletta, 2009). For example, when the relationship between hearing loss and memory was examined, a significant relationship was observed even after controlling for the effects of age and the type of task (e.g., verbal vs. visual), confirming hearing loss is an independent factor involved in cognitive decline (Lindenberger and Ghisletta, 2009).

Wayne and Johnsrude (2015) proposed a new framework with a common cause phenomena approach, where age-related changes of the brain result in decline of sensory (e.g., ARHL) and cognitive abilities. According to this hypothesis, the changes in sensory abilities increase the demand on the already strained cognitive system resulting in functional cognitive impairment (Wayne and Johnsrude, 2015).

Recently, a cognitive reserve hypothesis has been proposed to explain how individuals with similar neuropathological conditions differ substantially in their ability to make efficient use of brain reserve during tasks (Stern, 2002). Intelligence (Alexander et al., 1997) and higher education (Amieva et al., 2014), occupational level (Staff et al., 2004), participation in leisure activities (Scarmeas et al., 2001), and social networking (Fratiglioni et al., 2000) are considered to be contributing factors to the cognitive reserve. If sensory or cognitive demands exceeds the cognitive reserve capacity, it could lead to impairment in cognitive abilities (Wayne and Johnsrude, 2015). However, according to cognitive reserve hypothesis, the observed cognitive impairment could be reversed if the perceptual challenges were reduced (i.e., improved speech perception through amplification; Wayne and Johnsrude, 2015). In summary, no single hypothesis has been able to provide a comprehensive inference with regards to the causal relationship between ARHL and cognitive impairment.

Even though, ARHL is still not considered as a major risk factor for AD, it is a modifiable age-associated condition linked to dementia and late-life cognitive decline (Livingston et al., 2017). The Lancet International Commission on Dementia Prevention, Intervention, and Care has estimated that a third of AD diagnoses worldwide could be delayed or prevented, with early intervention and changes to lifestyle and utilizing available public health strategies (Livingston et al., 2017) such as post-lingual hearing loss correction. Furthermore, studies demonstrate that if the prevalence of each risk factor is reduced by only 10–20% per decade, the number of AD diagnoses could reduce by between 8.8 and 16.2 million worldwide by 2050 (Norton et al., 2014). As an example of potential changes in outcome measures following hearing loss correction, we have recently reported that cochlear implant recipients performed significantly better on cognitive function measures compared to similar individuals on a waiting list to receive cochlear implants (Jayakody et al., 2017). Others have also reported similar positive results following hearing loss improvement using hearing aids and cochlear implantation (Castiglione et al., 2016; Sonnet et al., 2017). Whether the correction of ARHL could significantly delay or arrest the late life pathological cognitive decline, dementia, or AD neurodegeneration is yet to be investigated. However, treating ARHL is extremely low risk, has significant health, social, and safety benefits beyond cognition, and there are enough evidence to discuss this option with ARHL patients (Golub, 2017).

ARHL is an important marker in frailty (Panza et al., 2015). Frailty has been defined as a clinical syndrome that contains three or more of the following symptoms: unintentional weight loss (10 lbs in past year), self-reported exhaustion, weakness (grip strength), slow walking speed, and low physical activity (Fried et al., 2001). Frailty increases the risk of detrimental health hazards such as falls (Speechley and Tinetti, 1991), institutionalization (Rockwood et al., 1996), hospitalization (Purser et al., 2006), and death (Ensrud et al., 2007). Interestingly, the risk of frailty in older adults with hearing loss increases by 63% and ARHL appears to be an independent risk factor for frailty with great risk of falls (Kamil et al., 2016). Further, ARHL has been shown to be associated with health-related outcomes linked to frailty, such as low level of physical activity (Gispen et al., 2014), slow gait speed (Li et al., 2013), and incident falls (Lin and Ferrucci, 2012). In addition, frailty is reported to be associated with cognitive impairment (Ávila-Funes et al., 2009; Yassuda et al., 2012), cognitive decline (Samper-Ternent et al., 2008; Auyeung et al., 2011) and incident dementia (Solfrizzi et al., 2013). See Robertson et al. (2013) for a review.

Social isolation and loneliness is considered one of the risk factors for ARHL. Mick and Lin (2013) found a significant association between ARHL and social isolation in the 60–69 year old age group [odds ratio = 2.4, 95% CI = (1.36, 4.23) p = 0.003]. The association was more pronounced in females than in males (Ramage-Morin, 2016). ARHL is significantly associated with late-life depression, anxiety, and stress (Jayakody et al., 2018b). Depression is considered an early manifestation of dementia and AD (Panza et al., 2010). Accumulated evidence suggest that depression accelerates the aging process by increasing the incident risk factors of age-associated diseases such as cardiovascular diseases (Frasure-Smith et al., 1993), metabolic disturbances (McIntyre et al., 2007), and cognitive impairment (Lee et al., 2011; Dawes et al., 2015). Further, a bidirectional association between depression and late-life frailty has also been reported (Mezuk et al., 2012). Freret et al. (2015) proposed that an imbalance between brain reserve, resilience, neuroplasticity, and impaired pathophysiological mechanisms associated with aging, stress, and synaptic plasticity may lead to late-life depression. Although the pathophysiological mechanisms underlying depression, ARHL, and dementia remain to be further elucidated, a putative mechanism underlying both depression and dementia linked to neuro-inflammation and perfusion deficits in older adults (Maes et al., 2009; see Popa-Wagner et al., 2014, for a review).

Based on the above reported findings, there seems to be a bidirectional association between ARHL and cognitive impairment, depression and social isolation, depression, and frailty (Figure 2). As shown in the schematic approach in Figure 2, primary factors including aging, genetic predispositions, environmental factors, and medical and health related conditions accompany cognitive decline, ARHL, and frailty. Furthermore, ARHL increases the risk of depression and frailty, both of which are associated with age-related cognitive decline and future risk of dementia. As ARHL is the primary modifiable risk factor for cognitive decline, depression, and social isolation in this model, its treatment may, to some degree, improve the associated outcomes and change the trajectory of cognitive impairment. The diagram shown in Figure 2 is a simplified account of the relationship between various known factors with the understanding that so many are yet to be determined. However, this model provides a basic rationale to argue that treating ARHL will provide multiple benefits, each with significant implications for future trajectory of cognitive functioning in older adults.

DJ and PF have no conflicts of interest to declare. RM is the founder and Chief Scientific Officer of the Alzhyme and KaRa Minds Institute. HS has, and continues to receive remuneration from activities with Pfizer and Takeda Pharmaceuticals. Although grants and funds for our research were provided by public and private organizations, they in no way influenced any part of this review, its conclusions or the decision to submit this manuscript for publication to this journal.