Purpose

Diabetes mellitus (DM) is a chronic health disorder with hyperglycaemia resulting from insulin resistance and/or insulin inadequacy. The atlas of the International Diabetes Federation estimated that 10.5% of the world’s population is currently living with DM [1]. Type 2 diabetes mellitus (T2DM) may lead to macro- and micro-vascular changes such as thickening of capillary basement membrane, loss of inner hair cells and outer hair cells (OHCs), atrophy of spiral ganglion cells and marginal cells, and oedematous changes of the intermediate cells [2]. It may also lead to neuropathic changes and hearing loss due to the effect on the auditory pathway. Various research projects have identified the association between diabetes and hearing loss (HL) and discovered a positive correlation between the two variables [3,4,5]. The auditory system consists of complex but well integrated afferent and efferent pathways, with feedback circuits from the primary level to the associated auditory cortex. The high metabolic activity of the auditory pathway mandates adequate glucose levels, deeming it vulnerable to small glycaemic changes, thus affecting normal functioning. Prolonged hyperglycaemia may damage neural auditory pathways [6].

Our study intends to discover the influence of diabetes on the auditory system in diabetics by measuring acoustic reflex (AR) parameters and the reflex decay test (RDT), and also the link between the AR parameters and RDT with duration and control of diabetes.

Methods

An analytical cross-sectional study was done in our tertiary care setup which included 42 subjects with T2DM (case group) aged between 30 and 60 years. This age range was chosen as the cut-off since age-related hearing loss is more common after 60 years and T2DM is known to often affect people after age 40. Age-matched 84 non-diabetic subjects (control group) who presented to the outpatient department with no ear complaints were assessed. Subjects with chronic middle ear disease, ear trauma, noise exposure, history of ototoxicity, facial nerve palsy, history of middle ear surgery, and familial history of congenital deafness were excluded. Subjects were selected by a non-random convenience sampling method. Basic details such as ear complaints, family history, duration, and treatment (with oral hypoglycaemic agents or insulin or both) of DM, random blood sugar and glycosylated haemoglobin (HbA1c) levels (up to 6.5% = normal, 6.6–8% = fair control and > 8% = poor control) were noted, followed by a thorough examination of ear, nose, and throat. Pure tone audiometry, tympanometry, speech identification scores (SIS), acoustic reflex tests (ART, ARL, ARA) and reflex decay tests (RDT) were performed on all the subjects. Pure tone audiometry was performed at various frequencies, i.e. 250, 500, 1000, 2000, 4000, and 8000 Hz and subjects with pure tone averages (PTA) < 25-decibel hearing loss (dBHL) were included in the study. Acoustic reflex threshold (ART) is defined as the lowest intensity at which acoustic reflex is elicited at each frequency and the acoustic reflex amplitude is explained as the maximum displacement of acoustic reflex for a given frequency. Acoustic reflex latency (ARL) describes the temporal characteristic of an acoustic reflex which illustrates the time course of the reflex. The reflex decay test (RDT) is defined as the estimation of the time at which reflex amplitude is 50% of the maximum amplitude [7]. According to ANSI (American National Standards Institute, 1982), the temporal characters of ARL are measured as initial latency (10% on) is defined as the time (in sec) from the beginning of an instantaneous immittance change to 10% of the measured steady-state immittance change, rise time (90% on) is defined as the time (in sec) from 10 to 90% of the measured steady-state immittance change, terminal latency (90% off) is defined as the time (in sec) from instantaneous termination of the initial immittance change to 90% of the measured steady-state immittance change and fall time (10% off) is defined as the time (in sec) from 90 to 10% of the measured steady-state immittance change after the initial immittance change is terminated [8]. AR parameters were evaluated at 500, 1000, and 2000 Hz and broad band noise (BBN). Approval was gained from the Institutional Ethics Committee (IECKMCMLR-09/2020/260) to conduct the study.

Statistical analysis

Data collected were subjected to analysis using IBM SPSS Statistics for Windows, Version 25.0. Armonk, NY: IBM Corp. The variables were compared in proportions and mean (standard deviation; SD). P value of ≤ 0.05 was considered statistically significant.

Results

The study included 126 individuals, with 42 cases and 84 controls. The case group included 23 (54.8%) males and 19 (45.2%) females and in the control group, there were 49 (58.3%) males and 35 females (41.7%).

An independent sample t test was performed to compare mean PTA and SIS between case and control groups. All 126 subjects showed PTA within normal limits, i.e., < 25dBHL; however, the case group showed increased PTA which was statistically significant in both ears (p value = right < 0.042, left < 0.001) when compared with the control group. No significant difference was noted in the mean SIS scores between both groups (p value right = 0.159, left = 0.689). All the subjects had an ‘A’ type tympanogram.

An independent sample t test was performed to compare ipsilateral and contralateral mean ART and ARA between cases and controls at 500, 1000, 2000 Hz and BBN. No significant difference was noted in the ipsilateral and contralateral mean ART between cases and controls. However, alterations of ARA (decreased amplitude) have been observed in the case group. There were a reduced number of reflexes (both ipsilateral and contralateral) in the case group when compared to the control group (Fig. 1a, b) and a significant difference was found in ARA at 500, 1000 Hz and BBN when compared between cases and controls (Fig. 1c, d).

Fig. 1
figure 1figure 1

a Comparison between the observed ipsilateral right and left ARA in cases and controls. b Comparison between the observed contralateral right and left ARA in cases and controls. c Descriptive analysis (mean and SD) of ipsilateral ARA in both groups. d Descriptive analysis (mean and SD) of contralateral ARA in both groups

Full size image

An independent sample t test was done to compare ARL between case and control groups at 500 and 1000 Hz. There was a significant difference in ARL in cases at 500, 1000 Hz (ipsilateral right 10% on) and at 500 Hz (ipsilateral left 90% on and left 90% off) and 1000 Hz (ipsilateral left 90% on and left 10% off) when compared with the control group (Fig. 2a–d). However, no significant difference was found between contralateral ARL (right and left) in both groups. All the subjects included in the study had no decay in the reflex decay test.

Fig. 2
figure 2figure 2

a Comparison between the observed ipsilateral right and left ARL in cases and controls. b Comparison between the observed contralateral right and left ARL in cases and controls. c Descriptive analysis (mean and SD) of ipsilateral ARL in both groups. d Descriptive analysis (mean and SD) of contralateral ARL in both groups

Full size image

One-way ANOVA test was performed to analyse the AR parameters and reflex decay tests in the case group based on the duration of T2DM and control of DM using HbA1c levels. The case group was subdivided into 3 groups; 5–9 years (61.9%), 10–14 years (33.3%) and ≥ 15 years (4.8%) based on the duration of diabetes (Fig. 3). Based on HbA1c levels, the case group was categorised into normal (HbA1c < 6 = 11.9%), fair control (HbA1c 6.5–7.9 = 21.4%) and poor control (HbA1c > 8 = 66.7%) (Fig. 4). Subgroup analysis was done; however, there was no significant difference in the AR parameter within the cases based on the duration and control of diabetes (Table 1, 2).

Fig. 3
figure 3

Subgroup Division of cases depicting their distribution based on the duration of diabetes mellitus (DM)

Full size image
Fig. 4
figure 4

Subgroup Division of cases depicting their distribution based on control of DM with glycosylated haemoglobin (HbA1c) levels

Full size image
Table 1 Subgroup analysis of AR parameters in comparison with control of DM based on HbA1c levels (ANOVA analysis)
Full size table
Table 2 Subgroup analysis of AR parameters in comparison with duration of DM (ANOVA analysis)
Full size table

Discussion

In the present study, mean PTA was comparatively greater in the cases than in controls although all the subjects included in the study had normal hearing thresholds (PTA < 25 dBHL) which demonstrates the positive association between T2DM and hearing loss. Kim MB et al. conducted a large cohort study which showed the development of bilateral hearing loss in diabetics [9]. A similar study was done by Dosemane et al., which postulated that bilateral SNHL is a complication of T2DM [4]. Akinpelu et al. conducted a meta-analysis and reviewed 18 articles which revealed an increased incidence of HL ranging from 44 to 69.7% and prolonged auditory brainstem response (ABR) wave V latencies in the subjects with T2DM which may be due to degeneration of hair cells in basal turn of cochlea and delay in the conduction of auditory signals within the brain stem with diabetes, respectively [10]. Another meta-analysis has been conducted by Mujica-Mota et al. which showed raised incidence of HL, lower oto-acoustic emissions (OAE) and prolonged latencies in ABR waves I, III and V in type 1 diabetes mellitus (T1DM) subjects [11]. In the present study, no significant difference was found in SIS between cases and controls. A study conducted by Huang et al. revealed significant decrease in SDS scores in diabetics which correlated with the high-frequency sensorineural hearing loss [12].

A study was done to assess the ART and RDT in the geriatric group by Ünsal et al. who concluded that although some changes were observed due to age, middle ear and stapedius work normally in geriatric category as no significant difference was found between geriatric and non-geriatric categories in ipsilateral and contralateral AR parameters and RDT [13]. Virtaniemi et al. concluded that decreased ARAs and prolonged ARLs in subjects with insulin-dependent diabetes mellitus (IDDM) were more probably attributed to the rigid middle ear structure than brainstem alterations [14]. Another similar study was done by Braite et al. who observed the absence of an inhibitory effect of medial olivocochlear reflex (MOC) with distortion product OAE (DPOAE) at 4000, 6000 and 8000 Hz in patients with T1DM as a result of early auditory dysfunction of the efferent pathway [6]. To the best of our understanding, no prior research has reported AR parameters in patients with T2DM. In the present study, we found a significant difference in ARA (decreased amplitudes at lower frequencies and BBN) and ARL in cases when compared to controls which may be indicative of damage to the neural auditory pathway. However, there was no clear explanation for the significant difference between the ipsilateral right and left ARL. There was no evidence of significant difference in the contralateral right and left ARL.

Our study showed no significant association found between AR parameters and other variables such as duration of DM and control of DM based on HbA1c levels. However, literature shows discrepancies; Mujica-Mota et al. demonstrated an increased risk of HL over time with an increase of prevalence by 1.7% per 1-year exposure of DM and another study conducted by Mishra et al. revealed increased severity of SNHL with duration of DM [11, 15]. On the contrary, analysis conducted by Kim et al. interestingly showed a stronger association of HL in the younger group (< 50 years) [9]. Various other investigators observed no association of HL with duration of DM [4, 6, 16, 17]. A positive correlation was noted between the severity of DM and degree of HL in a study by Mishra et al.; profound SNHL was highly prevalent in diabetics with FBS > 200 mg/dL [15]. Srinivas et al. evaluated the association between poorly controlled DM (HbA1c > 8) with SNHL; the prevalence of SNHL is more than 85% in the subjects with poor glycaemic control and duration of DM of more than 10 years [18]. Our study was unique as confounding factors known to cause HL were eliminated due to exclusion criteria and objective tests have been used to assess the effect of T2DM on the auditory pathway.

Conclusion

T2DM can lead to increased hearing thresholds, decreased ARA at lower frequencies and BBN. Hence, evaluating AR parameters in patients with T2DM may help detect the early effects of DM on the auditory pathway. However, there is no association between independent variables such as duration and glycaemic control of T2DM.