Elevated Triglyceride Glucose Index is Associated with Poor Hearing Recovery in Sudden Sensorineural Hearing Loss

Abstract

Importance:

Cochlea microcirculation dysfunction has been proposed as a vital etiology for sudden sensorineural hearing loss (SSNHL), wherein insulin resistance acted as a potential risk factor for microcirculatory dysfunctions. The triglyceride glucose (TyG) index, a validated surrogate of insulin resistance, has not been fully evaluated in SSNHL.

Objective:

To evaluate the association between the TyG index and hearing recovery in patients with SSNHL.

Design:

Retrospective cross-sectional study involving a large, clinically characterized cohort.

Setting:

Tertiary university hospital.

Participants:

A total of 944 adult SSNHL patients were included between January 2018 and March 2024. Eligibility required onset within 14 days and completion of standardized treatment and follow-up audiometry. Patients with conductive, recurrent, or secondary causes of hearing loss were excluded.

Exposures:

The TyG index was calculated as ln [fasting plasma glucose (mg/dL) × triglycerides (mg/dL)/2].

Main Outcomes and Measures:

Hearing recovery was assessed after 14 days of treatment using Siegel’s criteria and categorized as complete or incomplete recovery.

Results:

Among 944 patients (mean [SD] age, 49.3 [15.6] years; 458 men [48.5%]), 208 (22.0%) achieved complete recovery. Baseline hearing levels and audiogram pattern were similar across TyG tertiles. However, the proportion of patients achieving complete recovery was lower in the highest TyG tertile compared to the 2 lower tertiles (45 [14.3%] vs 77 [24.4%] vs 86 [27.4%]; P < .001, respectively). In fully adjusted logistic regression models, higher TyG index was independently associated with incomplete recovery with an odds ratio (95% CI) of 1.93 (1.14-3.29) in tertile 3 and 1.55 (1.05-2.29) for per unit increase in the TyG index.

Conclusion:

Higher TyG index values were independently associated with poorer hearing recovery in SSNHL.

Relevance:

The TyG index may help clinicians identify SSNHL patients at risk of poor cochlear recovery, supporting earlier and more targeted management.

Graphical Abstract.

Key Message

1. Elevated TyG index was associated with poorer hearing recovery in sudden sensorineural hearing loss.

2. The findings add evidence to the concept that insulin resistance and related microvascular vulnerability may influence cochlear outcomes.

3. In this cohort of 944 patients with SSNHL, the TyG index demonstrated potential as a simple metabolic measure to assist in identifying individuals at risk of incomplete recovery.

Introduction

Characterized by abrupt sensorineural hearing loss of more than 30 dB in at least 3 consecutive frequencies within 72 hours, sudden sensorineural hearing loss (SSNHL) is a common emergency and a major cause of sensorineural hearing impairment, affecting approximately 20 per 100 000 individuals in the United States.^1,2 Great heterogeneity has been observed among the SSNHL patients, despite about one-third of the patients have the potential for spontaneous recovery, ³ more patients remain unresponsive to standard therapeutic regimens, suffering from long-term severe hearing loss and tinnitus.^4,5 These sequelae can significantly impact patients’ quality of life, often leading to anxiety, depression, and even cognitive dysfunction.^{6 -8} Therefore, the identification of early biomarkers that can predict various hearing prognosis and facilitate timely interventions appears to be indispensable.

Although the precise etiology of SSNHL remains unclear, inner ear microcirculation disturbance and endothelial dysfunction have been widely recognized as primary pathogenesis.^2,9,10 As part of the systemic circulation, the microcirculation of the inner ear is also vulnerable to systemic metabolic and cardiovascular diseases, such as diabetes mellitus (DM), hypertension, and hyperlipidemia.^10,11

Notably, insulin resistance (IR), a precursor to DM, is known for the impaired sensitivity and responsiveness to insulin. ¹² Due to its extensive effects on not only glucose and lipid dysregulation, but also chronic inflammation and microcirculation, numerous studies demonstrated its association with cardiovascular diseases, including coronary artery disease, heart failure, and hypertension,^{12 -14} thus it is crucial to find a biomarker to measure IR state. Compared with homeostasis model assessment for insulin resistance, which is the gold standard but expensive and invasive, the triglyceride glucose (TyG) index has emerged as a simple and convenient tool to predict IR-related cardiovascular events.^15,16

Currently, the association between an elevated TyG index and risk of sensorineural hearing loss has been postulated. ¹⁷ However, the role of this index in predicting the hearing prognosis of SSNHL patients remains unexplored. In this large retrospective cross-sectional study of 944 patients with SSNHL, we aim to elucidate the potential role of the TyG index as a biomarker for predicting hearing prognosis in patients diagnosed with SSNHL.

Methods

Study Population

This was a single-center, retrospective, and observational study of 1304 patients diagnosed with SSNHL and admitted to our hospital from 1st January 2018 to 31st March 2024. SSNHL was defined as sensorineural hearing loss of ≥30 dB affecting at least 3 consecutive frequencies occurring in 72 hours according to the clinical practice guidelines.^1,2 As shown in Figure 1, among the 1304 patients, 360 were excluded for meeting the exclusion criteria: (1) ages less than 18 years; (2) incomplete or ambiguous admission data—including symptoms, medical history, and physical examination findings; (3) concomitant central nervous system diseases, head injury, or neurologic disorders; (4) prior otologic surgery, ototoxic drug usage, or Menère’s disease; (5) abnormal findings on head magnetic resonance imaging, including vestibular schwannoma, stroke, and other intracranial tumors; (6) bilateral simultaneous SSNHL; (7) missing fasting plasma glucose (FPG) and triglyceride (TG) measurements; (8) missing 14-day hearing examination data; and (9) treatment interruption or incompleteness. Ultimately, 944 patients were included in the final analysis. This study was conducted with the approval from the Ethics Committee of the First Affiliated Hospital with Nanjing Medical University (approval number: 2024-SR-163).

Figure 1.

Flow chart for study enrollment. FPG, fasting plasma glucose; MD, Meniere’s disease; SSNHL, sudden sensorineural hearing loss; TG, triglyceride; TyG, triglyceride glucose.

Data Collection and Definitions

Patients’ demographics, medical history, vital signs, clinical diagnoses, laboratory measurements, audiometry data, and medications were collected from the electronic medical record system. The body mass index (BMI) was calculated as weight (kg) divided by height squared (m²). Hypertension was defined as a history of hypertension or systolic blood pressure ≥140 mmHg and/or diastolic blood pressure ≥90 mmHg. DM was defined as self-reported history of diabetes, or FPG ≥126 mg/dL. The TyG index was calculated using the formula: TyG index = ln [FPG (mg/dL) × fasting TG (mg/dL)/2]. ¹⁶ According to their TyG index levels, patients were divided into 3 groups: Tertile 1 (TyG index < 8.162), Tertile 2 (8.162 ≤ TyG index < 8.703), and Tertile 3 (TyG index ≥ 8.703).

Audiologic Evaluation

Pure-tone audiometry was performed at baseline on the day of admission before treatment and repeated 14 days after completion of standard treatment. Air- and bone-conduction thresholds were measured at 250, 500, 1000, 2000, 4000, and 8000 Hz in both the affected and contralateral ears. Pure tone average (PTA) was calculated as the mean of air conduction thresholds at 500, 1000, 2000 and 4000 Hz. The severity of initial hearing loss was divided into 6 levels: mild (≤35 dB), moderate (35-49 dB), moderate to severe (50-64 dB), severe (65-79 dB), profound (80-94 dB), and total (≥95 dB). ¹⁸ For subgroup analyses, categories ≥65 dB were combined as severe to profound hearing loss to enhance model stability in some small subgroups. Regarding the audiogram pattern, ¹⁹ the ascending type refers to hearing loss primarily occurring in low frequencies (250 and 500 Hz), while the descending type occurs mainly in high frequencies (4000 and 8000 Hz). The flat type affects all frequencies with a PTA <80 dB, whereas the total type affects all frequencies with a PTA ≥80 dB. Contralateral hearing loss was defined as an initial contralateral PTA ≥25 dB.

Treatment Regimens

All patients underwent standardized treatments during their hospitalization. ¹⁹ Methylprednisolone was intravenously infused at a dosage of 80 mg/day on days 1 to 3, and 40 mg/day on days 4 to 6. Oral Ginkgo biloba extract was prescribed 3 times daily at a dosage of 40 mg each time. Batroxobin was administered systemically at a dose of 10 U initially and then 5 U every other day. For patients with severe to profound hearing loss, endoscopic injection of 5 mg dexamethasone into the middle ear was performed every other day, for a total of 3 times.

Primary Outcome: Hearing Recovery

The primary endpoint was hearing recovery in the affected ear, assessed at 14 days after treatment using Siegel’s criteria. ²⁰ Hearing recovery was evaluated based on audiometric outcomes in the affected ear. Complete recovery was defined as a final PTA of less than 25 dB regardless of the gain; partial recovery, a greater than 15 dB improvement with a final PTA of 25 to 45 dB; slight recovery, a greater than 15 dB improvement and a final PTA of greater than 45 dB; and no recovery, a less than 15 dB improvement and a final PTA of greater than 75 dB. Patients were divided into complete recovery and incomplete recovery (partial, slight, or no recovery) groups for further analysis.

Statistical Analysis

Continuous variables in a normal distribution were presented as mean [SD], otherwise as median [interquartile range]. Categorical variables were presented as frequencies [percentages]. The analysis of variance was performed to compare continuous variables in a normal distribution among 3 groups, or the Kruskal–Wallis test to those in a skewed distribution. The unpaired t test was conducted to compare normally distributed continuous variables between 2 groups. Categorical variables were compared by the χ² test among groups. Four multivariable binary logistic regression models were established to test the association of TyG index with incomplete hearing recovery: Model 1 was unadjusted; Model 2 was adjusted for age, sex, smoking history, and alcohol use history; Model 3 was adjusted for variables included in Model 2 plus BMI, history of hypertension, and history of DM; Model 4 was adjusted for variables included in Model 3 plus initial hearing loss severity, audiogram pattern, contralateral hearing loss, duration from onset to treatment, and presence of vertigo. The results were reported as odds ratios (ORs) and 95% confidence intervals (CIs). Restricted cubic spline (RCS) analysis was then performed to explore the dose–response relationship between TyG index and incomplete hearing recovery. Furthermore, we conducted subgroup analyses according to age (<65 vs ≥65 years), sex (male vs female), BMI (<24 vs ≥24 kg/m²), history of hypertension (no vs yes), history of DM (no vs yes), smoking history (no vs yes), alcohol use history (no vs yes), severe to profound hearing loss (no vs yes), and contralateral hearing loss (no vs yes). All statistical analyses were performed using IBM SPSS Statistics software (version 26.0) and R studio software (version 4.3.2), with P < .05 defined as statistically significant.

Results

Basic Demographics of Patients by TyG Tertiles

A total of 944 patients (mean [SD] age, 49.3 [15.6] years) were finally included, with 458 [48.5%] as males and the mean [SD] TyG index of 8.46 [0.57]. The basic clinical characteristics of the patients by TyG tertiles were summarized in Table 1. Higher TyG tertile was associated with older age, higher BMI, higher prevalence of hypertension and DM, and higher proportions of smokers (all P < .05), as well as elevated levels of TC, FPG, TG, and LDL-C (all P < .05).

Table 1.

Baseline Characteristics of SSNHL Patients Stratified into TyG Tertiles.

Characteristic	Total	Tertile 1 n = 314	Tertile 2 n = 315	Tertile 3 n = 315	P value
Age, mean (SD), y	49.3 (15.6)	44.4 (16.9)	50.1 (15.2)	53.6 (13.0)	<.001 a
Age group, No. (%)
<65	779 (82.5)	275 (87.6)	258 (81.9)	246 (78.1)	.032 b
≥65	165 (17.5)	39 (12.4)	57 (18.1)	69 (21.9)
Sex, No. (%)
Male	458 (48.5)	141 (44.9)	153 (48.6)	164 (52.1)	.199 b
Female	486 (51.5)	173 (55.1)	162 (51.4)	151 (47.9)
Hypertension, No. (%)
Yes	277 (29.3)	70 (22.3)	96 (30.5)	111 (35.2)	.002 b
No	667 (70.7)	244 (77.7)	219 (69.5)	204 (64.8)
DM, No. (%)
Yes	197 (20.9)	28 (8.9)	53 (16.8)	116 (36.8)	<.001 b
No	747 (79.1)	286 (91.1)	262 (83.2)	199 (63.2)
CAD, No. (%)
Yes	33 (3.5)	8 (2.5)	12 (3.8)	13 (4.1)	.522 b
No	911 (96.5)	306 (97.5)	303 (96.2)	302 (95.9)
BMI, mean (SD), kg/m2	23.9 (3.7)	23.2 (3.6)	24.0 (3.6)	24.4 (3.9)	<.001 a
BMI group, No. (%)
<24	499 (52.9)	190 (60.5)	156 (49.5)	153 (48.6)	.004 b
≥24	445 (47.1)	124 (39.5)	159 (50.5)	162 (51.4)
Smoking, No. (%)
Yes	98 (10.4)	23 (7.3)	29 (9.2)	46 (14.6)	.008 b
No	846 (89.6)	291 (92.7)	286 (90.8)	269 (85.4)
Alcohol, No. (%)
Yes	52 (5.5)	16 (5.1)	14 (4.4)	22 (7.0)	.349 b
No	892 (94.5)	298 (94.9)	301 (95.6)	293 (93.0)
Laboratory findings
TG level, mg/dL	101.0 (68.2, 143.5)	59.3 (46.9, 72.6)	102.7 (84.1, 123.1)	164.7 (134.6, 209.9)	<.001 c
FPG level, mg/dL	91.8 (78.7, 109.9)	86.9 (75.3, 104.0)	89.4 (78.0, 107.2)	99.3 (86.0, 127.6)	<.001 c
LDL-C level, mg/dL	117.9 (30.7)	107.7 (28.9)	120.8 (29.3)	125.3 (31.2)	<.001 a
HDL-C level, mg/dL	53.8 (12.0)	56.1 (12.2)	54.2 (11.8)	51.1 (11.5)	<.001 a
TC level, mg/dL	195.8 (43.0)	181.8 (39.4)	199.1 (41.1)	206.3 (44.7)	<.001 a
NLR	3.7 (2.2, 6.0)	3.8 (2.5, 6.3)	3.8 (2.2, 6.1)	3.5 (2.1, 5.2)	.047 c
PLR	125.6 (88.9, 176.0)	138.9 (93.0, 198.9)	116.7 (86.2, 173.7)	116.5 (88.9, 158.0)	<.001 c
LMR	4.5 (3.1)	4.8 (3.6)	4.2 (2.7)	4.3 (3.0)	.21 a

Abbreviations: BMI, body mass index; CAD, coronary artery disease; DM, diabetes mellitus; FPG, fasting plasma glucose; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; LMR, lymphocyte to monocyte ratio; NLR, neutrophil to lymphocyte ratio; PLR: platelet to lymphocyte ratio; TC, total cholesterol; TG, triglyceride; SSNHL, sudden sensorineural hearing loss; TyG, triglyceride glucose. Tertile 1: lowest TyG index group (TyG < 8.162); Tertile 2: middle TyG index group (8.162 ≤ TyG < 8.703); Tertile 3: highest TyG index group (TyG ≥ 8.703). SI conversion factors: To convert fasting plasma glucose to mmol/L, multiply values by 0.0555; to convert low-density lipoprotein cholesterol, high-density lipoprotein cholesterol, and total cholesterol to mmol/L, multiply values by 0.0259; to convert triglyceride to mmol/L, multiply values by 0.0113.

^aANOVA.

^bχ² test.

^cKruskal–Wallis test.

Correlation of TyG Index and Hearing Outcome

Chi-square test was performed to explore the association between TyG tertiles and hearing outcomes in SSNHL patients, as shown in Table 2. Notably, a total of 208 patients (22.0%) achieved complete hearing recovery after treatment, with 86 (27.4%) patients in the Tertile 1 group, 77 (24.4%) in the Tertile 2 group, and 45 (14.3%) in the Tertile 3 group, hinting a negative correlation between the TyG index and hearing recovery. However, no significant variations were observed in the baseline hearing levels and audiogram pattern across different TyG tertiles. Additionally, patients reporting concomitant symptoms such as tinnitus, aural fullness, and vertigo were similarly distributed among the groups.

Table 2.

Hearing Recovery of SSNHL Patients in TyG Tertiles.

Characteristic	Total	Tertile 1 n = 314	Tertile 2 n = 315	Tertile 3 n = 315	P value
Affected side, No. (%)
Left	519 (55.0)	174 (55.4)	176 (55.9)	169 (53.7)	.839 a
Right	425 (45.0)	140 (44.6)	139 (44.1)	146 (46.3)
Initial hearing, mean (SD), dB	76.8 (31.7)	75.8 (32.8)	75.2 (31.5)	79.3 (30.8)	.195 b
Final hearing, mean (SD), dB	56.2 (33.6)	53.1 (34.4)	56.1 (33.8)	59.4 (32.3)	.059 b
Initial hearing severity, No. (%)					.676 a
<35	103 (10.9)	36 (11.5)	40 (12.7)	27 (8.6)
35-49	114 (12.1)	42 (13.4)	37 (11.7)	35 (11.1)
50-64	144 (15.3)	52 (16.6)	42 (13.3)	50 (15.9)
65-79	137 (14.5)	39 (12.4)	50 (15.9)	48 (15.2)
80-94	142 (15.0)	42 (13.4)	51 (16.2)	49 (15.6)
≥95	304 (32.2)	103 (32.8)	95 (30.2)	106 (33.7)
Audiogram pattern, No. (%)					.517 a
Ascending type	153 (16.2)	59 (18.8)	47 (14.9)	47 (14.9)
Descending type	193 (20.4)	68 (21.7)	63 (20.0)	62 (19.7)
Flat type	226 (23.9)	68 (21.7)	85 (27.0)	73 (23.2)
Total type	372 (39.4)	119 (37.9)	120 (38.1)	133 (42.2)
Hearing recovery, No. (%)					<.001 a
Complete recovery	208 (22.0)	86 (27.4)	77 (24.4)	45 (14.3)
Partial recovery	120 (12.7)	29 (9.2)	39 (12.4)	52 (16.5)
Slight recovery	126 (13.3)	46 (14.6)	34 (10.8)	46 (14.6)
No recovery	490 (51.9)	153 (48.7)	165 (52.4)	172 (54.6)
Tinnitus, No. (%)					.681 a
Yes	833 (88.2)	281 (89.5)	275 (87.3)	277 (87.9)
No	111 (11.8)	33 (10.5)	40 (12.7)	38 (12.1)
Aural fullness, No. (%)					.974 a
Yes	474 (50.2)	156 (49.7)	159 (50.5)	159 (50.5)
No	470 (49.8)	158 (50.3)	156 (49.5)	156 (49.5)
Vertigo, No. (%)
Yes	228 (24.2)	77 (24.5)	81 (25.7)	70 (22.2)	.582 a
No	716 (75.8)	237 (75.5)	234 (74.3)	245 (77.8)
Duration from onset to treatment	6.5 (4.2)	6.2 (4.1)	6.6 (4.3)	6.9 (4.2)	.126 b
Initial contralateral hearing, mean (SD), dB	23.0 (23.0)	19.5 (20.4)	23.0 (21.7)	26.6 (25.9)	.001 b
Final contralateral hearing, mean (SD), dB	21.2 (20.6)	17.5 (18.2)	21.2 (19.9)	24.8 (22.7)	<.001 b
Initial contralateral hearing loss, No. (%)					.004 a
Yes	243 (25.7)	61 (19.4)	80 (25.4)	102 (32.4)
No	701 (74.3)	253 (80.6)	235 (74.6)	213 (67.6)

Abbreviations: SSNHL, sudden sensorineural hearing loss; SD, Standard deviation; TyG, triglyceride glucose. Tertile 1: lowest TyG index group (TyG < 8.162); Tertile 2: middle TyG index group (8.162 ≤ TyG < 8.703); Tertile 3: highest TyG index group (TyG ≥ 8.703).

^aχ² test.

^bANOVA.

To delve deeper into this potential linkage, the patients were divided into 2 groups according to hearing recovery: complete recovery group and incomplete recovery group. The baseline demographics of both groups were shown in Table 3. The incomplete hearing recovery group exhibited a significantly elevated TyG index, compared to the complete recovery group (8.30 ± 0.54 vs 8.51 ± 0.57, P < .001).

Table 3.

Baselines Characteristics of SSNHL Patients with Complete Hearing Recovery or Incomplete Hearing Recovery.

Characteristic	Complete recovery n = 208	Incomplete recovery n = 736	P value
TyG index, mean (SD)	8.30 (0.54)	8.51 (0.57)	<.001 a
TyG categories, No. (%)			<.001 b
Tertile 1	86 (41.3)	228 (31.0)
Tertile 2	77 (37.0)	238 (32.3)
Tertile 3	45 (21.6)	270 (36.7)
Age, mean (SD), y	41.1 (13.8)	51.7 (15.3)	<.001 a
Age group, No. (%)			<.001 b
<65	198 (95.2)	571 (77.6)
≥65	10 (4.8)	165 (22.4)
Sex, No. (%)			.704 b
Male	98 (47.1)	360 (48.9)
Female	110 (52.9)	376 (51.1)
BMI, mean (SD), kg/m2	23.2 (3.7)	24.1 (3.7)	.002 a
BMI group, No. (%)			.033 b
<24	124 (59.6)	375 (51.0)
≥24	84 (40.4)	361 (49.0)
Affected side, No. (%)			.893 b
Left	113 (54.3)	406 (55.2)
Right	95 (45.7)	330 (44.8)
Initial hearing, mean (SD), dB	49.3 (26.0)	84.5 (28.7)	<.001 a
Initial hearing severity, No. (%)			<.001 b
<35	82 (39.4)	21 (2.9)
35-49	30 (14.4)	84 (11.4)
50-64	38 (18.3)	106 (14.4)
65-79	22 (10.6)	115 (15.6)
80-94	27 (13.0)	115 (15.6)
≥95	9 (4.3)	295 (40.1)
Audiogram pattern, No. (%)			<.001 b
Ascending type	81 (38.9)	72 (9.8)
Descending type	33 (15.9)	160 (21.7)
Flat type	70 (33.7)	156 (21.2)
Total type	24 (11.5)	348 (47.3)
Tinnitus, No. (%)			.052 b
Yes	192 (92.3)	641 (87.1)
No	16 (7.7)	95 (12.9)
Aural fullness, No. (%)			.012 b
Yes	121 (58.2)	353 (48.0)
No	87 (41.8)	383 (52.0)
Vertigo, No. (%)			<.001 b
Yes	26 (12.5)	202 (27.4)
No	182 (87.5)	534 (72.6)
Hypertension, No. (%)			<.001 b
Yes	36 (17.3)	241 (32.7)
No	172 (82.7)	495 (67.3)
DM, No. (%)			<.001 b
Yes	23 (11.1)	174 (23.6)
No	185 (88.9)	562 (76.4)
CAD, No. (%)			.236 b
Yes	4 (1.9)	29 (3.9)
No	204 (98.1)	707 (96.1)
Smoking history, No. (%)			.590 b
Yes	19 (9.1)	79 (10.7)
No	189 (90.9)	657 (89.3)
Alcohol, No. (%)			.720 b
Yes	13 (6.3)	39 (5.3)
No	195 (93.8)	697 (94.7)
Duration from onset to treatment	6.0 (3.9)	6.7 (4.2)	.036 a
Laboratory findings
TG level, mg/dL	89.0 (60.2, 122.2)	108.1 (70.9, 147.0)	<.001 c
FPG level, mg/dL	89.5 (78.9, 107.9)	92.4 (78.7, 111.9)	.102 c
LDL-C level, mg/dL	117.6 (30.7)	118.0 (30.8)	.77 a
HDL-C level, mg/dL	55.0 (10.9)	53.5 (12.3)	.13 a
TC level, mg/dL	195.9 (41.8)	195.7 (43.3)	.702 a
NLR	3.7 (2.2, 5.7)	3.7 (2.2, 6.1)	.793 c
PLR	122.3 (88.9, 167.6)	125.9 (89.1, 179.9)	.591 c
LMR	4.8 (3.2)	4.4 (3.1)	.12 a

Abbreviations: BMI, body mass index; CAD, coronary artery disease; FPG, fasting plasma glucose; HDL-C, high-density lipoprotein-cholesterol; LDL-C, low-density lipoprotein cholesterol; LMR: lymphocyte to monocyte ratio; NLR, neutrophil to lymphocyte ratio; PLR: platelet to lymphocyte ratio; SSNHL, sudden sensorineural hearing loss; TC, total cholesterol; TG, triglyceride; TyG, triglyceride-glucose. Tertile 1: lowest TyG index group (TyG < 8.162); Tertile 2: middle TyG index group (8.162 ≤ TyG < 8.703); Tertile 3: highest TyG index group (TyG ≥ 8.703). SI conversion factors: To convert fasting plasma glucose to mmol/L, multiply values by 0.0555; to convert low-density lipoprotein cholesterol, high-density lipoprotein cholesterol, and total cholesterol to mmol/L, multiply values by 0.0259; to convert triglyceride to mmol/L, multiply values by 0.0113.

^at test.

^bχ² test.

^cWilcoxon rank sum test.

Predictive Value of TyG Index for Incomplete Hearing Recovery

Multivariate binary logistic regression analyses were performed to evaluate the predictive capacity of TyG index for incomplete hearing recovery. As shown in Table 4, in all 4 models, a higher TyG tertile was associated with a higher likelihood of incomplete hearing recovery (Model 1: OR 2.26, 95% CI 1.52-3.40, P < .001; Model 2: OR 2.07, 95% CI 1.38-3.14, P = .001; Model 3: OR 1.82, 95% CI 1.20-2.80, P = .008; Model 4: OR 1.93, 95% CI 1.14-3.29, P = .029). Meanwhile, when treated as a continuous variable, TyG index was also significantly associated with incomplete hearing recovery (Model 1: OR 1.91, 95% CI 1.44-2.55, P < .001; Model 2: OR 1.76, 95% CI 1.32-2.37, P < .001; Model 3: OR 1.58, 95% CI 1.17-2.16, P = .003; Model 4: OR 1.55, 95% CI 1.05-2.29, P = .027). The detailed results of the logistic regression analyses are provided in Supplementary Tables 1-2 .

Table 4.

Logistic Regression Analysis of the Association Between the TyG Index and Incomplete Hearing Recovery.

Variables	Model 1 OR (95% CI) a	Model 2 OR (95% CI) b	Model 3 OR (95% CI) c	Model 4 OR (95% CI) d
Continuous
TyG index	1.906 (1.435, 2.551)	1.764 (1.324, 2.367)	1.584 (1.170, 2.157)	1.546 (1.052, 2.287)
P value	<.001	<.001	.003	.027
Tertiles
Tertile 1	Reference	Reference	Reference	Reference
Tertile 2	1.166 (0.816, 1.668)	1.092 (0.759, 1.573)	1.024 (0.708, 1.480)	1.066 (0.665, 1.710)
Tertile 3	2.263 (1.522, 3.403)	2.072 (1.381, 3.142)	1.821 (1.195, 2.802)	1.925 (1.141, 3.285)
P value	<.001	.001	.008	.029

Abbreviations: BMI, body mass index; CI, confidence interval; OR, odds ratio; TyG, triglyceride glucose. Tertile 1: lowest TyG index group (TyG < 8.162); Tertile 2: middle TyG index group (8.162 ≤ TyG < 8.703); Tertile 3: highest TyG index group (TyG ≥ 8.703).

^aModel 1 was adjusted for no covariates.

^bModel 2 was adjusted for age, sex, smoking history, and alcohol use history.

^cMode 3 was adjusted for age, sex, BMI, smoking history, alcohol use history, hypertension, and diabetes mellitus.

^dModel 4 was adjusted for age, sex, BMI, smoking history, alcohol use history, hypertension, diabetes mellitus, initial hearing severity, audiogram pattern, contralateral hearing loss, duration from onset to treatment, and vertigo.

The RCS result was shown in Figure 2. After adjusting for all covariates in Model 4, a generally increasing trend in the odds of incomplete hearing recovery was observed across the TyG index (P for overall = .076, P for nonlinear = .530).

Figure 2.

Restricted cubic spline curve for the dose–response relationship between TyG index and incomplete hearing recovery. The restricted cubic spline analysis was performed upon Model 4 (adjusted for sex, age, smoking history, alcohol use history, BMI, hypertension, DM, initial hearing severity, audiogram pattern, contralateral hearing loss, duration from onset to treatment, and presence of vertigo). The blue line represents the odds ratio, and the shaded area represents the 95% confidence interval. BMI, body mass index; DM, diabetes mellitus; TyG, triglyceride glucose.

Subgroup Analysis

Finally, subgroup analyses were conducted according to age, sex, BMI, DM, hypertension, smoking and alcohol use history, severe to profound hearing loss, and contralateral hearing loss. As shown in Figure 3, the direction of the association between the TyG index and incomplete hearing recovery was generally consistent across subgroups, although the precision of subgroup-specific estimates varied. No statistically significant interactions were observed (all P for interaction > .05).

Figure 3.

Subgroup analysis for the association of TyG index with incomplete hearing recovery. The subgroup analysis was performed by using Model 4 (adjusted for sex, age, smoking history, alcohol use history, BMI, hypertension, DM, initial hearing severity, audiogram pattern, contralateral hearing loss, duration from onset to treatment, and presence of vertigo). The black vertical dashed line represents the OR value of 1. BMI, body mass index; CI, confidence interval; DM, diabetes mellitus; OR, odds ratio; TyG, triglyceride glucose.

Discussion

SSNHL is an otolaryngological emergency that affects 5 to 20 of 100 000 people in the US population.^1,2 Although the restoration of hearing has been observed in about one-third of patients, ³ the other two-thirds show no or only partial recovery, ⁵ of which a large proportion suffers from anxiety and depression,^{3 -5,7} thereby imposing a heavy global public health burden. In this retrospective study of 944 SSNHL patients, we investigated the association between the TyG index and hearing prognosis in a Chinese population. Our findings revealed that patients in a higher TyG index tertile were more likely to experience incomplete hearing recovery, despite comparable audiometric patterns and initial hearing severity across tertiles. After adjusting for relevant covariates, the TyG index remained an independent predictor of poor recovery (OR: 1.93, 95% CI 1.14-3.29). Subgroup analyses further showed generally similar directions of association across subgroups. These findings suggest that insulin resistance-related metabolic dysregulation may influence cochlear recovery, and that TyG index, a simple and widely available marker, may aid in early prognostic stratification in SSNHL.

Despite the various etiologies underlying SSNHL, poor microcirculation was recognized as a fundamental pathophysiological mechanism. The cochlea, as an end organ, relies heavily on a continuous supply of nutrients and oxygen to sustain its physiological functions, ²¹ making it pretty fragile to microcirculation and endothelial dysfunction. ²¹ Accordingly, alterations in blood viscosity, flow dynamics, and composition may impair microcirculation, thereby dysregulating the blood supply, oxygenation, and overall functions of delicate inner ear organs. In this context, extensive research has explored the correlation of hearing loss with cardiovascular risk factors, particularly DM, hypertension, and dyslipidemia. Two large cohort studies have reported that DM, due to its damage on the microvasculature within the cochlea, is associated with sensorineural hearing loss.^22,23 However, the predictive value of DM for hearing recovery in SSNHL patients remains controversial. On the one hand, several investigations have observed inferior hearing recovery among DM patients^24,25; on the other hand, other studies have observed similar or even favorable hearing prognosis in DM patients.^{26 -28} Similar dilemma aroused from the exploration of lipid metabolism and its association with SSNHL. While Zhang et al and Oreskovic et al have detected elevated levels of total cholesterols, triglycerides, and lipoproteins among SSNHL patients^29,30; another clinical study ³¹ and a systematic review ³² failed to identify these elevations. Apart from the inherent heterogeneity among study designs, these conflicting results may stem from the subtle yet cumulative damage on auditory organs induced by IR, which often precedes DM by several years. ¹³ Composing both glucose and lipid metabolism, TyG index is considered as a convenient and reliable marker of IR^15,16 and is associated with various cardiovascular events. TyG index has been proven useful for early identification of apparently healthy individuals at a risk of developing cardiovascular diseases.^33,34 In the present study, we found that the risk of incomplete hearing recovery increased in conjunction with TyG index, and that TyG index was an independent predictor of incomplete hearing recovery (OR: 1.93, 95% CI 1.14-3.29) after adjusting for traditional risk factors. Notably, the predictive value of TyG index has also been studied in healthy people without hypertension or DM. These findings support the notion that TyG index may serve as an independent marker to predict hearing recovery in both general and special populations.

The underlying mechanisms linking the TyG index and hearing outcomes are not fully understood. However, being part of the systemic circulation, the microcirculation of the inner ear may be influenced by IR through the following mechanisms. First, a continuous elevated glucose level induced by IR can abnormalize oxidative phosphorylation ³⁵ and inhibit the production of adenosine triphosphate (ATP), which in turn cause mitochondrial destruction and microvascular lesions, thus leading to marginal cell damage and necrosis. Studies on DM patients and animal models demonstrate capillary basement membrane thickening in the stria vascularis, internal auditory artery narrowing, ³⁶ Corti’s organ and spiral ganglion neuron loss. ³⁷ Additionally, abnormal triglyceride and HDL levels can increase blood viscosity, reduce blood flow to the inner ear, and potentially damage its delicate system.^38,39 This process may lead to excess lipid deposition in cochlear hair cells, harm cochlear nerve cells, impede nerve conduction, and increase capillary wall tension.^38,40 A high fat diet can increase cell loss in the spiral ganglion of the cochlea, and the ratio of vessel wall thickness to radius in the spiral vessel at the basal turn of the cochlea. ³⁸ Lastly, a critical balance of nitric oxide (NO) is essential for the normal functions of cochlear sensory and support cells.^41,42 Located in key cochlear blood vessels, NO plays a pivotal role in regulating cochlear vascular tone, blood flow, and antithrombotic activity.^41,42 Under IR and hyperglycemia, advanced glycosylation end-products and excessive reactive oxygen species production, ⁴³ NO bioavailability may be reduced, further impairing cochlear function.

Our study also found that TyG index was an independent predictor of hearing recovery, even after adjusting for audiogram pattern and initial hearing severity, which are established prognostic indicators for SSNHL.^19,44 A comprehensive, multicenter cohort study involving 1024 SSNHL patients has revealed variations in hearing recovery across distinct audiogram pattern, with ascending patterns exhibiting the most favorable and descending patterns yielding the least favorable prognosis; ⁴⁴ besides, initial hearing severity is also associated with hearing recovery. In this study, the distributions of audiogram pattern and hearing levels were normal among the TyG tertiles. Moreover, after adjusting for audiogram pattern, initial hearing severity, and other relevant characteristics, TyG index remained as an independent predictor of incomplete hearing recovery.

In addition to hearing thresholds in the affected ear, contralateral hearing status has been suggested as a prognostic factor in SSNHL and may reflect broader systemic vascular or metabolic vulnerability. ⁴⁵ In a retrospective analysis of 762 patients with SSNHL, Tsuzuki et al ⁴⁵ reported that several atherosclerosis-related factors were associated with hearing thresholds on both the affected and contralateral (healthy) ears, and higher contralateral-ear thresholds were associated with lack of recovery. In the present study, contralateral hearing loss was more common across higher TyG index tertiles (T1 vs T2 vs T3: 61 [19.4%] vs 80 [25.4%] vs 102 [32.4%]; P < .001), indicating a potential relationship between the TyG index and contralateral hearing loss. Importantly, the TyG index remained independently associated with incomplete hearing recovery after adjustment for contralateral hearing loss and other covariates, suggesting that it may provide additional prognostic information beyond baseline hearing status in the nonaffected ear.

To our knowledge, this is the first research to explore the association between TyG index and hearing recovery of SSNHL patients based on a large-scale Chinese population, which provides a novel and interesting perspective into the role of the TyG index in hearing loss. SSNHL patients with a high TyG index level may benefit from lifestyle adjustments or tailored treatment strategies.

There were several limitations in this study. To begin with, the cross-sectional study design hindered the explanation of the causal relationship between TyG index and hearing loss prognosis. In addition, as this was a single-center study, potential data bias might persist despite correction for multiple confounding factors. Third, prognosis was assessed 2 weeks after treatment, and the findings therefore reflect short-term hearing outcomes. Future prospective cohort studies are warranted to validate these findings.

Conclusion

In summary, this study indicated that elevated TyG index is independently associated with incomplete hearing recovery in patients with SSNHL. This finding suggests that metabolic status may play a role in cochlear prognosis, and that TyG index could serve as a simple tool for early risk stratification in clinical practice.