Association Between Triglyceride-Glucose Index and Hearing Threshold Shifts of Adults in the United States: National Health and Nutrition Examination Survey, 2015–2016

Figures & data

Figure 1 Flow chart of the selection.

Notes: Figure 1 described the selection procedure of study participants. Audiometry examinations were performed in adults aged 20–69 years from the 2015–2016 NHANES data. Participants without complete data on hearing levels, otoscopic test, tympanogram test, or Triglyceride Glucose index were excluded, as well as participants with abnormal otoscopic results, poor-quality tympanogram results, or tympanogram with compliance of ≤ 0.3mL. Ultimately, the study included 1226 adults.

Table 1 The Weighted Demographic Characteristics of Study Participants

Variable	Overall	Tertiles of Triglyceride-Glucose Index			P value
		Tertile 1 (5.85–8.29)	Tertile 2 (8.30–8.90)	Tertile 3 (8.91–11.47)
Continuous variables, mean ± SD
Age (years)	44.44 ± 14.11	41.76 ± 14.58	43.79 ± 13.65	47.97 ± 13.33	<0.0001
Low-frequency PTA (dB)	8.22 ± 8.49	7.11 ± 9.71	7.62 ± 7.10	10.03 ± 8.18	<0.0001
Speech-frequency PTA (dB)	10.36 ± 9.57	8.93 ± 10.90	9.40 ± 8.02	12.90 ± 9.08	<0.0001
High-frequency PTA (dB)	20.87 ± 17.13	18.88 ± 18.08	18.57 ± 14.88	25.45 ± 17.45	<0.0001
HDL-Cholesterol (mg/dL)	56.68 ± 20.56	67.64 ± 19.62	57.28 ± 20.99	44.48 ± 12.92	<0.0001
CRP (mg/dL)	3.83 ± 7.18	2.38 ± 5.61	4.12 ± 8.24	5.05 ± 7.17	<0.0001
Insulin (uU/mL)	13.00 ± 14.57	7.63 ± 6.37	11.81 ± 12.67	19.95 ± 19.19	<0.0001
Total Cholesterol (mg/dL)	193.06 ± 39.91	176.02 ± 32.89	195.32 ± 37.42	208.62 ± 42.26	<0.0001
HbA1c (%)	5.59 ± 0.83	5.30 ± 0.40	5.49 ± 0.49	5.99 ± 1.20	<0.0001
Categorical variables, %
Gender					<0.0001
Male	45.87	36.31	45.96	55.86
Female	54.13	63.69	54.04	44.14
Race/Ethnicity					0.0008
Mexican American	8.12	6.03	8.65	9.76
Non-Hispanic White	67.18	67.03	66.09	68.49
Non-Hispanic Black	9.96	14.66	9.33	5.66
Other Races	14.75	12.27	15.93	16.09
Education level					0.0003
Below high school	12.10	10.97	9.76	15.81
High school	20.91	15.51	25.81	21.36
Above high school	66.98	73.52	64.43	62.83
BMI status					<0.0001
Underweight (<18.5)	1.52	2.94	1.20	0.39
Normal (≥18.5,<25)	26.33	45.20	24.49	8.40
Overweight (≥25,<30)	30.90	29.88	33.51	29.16
Obesity (≥30)	41.25	21.98	40.80	62.05
Smoking status					<0.0001
Never	55.80	66.70	49.27	51.43
Former	25.09	14.63	30.77	29.86
Current	19.11	18.67	19.96	18.71
Hypertension	26.71	16.73	23.55	40.64	<0.0001
Diabetes	8.64	2.63	6.40	17.37	<0.0001
Exposed to loud noise/music	15.62	17.52	17.25	11.86	0.0456
Hearing loss	14.03	10.95	11.62	19.85	0.0003

Abbreviations: PTA, pure-tone average; HDL, high density lipoprotein; CRP, C-reactive protein; BMI, body mass index.

Table 2 The Univariate Analysis of Comparison of Variables in Hearing Threshold Group

Variables	N (%)/Mean ± SD	Low-Frequency PTA	Speech-Frequency PTA	High-Frequency PTA
		β/OR (95% CI) p value	β/OR (95% CI) p value	β/OR (95% CI) p value
Age	43.91 ± 14.00	0.24 (0.21, 0.27) <0.0001	0.35 (0.31, 0.38) <0.0001	0.75 (0.70, 0.80) <0.0001
GENDER
Male	549 (44.78%)	Reference	Reference	Reference
Female	677 (55.22%)	−0.20 (−1.16, 0.75) 0.6752	−2.45 (−3.52, −1.38) <0.0001	−6.04 (−7.94, −4.15) <0.0001
Race/Ethnicity
Mexican American	210 (17.13%)	Reference	Reference	Reference
Non-Hispanic White	405 (33.03%)	2.35 (0.59, 4.11) 0.0090	3.00 (1.03, 4.98) 0.0030	6.77 (3.27, 10.27) 0.0002
Non-Hispanic Black	242 (19.74%)	0.37 (−1.87, 2.61) 0.7433	−0.35 (−2.87, 2.16) 0.7832	−1.82 (−6.28, 2.63) 0.4227
Other Races	369 (30.10%)	1.61 (−0.46, 3.68) 0.1275	1.61 (−0.72, 3.93) 0.1761	1.35 (−2.77, 5.47) 0.5221
Education level
Below high school	244 (19.90%)	Reference	Reference	Reference
High school	266 (21.70%)	−2.57 (−4.28, −0.86) 0.0032	−2.55 (−4.47, −0.63) 0.0095	−1.33 (−4.79, 2.12) 0.4492
Above high school	716 (58.40%)	−2.87 (−4.35, −1.39) 0.0001	−3.53 (−5.19, −1.87) <0.0001	−3.72 (−6.71, −0.74) 0.0147
BMI status
Underweight (<18.5)	21 (1.71%)	Reference	Reference	Reference
Normal (≥18.5,<25)	318 (25.94%)	0.91 (−3.03, 4.84) 0.6519	2.60 (−1.83, 7.03) 0.2499	11.30 (3.36, 19.25) 0.0054
Overweight (≥25,<30)	386 (31.48%)	1.71 (−2.21, 5.62) 0.3934	3.72 (−0.69, 8.14) 0.0983	12.41 (4.50, 20.32) 0.0022
Obesity (≥30)	501 (40.86%)	3.38 (−0.51, 7.27) 0.0891	5.36 (0.97, 9.74) 0.0168	14.23 (6.37, 22.10) 0.0004
Smoking status
Never	715 (58.32%)	Reference	Reference	Reference
Former	264 (21.53%)	0.96 (−0.18, 2.10) 0.0994	1.93 (0.65, 3.22) 0.0032	4.34 (2.05, 6.64) 0.0002
Current	244 (19.90%)	1.69 (0.43, 2.95) 0.0086	2.21 (0.80, 3.62) 0.0022	2.45 (−0.08, 4.99) 0.0575
Hypertension	361 (29.45%)	4.41 (3.37, 5.46) <0.0001	5.68 (4.51, 6.85) <0.0001	10.41 (8.33, 12.50) <0.0001
Diabetes	138 (11.26%)	4.71 (3.04, 6.38) <0.0001	6.52 (4.65, 8.39) <0.0001	12.23 (8.88, 15.58) <0.0001
Exposed to loud noise/music	182 (14.85%)	−2.09 (−3.40, −0.79) 0.0017	−2.90 (−4.37, −1.43) 0.0001	−6.90 (−9.52, −4.29) <0.0001
HDL-Cholesterol	55.33 ± 17.79	0.01 (−0.02, 0.03) 0.6424	−0.01 (−0.04, 0.01) 0.3709	−0.02 (−0.07, 0.02) 0.3416
CRP	4.20 ± 9.11	0.10 (0.03, 0.16) 0.0043	0.09 (0.02, 0.16) 0.0185	0.11 (−0.02, 0.25) 0.0992
Insulin	13.79 ± 17.66	0.07 (0.03, 0.10) <0.0001	0.07 (0.03, 0.11) 0.0002	0.08 (0.02, 0.15) 0.0142
Total Cholesterol (mg/dL)	192.13 ± 40.33	0.01 (−0.00, 0.02) 0.1417	0.01 (−0.00, 0.02) 0.1740	0.02 (−0.01, 0.04) 0.1864
HbA1c (%)	5.71 ± 1.00	2.01 (1.44, 2.57) <0.0001	2.85 (2.22, 3.48) <0.0001	5.37 (4.24, 6.49) <0.0001
TyG index	8.63 ± 0.70	1.79 (1.09, 2.48) <0.0001	2.61 (1.83, 3.39) <0.0001	4.84 (3.45, 6.23) <0.0001
TyG index (Tertiles)
Tertile 1 (5.85–8.29)	403 (32.87%)	Reference	Reference	Reference
Tertile 2 (8.30–8.90)	410 (33.44%)	0.51 (−0.63, 1.65) 0.3827	0.47 (−0.80, 1.75) 0.4679	−0.31 (−2.60, 1.98) 0.7888
Tertile 3 (8.91–11.47)	410 (33.44%)	2.92 (1.76, 4.09) <0.0001	3.97 (2.67, 5.28) <0.0001	6.57 (4.24, 8.90) <0.0001

Abbreviations: BMI, body mass index; HDL, high density lipoprotein; CRP, C-reactive protein; TyG, index, triglyceride–glucose index.

Table 3 Multivariable Linear Regression Models for Outcome of Hearing Thresholds

TyG Index	β (95% CI), P value of PTA levels, dB
	Crude Model	Model I	Model II
Low-frequency PTA
TyG index	1.79 (1.09, 2.48) <0.001	0.63 (−0.04, 1.30) 0.066	−0.81 (−2.31, 0.70) 0.296
Tertile 1	Reference	Reference	Reference
Tertile 2	0.51 (−0.63, 1.65) 0.383	−0.01 (−1.07, 1.05) 0.982	0.10 (−0.89, 1.08) 0.850
Tertile 3	2.92 (1.76, 4.09) <0.001	1.37 (0.27, 2.48) 0.015	0.44 (−1.08, 1.97) 0.569
P for trend	< 0.001	0.016	0.223
Speech-frequency PTA
TyG index	2.61 (1.83, 3.39) <0.001	0.64 (−0.05, 1.34) 0.071	−1.03 (−2.44, 0.39) 0.157
Tertile 1	Reference	Reference	Reference
Tertile 2	0.47 (−0.80, 1.75) 0.468	−0.51 (−1.61, 0.59) 0.366	−0.46 (−1.39, 0.47) 0.334
Tertile 3	3.97 (2.67, 5.28) <0.001	1.26 (0.11, 2.41) 0.032	0.08 (−1.35, 1.52) 0.909
P for trend	< 0.001	0.036	0.375
High-frequency PTA
TyG index	4.84 (3.45, 6.23) <0.001	0.42 (−0.70, 1.54) 0.463	−1.56 (−4.20, 1.07) 0.245
Tertile 1	Reference	Reference	Reference
Tertile 2	−0.31 (−2.60, 1.98) 0.789	−2.56 (−4.33, −0.80) 0.004	−2.48 (−4.20, −0.76) 0.005
Tertile 3	6.57 (4.24, 8.90) <0.001	0.41 (−1.43, 2.25) 0.666	−1.06 (−3.72, 1.60) 0.435
P for trend	< 0.001	0.726	0.732

Notes: Crude Model: unadjusted. Model I: Crude Model+gender, age. Model II: Model I+ HDL-cholesterol, CRP, insulin, total cholesterol, HbA1c, race, education level, BMI status, smoking status, hypertension, diabetes, exposed to loud noise/music.

Abbreviations: TyG, index, triglyceride–glucose index; CI, confidence interval; PTA, pure tone average.

Table 4 The Results of Two-Piecewise Linear Regression Model Between TyG Index and Hearing Thresholds

Exposure Variables	Low-Frequency PTA	Speech-Frequency PTA	High-Frequency PTA
Cut off point of TyG index	8.14	8.35	8.57
< Cut off point of TyG index	−2.00 (−4.12, 0.12) 0.0651	−1.71 (−3.26, −0.17) 0.0297	−2.75 (−5.11, −0.39) 0.0228
≥ Cut off point of TyG index	0.81 (−0.18, 1.81) 0.1085	1.17 (0.16, 2.18) 0.0228	2.73 (0.67, 4.80) 0.0096
Log-likelihood ratio test	0.022	0.003	<0.001

Abbreviations: TyG, index, triglyceride–glucose index; PTA, pure tone average.

Figure 2 The relationship between TyG index and hearing threshold shifts: (A) low-frequency PTA, (B) speech-frequency PTA and (C) high-frequency PTA.

Notes: Figure 2 used smooth curve fittings to explore the relationship between TyG index and hearing threshold shifts, after the adjustment of potential confounders. Finally, TyG index shows a non-linear relevance to the speech-frequency and high-frequency PTAs, but not to low-frequency PTA. The fitted smooth curves were approximately U-shaped.

Abbreviations: PTA, pure tone average; TyG index, triglyceride-glucose index.