Associations of thyroid feedback quantile-based index with diabetes in euthyroid adults in the United States and China

Figures & data

Figure 1. Flowchart of the study participants. (a) The NHANES. (b) The SPEED-Shunde.

Table 1. General characteristics of participants in the study.

	NHANES	SPEED-Shunde
N	2296	8319
Age, year	44.56 (16.14)	46.48 (12.11)
Men, %	51.44	38.19
Race/ethnicity
Mexican American	9.16	–
Other Hispanic	5.52	–
Non-Hispanic White	68.78	–
Non-Hispanic Black	9.81	–
Others	6.73	–
Education, %
Less than high school	29.68	39.07
High school	31.20	24.10
Beyond high school	39.12	36.83
Abusive drink, %	0.19	2.95
Smoking, %
Non-smokers	55.20	85.88
Former smokers	23.54	2.97
Current smokers	21.26	11.16
Diabetes, %	6.62	12.30
BMI, kg/m^2	27.99 (5.97)	23.66 (3.44)
SBP, mmHg	120.34 (16.12)	123.43 (17.07)
DBP, mmHg	70.01 (11.84)	82.14 (17.91)
FPG, mmol/L	5.56 (0.78)	4.79 (1.01)
HbA1c, %	5.41 (0.48)	5.68 (0.62)
PPG, mmol/L	6.37 (2.53)	7.77 (2.90)
HOMA-IR	2.29 (1.46, 3.69)	3.92 (7.83)
FT3, pmol/L	4.97 (0.47)	5.35 (0.58)
FT4, pmol/L	10.23 (1.56)	11.26 (1.42)
TSH, mIU/L	1.67 (1.17, 2.36)	1.74 (1.24,2.43)
FT3/FT4 ratio	0.50 (0.08)	0.48 (0.07)
TFQIFT3	−0.02 (0.40)	−0.04 (0.39)
TFQIFT4	−0.04 (0.39)	−0.04 (0.38)
TPOAb, IU/ml	0.60 (0.30, 1.50)	1.00 (0.50;3.23)
TgAb, IU/ml	0.60 (0.60, 0.60)	0.10 (0.07;0.64)
eGFR, mL/min/m^2	98.35 (19.74)	96.26 (15.76)
TG, mmol/L	1.19 (0.85, 1.71)	1.21 (0.86,1.80)
TC, mmol/L	5.07 (1.03)	5.38 (1.58)
HDL-C, mmol/L	1.39 (0.39)	1.45 (0.33)
LDL-C, mmol/L	3.03 (0.89)	3.01 (0.74)

Notes: The baseline characteristics of participants were summarized as mean (standard deviation) or median (interquartile range) for continuous variables and proportion (%) for categorical variables according to with or without diabetes. Appropriate sampling weights were used when calculating the data in the NHANES. BMI: body mass index; SBP: systolic blood pressure; DBP: diastolic blood pressure; FPG: fasting plasma glucose; HbA1c: Glycated haemoglobin; PPG: postprandial glucose: FT3: free triiodothyronine; FT4: free thyroxine; TSH: thyroid-stimulating hormone; FT3/FT4 ratio: FT3 to FT4 ratio; TFQI_FT3: the thyroid feedback quantile-based index calculated by FT3; TFQI_FT4: the thyroid feedback quantile-based index calculated by FT4; TPOAb: thyroid peroxidase antibody; TgAb: thyroglobulin antibody; eGFR: estimated glomerular filtration rate; TG: triglyceride; TC: total cholesterol; HDL-C: high-density lipoprotein-cholesterol; LDL-C: low-density lipoprotein-cholesterol.

Table 2. Logistic regression analysis for the association between thyroid sensitivity indices and diabetes in the euthyroid population.

	NHANES			SPEED-Shunde
	Range	OR (95%CI)	p	Range	OR (95%CI)	p
FT3/FT4 ratio
P for trend			.975			.252
Q4	0.55≤ Q4	0.89 (0.53, 1.50)	.650	0.53≤ Q4	1.07 (0.87, 1.30)	.535
Q3	0.49 ≤ Q3 < 0.55	1.31 (0.72, 2.38)	.356	0.48 ≤ Q3 < 0.53	1.16 (0.95, 1.41)	.148
Q2	0.44≤ Q2 < 0.49	1.05 (0.60, 1.83)	.854	0.43≤ Q2 < 0.48	0.96 (0.78, 1.18)	.710
Q1	Q1 < 0.44	Ref.		Q1 < 0.43	Ref.
TFQIFT3
P for trend			.011			.001
Q4	0.24≤ Q4	2.12 (1.18, 3.81)	.014	0.23≤ Q4	1.36 (1.11, 1.66)	.003
Q3	−0.02 ≤ Q3 < 0.24	2.22 (1.29, 3.83)	.006	−0.04 ≤ Q3 < 0.23	1.06 (0.87, 1.30)	.572
Q2	−0.32≤ Q2 < −0.02	1.92 (0.98, 3.74)	.056	−0.34≤ Q2 < −0.04	0.96 (0.78, 1.17)	.660
Q1	Q1 < −0.32	Ref.		Q1 < −0.34	Ref.
TFQIFT4
P for trend			.021			.137
Q4	0.19≤ Q4	1.79 (0.95, 3.37)	.071	0.20≤ Q4	1.19 (0.98, 1.46)	.083
Q3	−0.06 ≤ Q3 < 0.19	2.31 (1.18, 4.53)	.017	−0.04 ≤ Q3 < 0.20	1.15 (0.95, 1.40)	.161
Q2	−0.34≤ Q2 < −0.06	1.49 (0.65, 3.41)	.339	−0.32≤ Q2 < −0.04	1.22 (1.00, 1.48)	.046
Q1	Q1 < −0.34	Ref.		Q1 < −0.32	Ref.

Notes: The model was adjusted for sex, age, BMI, eGFR, hypertension, thyroiditis, dyslipidemia, abused drink, education, smoking status and HOMA-IR in the SPEED-Shunde (further adjusted for race/ethnicity in the NHANES).

Figure 2. Non-linear relationships of thyroid sensitivity indices with diabetes. Restricted cubic spline (RCS) analysis was conducted. (a) FT3/FT4 ratio and diabetes in the NHANES (b) TFQI_FT3 and diabetes in the NHANES (c) TFQI_FT4 and diabetes in the NHANES (d) FT3/FT4 ratio and diabetes in the SPEED-Shunde (e) TFQI_FT3 and diabetes in the SPEED-Shunde (f) TFQI_FT4 and diabetes in the SPEED-Shunde. The model was adjusted for sex, age, BMI, eGFR, hypertension, thyroiditis, dyslipidemia, abused drink, education, smoking status and HOMA-IR in the SPEED-Shunde (further adjusted for race/ethnicity in the NHANES).

Figure 3. Associations of thyroid sensitivity indices with diabetes stages using general ordinal logit regression. Thyroid sensitivity indices quartiles and diabetes (a) in the NHANES and (b) in the SPEED-Shunde. The model was adjusted for sex, age, BMI, eGFR, hypertension, thyroiditis, dyslipidemia, abused drink, education, smoking status and HOMA-IR in the SPEED-Shunde (further adjusted for race/ethnicity in the NHANES).

Figure 4. Stratification analysis stratified by sex, age and BMI. Associations of (a) FT3/FT4 ratio, (b) TFQI_FT3 and (c) TFQI_FT4 with diabetes in the NHANES. Associations of (d) FT3/FT4 ratio, (e) TFQI_FT3 and (f) TFQI_FT4 with diabetes in the SPEED-Shunde. The model was adjusted for sex, age, BMI, eGFR, hypertension, thyroiditis, dyslipidemia, abused drink, education, smoking status and HOMA-IR in the SPEED-Shunde (further adjusted for races in the NHANES).

Data availability statement

The data supporting the study findings are available from the corresponding authors upon reasonable request.