Association Between Hypertension and New-Onset Non-Alcoholic Fatty Liver Disease in Chinese Non-Obese People: A Longitudinal Cohort Study

Figures & data

Figure 1 Flowchart of study participants. The process of screening participants.

Table 1 Baseline Characteristics Before Propensity-Score Matching in the Entire Study Population

Characteristic	Non-Hypertension(n=13,726)	Hypertension(n=2427)	SD (100%)
AGE (year)	42.53 ± 14.65	47.19 ± 16.03	30.4
SEX			8.9
Female	6618 (48.22%)	1063 (43.80%)
Male	7108 (51.78%)	1364 (56.20%)
BMI (kg/m^2)	21.21 ± 2.03	22.33 ± 1.86	57.3
ALP (U/L)	70.88 ± 22.12	80.19 ± 23.86	40.5
ALB (g/L)	44.47 ± 2.67	44.16 ± 2.91	10.8
AST (U/L)	22.59 ± 8.74	25.12 ± 8.96	28.6
GLB (g/L)	29.32 ± 3.78	30.36 ± 4.20	26.0
BUN (mg/dL)	12.52 ± 3.46	14.57 ± 5.18	46.4
Scr (mg/dL)	0.87 ± 0.21	1.01 ± 0.53	36.3
UA (mg/dL)	4.59 ± 1.40	5.32 ± 1.52	50.0
FPG (mg/dL)	91.00 ± 11.93	101.48 ± 20.49	62.5
TC (mg/dL)	177.66 ± 28.23	185.24 ± 30.64	25.7
HDL-c (mg/dL)	56.86 ± 13.93	55.00 ± 14.69	13.0
LDL-c (mg/dL)	87.00 ± 17.93	90.58 ± 17.99	20.0
GGT (U/L)	21.00 (16.00–30.11)	26.08 (20.00–41.00)	32.3
ALT (U/L)	16.41 (12.00–23.00)	18.00 (13.62–25.00)	10.5
TG (mg/dL)	92.12 (69.09–129.32)	115.15 (86.80–162.98)	33.8
PS	0.09 (0.05–0.16)	0.24 (0.14–0.39)	98.3

Note: Values are n (%) or mean ±standard deviation or median (quartile).

Abbreviations: SD, Standardized differences; ALP, Alkaline phosphatase; BMI, Body mass index; TC, Total cholesterol; GGT, γ-glutamyl transpeptidase; TG, Triglyceride; ALB, albumin; HDL-C, High-density lipoprotein cholesterol; ALT, Alanine aminotransferase; GLB, globulin; AST, Aspartate aminotransferase; LDL-C, Low-density lipid cholesterol; BUN, Serum urea nitrogen; Scr, Serum creatinine; UA, uric acid; FPG, Fasting plasma glucose; PS, propensity score.

Figure 2 The ROC curve of the propensity score to predict hypertension. The logistic model was used to estimate the propensity score, which yielded a c-statistic of 0.795.

Figure 3 Distribution of propensity score in hypertension and non-hypertension groups before matching. The probability density functions of the propensity score for hypertensive and non-hypertensive participants.

Table 2 Proportion of NAFLD Among 16,153 NAFLD-Free Non-Obese Individuals Who Were Hypertension or Not, According to Percentiles of the Propensity Score for the Entire Study Population

Percentile	Hypertension (n=2427)				Non-Hypertension (n=13,726)				Empirical OR*
	Score^#	No.	NAFLD		Score^#	No.	NAFLD
			No.	%			No.	%
99 to 100	0.8837	113	45	39.8	0.8997	49	17	34.7	1.25
95 to <99	0.5475	334	128	38.3	0.5281	312	103	33.0	1.26
90 to <95	0.3768	359	89	24.8	0.3680	449	148	33.0	0.67
75 to <90	0.2503	661	180	27.2	0.2369	1762	462	26.2	1.05
50 to <75	0.1457	598	121	20.2	0.1370	3440	631	18.3	1.13
25 to <50	0.0818	268	36	13.4	0.0755	3770	292	7.7	1.85
10 to <25	0.0446	68	4	5.9	0.0427	2354	52	2.2	2.77
5 to <10	0.0301	16	1	6.3	0.0284	792	7	0.9	7.48
1 to <5	0.0183	10	0	0	0.0198	636	2	0.3
0 to <1		0	0	0	0.0119	162	3	1.9
Overall	0.2351	2427	604	24.9	0.0895	13,726	1717	12.5	2.32

Notes: *Propensity-stratum-specific-treatment–mortality odds ratio. ^#Median propensity score in percentile.

Abbreviations: NAFLD, Non-alcoholic fatty liver disease; OR, Odds ratio; No., Number.

Table 3 Baseline Characteristics After Propensity-Score Matching in the Matching Study Population

Characteristic	Non-Hypertension(n=2316)	Hypertension(n=2316)	SD (100%).
AGE (year)	46.77 ± 15.95	46.95 ± 15.93	1.1
SEX			1.3
Female	999 (43.13%)	1014 (43.78%)
Male	1317 (56.87%)	1302 (56.22%)
BMI (kg/m^2)	22.29 ± 1.84	22.30 ± 1.87	0.4
ALP (U/L)	79.46 ± 29.61	79.55 ± 23.34	0.3
ALB (g/L)	44.32 ± 2.77	44.23 ± 2.83	3.2
AST (U/L)	24.96 ± 11.26	24.88 ± 8.25	0.9
GLB (g/L)	30.17 ± 4.20	30.27 ± 4.04	2.5
BUN (mg/dL)	14.17 ± 4.25	14.27 ± 4.45	2.5
Scr (mg/dL)	0.97 ± 0.29	0.99 ± 0.40	5.0
UA (mg/dL)	5.30 ± 1.52	5.29 ± 1.49	1.0
FPG (mg/dL)	98.59 ± 21.72	99.91 ± 17.43	6.7
TC (mg/dL)	185.73 ± 30.23	185.06 ± 30.27	2.2
HDL-c (mg/dL)	55.17 ± 14.21	55.07 ± 14.54	0.7
LDL-c (mg/dL)	91.26 ± 17.96	90.72 ± 17.79	3.0
GGT (U/L)	26.00 (19.00–40.00)	26.00 (19.00–40.09)	0.7
ALT (U/L)	18.00 (13.02–25.00)	18.00 (13.49–25.00)	1.7
TG (mg/dL)	113.37 (81.49–161.20)	115.15 (85.92–162.09)	1.0
PS	0.22 (0.12–0.35)	0.22 (0.13–0.35)	4.7

Note: Values are n (%) or mean ±standard deviation or median (quartile).

Abbreviations: SD, Standardized differences; ALP, Alkaline phosphatase; BMI, Body mass index; TC, Total cholesterol; GGT, γ-glutamyl transpeptidase; TG, Triglyceride; ALB, albumin; HDL-C, High-density lipoprotein cholesterol; ALT, Alanine aminotransferase; GLB, globulin; AST, Aspartate aminotransferase; LDL-C, Low-density lipid cholesterol; BUN, Serum urea nitrogen; Scr, Serum creatinine; UA, uric acid; FPG, Fasting plasma glucose; PS, propensity score.

Figure 4 Distribution of propensity score in hypertension and non-hypertension groups after matching. The probability density functions of the propensity score for hypertensive and non-hypertensive participants after matching.

Figure 5 Kaplan–Meier event-free survival curve based on hypertension in the original cohort. Kaplan–Meier analysis of the probability of NAFLD-free survival based on hypertension (HBP) and non-hypertension (NHBP) in the original cohort (log-rank, P < 0.0001).

Table 4 Associations Between Hypertension and NAFLD in the Crude Analysis, Multivariable Analysis, and Propensity-Score Analyses

Analysis	NAFLD	P-value
No.of events/ no. of patients at risk (%)
Hypertension	604/2427 (24.9)
No-hypertension
Crude analysis — hazard ratio (95% CI)	1717/13,726 (12.5)	<0.001
Multivariable analysis — hazard ratio (95% CI) *	2.05 (1.87, 2.25)	0.005
Propensity-score analyses — hazard ratio (95% CI)	1.15 (1.04, 1.28)
With matching^‡	1.09 (0.97, 1.23)	0.145
Adjusted for propensity score^§	1.19 (1.08, 1.32)	0.001
With IPTW^†	1.59 (1.43, 1.77)	<0.001
With SMR weighting^&	2.24 (2.05, 2.44)	<0.001

Notes: The above models adjusted for Sex, Age, ALP, GGT, ALT, ALB, AST, GLB, BUN, Scr, UA, GLU, TC, TG, HDL, LDL, and BMI. *Shown is the hazard ratio from the multivariable Cox proportional-hazards model. The analysis included all 16,153 participants. ^‡Shown is the hazard ratio from a multivariable Cox proportional-hazards model in the participants with the propensity score matching. The analysis included 4632 participants (2316 with hypertension and 2316 were not). ^§Shown is the hazard ratio from a multivariable Cox proportional-hazards model, with additional adjustment for the propensity score. The analysis included all the participants. ^†Shown is the primary analysis with a hazard ratio from the multivariable Cox proportional-hazards model with IPTW. The analysis included all the participants. ^&Shown is the primary analysis with a hazard ratio from the multivariable Cox proportional-hazards model with SMR weighting. The analysis included all the participants.

Abbreviations: HR, hazard ratio; CI, confidence interval; IPTW, inverse-probability of treatment weighted; SMR, standardized mortality ratio.

Table 5 Associations Between Hypertension and NAFLD in the Crude Analysis, Multivariable Analysis, and Propensity-Score Analyses (Restriction to Participants Whose Propensity Score is Above 0.08)

Analysis	NAFLD	P-value
No. of events/ no. of patients at risk (%)
Hypertension	580/2205 (26.3)
No-hypertension
Crude analysis — hazard ratio (95% CI)	1522/7534 (20.2)	<0.001
Multivariable analysis — hazard ratio (95% CI)*	1.34 (1.22, 1.47)	0.008
Propensity-score analyses — hazard ratio (95% CI)	1.15 (1.04, 1.27)
With matching^‡	1.04 (0.92, 1.17)	0.548
Adjusted for propensity score^§	1.11 (1.00, 1.24)	0.041
With IPTW^†	1.38 (1.25, 1.52)	<0.001
With SMR weighting^&	1.80 (1.64, 1.97)	<0.001

Notes: The above models adjusted for Sex, Age, ALP, GGT, ALT, ALB, AST, GLB, BUN, Scr, UA, GLU, TC, TG, HDL, LDL, and BMI. *Shown is the hazard ratio from the multivariable Cox proportional-hazards model. The analysis included all 9739 participants. ^‡Shown is the hazard ratio from a multivariable Cox proportional-hazards model in the participants with the propensity score matching. The analysis included 4130 participants (2065 with hypertension and 2065 were not). ^§Shown is the hazard ratio from a multivariable Cox proportional-hazards model, with additional adjustment for the propensity score. The analysis included all the participants. ^†Shown is the primary analysis with a hazard ratio from the multivariable Cox proportional-hazards model with IPTW. The analysis included all the participants. ^&Shown is the primary analysis with a hazard ratio from the multivariable Cox proportional-hazards model with SMR weighting. The analysis included all the participants.

Abbreviations: HR, hazard ratio; CI, confidence interval; IPTW, inverse-probability of treatment weighted; SMR, standardized mortality ratio.

Table 6 Associations Between Hypertension and NAFLD in Different Subgroups

Analysis	NAFLD
Male
Crude analysis — hazard ratio (95% CI)	1.911 (1.689, 2.163) <0.001
Multivariable analysis — hazard ratio (95% CI) *	1.177 (1.032, 1.343) 0.0153
Propensity-score analyses — hazard ratio (95% CI)
With matching^‡	1.014 (0.822, 1.250) 0.8963
Adjusted for propensity score^§	1.186 (1.035, 1.359) 0.0139
With IPTW^†	1.822 (1.580, 2.101) <0.001
With SMR weighting^&	3.244 (2.878, 3.656) <0.001
Female
Crude analysis — hazard ratio (95% CI)	2.207 (1.917, 2.539) <0.001
Multivariable analysis — hazard ratio (95% CI)*	1.141 (0.978, 1.332) 0.0939
Propensity-score analyses — hazard ratio (95% CI)
With matching^‡	1.223 (0.902, 1.659) 0.1955
Adjusted for propensity score^§	1.171 (0.992, 1.383) 0.0629
With IPTW^†	1.495 (1.274, 1.754) <0.001
With SMR weighting^&	1.525 (1.339, 1.737) <0.001
Age<50 years
Crude analysis — hazard ratio (95% CI)	2.277 (2.209, 2.555) <0.001
Multivariable analysis — hazard ratio (95% CI)*	1.184 (1.046, 1.340) 0.0077
Propensity-score analyses — hazard ratio (95% CI)
With matching^‡	1.112 (0.911, 1.357) 0.2978
Adjusted for propensity score^§	1.275 (1.118, 1.453) <0.001
With IPTW^†	1.576 (1.376, 1.805) <0.001
With SMR weighting^&	1.688 (1.516, 1.879) <0.001
Age≥50 years
Crude analysis — hazard ratio (95% CI)	1.629 (1.392, 1.907) <0.001
Multivariable analysis — hazard ratio (95% CI)*	1.136 (0.961, 1.342) 0.1354
Propensity-score analyses — hazard ratio (95% CI)
With matching^‡	1.049 (0.771, 1.427) 0.7611
Adjusted for propensity score^§	1.057(0.887, 1.258) 0.536
With IPTW^†	1.997 (1.678, 2.376) <0.001
With SMR weighting^&	3.676 (3.144, 4.298) <0.001
BMI<24 kg/m^2
Crude analysis — hazard ratio (95% CI)	2.072 (1.844, 2.329) <0.001
Multivariable analysis — hazard ratio (95% CI) *	1.463 (1.288, 1.661) <0.001
Propensity-score analyses — hazard ratio (95% CI)
With matching^‡	1.080 (0.905, 1.287) 0.3934
Adjusted for propensity score^§	1.467 (1.287, 1.67) <0.001
With IPTW^†	2.014 (1.780, 2.278) <0.001
With SMR weighting^&	2.372 (2.131, 2.640) <0.001
BMI≥24 kg/m^2
Crude analysis — hazard ratio (95% CI)	1.138 (0.975, 1.327) 0.1013
Multivariable analysis — hazard ratio (95% CI)*	1.071 (0.907, 1.263) 0.4188
Propensity-score analyses — hazard ratio (95% CI)
With matching^‡	0.791 (0.525, 1.192) 0.2631
Adjusted for propensity score^§	1.069 (0.905, 1.263) 0.4293
With IPTW^†	1.056 (0.953, 1.170) 0.2951
With SMR weighting^&	1.132 (0.933, 1.373) 0.2083

Notes: Note 1: Above models were adjusted for Sex, Age, ALP, GGT, ALT, ALB, AST, GLB, BUN, Scr, UA, GLU, TC, TG, HDL, LDL, and BMI. Note 2: In each case, the model is not adjusted for the stratification variable. *Shown is the hazard ratio from the multivariable Cox proportional-hazards model in the original participants. ^‡Shown is the hazard ratio from a multivariable Cox proportional-hazards model in the participants with the propensity score matching. ^§Shown is the hazard ratio from a multivariable Cox proportional-hazards model, with additional adjustment for the propensity score. Analysis in the original participants. ^†Shown is the primary analysis with a hazard ratio from the multivariable Cox proportional-hazards model with inverse probability weighting according to the propensity score. Analysis in the original participants. ^&Shown is the primary analysis with a hazard ratio from the multivariable Cox proportional-hazards model with standardized mortality ratio weighting according to the propensity score. Analysis in the original participants.

Abbreviations: HR, hazard ratio; CI, confidence interval; IPTW, inverse-probability of treatment weighted; SMR, standardized mortality ratio.

Table 7 Baseline Characteristics After Genetic Matching

Characteristic	Non-Hypertension(n=16,153)	Hypertension(n=16,153)	SD (100%)
AGE (year)	43.00 ± 14.68	43.53 ± 13.79	3.7
SEX			0.7
Female	7627 (47.22%)	7574 (46.89%)
Male	8526 (52.78%)	8579 (53.11%)
BMI (kg/m^2)	21.30 ± 2.02	22.11 ± 1.86	41.6
ALP (U/L)	71.86 ± 21.91	72.64 ± 19.40	3.7
ALB (g/L)	44.45 ± 2.63	44.38 ± 2.42	3.0
AST (U/L)	22.84 ± 8.65	22.97 ± 7.43	1.5
GLB (g/L)	29.44 ± 3.80	29.58 ± 3.51	3.8
BUN (mg/dL)	12.78 ± 3.73	12.91 ± 3.63	3.5
Scr (mg/dL)	0.88 ± 0.26	0.89 ± 0.28	2.9
UA (mg/dL)	4.69 ± 1.42	4.71 ± 1.36	1.6
FPG (mg/dL)	92.15 ± 13.44	94.25 ± 12.91	15.9
TC (mg/dL)	178.55 ± 28.27	179.43 ± 26.22	3.2
HDL-c (mg/dL)	56.45 ± 13.80	55.90 ± 12.97	4.1
LDL-c (mg/dL)	87.59 ± 17.86	88.41 ± 16.44	4.8
GGT (U/L)	22.00 (16.00–31.36)	23.00 (18.00–32.00)	3.5
ALT (U/L)	17.00 (12.07–23.30)	16.00 (13.00–22.00)	4.9
TG (mg/dL)	95.66 (70.86–132.86)	100.97 (77.95–131.98)	3.1

Note: Values are n (%) or mean ±standard deviation or median (quartile).

Abbreviations: SD, Standardized differences; ALP, Alkaline phosphatase; BMI, Body mass index; TC, Total cholesterol; GGT, γ-glutamyl transpeptidase; TG, Triglyceride; ALB, albumin; HDL-C, High-density lipoprotein cholesterol; ALT, Alanine aminotransferase; GLB, globulin; AST, Aspartate aminotransferase; LDL-C, Low-density lipid cholesterol; BUN, Serum urea nitrogen; Scr, Serum creatinine; UA, uric acid; FPG, Fasting plasma glucose.

Table 8 Associations Between Hypertension and NAFLD After Genetic Matching

Exposure	Crude Model (HR, 95% CI, P)	Model I (HR, 95% CI, P)	Model II (HR, 95% CI, P)
Non-hypertension	Ref.	Ref.	Ref.
Hypertension	1.33 (1.26, 1.40) <0.001	1.32 (1.25, 1.40) <0.001	1.06 (1.01, 1.13) 0.030

Notes: Crude model: we did not adjust other covariates. Model I: we adjust age, gender. Model II: we adjust age, gender, BMI, FPG, TC, TG, HDL-c, LDL-c, ALT, AST, BUN, Scr, GGT, GLB, ALB, UA, ALP.

Abbreviations: HR, Hazard ratios; CI, Confidence interval; Ref, Reference.