Association Between Psoriasis and Nonalcoholic Fatty Liver Disease Among Outpatient US Adults

This cross-sectional study uses data from the National Health and Nutrition Examination Survey to examine the association between psoriasis and nonalcoholic fatty liver disease among outpatient US adults.

Key Points

Question

Is psoriasis associated with a higher risk of nonalcoholic fatty liver disease (NAFLD) in the outpatient US population?

Findings

In this population-based cross-sectional study that included 5672 US adults 20 to 59 years of age, psoriasis was associated with NAFLD in adjusted models.

Meaning

Because some antipsoriatic agents are potentially hepatotoxic, the association between psoriasis and NAFLD in US adults found in this study may be worth considering in psoriasis management.

Abstract

Importance

Recent studies have shown an association between psoriasis and nonalcoholic fatty liver disease (NAFLD) in US inpatients, but the association is still unclear in the outpatient US population.

Objective

To assess whether psoriasis is associated with NAFLD in outpatient US adults.

Design, Setting, and Participants

This population-based cross-sectional study used data on US adults aged 20 to 59 years from the National Health and Nutrition Examination Survey (NHANES) 2003-2006 and 2009-2014 cycles. Data were analyzed from June to September 2021.

Exposures

Self-reported psoriasis.

Main Outcomes and Measures

The main outcome was NAFLD, defined as a US fatty liver index score greater than 30. Sampling weights were calculated according to NHANES guidelines.

Results

Among 5672 adults included in this study (mean age, 38.9 years [95% CI, 38.4-39.3 years]; 2999 [51.1%] female), 148 (3.0%) had psoriasis and 5524 (97.0%) did not have psoriasis. A total of 1558 participants (26.8%) were classified as having NAFLD. Compared with participants without psoriasis, those with psoriasis had a higher prevalence of NAFLD (32.7% [52] vs 26.6% [1506]). In a multivariable logistic regression model adjusted for age, sex, race and ethnicity, educational level, family income, marital status, NHANES cycles, diabetes, metabolic syndrome, and smoking and alcohol drinking status, psoriasis was associated with NAFLD (odds ratio [OR], 1.67; 95% CI, 1.03-2.70). In subgroup analyses, psoriasis was associated with NAFLD among men (OR, 2.16; 95% CI, 1.10-4.24), among those aged 20 to 39 years (OR, 2.48; 95% CI, 1.09-5.67), and among those without diabetes (1.70; 95% CI, 1.05-2.76). An association between psoriasis and NAFLD was found in sensitivity analyses that excluded potential hepatotoxic medication use (OR, 1.72; 95% CI, 1.01-2.95) or non-Hispanic Black participants (OR, 1.76; 95% CI, 1.07-2.87), redefined NAFLD based on the hepatic steatosis index score (OR, 1.59; 95% CI, 1.01-2.50), and used inverse probability of treatment weighting (OR, 1.43; 95% CI, 1.09-1.86).

Conclusions and Relevance

In this cross-sectional study, psoriasis was associated with NAFLD in the outpatient US adult population in adjusted models. This association may be important to consider in the context of clinicians prescribing potentially hepatotoxic medication for psoriasis management.

Introduction

Psoriasis is a chronic, relapsing dermatosis that affects approximately 3% of adults in the US,¹ and its prevalence has been gradually increasing.^2,3 Recently, psoriasis has been regarded as a systemic inflammatory disease with multiple organ systems involved that manifests as cardiovascular diseases, diabetes, metabolic syndrome, inflammatory bowel disease, and nonalcoholic fatty liver disease (NAFLD).^4,5

NAFLD is defined as the excessive accumulation of lipids in liver tissues without other clear causes (such as considerable alcohol consumption).⁶ The prevalence of NAFLD has been increasing in the US,^7,8,9,10,11 and the number of individuals with NAFLD requiring liver transplant has continued to increase.¹² A model of NAFLD in 2018 estimated that approximately 800 000 liver-related deaths will occur from 2015 to 2030,¹³ resulting in a large health care burden.^14,15

Previous studies reported a high prevalence of NAFLD among patients with psoriasis, ranging from 44.3% to 65.6%.^{16,17,18,19,20} The association between psoriasis and NAFLD appears to be biologically plausible. In addition to being common comorbidities of psoriasis, metabolic syndrome and its components are associated with development of NAFLD.²¹ Furthermore, some cytokines (interleukin [IL]–23, IL-17, and tumor necrosis factor [TNF]–α) have been known to be mechanisms in the development of psoriasis and to contribute to insulin resistance, which is associated with NAFLD development.

To our knowledge, 2 studies have evaluated the association between psoriasis and NAFLD in US inpatients, but the association in the outpatient US population remains unclear.^22,23 In this cross-sectional study, we aimed to explore the association between psoriasis and NAFLD using data from the National Health and Nutrition Examination Survey (NHANES).

Methods

Data Sources

NHANES is a nationally representative survey conducted by the National Center for Health Statistics (NCHS)²⁴ in which stratified, multistage probability cluster sampling has been used to assess the health or nutritional status of the noninstitutionalized US population. Information on psoriasis was only provided in the NHANES 2003-2006 cycles for adults 20 to 59 years of age and in the 2009-2014 cycles for individuals 16 to 80 years of age. In this cross-sectional study, deidentified data for participants 20 to 59 years of age were extracted from the 2003-2006 and 2009-2014 NHANES cycles. Data were analyzed from June to September 2021. The NCHS research ethics review board approved the NHANES study protocol, and participants provided written informed consent at enrollment. The Shantou University Medical College institutional review board determined the study to be exempt because it used publicly available deidentified data, and informed consent was waived. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.

Study Design and Population

Our analyses were based on data collected from participants during five 2-year NHANES cycles (2003-2006 and 2009-2014). To restrict our analysis to participants with NAFLD, the following exclusion criteria were used²⁵: missing data for alcohol consumption or presence of considerable alcohol consumption (>21 drinks per week for men and >14 drinks per week for women²⁶); missing data for viral hepatitis or presence of viral hepatitis (positive for serum hepatitis B surface antigen or for serum hepatitis C antibody); or missing data needed to calculate the US fatty liver index (USFLI) score, including race and ethnicity, γ-glutamyltransferase level, waist circumference, fasting glucose level, and insulin level.²⁷ The flowchart for participant enrollment is presented in Figure 1.

Figure 1. Flow Diagram of the Screening and Enrollment of Study Participants.

NAFLD indicates nonalcoholic fatty liver disease.

Diagnosis of Psoriasis and NAFLD

Psoriasis was defined as an affirmative response to the question, “Have you ever been told by a health care provider that you had psoriasis?” The USFLI is a noninvasive method for assessing NAFLD and has been validated and used in previous studies.^11,28 Scores range from 0 to 100. As suggested by Ruhl and Everhart,²⁷ the presence of NAFLD was defined in this study as a USFLI score of greater than 30, with an area under the receiver operating characteristic curve (AUROC) of 0.80 (sensitivity, 62%; specificity, 88%).

Covariates

Based on the literature, the following covariates were included: age, sex, race and ethnicity, educational level, family income, marital status, diabetes, metabolic syndrome, smoking status, and alcohol drinking status.^19,22,29 In NHANES, information on self-reported race and ethnicity were derived from responses to survey questions on race and Hispanic ethnicity. As used by NHANES, we categorized the participants into the following 6 races and ethnicities: Mexican American, non-Hispanic Asian, non-Hispanic Black, non-Hispanic White, other Hispanic, and Other (including multiracial). Educational level was divided into 3 levels (high school or less, some college, and college graduate or higher). As used by US government departments to report NHANES dietary and health data,³⁰ we categorized family income into the following 3 levels based on the family poverty income ratio: low income (≤1.3), medium income (>1.3 to 3.5), and high income (>3.5). Marital status was categorized into the following 4 groups: married, never married, living with a partner, and other (including widowed, divorced, or separated). Diabetes was defined as using antidiabetic medication or a fasting glucose level of 126 mg/dL or more (to convert to mmol/L, multiply by 0.0555). Metabolic syndrome was defined based on the Adult Treatment Panel III criteria in 2005.³¹ Smoking status was categorized into the following 3 groups: never smoked (or smoked <100 cigarettes), former smoker (smoked at least 100 cigarettes but has quit), and current smoker. Alcohol drinking status was determined by the survey question, “In any 1 year, have you had at least 12 drinks of any type of alcoholic beverage?” Participants who answered “yes” were defined as alcohol drinkers.

Statistical Analysis

According to NHANES analytic guidelines, complex sampling design and sampling weights were considered in our analyses.³² The sampling weight was calculated using the following formula: fasting subsample 10-year mobile examination center (MEC) weight = fasting subsample 2-year MEC weight/5. The characteristics of participants are described as means (95% CIs) for continuous variables and percentage frequencies (95% CIs) for categorical variables. Continuous data were compared using t tests, and categorical data were compared by the χ² test. Because the percentage of missing data was small (missing rate varied from 0% to 6.1%) for any variable, no imputation method was used.

Odds ratios (ORs) and 95% CIs were calculated to assess the association between NAFLD and psoriasis using logistic regression models. Multicollinearity was tested using the variance inflation factor (VIF) method, with a VIF of 5 or more indicating the presence of multicollinearity. Model 1 was the fully adjusted model and included sociodemographic variables (age, sex, race and ethnicity, educational level, family income, and marital status), NHANES cycles, metabolic syndrome, diabetes, smoking status, and alcohol drinking status. Model 2 was the crude model with no covariates adjusted. Model 3 was adjusted for sociodemographic variables and NHANES cycles. Model 4 was based on model 3 and metabolic syndrome. Model 5 was based on model 4 and diabetes. To avoid overadjustment, metabolic syndrome was inputted as a single covariate in models 4 and 5. Of note, the calculation of USFLI was based on variables that were associated with obesity. To avoid overadjustment, obesity was not included as a covariate in many previous studies based on the USFLI.^33,34,35,36 This was also the case in our study. However, we conducted a subgroup analysis on obesity (defined as body mass index ≥30 [calculated as weight in kilograms divided by height in meters squared]) to assess the possible influence of obesity on the association between psoriasis and NAFLD. In addition, interaction and subgroup analyses were performed according to age group, sex, diabetes, and metabolic syndrome using logistic regression models.

To assess the robustness of our findings, several sensitivity analyses were performed. First, considering the potential association of medication with NAFLD, we conducted a sensitivity analysis that excluded participants who were treated with methotrexate,³⁷ acitretin,^38,39 pioglitazone,^40,41 liraglutide,⁴² semaglutide,⁴³ atorvastatin,⁴⁴ or aspirin.^45,46 Second, because of the lower prevalence of NAFLD and the lower diagnostic accuracy of the USFLI among Black individuals,^27,47,48 we conducted a sensitivity analysis excluding non-Hispanic Black participants. Third, although the AUROC for USFLI is 0.80, which represents a moderate accuracy, misclassification was still inevitable. To verify our results, we conducted a sensitivity analysis based on the hepatic steatosis index (HSI), a previously reported diagnostic algorithm for NAFLD with an AUROC of 0.812 (95% CI, 0.801-0.824). Specifically, NAFLD was defined as an HSI score greater than 36 in this model, with a specificity of 92.4% and a positive likelihood ratio of 6.069.⁴⁹ Fourth, we performed an analysis using inverse probability of treatment weighting (IPTW) to address potential confounding. The following variables were used to generate the model: age, sex, race and ethnicity, educational level, family income, marital status, NHANES cycles, diabetes, and metabolic syndrome.

All statistical analyses were performed with R, version 4.0.5 (R Project for Statistical Computing) using the survey package, version 4.1-1 and with Free Software Foundation statistics software, version 1.3. In all tests, P < .05 (2-sided) was considered to indicate statistical significance.

Results

Characteristics of the Participants

Of 18 391 participants 20 to 59 years of age (18 368 had data on psoriasis, and 17 744 underwent laboratory examination), 12 719 were excluded owing to missing or implausible data regarding psoriasis and NAFLD, leaving 5672 participants for analysis. A total of 148 participants (3.0%; 95% CI, 2.4%-3.7%) had psoriasis, and 5524 (97.0%; 95% CI, 96.3%-97.6%) did not have psoriasis. The baseline characteristics of participants excluded and included are given in the eTable in the Supplement. Based on the weighted analyses, the mean age of the 5672 participants was 38.9 years (95% CI, 38.4-39.3 years) and 2999 were female (51.1%; 95% CI, 49.9%-52.4%) (Table 1). Based on the USFLI score, 1558 participants (26.8%; 95% CI, 24.9%-28.7%) were classified as having NAFLD. The prevalence of NAFLD was higher among participants with psoriasis (52 [32.7%; 95% CI, 23.3%-42.1%]) than among participants without psoriasis (1506 [26.6%; 95% CI, 24.7%-28.5%]). In addition, participants with psoriasis were older (mean age, 41.4 years; 95% CI, 39.1-43.8 years) than those without psoriasis (mean age, 38.8 years; 95% CI, 38.4-39.2 years) and were more likely to be non-Hispanic White individuals (91 [82.3%; 95% CI, 75.6%-88.9%] vs 2396 [67.3%; 95% CI, 64.2%-70.5%]) and to have an educational level of college or higher (53 [38.3%; 95% CI, 29.8%-46.9%] vs 1430 [31.2%; 95% CI, 28.7%-33.7%]).

Table 1. Characteristics of Participants in the NHANES 2003-2006 and 2009-2014 Cycles.

Characteristic	Participantsa			P value
	Total (N = 5672)	Without psoriasis (n = 5524)	With psoriasis (n = 148)
USFLI score, mean (95% CI)b	21.9 (21.0-22.8)	21.8 (20.9-22.7)	25.4 (21.0-29.7)	.10
Age, mean (95% CI), y	38.9 (38.4-39.3)	38.8 (38.4-39.2)	41.4 (39.1-43.8)	.03
Sex
Female	2999 (51.1) [49.9-52.4]	2918 (51.0) [49.8-52.3]	81 (55.5) [45.1-65.8]	.30
Male	2673 (48.9) [47.6-50.1]	2606 (49.0) [47.7-50.2]	67 (44.5) [34.2-54.9]
Race and ethnicityc
Mexican American	1018 (9.1) [7.5-10.8]	1005 (9.3) [7.7-11.0]	13 (3.3) [1.3-5.2]	<.001
Non-Hispanic
Asian	331 (2.1) [1.5-2.6]	318 (2.1) [1.5-2.6]	13 (2.2) [0.5-3.9]
Black	1125 (11.5) [9.9-13.0]	1107 (11.7) [10.1-13.3]	18 (6.0) [2.7-9.4]
White	2487 (67.8) [64.7-70.9]	2396 (67.3) [64.2-70.5]	91 (82.3) [75.6-88.9]
Other Hispanic	473 (5.0) [3.8-6.3]	464 (5.1) [3.8-6.4]	9 (2.8) [0.3-5.3]
Otherd	238 (4.4) [3.7-5.2]	234 (4.5) [3.8-5.2]	4 (3.4) [-0.4-7.3]
Family incomee
Low	1626 (20.7) [18.9-22.6]	1579 (20.6) [18.8-22.5]	47 (23.2) [16.0-30.5]	.70
Medium	1872 (34.6) [32.5-36.8]	1834 (34.8) [32.6-36.9]	38 (31.2) [22.3-40.1]
High	1828 (44.6) [41.9-47.4]	1775 (44.6) [41.9-47.3]	53 (45.5) [34.8-56.2]
Educational level
High school or less	2360 (35.1) [32.6-37.5]	2314 (35.4) [33.0-37.8]	46 (25.4) [17.0-33.7]	.04
Some college	1827 (33.5) [31.8-35.3]	1778 (33.4) [31.6-35.3]	49 (36.3) [29.1-43.5]
College graduate or higher	1483 (31.4) [29.0-33.8]	1430 (31.2) [28.7-33.7]	53 (38.3) [29.8-46.9]
Marital status
Married	2935 (55.6) [53.5-57.7]	2857 (55.5) [53.4-57.7]	78 (56.9) [49.3-64.6]	.90
Never married	764 (12.8) [11.6-14.0]	742 (12.9) [11.7-14.1]	22 (11.0) [5.0-16.9]
Living with partner	1383 (22.4) [20.7-24.2]	1352 (22.4) [20.7-24.2]	31 (22.7) [14.5-31.0]
Otherf	587 (9.1) [7.8-10.5]	570 (9.1) [7.8-10.5]	17 (9.3) [3.5-15.1]
NAFLD defined by USFLI scoreb
No	4114 (73.2) [71.3-75.1]	4018 (73.4) [71.5-75.3]	96 (67.3) [57.9-76.7]	.17
Yes	1558 (26.8) [24.9-28.7]	1506 (26.6) [24.7-28.5]	52 (32.7) [23.3-42.1]
Diabetes
No	5243 (93.8) [93.0-94.7]	5108 (93.7) [92.8-94.6]	135 (96.5) [94.3-98.7]	.06
Yes	429 (6.2) [5.3-7.0]	416 (6.3) [5.4-7.2]	13 (3.5) [1.3-5.7]
Metabolic syndrome
No	4114 (73.2) [71.3-75.1]	4018 (73.4) [71.5-75.3]	96 (67.3) [57.9-76.7]	.19
Yes	1558 (26.8) [24.9-28.7]	1506 (26.6) [24.7-28.5]	52 (32.7) [23.3-42.1]
Alcohol drinkerg
No	1617 (22.8) [20.6-25.0]	1578 (22.8) [20.6-25.1]	39 (21.9) [12.2-31.5]	.80
Yes	4055 (77.2) [75.0-79.4]	3946 (77.2) [74.9-79.4]	109 (78.1) [68.5-87.8]
Smoking status
Never	3501 (59.6) [57.7-61.6]	3419 (59.9) [57.9-61.9]	82 (52.0) [41.8-62.1]	.20
Former	962 (18.5) [17.1-19.9]	934 (18.4) [16.9-19.8]	28 (22.9) [15.3-30.6]
Current	1208 (21.8) [20.2-23.5]	1170 (21.7) [20.1-23.4]	38 (25.1) [17.1-33.2]

Abbreviations: NAFLD, nonalcoholic fatty liver disease; NHANES, National Health and Nutrition Examination Survey; USFLI, US fatty liver index.

^a Data are presented as unweighted number (weighted percentage) [95% CI] unless otherwise indicated.

^b Scores range from 0 to 100, with a USFLI score greater than 30 used to indicate NAFLD in this study.

^c Race and ethnicity were self-reported.

^d Included multiracial participants. NHANES did not provide a detailed list of all races and ethnicities included in this category.

^e Categorized into the following 3 levels based on the family poverty income ratio: low income (≤1.3), medium income (>1.3 to 3.5), and high income (>3.5).

^f Included widowed, divorced, or separated individuals.

^g Determined by answering the following question: “In any 1 year, have you had at least 12 drinks of any type of alcoholic beverage?”

Multivariable Regression Analyses

The results from sample-weighted multivariable regression analyses are presented in Table 2. An association between psoriasis and NAFLD based on the USFLI was revealed in model 1 (the fully adjusted model) (OR, 1.67; 95% CI, 1.03-2.70) and model 5 (OR, 1.70; 95% CI, 1.06-2.72).

Table 2. Association of Psoriasis With NAFLD Among Participants in the NHANES 2003-2006 and 2009-2014 Cycles.

Model	Unweighted participants/total participants, No.		NAFLD defined by USFLIa
	Without psoriasis	With psoriasis	OR (95% CI)	P value
Model 1b	1351/4900	48/128	1.67 (1.03-2.70)	.04
Model 2c	1506/5524	52/148	1.34 (0.88-2.05)	.17
Model 3d	1407/5188	48/138	1.39 (0.88-2.20)	.15
Model 4e	1351/4900	48/128	1.58 (0.97-2.59)	.06
Model 5f	1351/4900	48/128	1.70 (1.06-2.72)	.03

Abbreviations: NAFLD, nonalcoholic fatty liver disease; NHANES, National Health and Nutrition Examination Survey; OR, odds ratio; USFLI, US fatty liver index.

^a Scores range from 0 to 100, with a USFLI score greater than 30 used to indicate NAFLD in this study.

^b Adjusted for sociodemographic variables (age, sex, race and ethnicity, educational level, family income, and marital status), NHANES cycles, metabolic syndrome, diabetes, and smoking and alcohol drinking status.

^c Crude model.

^d Adjusted for sociodemographic variables and NHANES cycles.

^e Adjusted for sociodemographic variables, NHANES cycles, and metabolic syndrome.

^f Adjusted for sociodemographic variables, NHANES cycles, metabolic syndrome, and diabetes.

Subgroup Analyses

The results of subgroup analyses are presented in Figure 2. Psoriasis was associated with NAFLD among participants aged 20 to 39 years (OR, 2.48; 95% CI, 1.09-5.67), men (OR, 2.16; 95% CI, 1.10-4.24), and those without diabetes (OR, 1.70; 95% CI, 1.05-2.76). There was no association among women, participants aged 40 to 59 years, or those with diabetes or in the obesity and metabolic syndrome subgroups.

Figure 2. Association Between Psoriasis and Nonalcoholic Fatty Liver Disease (NAFLD).

Each stratification was adjusted for age, sex, race and ethnicity, educational level, family income, marital status, National Health and Nutrition Examination Survey cycles, metabolic syndrome, and diabetes except the stratification factor itself. Squares indicate odds ratios (ORs), with horizontal lines indicating 95% CIs. Diamonds indicate overall ORs, with outer points of the diamonds indicating 95% CIs.

Sensitivity Analyses

Table 3 summarizes the results of sensitivity analyses. After excluding participants treated with methotrexate, acitretin, pioglitazone, liraglutide, semaglutide, atorvastatin, and aspirin (unweighted n = 4739), psoriasis was associated with NAFLD (OR, 1.72; 95% CI, 1.01-2.95). After excluding non-Hispanic Black participants (unweighted n = 4036), the OR was 1.76 (95% CI, 1.07-2.87). In addition, when NAFLD was defined by HSI, after adjusting for sociodemographic characteristics, NHANES cycles, and metabolic syndrome (unweighted n = 5075), the association between psoriasis and NAFLD remained (OR, 1.59; 95% CI, 1.01-2.50). After IPTW (unweighted n = 5028), the OR was 1.43 (95% CI, 1.09-1.86).

Table 3. Sensitivity Analyses.

Analysis	Unweighted participants/total participants, No.	Adjusted OR (95% CI)a	P value
Excluding participants taking potential hepatotoxic medicationsb
Without psoriasis	1193/4621	1 [Reference]	.046
With psoriasis	41/118	1.72 (1.01-2.95)
Excluding non-Hispanic Black participants
Without psoriasis	1183/3924	1 [Reference]	.03
With psoriasis	44/112	1.76 (1.07-2.87)
NAFLD defined by HSI scorec
Without psoriasis	2769/4948	1 [Reference]	.04
With psoriasis	89/127	1.59 (1.01-2.50)
Inverse probability treatment weighted analysesd
Without psoriasis	1351/4900	1 [Reference]	.01
With psoriasis	48/128	1.43 (1.09-1.86)

Abbreviations: HSI, hepatic steatosis index; NAFLD, nonalcoholic fatty liver disease; NHANES, National Health and Nutrition Examination Survey; OR, odds ratio.

^a Adjusted for age, sex, race and ethnicity, NHANES cycles, educational level, family income, marital status, metabolic syndrome, and diabetes.

^b Included the following medications that may affect liver function: methotrexate, acitretin, pioglitazone, liraglutide, semaglutide, atorvastatin, and aspirin.

^c NAFLD was defined as an HSI score greater than 36. This model was adjusted for age, sex, race and ethnicity, NHANES cycles, educational level, family income, marital status, and metabolic syndrome.

^d The following variables were used to generate the model: age, sex, race and ethnicity, NHANES cycles, educational level, family income, marital status, diabetes, and metabolic syndrome.

Discussion

In this nationally representative cross-sectional study, psoriasis was found to be associated with NAFLD in the outpatient US population in adjusted models. After excluding potential hepatotoxic medication use, excluding non-Hispanic Black participants (because of the lower prevalence of NAFLD and concomitant reduced diagnostic accuracy of NAFLD in this population), redefining NAFLD based on the HSI score, and conducting a sensitivity analysis with IPTW, the association remained. Subgroup analysis indicated that the OR for the association between psoriasis and NAFLD was higher among patients with psoriasis and diabetes.

A previous meta-analysis showed that psoriasis was positively associated with NAFLD.⁵⁰ However, few studies have evaluated the US population. In our study, participants with psoriasis had a higher risk of NAFLD (OR, 1.67; 95% CI, 1.03-2.70) after adjusting for age, sex, race and ethnicity, educational level, family income, marital status, NHANES cycles, diabetes, metabolic syndrome, and smoking and alcohol drinking status. When NAFLD was defined by USFLI scores, its prevalence among participants with psoriasis was 42.4%. This is comparable with previous findings in a cohort study in Houston, Texas, which found the prevalence of NAFLD (diagnosed based on ultrasonography findings) among patients with psoriasis to be 46.6%.⁵¹ In a meta-analysis conducted by Candia et al⁵⁰ in 2015, participants with psoriasis had twice the risk of NAFLD compared with those without psoriasis (4 case-control studies; OR, 2.16, [95% CI, 1.19-3.93]; I² = 84%). In addition, a pooled meta-analysis of adjusted studies by Phan et al⁵² in 2019 supported a positive association between psoriasis and NAFLD (OR, 1.95; 95% CI, 1.35-2.83; I² = 91%).

To our knowledge, 2 studies^22,23 have investigated the association between psoriasis and NAFLD in the US. One study, conducted in 2020 using data from the Nationwide Inpatient Sample, reported an OR for the association between psoriasis and NAFLD of 2.22 (95% CI, 2.1-2.3) after adjusting for age, sex, race and ethnicity, income, insurance status, obesity, smoking, and alcohol use.²³ The other study found that psoriasis was positively associated with NAFLD (OR, 1.92; 95% CI, 1.83-2.01) after controlling for age, sex, income quartile, race and ethnicity, tobacco use, alcoholism, metabolic syndrome, HIV infection or AIDS, and type 2 diabetes.²² These results are consistent with our findings. However, only inpatients were recruited in those 2 studies, which limits their external validity. Our study focused on the outpatient US population and may complement the previous work. Because we accounted for the NHANES design to obtain US nationally representative estimates, our results should be generalizable to the outpatient population of US adults.

Previous studies did not investigate the association and interaction between psoriasis and NAFLD in different subgroups. Our subgroup analysis showed that patients with psoriasis and diabetes had a higher risk of NAFLD. Owing to the limited sample size of participants with psoriasis and diabetes in our study, this result should be interpreted with caution, and more well-designed prospective studies in this field are warranted.

The association between psoriasis and NAFLD may be bidirectional owing to their similar biological mechanisms. The cytokine TNF-α is involved in psoriasis.^53,54,55,56 Elevated TNF-α levels can also exacerbate insulin resistance among patients with psoriasis, which contributes to the generation and release of free fatty acids and enhances the deposition of free fatty acids in the liver.²¹ Interleukin-23 is involved in the pathogenesis of psoriasis.^{57,58,59,60,61,62,63,64} A recent study in mice suggested that the imbalance of regulatory T cell and helper T cell 17 and their associated cytokines (IL-17 and IL-23) are involved in NAFLD.⁶⁴ Unhealthy lifestyles may be associated with some common comorbidities, including diabetes and metabolic syndrome, which may contribute to the pathogenesis of both psoriasis and NAFLD.²¹

Limitations

The present study has several limitations. First, as a result of the cross-sectional nature of the study, we could not determine the temporal association between psoriasis and NAFLD. Additional well-designed cohort studies are needed. Second, residual confounders, such as obesity, could have been present. We adjusted for some potential confounders and used IPTW to address potential confounding. Because diagnosis of NAFLD was based on the USFLI in our study, we did not treat obesity as a covariate, and thus it was not used for adjustment in the regression model. However, we conducted a subgroup analysis with obesity to explore the potential influence of obesity on the association between psoriasis and NAFLD; the ORs in the groups with and without obesity were similar, but no interaction was detected. We also conducted a sensitivity analysis to reduce the influence of medication use on our results. Third, liver biopsy is the gold standard for the diagnosis of NAFLD, but it is difficult to apply liver biopsy as a criterion in a population-based study. Thus, we conducted a sensitivity analysis using another NAFLD diagnostic criterion (HSI) to verify our results. Fourth, psoriasis was diagnosed based on self-reported data, which can lead to recall bias. However, previous studies indicated that self-reporting of psoriasis had high specificity.^65,66 The NHANES 2003-2004 cycle contains psoriasis data ascertained by both self-reported data and physician evaluation. A study using data from NHANES 2003-2004 cycle showed that self-reported data may underestimate the prevalence of psoriasis,⁶⁷ which may have led to lower ORs for the association between psoriasis and NAFLD in our study.

Conclusions

The findings of this cross-sectional study suggest that psoriasis is positively associated with NAFLD in the outpatient US adult population. Because some antipsoriatic agents are potentially hepatotoxic, this finding may be important for clinicians to consider for psoriasis management.

Supplement.

eTable. Characteristics of Included and Excluded Participants