Depression Is Associated With Non-alcoholic Fatty Liver Disease Among Adults in the United States

Donghee Kim; Eric R. Yoo; Andrew A. Li; Sean P. Tighe; George Cholankeril; Stephen A. Harrison; Aijaz Ahmed


Aliment Pharmacol Ther. 2019;50(5):590-598. 

In This Article


In this population-based study with a nationally representative sample of American adults, the prevalence of depression was significantly higher in subjects with NAFLD than in those without NAFLD. Depression has a significant effect on NAFLD, independent of other known risk factors, confirming a relevant clinical relationship between these two diseases entities.

In the past, several studies have demonstrated a conflicting association between depression and NAFLD.[13,14,16] Elwing et al showed that subjects with NASH had significantly increased lifetime rates of depression, although this study was limited by small sample size.[16] Weinstein et al reported that the prevalence of depression was higher in patients with NAFLD (27.2%) than in those with hepatitis B virus infection (3.7%), but this study was limited by the fact that depression was diagnosed based on self-reported history of depression and use of antidepressants; but, without any efforts to diagnose mild depression.[13] Lee et al showed that prevalence of depression was not different between subjects with and without NAFLD and that depression was not associated with NAFLD in their multivariate analysis.[14] This study, however, used elevated ALT to define NAFLD, the true prevalence of NAFLD was likely underestimated and misclassified. A US study using 567 biopsy-proven NAFLD cohort reported that depression was associated with hepatocyte ballooning (P = 0.020) and portal fibrosis (P = 0.087) after adjusting for age, sex, Race, BMI, diabetes and hypertension.[17] On the contrary, a Japanese study using 258 biopsy-proven NAFLD cohort found an association between depression and histological steatosis and NAFLD activity score.[18] In these retrospective studies, subjects underwent a liver biopsy at a university hospital; thus, there may be a selection bias embedded in these studies. We confirmed the independent association between depression and NAFLD in a large population-based sample that included various ethnicities. Our findings are therefore more generalisable to the US and Western populations than in any previously published reports.

Currently, the management of NAFLD is focused on lifestyle modifications with no approved pharmacological therapy.[36,37] Reliance on adherence to lifestyle modifications is difficult for patients as psychological features of NAFLD-related comorbidities such as, depression can lead to poor adherence to dietary counselling and exercise.[18] Moreover, depression is associated with poor compliance with medical treatment recommendations for general chronic medical illnesses.[38] Psychological stress in the setting of anxiety and depression was associated with increased risk of liver-related mortality, particularly in patients NAFLD and/or alcoholic liver disease.[39] To address the impact of depression and its effects on the management and adherence of NAFLD, we believe that it is reasonable to pro-actively screen and promptly treat concurrent depression in patients with NAFLD.

We believe there are underlying mechanisms that support an association between depression and NAFLD. Increase in overall body stress due to underlying obesity, diabetes and/or metabolic syndrome leads chronic inflammatory state resulting from the increased production of pro-inflammatory cytokines, cortisol and epinephrine with further amplification by depression and increased predisposition to the pathogenesis of NAFLD.[10,40,41] A recent study suggests that inflammation in the central and peripheral immune system may relate metabolic abnormalities to major depression.[41] Pathophysiological factors associated with depression, such as increased monoamine oxidase-A enzyme activity, might enhance cellular oxidative stress and thus adversely impact NAFLD.[17] Chronic inflammation and oxidative stress lead to NAFLD.[42] Another possible explanation is the association between NAFLD and obesity or diabetes, which are closely correlated with depression.[12] After further adjustment for diabetes and general/abdominal obesity, the association was attenuated and remained significant suggesting that the link between depression and NAFLD is partly mediated by diabetes and obesity. Emerging evidence of the involvement of insulin signalling on brain mechanisms linked to depression indicates that insulin resistance, which may be one of the main pathogenic driver for NAFLD, could develop in the brains of patients with depression.[43] When we further adjusted for insulin resistance, the association was attenuated and remained significant suggesting that the association between depression and NAFLD may be partly mediated by insulin resistance. Another plausible explanation for this association may be an unhealthy lifestyle including smoking, alcohol consumption, sedentary lifestyle and poor compliance with recommendations for primary prevention in subjects with depression.[11] In our study, depression was independently associated with the risk of NAFLD after further adjustment for smoking and lack of physical activity.

The strengths of this study are the use of high-quality data and questionnaires (PHQ-9) collected by trained personnel with a systematic protocol, the wealth of metabolic variables, and a large number of subjects who represent the general US population. Therefore, our findings are generalizable to the US population regardless of Race/ethnicity. Additionally, because patients with major depression diagnosed by a psychiatrist tend to have more severe depression than those defined as 'clinically depressed' on symptom questionnaires (PHQ-9), we utilised a broader definition of depression which included use of antidepressant medications and report of functional impairment. This study also has limitations. First, we utilised three non-invasive markers to define NAFLD in the absence of known liver disease and history of exposure to steatogenic medication, which may misclassify and under/overestimate the true prevalence of NAFLD. Also, it is important to stress that the three non-invasive markers do not have an absolute cut-off and it may unavoidable to misclassify the presence of NAFLD. However, by employing three non-invasive markers we expect the likelihood of misclassification was minimised. Sensitivity analyses using different non-invasive markers determined that our results were robust and consistent. The USFLI, FLI and HSI have been established as reliable non-invasive markers for NAFLD in the US population and other populations, and can independently predict liver-related and all-cause mortality.[23–29,44,45] However, prospective studies using imaging or histology are needed to reproduce and reconfirm these observations in future. Second, the NHANES data lack radiological or histological information, which is considered to be the gold standard for the diagnosis of NAFLD and advanced fibrosis. Although the non-invasive markers for diagnosing NAFLD and advanced fibrosis were validated against liver biopsies or imaging in various Race/ethnicities,[24,28,29,46] it is possible that the NAFLD and advanced fibrosis due to NAFLD may have been misclassified. Third, the temporal causality of the observed association could not be established due to the cross-sectional study design. Some study reports indicate a relationship between diabetes and depression with a bidirectional increased risk between two diseases.[43] We have discussed evidence regarding linkage depression to NAFLD as above. However, given that the bidirectional association between depression and NAFLD is unknown, a future prospective cohort study is needed to investigate the impact of depression on NAFLD to confirm our results. Fourth, although we attempted to adjust known risk factors to investigate an independent association between two diseases, residual confounding must be considered as a possible explanation for at least part of the association. As diet information was not available for all studied NHANES participants, we are unable to evaluate the impact of nutrition on the association between two diseases. Fifth, we were unable to exclude other chronic liver diseases such as primary biliary cholangitis, primary sclerosing cholangitis, haemochromatosis, etc. While PHQ-9 questionnaire was widely validated, the lack of externally validated depression and functional impairment questionnaires is another limitation in our study.

In summary, although depression is associated with metabolic risk factors, our study suggests that depression may be an independent risk factor for NAFLD. While the pathophysiology underlying these associations remain undetermined, these data suggest that appropriate screening and prompt treatment of depression may prevent the development or progression of NAFLD. Further prospective data are needed to confirm our conclusions.