Food Insecurity is Increasing and is More Common Among Persons with Chronic Liver Disease

Background: Effective interventions for metabolic liver disease include optimized nutritional intake. It is increasingly clear, however, that many patients with metabolic liver disease lack the resources to execute nutritional advice. Data on the trends of food insecurity are needed to prioritize public health strategies to address the burden of liver disease. Methods: Cross-sectional analysis of six waves of data from the 2007–2018, 24,847 subjects aged ≥20 years from the 2017–2018 National Health and Nutrition Examination Survey. Food security was measured using the US Department of Agriculture’s Core Food Security Module. Liver disease was defined as elevated liver enzymes and a risk factor: elevated BMI, diabetes, and/or excess alcohol consumption. Models were adjusted using age, sex, race/ethnicity, education, poverty-income ratio, smoking, physical activity, alcohol intake, sugary beverage intake, Healthy Eating Inex-2015 score. Advanced liver disease was estimated using FIB-4 >2.67. Results: The overall prevalence of liver disease was 24.6%, ranging from 21.1% (2017–2018) to 28.3% (2015–2016) (P-trend=0.85). 3.4% of participants had possible advanced liver disease, ranging from 1.9% (2007–2008) to 4.2% (2015–2016)(P-trend=0.07). Among those with liver disease, the prevalence of food insecurity was 13.6% in 2007–2008, which rose steadily to 21.6% in 2015–2016, before declining to 18.0% in 2017–2018 (P-trend=0.0004). Food insecurity rose more sharply for adults aged <50 years (2007–2008: 17.6%, 2015–2016: 28.0%, P-trend=0.004) compared to adults aged ≥50 years (2007–2008: 9.5%, 2015–2016: 16.5%, P-trend<0.0001). Food insecurity was more common among women, those with high BMI, and those with diabetes Conclusion : Food insecurity is increasingly common among those with liver disease.


Introduction
Chronic liver disease (CLD) is a major threat to our public health. 1Roughly 60,000 people die from CLD annually. 2CLD accounts for more than 1 million outpatient visits and CLD-related hospitalizations are increasing annually, overtaking hospitalization-rates for heart failure and lung disease. 3,4Annual CLD healthcare costs exceed $29.9 billion. 57][8] Even among persons with excess alcohol intake and alcohol related liver disease (ALD), diet and obesity in uence the risk of liver disease -the so-called 'MetALD'.Effective interventions for metabolic liver disease include optimized nutritional intake.It is increasingly clear, however, that many patients with metabolic liver disease lack the resources to execute nutritional advice. 9healthy diet presupposes accessibility and affordability of foods consistent with evidence-based diet patterns.1][12] Food insecurity, a condition of limited access to healthy food, has been associated with increased risks of obesity, diabetes, cardiovascular disease and indeed with metabolic liver disease. 12,13,13 ur previous study showed that food insecurity is associated with an increased risk of cirrhosis for persons aged > 50 years in a nationally representative sample. 9Kardashian has shown that food insecurity increased mortality for those with chronic liver disease. 14Food insecurity is an urgent problem.
Given that food insecurity is associated with metabolic liver disease, data regarding trends in food insecurity are needed.Such data would inform policy and justify funding for interventions which can both alleviate food insecurity and prevent progressive liver disease.We therefore used data from the 2007-2018 National Health and Nutrition Examination Survey (NHANES) to identify trends in food insecurity among persons with and without CLD over a two-year period.

Study population
We performed a serial cross-sectional analysis of six waves of data from the 2007-2018 NHANES.The NHANES is a nationally representative cross-sectional study conducted by the National Center for Health Statistics (NCHS) at the Centers for Disease Control and Prevention.NHANES enrolls participants using a strati ed multistage probability with an oversampling design of certain age and racial/ethnic groups to allow for weighted analyses of the civilian non-institutionalized US population. 15All participants are interviewed for demographic, socioeconomic, dietary, and health information, and physical examinations and laboratory tests are conducted on most study participants.In the present study, the analytic sample included 24,847 adults ≥ 20 years, not pregnant at the time of the survey, and with complete data on food insecurity.We also excluded participants with chronic viral hepatitis based on positive hepatitis C virus (HCV) RNA or hepatitis B surface antigen (HBsAg).NHANES data are de-identi ed and publicly available; thus, no further IRB review was deemed necessary.

Liver disease
We de ned the presence of probably liver disease as having elevated liver enzymes and the presence of additional risk factors.Elevated liver enzymes were de ned as high ALT or high AST (men ≥ 30 U/L and women ≥ 20 U/L).Additional risk factors included having a body mass index (BMI) ≥ 30 kg/m 2 , selfreported prior diagnosis of diabetes, or heavy alcohol consumption (current intake ≥ 7 drinks/week for women, ≥ 14 drinks/week for men).BMI was calculated from measured height and weight in the Mobile Examination Center (MEC).As secondary de nitions, we also used advanced brosis on the basis of brosis-4 (FIB-4) index > 2.67.This was calculated as age (year) × AST (U/L)/(platelet count [109/L] × square root(ALT U/L). 14

Food Insecurity
Household food security over the past year was de ned using the US Department of Agriculture's (USDA) Core Food Security Module.This module includes 18 questions about food security addressing anxiety over the food supply, ability to eat a balanced meal, and behavioral manifestations of food rationing at the household, adult, and child levels.Individuals with > 3 a rmative responses are de ned as food insecure; individuals with 0-2 a rmative responses are de ned as food secure.

Statistical analysis
NHANES MEC sampling weights were recalculated and used in all analyses to account for differential selection probabilities, nonresponse, and to make nationally representative estimates over the 12-year period.First, we estimated proportions of adults with food insecurity, liver disease, and demographic and health covariates strati ed by survey wave, Chi-squared tests were used to examine differences in demographic and health covariates across survey waves.We calculated frequencies in food insecurity among adults with liver disease across survey waves in the overall population and strati ed by sociodemographic and health covariates.We then estimated trends in food insecurity among adults with and without liver disease across survey waves, adjusted for covariates.In all analyses, we tested the signi cance of time trends by including survey wave as an ordinal variable in multivariate regression models.Finally, we t a multivariable logistic regression model for the outcome of food insecurity including all sociodemographic and health covariates among adults with liver disease.All statistical analyses were performed using SAS, version 9.4.

Trends in food insecurity among adults with liver disease
In Table 2, we highlight the overall trends in food insecurity among adults with liver disease and strati ed by demographic and health subgroups.Among those with liver disease, the prevalence of food insecurity was 13.6% in 2007-2008, which rose steadily to 21.6% in 2015-2016, before declining to 18.0% in 2017-2018 (P-trend = 0.0004).Figure 1 shows trends in food insecurity among those with and without liver disease, after adjusting for sociodemographic covariates.Among adults with liver disease, the adjusted prevalence of food insecurity rose from 16.9% in 2007-2008 to 24.0% in 2017-2018 (P-trend = 0.0003).Among adults without liver disease, the adjusted prevalence of food insecurity increased from 13.6% in 2007-2008 to 20.4% in 2017-2018 (P-trend < 0.0001).At each survey wave, the prevalence of food insecurity was higher among those with liver disease than those without liver disease.

Predictors of food insecurity among adults with liver disease
In Table 3, results from the multivariable model showed that the odds of food insecurity were signi cantly lower among those with older age (≥ 50 years), with higher educational attainment (OR 0.77, 95% CI 0.65, 0.92), and with higher incomes (OR 0.22, 95% CI 0.17, 0.29), and signi cantly higher among females (OR

Discussion
Metabolic liver disease is a large and growing public health problem.Dietary modi cation is essential to prevent progression to cirrhosis and its complications.Common recommendations include eliminating excess carbohydrates in favor of whole foods. 6,16,17This strategy is challenged in the context of increasingly prevalent food insecurity.In this nationally representative study, we showed that food insecurity was prevalent among those with liver disease, and their levels increased signi cantly over time.Food insecurity poses a major barrier to the effectiveness of interventions to improve liver health and our data highlight both the increasing burden overall and within high-risk subgroups.

New ndings
Our study extends the literature with three key ndings.First, we show that food insecurity is rising among persons with metabolic liver disease, from 16.9% in 2007-2008 and peaking at 28.7% in 2015-2016.Second, we show that among those with metabolic liver disease the subgroups with the starkest increases in food insecurity include younger-aged adults (< 50 years), women, Hispanic persons, people with less than a college education, unmarried/unpartnered persons, and those with BMI > 30 kg/m 2 .
Third, we show that among adults with liver disase, food insecurity is strongly predicted by markers of economic disadvantage, including lower educational attainment, lower income, and SNAP participation, and by correlated health risks, such as elevated BMI and glycohemoglobin.

Con rmatory ndings
Our data con rms and extends prior ndings showing that persons with chronic disease, such as those with cardiovascular disease, are more likely to have food insecurity. 18In contrast to cardiovascular disease, however, metabolic liver disease is both caused by malnutrition and, itself, independently associated with cardiovascular disease morbidity and mortalty. 19Fluid retention causes nausea, anorexia, and is associated with poor nutritional intake.Liver disease related cholestasis, microbiome alterations, and shunting leads to decreased absorption of macro and micro-nutrients.Patients with liver disease are managed with further dietary restrictions including salt and uid limits leading to additional challenges obtaining nutritionally adequate and palatable diets. 20Comprehensive treatment of CLD must include targeted interventions for malnutrition to address both these causes and effects.
Recently, Ochoa-Allemant et al showed that among persons with metabolic liver disease, those with MetALD have a higher prevalence of food insecurity (42.1% vs. 27.7%)compared to those with MASLD. 21od insecurity is also associated with increased alcohol intake, 22 potentiating liver disease.Along with treating alcohol use disorder nutrition must also be targeted in a MetALD management strategy to decrease the rising burden of alcohol related liver disease.

Next steps
Our data suggest that the burden of food insecurity is increasing necessitating targeted intervention.Targeted dietary therapies including those commonly prescribed for MASLD may be insu cient to address those with food insecurity.Patients with liver disease and food insecurity should be offered social work counseling and potential enrollment in federal, state and/or local programs to improve access to food.Our data demonstrates that SNAP enrollment may be insu cient and additional supplementation and counseling may be needed speci cally targeting those with CLD.Those most at risk including patients < 50, women, Hispanic persons, people with less than a college education, unmarried/unpartnered persons, and those with BMI > 30 kg/m 2 merit additional awareness for conscious targeted intervention.More studies are needed to inform what additional interventions may be most effective for these groups of patients with CLD.

Figures Figure 1
Figures

Table 1
CI: con dence interval, SNAP: Supplemental Nutrition Assistance Program, OR = odds ratio CI: con dence interval, SNAP: Supplemental Nutrition Assistance Program, OR = odds ratio

Table 4
Predictors of Food Insecurity Among Participants with Liver Disease CI: con dence interval, SNAP: Supplemental Nutrition Assistance Program, OR = odds ratio CI: con dence interval, SNAP: Supplemental Nutrition Assistance Program, OR = odds ratio