You are currently viewing a new version of our website. To view the old version click .
MDPIMDPI
Search all articles
Nutrients
Download PDF
  • Article
  • Open Access

25 May 2024

The Association between Diet–Exercise Patterns and Cirrhosis: A Cross-Sectional Study from NHANES 2017-March 2020

,
,
,
,
,
,
and
Department of Biostatistics, School of Public Health, Harbin Medical University, Harbin 150088, China
*
Author to whom correspondence should be addressed.
†
These authors contributed equally to this work.
Nutrients2024, 16(11), 1617;https://doi.org/10.3390/nu16111617
This article belongs to the Section Nutritional Epidemiology
Version Notes
Order Reprints

Abstract

Background: Liver cirrhosis (LC) is one of the most significant causes of morbidity and mortality in patients with chronic liver disease worldwide. Nutrition may be an important component of primary prevention of chronic liver disease. Diet–exercise patterns frame the eating behaviors and exercise habits of people through statistical methods related to nutritional epidemiology, which can explore the relationship between living habits and diseases among diverse populations. The purpose of this study was to explore the association between diet–exercise patterns and cirrhosis, and provide guidance on preventive diets for liver patients. Methods: This study identified diet–exercise patterns via clustering analysis of principal components and assessed their association with cirrhosis through the population samples of the National Health and Nutrition Examination Survey (NHANES) from 2017 to March 2020. Results: We identified two diet–exercise patterns that were named the “prudent pattern” (consumption of various staple foods, eggs, meat, fruits and vegetables; less sedentary) and the “dangerous pattern” (higher consumption of desserts, nuts, milk, meat, alcoholic beverages; recreational activities). The t-test demonstrated a significant relationship between patterns and multiple foods. The simple logistic regression test showed a lower risk of cirrhosis in those in the “prudent pattern” (OR = 0.73, 95%CI = 0.59–0.93). Conclusions: Two diet–exercise patterns associated with cirrhosis were identified: “prudent pattern” and “dangerous pattern”. The results of this study may be useful for suggesting preventive diets for people at risk of cirrhosis.

1. Introduction

Liver cirrhosis is the terminal stage of most chronic liver diseases (CLDs), having the highest incidence and mortality among liver-related diseases, which has an inferior prognosis [1,2,3]. In 2019, LC was associated with 2.4% of global deaths, and cirrhosis deaths are predicted to increase over the next decade, with the rate gradually increasing in some countries [4,5]. Steatosis is associated with the progression of liver fibrosis and cirrhosis, which is the process of diffuse liver damage [6,7]. Simple steatosis (NAFL) and nonalcoholic steatohepatitis (NASH) are closely related to a series of liver injuries caused by lipotoxicity, oxidative stress and other factors [8]. Biopsy studies have proved that NAFL can progress to advanced fibrosis [9]. It eventually leads to liver regeneration and fibrogenesis, enhances collagen and extracellular matrix deposition, and gradually progresses to cirrhosis [6,10]. Cirrhosis gives rise to numerous systemic complications, including ascites, malnutrition, sarcopenia and various others [3,5,11]. Malnutrition is a prominent characteristic and a significant complication of cirrhosis. Accumulating data have suggested that patients with LC often develop protein–energy malnutrition (PEM) at a rate of 25.1–65.5%, and PEM plays a crucial role in their poor survival [12]. Therefore, there is a need for more interventions to promote primary prevention, early detection and early treatment for liver disease-related patients [4].
Risk factors for LC include age, gender, obesity, smoking, alcohol consumption, genetics, the dysregulation of gut microbes, etc. [13]. Previous studies have demonstrated that sedentary behavior is also a risk factor for fibrosis progression in chronic liver diseases. In addition, cirrhosis is associated with significant nutritional risks. Assessing the impact of a single food, a food group or a nutrient on diseases is often not conclusive, because overall dietary habits of people include a variety of foods and the components of the diet are closely intertwined [14,15]. Dietary pattern analysis, of which common methods include index analysis, factor analysis and so on, is an alternative and complementary approach to analyze the relationship between nutrition and chronic disease risk, with the advantage of taking into account possible correlations between foods and food groups, as well as between nutrients [15]. Therefore, dietary pattern analysis is a better choice to capture the complex interactions between nutrition, food and other relevant factors.
Thus, a variety of food items were included in this analysis, and we further added exercise items to increase the accuracy of the clustering. Through an unsupervised machine learning approach, we explored the diet–exercise patterns associated with cirrhosis, and assessed the association between the patterns and cirrhosis. Finally, we discussed these results, drew conclusions based on the analyses we performed and provided guidance on preventive diets and activities for liver disease patients.

2. Materials and Methods

2.1. Study Population

The data from the National Health and Nutrition Examination Survey (NHANES) from 2017 to March 2020 were used for this study. The NHANES database is a population-based national survey that was published by the National Center for Health Statistics (NCHS), US Centers for Disease Control and Prevention (CDC). It obtains a nationally representative sample of approximately 5000 individuals per year through a complex, multi-stage sampling design and updates the database every 2 years, which can be used to assess the health and nutritional status of the United States population through information from questionnaires, physical examinations and laboratory tests [16,17]. And the NHANES protocol was reviewed and approved by the National Center for Health Statistics Research Ethics Review Board [18]. All relative data are publicly available at www.cdc.gov/nchs/nhanes/, accessed on 5 September 2023.

2.2. Outcome Indexes

Outcome indexes: The Mobile Examination Center (MEC) performed vibration-controlled transient elastography (VCTE) on the study population using the FibroScan Model 502 V2 Touch (Echosens, Paris, France) equipped with medium (M) and extra-large (XL) probes [19]. VCTE is a widely used clinical method for evaluating hepatic steatosis and fibrosis through controlled attenuation parameters (CAPs) and liver stiffness measurements (LSMs), respectively, obtained by measuring the shear wave velocities in liver tissue [19,20]. According to Tsochatzis et al. [21], VCTE recognizes cirrhosis with a high overall sensitivity and specificity (0.83 and 0.89, respectively). Sasso et al. [22] showed that using CAPs was a non-invasive, immediate, objective and effective method for the detection and quantification of steatosis, and can effectively separate several steatosis grades (compared with grade 1, AUC = 0.89, 0.91 and 0.95). Eddowes et al. [23] found that CAPs and LSMs can evaluate hepatic steatosis and fibrosis, respectively, with AUROC values ranging from 0.7 to 0.89. For VCTE results, examinations were considered reliable only if at least 10 LSM values were obtained after fasting for at least 3 h, with an interquartile range (IQR)/median < 30% [24].
Diagnostic criteria: We adopted the recent findings of Eddowes et al. [23] and Ciardullo et al. [25] as the disease grading criteria in this study (Table 1). Median CAP values and median LSM values were, respectively, divided into four grades.
Table 1. Grading criteria of hepatic steatosis and fibrosis.
Auxiliary indexes: Through examination and laboratory testing, the measurements of alanine aminotransferase (ALT), aspartate aminotransferase (AST), gamma-glutamyltransferase (GGT) and alkaline phosphatase (ALP) were obtained, and the NAFLD-4 (FIB-4) index and AST-to-platelet ratio index (APRI) [26] were calculated. The calculation formula was as follows: FIB-4 = (age×AST level)/(platelet count×ALT1/2). Cirrhosis/severe fibrosis was defined as FIB-4 > 3.25 [27]. APRI = (AST level/AST upper limit of normal) × 100/platelet count. Cirrhosis was defined as APRI > 2 [28].

2.3. Assessing Diet and Exercise and Constructing Patterns

In the NHANES, dietary and nutrient intake were assessed by reliable 24 h dietary recall interviews [29]. The data obtained from the 24 h recall were used to estimate individual characteristics of nutrient intake according to the Food and Nutrient Database for Dietary Studies (FNDDS) supplied by the US Department of Agriculture (USDA). The FNDDS contains a variety of food and beverage items and is updated biannually [30]. Dietary studies of the FNDDS database were used to encode and analyze dietary intakes collected in the WWEIA portion of the NHANES survey [31].
Physical activity was measured by using the global physical activity questionnaire that includes questions about frequency, intensity and duration [30]. To assess physical activities, we calculated the weekly metabolic equivalent (MET) score for each activity and intensity of the participants, referring to the methodology recommended by the NHANES [16]. With reference to the FNDDS database data, we divided all of the original food items into 53 food groups and 6 exercise groups (Table S1).
Dietary pattern analysis is an alternative and complementary approach to analyzing the relationship between nutrition and chronic disease. To determine dietary patterns, we can use priori approaches (i.e., indices or scores) to assess the conformity of diets with nutrition guidelines or established patterns. However, most analyses use posterior methods at present, such as cluster analysis and factor analysis based on exploratory data [32]. To obtain different dietary patterns, this study applied a new method developed by Maugeri et al. [33], which was defined as clustering of principal components. The method includes two multivariate techniques that both reduce the dimension of the diet datasets and provide a clustering solution [34]. Principal component analysis (PCA) is applied to reduce the data dimension of diet–exercise datasets. As a dimension reduction method commonly used in various fields including nutritional epidemiology, PCA usually selects principal components based on a cumulative variance greater than 80% or an eigenvalue greater than 1. However, for dietary data, the percentage of variance explained is typically between 10% and 30% [33]. Thus, choosing principal components based on eigenvalues is more suitable for the diet–exercise data in this study. In this step, the number of principal components retained were selected through the eigenvalues. The absolute values of factor loads were used to determine the contribution of each category to the composition. For each principal component (PC), the factor scores were calculated as the sum of the products between energy intakes and factor loads. We then performed K-means clustering on the retained PCs. Finally, the number of clusters to be considered were selected according to the clustering elbow diagram and the silhouette coefficients diagram.

2.4. Other Laboratory Tests and Clinical Data

Demographic information on age, gender, race/ethnicity, education, marriage, drinking and smoking status was collected through interviewer questionnaires. Race/ethnicity was divided into Mexican American, other Hispanic, non-Hispanic White, non-Hispanic Black, non-Hispanic Asian and others [35]. Educational attainment was classified as less than high school, high school graduate, some college or AA degree, college graduate or above. Marital status was classified as married, divorced/widowed/separated or unmarried. Drinking or not depended on whether the individual had consumed any alcoholic beverages or not. Smoking was defined as having smoked at least 100 cigarettes. Hypertension, diabetes and high cholesterol levels were determined based on questionnaires [36]. Also based on the questionnaires, we obtained information on whether the participants had anemia, kidney disease, cancer, hepatitis B, hepatitis C and other diseases. Other covariates in relation to the clinical and laboratory assessments were examined by trained medical experts [16].

2.5. Statistical Analysis

All statistical analyses accounted for complex survey design factors of the NHANES, including sample weights, stratification, and clustering, following the NHANES analytic and reporting guidelines. We used the t-test and rank sum test to evaluate the relationship between the patterns and the original food, the nutrients and cirrhosis-related indicators. Chi-square tests were used to detect the relationship between patterns and population characteristics, and between patterns and other diseases. Simple logistic regression tests were used to assess the association of cirrhosis with diet–exercise patterns, age (continuous), gender, education level, marital status, BMI (continuous), smoking status (smoking and non-smoking) and alcohol consumption (drinking and non-drinking). Logistic regression results were reported as odds ratios (ORs) and 95% confidence intervals (95%CIs). All statistical tests were conducted at the significance level of α = 0.05, and p < 0.05 indicated statistical significance [37]. All analyses were performed by using R software, version 4.2.0.

3. Results

3.1. Study Population

We investigated the association between diet–exercise patterns and the risk of cirrhosis in 7869 participants aged 18 years or older in the NHANES database. Participants with the absence of liver elastography information, diet or exercise recall information, and other covariates were excluded. A total of 5757 participants were included in this study (Figure S1).

3.2. Clustering of Principal Components

We applied PCA to standardized and energy-adjusted dietary and exercise data. PCA produced 22 PCs with eigenvalues greater than 1, of which 4 PCs had eigenvalues greater than 1.2. The first 4 PCs together explained 15.27% of the total variance, and the first 22 PCs together explained 52.03% of the total variance. Figure 1 displays the factor loads related to 22 PCs. Therefore, according to the selection principle of eigenvalues greater than 1, we selected the first 22 PCs for the next step of clustering research. This choice not only greatly reduced the dimension of the original data, which was more convenient for our subsequent analysis, but also greatly retained 52.03% of the original information in the diet–exercise data, which made our clustering results more accurate in interpreting the original information. Meanwhile, we also display the features of the first four principal components so that we can observe the distribution of the original features. Table 2 and Table S2 describe the first four PC-related explanatory variances and factor loads, while Figure S2 illustrates the distribution of the samples across the principal components. PC1 was mainly characterized by less sedentary activity, but insufficient intake of water, fruits and vegetables. PC2 mainly showed a higher intake of fruit and salad dressing, as well as less engagement in work activities. PC3 mainly consumed rice and vegetables and engaged in more recreational activities. PC4 mainly consumed milk and uncooked grains.
Figure 1. Factor loadings for each category and principal component.
Table 2. Total variance explained by principal components.
Next, we performed clustering on 22 principal components to obtain the clustering of samples. We also drew an elbow curve of clustering and calculated the silhouette scores of different clustering solutions, choosing the solution with two clustering features (Figure 2). So, at last, we obtained two diet–exercise patterns (Figure 3).
Figure 2. Elbow curve and silhouette scores for different cluster solutions. (A) Elbow curve. (B) Silhouette scores.
Figure 3. Distribution of participants by principal components and clusters.

3.3. Diet and Exercise Characteristics between Two Different Patterns

We compared the differences between the patterns for each diet–exercise category (Table S3). Figure 4 shows the average Z-scores for each diet–exercise category across patterns. Participants of pattern 1 consumed mostly noodles, rice, non-carbonated soda, poultry, potatoes, tomatoes, meat mixes, dark green vegetables, ice cream, fats, orange vegetables, sandwiches and lamb and tended to engage in moderate workout activities. Participants of pattern 2 were characterized by their intake of milk, cereals, biscuits, pancakes, pork, rising bread, sausages, other vegetables, other fruits, non-alcoholic beverages, alcoholic beverages, fast bread, egg mixes, desserts, nuts, beef, citrus fruits and oils, and tended to engage in moderate recreational activities and sedentary activities.
Figure 4. Comparison of diet–exercise items between patterns (Orange color represents a positive z-score and blue represents a negative z-score).
In terms of nutrition, participants of pattern 1 had higher intakes of protein, carbohydrates, total fat, total monounsaturated fatty acids, alpha-tocopherol vitamin E, retinol, beta-carotene, beta-zeaxanthin, lycopene, folate, vitamin B12, vitamin D, calcium, phosphorus, magnesium, zinc, sodium and selenium. Accordingly, participants of pattern 2 had higher intakes of sugar, total saturated fatty acids, total polyunsaturated fatty acids, cholesterol, vitamin E, vitamin A, alpha-carotene, vitamin B1, niacin, vitamin B6, total folic acid, dietary folic acid, vitamin C, vitamin K, iron, copper, potassium, caffeine and water (Table S4 and Figure 5).
Figure 5. Comparison of nutrient intakes between patterns (Orange color represents a positive z-score and blue represents a negative z-score).

3.4. Population Characteristics between Two Patterns

We then compared the basic characteristics of the two patterns of people. There were fewer people younger than 30 years in pattern 1 and more people older than 65 years in pattern 2 (p = 0.002). There were more women in pattern 1 (p < 0.001). Participants in pattern 1 were less educated (p < 0.001) and less often married compared to the dangerous pattern (p < 0.001). In addition, pattern 2 reported more participants with smoking and alcohol consumption (p = 0.036; p < 0.001). No statistically significant difference in BMI was found between the patterns (Table 3).
Table 3. The distribution of the sociodemographic characteristics of the study population across the two patterns.
We then analyzed how diseases were related to diet–exercise habits in different demographic subgroups. We found that although there was no significant relationship between CAPs and patterns in the overall population, there was a connection between the hepatic steatosis degree and diet–exercise patterns among people older than 65 years, people with a BMI under 25 and people with BMI of 30~40. In the overall population, the LSM was associated with the patterns. Moreover, the association between LSMs and diet–exercise patterns varied across demographic variables. For example, diet–exercise patterns had a significant effect on LSMs in people above 65 years old. However, in people under 65 years old, diet–exercise patterns could not be considered to have a significant effect on LSMs (Figure 6). We think that those with unhealthy diet–exercise habits may be more likely to suffer from diseases.
Figure 6. Forest plots of subgroup analyses. (A) Subgroup analysis between CAPs and patterns. (B) Subgroup analysis between LSMs and patterns.

3.5. Risk Factors of CAPs

Next, we assessed the main factors associated with CAPs (Table 4). According to our single-factor analysis, the risk factors associated with CAPs were age, gender, race, education, marital status and BMI. Our multi-factor analysis of these factors also confirmed these results (stepwise according to AIC). At the same time, our analysis showed that hypertension, diabetes, high cholesterol level and hepatitis B were all important influencing factors for hepatic steatosis in the US population.
Table 4. Factors associated with CAPs.

3.6. Risk Factors of LSMs

The results of univariate and multivariate analyses of LSMs are shown in Table 5. For the cluster and sociodemographic characteristics, the main factors associated with LSMs were diet–exercise patterns, age, gender, education level, BMI and smoking. People belonging to the first cluster had a lower probability of liver fibrosis and cirrhosis than those belonging to the second cluster (OR = 0.73; 95%CI = 0.59–0.93; p = 0.003). The diet–exercise patterns of these two clusters were named “prudent pattern” and “dangerous pattern” based on the risk of cirrhosis and the diet–exercise preferences. In addition, the risk of cirrhosis was higher in people with a higher BMI level (OR = 1.13; 95%CI = 1.11–1.15; p < 0.001). Hypertension, diabetes, high cholesterol level and hepatitis C were also important influencing factors for liver fibrosis and cirrhosis. We conducted a multi-factor analysis with a stepwise selection of cluster and sociodemographic characteristics, which obtained results that were almost consistent with the single-factor analysis. However, the results of the multi-factor analysis of cluster characteristics, sociodemographic characteristics and related disease characteristics showed that the influence of cluster characteristics on LSMs was no longer statistically significant.
Table 5. Factors associated with LSMs.
To further support our results, we compared the differences in liver-related indicators between different patterns. ALT, AST, GGT, ALP and FIB4 scores showed statistically significant differences between the two patterns (Table 6). The FIB4 index (OR = 0.911; 95%CI = 0.885–0.937; p < 0.001) and APRI index (OR = 0.996; 95%CI = 0.993–1.00; p < 0.001) confirmed the relationship between cirrhosis and the two patterns. Specifically, the prudent pattern was found to be less likely to develop cirrhosis compared to the dangerous pattern.
Table 6. The related indicators of liver cirrhosis with the two clusters.
Thus, we believe that diet–exercise pattern is an important influencing factor of liver cirrhosis. However, the reason for the difference between the single-factor analysis and the multi-factor analysis requires further analysis.

3.7. Other Diseases with Different Patterns

Table 7 shows the association between the two diet–exercise patterns and other diseases, indicating that there are significant differences in the prevalence of hypertension, diabetes and anemia between the two patterns (p < 0.001), and high cholesterol levels are also associated with the two diet–exercise patterns (p = 0.036). Since hypertension, diabetes and high cholesterol levels are important influencing factors of liver cirrhosis and closely related to the diet–exercise patterns, we analyzed whether the role of diet–exercise patterns in cirrhosis may be influenced by these diseases.
Table 7. Relations of other diseases with the two clusters.

3.8. The Medication Effect of the Disease

In order to address the uncertainties arising from the results of the previous two sections, we utilized these associated diseases (hypertension, diabetes, anemia, high cholesterol levels) as mediating variables to examine the impact of diet–exercise patterns on liver cirrhosis. The results of the mediation analysis showed that two diseases, hypertension and diabetes, were important mediating factors between diet–exercise patterns and liver cirrhosis. Their mediating effects were, respectively, 24.48% (p < 0.001) and 19.51% (p < 0.001). The results of their chain mediating effect are shown in Figure 7.
Figure 7. The medication effect of hypertension and diabetes.

4. Discussion

By utilizing a relatively new unsupervised machine learning method to analyze the sample population of the NHANES 2017-March 2020, we identified two diet–exercise patterns: “prudent pattern” and “dangerous pattern”. Of these, people in the “prudent pattern” tended to eat a variety of staple foods, eggs, meat, vegetables and fruits, were and less sedentary. People in the “dangerous pattern” ate more desserts, nuts, milk, meat, alcoholic beverages, and did more recreational activities. We then further demonstrated a strong link between diet–exercise habits, as indicated by these two diet–exercise patterns, and their risk of developing cirrhosis. This suggested that individuals following the “prudent pattern” had a lower risk.

4.1. Selection of Pattern Factors

In this study, diet items and exercise items were included to classify patterns affecting disease outcomes. Based on this, we analyzed the patterns and diseases. In previous studies, patterns typically included only dietary information. CORRAO et al. [38,39] conducted a principal component analysis on dietary data, proving the relationship between certain nutritional factors and the risk of LC. Guo et al. [40] also conducted a principal component analysis of dietary data, identifying two main dietary patterns. FARCHI et al. [14] conducted a cluster analysis on the sample population and divided it into four categories. These studies suggested that food intakes might be associated with the risk of chronic liver disease. Then, we considered that both diet items and exercise items had positive effects on multiple chronic diseases [41]. Exercise training has been shown to improve the health of patients with cirrhosis [42]. Therefore, we believed that it was necessary to incorporate exercise into the calculation of dietary patterns. In our results, the exercise factors do play an important role in the patterns, especially sedentary and recreational activities.

4.2. The Related Factors of Liver Cirrhosis

There is a large variation in the tendency of individuals to develop cirrhosis, which may be closely related to individual genetic factors and lifestyle habits. Previous studies have explored the association between cirrhosis and individual foods or nutrients. For instance, BRIDGES et al. [43] found a positive correlation between pork and beef and cirrhosis. The mechanism of the effect of protein, fiber and coffee on liver cirrhosis may be related to the duodenal microflora. Reducing the inflammatory potential of one’s diet may provide novel strategies to prevent the progression of cirrhosis [44,45]. In addition, the liver is a major organ responsible for zinc metabolism, and zinc deficiency may also alter liver cell function and immune response in inflammatory liver disease [2]. The protective and risk factors that these results point to are consistent with the items included in our prudent and dangerous patterns. Moreover, other sub-items in our patterns, particularly macro-nutrients and micro-nutrients, are likely to have significant effects on disease. Improving individual living habits, whether from the perspective of individual factors or from the level of overall dietary patterns, can have a positive impact on health conditions.
We further investigated the association between characteristics of the general population and cirrhosis. The results showed that low education, smoking, alcohol consumption and other factors are all related to the severity of cirrhosis progression. There are findings that support the causal protective effect of education on chronic liver disease [46,47]. There is evidence suggesting that smoking increases the risk of fibrosis progression and the development of hepatocellular carcinoma [48,49]. Also, individual differences in liver sensitivity to alcohol may influence the risk of developing this condition [49,50]. However, it has been suggested that women are more susceptible to the harmful effects of alcohol and have a higher risk of developing cirrhosis and related complications. Additionally, women also exhibit a higher rate of fibrosis progression after liver transplantation [51,52]. Our study suggests that women have a lower risk of cirrhosis than men, which is inconsistent with previous studies. This may be because many of the men in our sample had smoking and drinking habits.

4.3. Other Discussions about Disease Prevention

In summary, we obtained two diet–exercise patterns. These two patterns are not associated with CAPs but are significantly associated with LSMs. We believe that the combination of these items has little impact on early liver lesions, but is very useful for the disease outcome of patients with advanced liver lesions, especially cirrhosis. Lifestyle modifications targeting these combinations may not change disease development or exacerbation in patients with hepatic steatosis, but they may have a positive effect on the condition of patients with hepatic fibrosis. The results of our single-factor logistic regression analysis and multi-factor logistic regression analysis including demographic characteristics showed that the prudent pattern was a protective factor for LSMs. However, according to the multi-factor logistic regression analysis after adding hypertension, diabetes and other related diseases, the effect of the patterns on cirrhosis was not statistically significant. In order to determine the reason for this situation, we conducted a further mediation analysis and discovered that hypertension and diabetes were significant mediating factors between diet–exercise patterns and liver cirrhosis. Therefore, we discussed that the effect of the diet–exercise patterns was largely due to their impact on hypertension and diabetes, which led to the progression of liver cirrhosis. Thus, we suggest that adjusting diet and exercise, along with controlling blood pressure and glucose, can play a better role in preventing liver cirrhosis.
The main strength of our study was that this study is the first to address a specific target, population-wide U.S. diet–exercise patterns of cirrhosis risk. First of all, in the studies on chronic liver diseases, most diet studies have tended to focus on changes in the related traits of liver fat. However, our study added to the current knowledge on the impact of diet and exercise on liver cirrhosis, emphasizing the relationship between an unhealthy lifestyle and the possibility of increasing the degree of liver fibrosis and cirrhosis. We also incorporated exercise information and obtained more accurate analysis results on the living habits of the population. Secondly, in the study of dietary habits in liver cirrhosis, more studies are based on a priori pattern analysis or single-nutrient analysis, which often could obtain accurate and beneficial results. Of course, these results always cover what can be explained by posterior analyses. For example, whether a priori or a posteriori, we always obtain the conclusion that moderate vitamin D intake is likely to be good for health, which is often constant and consistent with our life experience. Nevertheless, we believe that a posteriori analysis—the statistical -partitioning of dietary patterns—is useful because it takes into account the relationship between food and nutrients, food interactions and nutrient interactions. For example, in the “prudent pattern” and “dangerous pattern” divided by the method of clustering of principal components, the intake of some vitamins is higher in the “prudent pattern” group, and the intake of others is higher in the “dangerous pattern” group. We cannot directly determine whether a certain vitamin is beneficial or harmful to the human body. This is because they interact with various other foods during ingestion. And how to make a better combination between various dietary factors and exercise factors, which is also the significance of posterior partition patterns. In the process of pattern division, different methods have different directivity and accuracy. We think that the method of this study is suitable for such analyses. The application of a relatively new method to analyze dietary patterns in cirrhosis is also one of the strengths of our study. This method overcomes the limitations of classical methods, which only focus on food itself and ignore individual living habits. Thirdly, we applied it to the data of the NHANES, which was a national sample survey in the United States. The research results of the NHANES could be radiated to the national population of the United States through the calculation of specific survey weights, which also made the research results more population-specific and made the conclusion credible. The two diet–exercise patterns that we obtained have confirmed an association with cirrhosis, and the differences in lifestyle habits between the two patterns can provide guidance on preventive diets for people at risk of cirrhosis. At the same time, these two patterns were also related to other diseases, such as diabetes and anemia, indicating that changing the dietary habits of the population can lead to more positive improvements in their health status. Finally, we conducted an effective mediation analysis of liver cirrhosis, which revealed the mediating role of diabetes and hypertension in the progression of liver cirrhosis. This may remind us of the relationship between diet and comorbidities.
The limitation of our study was that the selective elimination of the principal components may result in the loss of some original information from the food groups, potentially affecting the accuracy of the results. And we only briefly discussed the role of some related diseases in mediating the relationship between patterns and cirrhosis. The exploration of comorbidities between cirrhosis and related diseases, as well as the possible effects of diet–exercise patterns on these comorbidities, especially liver cancer, which may have a longitudinal relationship with the progression of cirrhosis, are also questions we need to consider for in-depth analysis in our upcoming studies. Another limitation of this study was the potential issue with data sources. We used the 24 h dietary recall data from the questionnaires in the NHANES database. Similarly, our physical activity data were also obtained from the recall interviews on the amount of physical activity performed in a week/day. However, there is inevitably some recall bias in the data obtained from these recall interviews. We have not been able to solve the problem. How to reduce the impact of these biases on lifestyle patterns and other research outcomes is also a topic that needs further research in the future.

5. Conclusions

The method of clustering of principal components describes the two main diet–exercise patterns associated with cirrhosis in the NHANES population: the “prudent pattern” and the “dangerous pattern”. The risk of disease in the “dangerous pattern” is relatively higher than that in the “prudent pattern”. The results of this study may help with preventive diets for people at risk of cirrhosis.

Supplementary Materials

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/nu16111617/s1, Figure S1: Flowchart of the study subjects; Figure S2: Score plots of first four principal components obtained after principal component analysis; Table S1: List of 59 food and activity categories; Table S2: Factor loadings for categories and four principal components; Table S3: Characteristics across two diet–exercise patterns; Table S4: Comparison of nutrient intakes between patterns.

Author Contributions

Conceptualization: Q.Z. and J.L.; methodology: Q.Z., J.L. and L.C.; project administration: Q.Z.; data curation: X.H.; formal analysis: J.L.; software: J.L. and X.H.; visualization: J.L. and L.C.; writing—original draft: J.L.; writing—review and editing: Q.Z., J.L., X.H., L.C., L.M., X.S., Y.D. and B.W.; supervision: Q.Z., L.M., X.S., Y.D. and B.W.; funding acquisition: Q.Z. All authors have read and agreed to the published version of the manuscript.

Funding

This study was funded by the National Natural Science Foundation of China (Funder: Qiuju Zhang; Funding Number: 82073666).

Institutional Review Board Statement

This study is an analysis derived from publicly available NHANES data which were acquired with the approval of the Ethics Review Board of the National Center for Health Statistics (https://www.cdc.gov/nchs/nhanes/irba98.htm), with the NCHS IRB/ERB Protocol Number: Protocol #2018-01 (Effective beginning 26 October 2017). Given that the NHANES publicly available data do not contain individually identifiable information, ethical review and approval were waived for this study.

Data Availability Statement

All relative original data are publicly available at www.cdc.gov/nchs/nhanes/, accessed on 5 September 2023.

Conflicts of Interest

The authors declare no conflicts of interest.

References

  1. Topan, M.M.; Sporea, I.; Dănilă, M.; Popescu, A.; Ghiuchici, A.M.; Lupușoru, R.; Șirli, R. Association between Dietary Habits and Sarcopenia in Patients with Liver Cirrhosis. J. Clin. Med. 2023, 12, 4693. [Google Scholar] [CrossRef] [PubMed]
  2. Grüngreiff, K.; Reinhold, D.; Wedemeyer, H. The role of zinc in liver cirrhosis. Ann. Hepatol. 2016, 15, 7–16. [Google Scholar] [CrossRef] [PubMed]
  3. Daftari, G.; Tehrani, A.N.; Pashayee-Khamene, F.; Karimi, S.; Ahmadzadeh, S.; Hekmatdoost, A.; Salehpour, A.; Saber-Firoozi, M.; Hatami, B.; Yari, Z. Dietary protein intake and mortality among survivors of liver cirrhosis: A prospective cohort study. BMC Gastroenterol. 2023, 23, 227. [Google Scholar] [CrossRef] [PubMed]
  4. Huang, D.Q.; Terrault, N.A.; Tacke, F.; Gluud, L.L.; Arrese, M.; Bugianesi, E.; Loomba, R. Global epidemiology of cirrhosis—Aetiology, trends and predictions. Nat. Rev. Gastroenterol. Hepatol. 2023, 20, 388–398. [Google Scholar] [CrossRef] [PubMed]
  5. Yao, C.K.; Fung, J.; Chu, N.H.S.; Tan, V.P.Y. Dietary Interventions in Liver Cirrhosis. J. Clin. Gastroenterol. 2018, 52, 663–673. [Google Scholar] [CrossRef] [PubMed]
  6. Enomoto, H.; Bando, Y.; Nakamura, H.; Nishiguchi, S.; Koga, M. Liver fibrosis markers of nonalcoholic steatohepatitis. World J. Gastroenterol. 2015, 21, 7427–7435. [Google Scholar] [CrossRef] [PubMed]
  7. Smith, A.; Baumgartner, K.; Bositis, C. Cirrhosis: Diagnosis and Management. Am. Fam. Physician 2019, 100, 759–770. [Google Scholar]
  8. Schuppan, D.; Surabattula, R.; Wang, X.Y. Determinants of fibrosis progression and regression in NASH. J. Hepatol. 2018, 68, 238–250. [Google Scholar] [CrossRef] [PubMed]
  9. Singh, S.; Allen, A.M.; Wang, Z.; Prokop, L.J.; Murad, M.H.; Loomba, R. Fibrosis progression in nonalcoholic fatty liver vs nonalcoholic steatohepatitis: A systematic review and meta-analysis of paired-biopsy studies. Clin. Gastroenterol. Hepatol. 2015, 13, 643–654.e9. [Google Scholar] [CrossRef]
  10. Kumar, S.; Duan, Q.; Wu, R.; Harris, E.N.; Su, Q. Pathophysiological communication between hepatocytes and non-parenchymal cells in liver injury from NAFLD to liver fibrosis. Adv. Drug Deliv. Rev. 2021, 176, 113869. [Google Scholar] [CrossRef]
  11. Sajja, K.C.; Mohan, D.P.; Rockey, D.C. Age and ethnicity in cirrhosis. J. Investig. Med. 2014, 62, 920–926. [Google Scholar] [CrossRef] [PubMed]
  12. Toshikuni, N.; Arisawa, T.; Tsutsumi, M. Nutrition and exercise in the management of liver cirrhosis. World J. Gastroenterol. 2014, 20, 7286–7297. [Google Scholar] [CrossRef]
  13. Huang, D.Q.; Mathurin, P.; Cortez-Pinto, H.; Loomba, R. Global epidemiology of alcohol-associated cirrhosis and HCC: Trends, projections and risk factors. Nat. Rev. Gastroenterol. Hepatol. 2023, 20, 37–49. [Google Scholar] [CrossRef] [PubMed]
  14. Farchi, G.; Mariotti, S.; Menotti, A.; Seccareccia, F.; Torsello, S.; Fidanza, F. Diet and 20-y mortality in two rural population groups of middle-aged men in Italy. Am. J. Clin. Nutr. 1989, 50, 1095–1103. [Google Scholar] [CrossRef] [PubMed]
  15. Choy, M.J.Y.; Brownlee, I.; Murphy, A.M. Data-Driven Dietary Patterns, Nutrient Intake and Body Weight Status in a Cross-Section of Singaporean Children Aged 6–12 Years. Nutrients 2021, 13, 1335. [Google Scholar] [CrossRef] [PubMed]
  16. Peng, P.; Gao, S.; Xiao, F.; He, M.; Fang, W.; Zhang, Y.; Wei, Q. Association of macronutrients intake distribution with osteoarthritis risk among adults in NHANES, 2013–2016. Front. Nutr. 2023, 10, 976619. [Google Scholar] [CrossRef]
  17. Han, K.; Tan, K.; Shen, J.; Gu, Y.; Wang, Z.; He, J.; Kang, L.; Sun, W.; Gao, L.; Gao, Y. Machine learning models including insulin resistance indexes for predicting liver stiffness in United States population: Data from NHANES. Front. Public Health 2022, 10, 1008794. [Google Scholar] [CrossRef] [PubMed]
  18. Matsunaga, M.; Lim, E.; Davis, J.; Chen, J.J. Dietary Quality Associated with Self-Reported Diabetes, Osteoarthritis, and Rheumatoid Arthritis among Younger and Older US Adults: A Cross-Sectional Study Using NHANES 2011–2016. Nutrients 2021, 13, 545. [Google Scholar] [CrossRef]
  19. Vilar-Gomez, E.; Nephew, L.D.; Vuppalanchi, R.; Gawrieh, S.; Mladenovic, A.; Pike, F.; Samala, N.; Chalasani, N. High-quality diet, physical activity, and college education are associated with low risk of NAFLD among the US population. Hepatology 2022, 75, 1491–1506. [Google Scholar] [CrossRef]
  20. Sandrin, L.; Fourquet, B.; Hasquenoph, J.M.; Yon, S.; Fournier, C.; Mal, F.; Christidis, C.; Ziol, M.; Poulet, B.; Kazemi, F.; et al. Transient elastography: A new noninvasive method for assessment of hepatic fibrosis. Ultrasound Med. Biol. 2003, 29, 1705–1713. [Google Scholar] [CrossRef] [PubMed]
  21. Tsochatzis, E.A.; Gurusamy, K.S.; Ntaoula, S.; Cholongitas, E.; Davidson, B.R.; Burroughs, A.K. Elastography for the diagnosis of severity of fibrosis in chronic liver disease: A meta-analysis of diagnostic accuracy. J. Hepatol. 2011, 54, 650–659. [Google Scholar] [CrossRef] [PubMed]
  22. Sasso, M.; Beaugrand, M.; de Ledinghen, V.; Douvin, C.; Marcellin, P.; Poupon, R.; Sandrin, L.; Miette, V. Controlled attenuation parameter (CAP): A novel VCTE™ guided ultrasonic attenuation measurement for the evaluation of hepatic steatosis: Preliminary study and validation in a cohort of patients with chronic liver disease from various causes. Ultrasound Med. Biol. 2010, 36, 1825–1835. [Google Scholar] [CrossRef] [PubMed]
  23. Eddowes, P.J.; Sasso, M.; Allison, M.; Tsochatzis, E.; Anstee, Q.M.; Sheridan, D.; Guha, I.N.; Cobbold, J.F.; Deeks, J.J.; Paradis, V.; et al. Accuracy of FibroScan Controlled Attenuation Parameter and Liver Stiffness Measurement in Assessing Steatosis and Fibrosis in Patients With Nonalcoholic Fatty Liver Disease. Gastroenterology 2019, 156, 1717–1730. [Google Scholar] [CrossRef] [PubMed]
  24. Heredia, N.I.; Zhang, X.; Balakrishnan, M.; Daniel, C.R.; Hwang, J.P.; McNeill, L.H.; Thrift, A.P. Physical activity and diet quality in relation to non-alcoholic fatty liver disease: A cross-sectional study in a representative sample of U.S. adults using NHANES 2017–2018. Prev. Med. 2022, 154, 106903. [Google Scholar] [CrossRef] [PubMed]
  25. Ciardullo, S.; Perseghin, G. Prevalence of NAFLD, MAFLD and associated advanced fibrosis in the contemporary United States population. Liver Int. 2021, 41, 1290–1293. [Google Scholar] [CrossRef]
  26. Huang, S.W.; Chen, C.; Kong, H.Y.; Huang, J.Q. Prevalence of Cirrhosis/Advanced Fibrosis Among HBsAg-Negative and HBcAb-Positive US Adults: A Nationwide Population-Based Study. Infect. Dis. Ther. 2022, 11, 1901–1916. [Google Scholar] [CrossRef] [PubMed]
  27. Udompap, P.; Mannalithara, A.; Heo, N.Y.; Kim, D.; Kim, W.R. Increasing prevalence of cirrhosis among U.S. adults aware or unaware of their chronic hepatitis C virus infection. J. Hepatol. 2016, 64, 1027–1032. [Google Scholar] [CrossRef]
  28. Kabbany, M.N.; Conjeevaram Selvakumar, P.K.; Watt, K.; Lopez, R.; Akras, Z.; Zein, N.; Carey, W.; Alkhouri, N. Prevalence of Nonalcoholic Steatohepatitis-Associated Cirrhosis in the United States: An Analysis of National Health and Nutrition Examination Survey Data. Am. J. Gastroenterol. 2017, 112, 581–587. [Google Scholar] [CrossRef]
  29. Mazidi, M.; Pennathur, S.; Afshinnia, F. Link of dietary patterns with metabolic syndrome: Analysis of the National Health and Nutrition Examination Survey. Nutr. Diabetes 2017, 7, e255. [Google Scholar] [CrossRef]
  30. Akbar, Z.; Shi, Z. Dietary Patterns and Circadian Syndrome among Adults Attending NHANES 2005–2016. Nutrients 2023, 15, 3396. [Google Scholar] [CrossRef]
  31. Lau, C.S.; Fulgoni, V.L., 3rd; Van Elswyk, M.E.; McNeill, S.H. Trends in Beef Intake in the United States: Analysis of the National Health and Nutrition Examination Survey, 2001–2018. Nutrients 2023, 15, 2475. [Google Scholar] [CrossRef]
  32. Gherasim, A.; Arhire, L.I.; Niță, O.; Popa, A.D.; Graur, M.; Mihalache, L. The relationship between lifestyle components and dietary patterns. Proc. Nutr. Soc. 2020, 79, 311–323. [Google Scholar] [CrossRef] [PubMed]
  33. Maugeri, A.; Barchitta, M.; Favara, G.; La Mastra, C.; La Rosa, M.C.; Magnano San Lio, R.; Agodi, A. The Application of Clustering on Principal Components for Nutritional Epidemiology: A Workflow to Derive Dietary Patterns. Nutrients 2022, 15, 195. [Google Scholar] [CrossRef]
  34. Barchitta, M.; Magnano San Lio, R.; La Rosa, M.C.; La Mastra, C.; Favara, G.; Ferrante, G.; Galvani, F.; Pappalardo, E.; Ettore, C.; Ettore, G.; et al. The Effect of Maternal Dietary Patterns on Birth Weight for Gestational Age: Findings from the MAMI-MED Cohort. Nutrients 2023, 15, 1922. [Google Scholar] [CrossRef] [PubMed]
  35. Spaur, M.; Nigra, A.E.; Sanchez, T.R.; Navas-Acien, A.; Lazo, M.; Wu, H.C. Association of blood manganese, selenium with steatosis, fibrosis in the National Health and Nutrition Examination Survey, 2017–2018. Environ. Res. 2022, 213, 113647. [Google Scholar] [CrossRef] [PubMed]
  36. Du, R.; Tang, X.Y.; Yang, C.; Gao, W.H.; Gao, S.J.; Xiang, H.J.; Yang, L. Marijuana use is inversely associated with liver steatosis detected by transient elastography in the general United States population in NHANES 2017–2018: A cross-sectional study. PLoS ONE 2023, 18, e0284859. [Google Scholar] [CrossRef]
  37. Qarmiche, N.; El Kinany, K.; Otmani, N.; El Rhazi, K.; Chaoui, N.E.H. Cluster analysis of dietary patterns associated with colorectal cancer derived from a Moroccan case-control study. BMJ Health Care Inf. 2023, 30, e100710. [Google Scholar] [CrossRef]
  38. Corrao, G.; Zambon, A.; Bagnardi, V.; Aricò, S.; Loguercio, C.; D’Amicis, A. Nutrient intakes, nutritional patterns and the risk of liver cirrhosis: An explorative case-control study. Eur. J. Epidemiol. 2004, 19, 861–869. [Google Scholar] [CrossRef]
  39. Corrao, G.; Ferrari, P.A.; Galatola, G. Exploring the role of diet in modifying the effect of known disease determinants: Application to risk factors of liver cirrhosis. Am. J. Epidemiol. 1995, 142, 1136–1146. [Google Scholar] [CrossRef]
  40. Guo, W.; Ge, X.; Lu, J.; Xu, X.; Gao, J.; Wang, Q.; Song, C.; Zhang, Q.; Yu, C. Diet and Risk of Non-Alcoholic Fatty Liver Disease, Cirrhosis, and Liver Cancer: A Large Prospective Cohort Study in UK Biobank. Nutrients 2022, 14, 5335. [Google Scholar] [CrossRef]
  41. Johnston, H.E.; Takefala, T.G.; Kelly, J.T.; Keating, S.E.; Coombes, J.S.; Macdonald, G.A.; Hickman, I.J.; Mayr, H.L. The Effect of Diet and Exercise Interventions on Body Composition in Liver Cirrhosis: A Systematic Review. Nutrients 2022, 14, 3365. [Google Scholar] [CrossRef]
  42. Williams, F.R.; Berzigotti, A.; Lord, J.M.; Lai, J.C.; Armstrong, M.J. Review article: Impact of exercise on physical frailty in patients with chronic liver disease. Aliment. Pharmacol. Ther. 2019, 50, 988–1000. [Google Scholar] [CrossRef]
  43. Bridges, F.S. Relationship between dietary beef, fat, and pork and alcoholic cirrhosis. Int. J. Environ. Res. Public Health 2009, 6, 2417–2425. [Google Scholar] [CrossRef] [PubMed]
  44. Ibrahim, M.K.; Wilechansky, R.M.; Challa, P.K.; Zhang, X.; Giovannucci, E.; Stampfer, M.; Chan, A.T.; Simon, T.G. The empirical dietary inflammatory pattern score and the risk of nonalcoholic fatty liver disease and cirrhosis. Hepatol. Commun. 2023, 7, e0263. [Google Scholar] [CrossRef]
  45. Hussain, S.K.; Dong, T.S.; Agopian, V.; Pisegna, J.R.; Durazo, F.A.; Enayati, P.; Sundaram, V.; Benhammou, J.N.; Noureddin, M.; Choi, G.; et al. Dietary Protein, Fiber and Coffee Are Associated with Small Intestine Microbiome Composition and Diversity in Patients with Liver Cirrhosis. Nutrients 2020, 12, 1395. [Google Scholar] [CrossRef] [PubMed]
  46. Wang, Y.; Kong, L.; Ye, C.; Dou, C.; Zheng, J.; Xu, M.; Xu, Y.; Li, M.; Zhao, Z.; Lu, J.; et al. Causal impacts of educational attainment on chronic liver diseases and the mediating pathways: Mendelian randomization study. Liver Int. 2023, 43, 2379–2392. [Google Scholar] [CrossRef] [PubMed]
  47. Stroffolini, T.; Sagnelli, E.; Sagnelli, C.; Morisco, F.; Babudieri, S.; Furlan, C.; Pirisi, M.; Russello, M.; Smedile, A.; Pisaturo, M.; et al. The association between education level and chronic liver disease of any etiology. Eur. J. Intern. Med. 2020, 75, 55–59. [Google Scholar] [CrossRef] [PubMed]
  48. Juran, B.D.; Lazaridis, K.N. Environmental factors in primary biliary cirrhosis. Semin. Liver Dis. 2014, 34, 265–272. [Google Scholar] [CrossRef]
  49. Hagström, H. Alcohol, smoking and the liver disease patient. Best Pract. Res. Clin. Gastroenterol. 2017, 31, 537–543. [Google Scholar] [CrossRef] [PubMed]
  50. Dam, M.K.; Flensborg-Madsen, T.; Eliasen, M.; Becker, U.; Tolstrup, J.S. Smoking and risk of liver cirrhosis: A population-based cohort study. Scand. J. Gastroenterol. 2013, 48, 585–591. [Google Scholar] [CrossRef]
  51. Bizzaro, D.; Becchetti, C.; Trapani, S.; Lavezzo, B.; Zanetto, A.; D’Arcangelo, F.; Merli, M.; Lapenna, L.; Invernizzi, F.; Taliani, G.; et al. Influence of sex in alcohol-related liver disease: Pre-clinical and clinical settings. United Eur. Gastroenterol. J. 2023, 11, 218–227. [Google Scholar] [CrossRef] [PubMed]
  52. Rubin, J.B.; Sundaram, V.; Lai, J.C. Gender Differences Among Patients Hospitalized With Cirrhosis in the United States. J. Clin. Gastroenterol. 2020, 54, 83–89. [Google Scholar] [CrossRef] [PubMed]
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Article Metrics

Citations

Article Access Statistics

Journal Statistics
Multiple requests from the same IP address are counted as one view.
Download PDF
Yerba Mate (Ilex paraguariensis) Reduces Colitis Severity by Promoting Anti-Inflammatory Macrophage Polarization
Previous
Exploring Family Perspectives on a Group-Based Hands-on Advanced Carbohydrate Counting Education Program for Children and Adolescents with Type 1 Diabetes: A Qualitative Study
Next