Hostname: page-component-7bb8b95d7b-l4ctd Total loading time: 0 Render date: 2024-10-04T08:44:44.791Z Has data issue: false hasContentIssue false

Added sugar intake and its forms and sources in relation to risk of non-alcoholic fatty liver disease: results from the Tianjin Chronic Low-grade Systemic Inflammation and Health cohort study

Published online by Cambridge University Press:  26 September 2022

Shunming Zhang
Affiliation:
Nutrition and Radiation Epidemiology Research Center, Institute of Radiation Medicine, Chinese Academy of Medical Sciences & Peking Union Medical College, Tianjin, People’s Republic of China Nutritional Epidemiology Institute and School of Public Health, Tianjin Medical University, Tianjin, People’s Republic of China Nutritional Epidemiology, Department of Clinical Sciences Malmö, Lund University, Malmö, Sweden
Huiping Li
Affiliation:
Nutritional Epidemiology Institute and School of Public Health, Tianjin Medical University, Tianjin, People’s Republic of China
Ge Meng
Affiliation:
Nutritional Epidemiology Institute and School of Public Health, Tianjin Medical University, Tianjin, People’s Republic of China Department of Toxicology and Sanitary Chemistry, School of Public Health, Tianjin Medical University, Tianjin, People’s Republic of China
Qing Zhang
Affiliation:
Health Management Centre, Tianjin Medical University General Hospital, Tianjin, People’s Republic of China
Li Liu
Affiliation:
Health Management Centre, Tianjin Medical University General Hospital, Tianjin, People’s Republic of China
Hongmei Wu
Affiliation:
Nutritional Epidemiology Institute and School of Public Health, Tianjin Medical University, Tianjin, People’s Republic of China
Yeqing Gu
Affiliation:
Nutrition and Radiation Epidemiology Research Center, Institute of Radiation Medicine, Chinese Academy of Medical Sciences & Peking Union Medical College, Tianjin, People’s Republic of China
Tingjing Zhang
Affiliation:
Nutritional Epidemiology Institute and School of Public Health, Tianjin Medical University, Tianjin, People’s Republic of China
Xuena Wang
Affiliation:
Nutritional Epidemiology Institute and School of Public Health, Tianjin Medical University, Tianjin, People’s Republic of China
Juanjuan Zhang
Affiliation:
Nutritional Epidemiology Institute and School of Public Health, Tianjin Medical University, Tianjin, People’s Republic of China
Jun Dong
Affiliation:
Nutritional Epidemiology Institute and School of Public Health, Tianjin Medical University, Tianjin, People’s Republic of China
Xiaoxi Zheng
Affiliation:
Nutritional Epidemiology Institute and School of Public Health, Tianjin Medical University, Tianjin, People’s Republic of China
Zhixia Cao
Affiliation:
Nutritional Epidemiology Institute and School of Public Health, Tianjin Medical University, Tianjin, People’s Republic of China
Xu Zhang
Affiliation:
Nutritional Epidemiology Institute and School of Public Health, Tianjin Medical University, Tianjin, People’s Republic of China
Xinrong Dong
Affiliation:
Nutritional Epidemiology Institute and School of Public Health, Tianjin Medical University, Tianjin, People’s Republic of China
Shaomei Sun
Affiliation:
Health Management Centre, Tianjin Medical University General Hospital, Tianjin, People’s Republic of China
Xing Wang
Affiliation:
Health Management Centre, Tianjin Medical University General Hospital, Tianjin, People’s Republic of China
Ming Zhou
Affiliation:
Health Management Centre, Tianjin Medical University General Hospital, Tianjin, People’s Republic of China
Qiyu Jia
Affiliation:
Health Management Centre, Tianjin Medical University General Hospital, Tianjin, People’s Republic of China
Kun Song
Affiliation:
Health Management Centre, Tianjin Medical University General Hospital, Tianjin, People’s Republic of China
Yan Borné
Affiliation:
Nutritional Epidemiology, Department of Clinical Sciences Malmö, Lund University, Malmö, Sweden
Emily Sonestedt
Affiliation:
Nutritional Epidemiology, Department of Clinical Sciences Malmö, Lund University, Malmö, Sweden
Lu Qi*
Affiliation:
Department of Epidemiology, School of Public Health and Tropical Medicine, Tulane University, New Orleans, LA, USA Department of Nutrition, Harvard T.H. Chan School of Public Health, Boston, MA, USA
Kaijun Niu*
Affiliation:
Nutrition and Radiation Epidemiology Research Center, Institute of Radiation Medicine, Chinese Academy of Medical Sciences & Peking Union Medical College, Tianjin, People’s Republic of China Nutritional Epidemiology Institute and School of Public Health, Tianjin Medical University, Tianjin, People’s Republic of China Health Management Centre, Tianjin Medical University General Hospital, Tianjin, People’s Republic of China Tianjin Key Laboratory of Environment, Nutrition and Public Health, Tianjin, People’s Republic of China Center for International Collaborative Research on Environment, Nutrition and Public Health, Tianjin, People’s Republic of China
*
* Corresponding authors: Lu Qi, email lqi1@tulane.edu; Kaijun Niu, email nkj0809@gmail.com
* Corresponding authors: Lu Qi, email lqi1@tulane.edu; Kaijun Niu, email nkj0809@gmail.com
Rights & Permissions [Opens in a new window]

Abstract

It has been suggested that added sugar intake is associated with non-alcoholic fatty liver disease (NAFLD). However, previous studies only focused on sugar-sweetened beverages; the evidence for associations with total added sugars and their sources is scarce. This study aimed to examine the associations of total added sugars, their physical forms (liquid v. solid) and food sources with risk of NAFLD among adults in Tianjin, China. We used data from 15 538 participants, free of NAFLD, other liver diseases, CVD, cancer or diabetes at baseline (2013–2018 years). Added sugar intake was estimated from a validated 100-item FFQ. NAFLD was diagnosed by ultrasonography after exclusion of other causes of liver diseases. Multivariable Cox proportional hazards models were fitted to calculate hazard ratios (HR) and corresponding 95 % CI for NAFLD risk with added sugar intake. During a median follow-up of 4·2 years, 3476 incident NAFLD cases were documented. After adjusting for age, sex, BMI and its change from baseline to follow-up, lifestyle factors, personal and family medical history and overall diet quality, the multivariable HR of NAFLD risk were 1·18 (95 % CI 1·06, 1·32) for total added sugars, 1·20 (95 % CI 1·08, 1·33) for liquid added sugars and 0·96 (95 % CI 0·86, 1·07) for solid added sugars when comparing the highest quartiles of intake with the lowest quartiles of intake. In this prospective cohort of Chinese adults, higher intakes of total added sugars and liquid added sugars, but not solid added sugars, were associated with a higher risk of NAFLD.

Type
Research Article
Copyright
© The Author(s), 2022. Published by Cambridge University Press on behalf of The Nutrition Society

Non-alcoholic fatty liver disease (NAFLD) is emerging as the leading chronic liver disease worldwide with an estimated prevalence of 25–30 % of adults in many countries(Reference Powell, Wong and Rinella1,Reference Loomba, Friedman and Shulman2) . The prevalence of NAFLD was estimated to be 29·8 % in China(Reference Lee and Wong3). NAFLD has traditionally been viewed as a liver disease with a high risk of developing liver-related disorders such as liver fibrosis, cirrhosis and hepatocellular carcinoma. However, convincing evidence indicates that NAFLD is a multisystem disease(Reference Targher, Tilg and Byrne4), related to increased risks of CVD, type 2 diabetes and all-cause mortality(Reference Targher, Tilg and Byrne4). Currently, there are no approved pharmacotherapies for the management of NAFLD(Reference Petroni, Brodosi and Bugianesi5). Thus, the identification of potentially modifiable risk factors (e.g. diet) for NAFLD becomes highly relevant for prevention and management of this condition.

Recently, the influence of sugar intake on human health has attracted growing scientific and media attention(Reference Tybor, Beauchesne and Niu6). Sugars naturally exist in foods in low amounts, such as fruit and vegetables, and are usually consumed with dietary fibre and other healthy nutrients. Therefore, these types of sugars are not a major concern because they are within the food matrix. However, concerns arise with added sugars, which are defined as all sugars used in prepared or processed foods or added at the table(Reference Bowman7). While added sugars have been consistently related to the detrimental effects on liver health(Reference Jensen, Abdelmalek and Sullivan8,Reference Ma, Karlsen and Chung9) , the previous studies primarily focused on NAFLD and sugar-sweetened beverages(Reference Asgari-Taee, Zerafati-Shoae and Dehghani10Reference Park, Yiannakou and Petersen13), rather than added sugars per se. The study of added sugar intake and NAFLD may help to understand the role of added sugars in contributing to the epidemic of NAFLD. In addition, different food sources of added sugars may vary in composition, energy density and absorption. Thus, it is important to distinguish between different sources of added sugars when studying their associations with NAFLD. Further, considering the sources of added sugars is meaningful to inform dietary guidance. On the other hand, emerging evidence has suggested that liquid and solid sugars are metabolised differently, and thus they might have a distinct effect on the risk of NAFLD(Reference Eng and Estall14). However, to our knowledge, evidence on the association between added sugar intake per se and the risk of NAFLD is lacking.

To address this research gap, we prospectively investigated the associations of added sugars, their physical forms (liquid and solid) and food sources with the risk of NAFLD in a general Chinese adult population.

Methods

Study population

The Tianjin Chronic Low-grade Systemic Inflammation and Health (TCLSIH) cohort study is a prospective dynamic cohort study established in Tianjin, China. Details of the study design have been published previously(Reference Zhang, Gan and Zhang15). In brief, participants (aged ≥ 18 years) were randomly selected since May 2013 from the general population residing in Tianjin. All participants were invited to undergo in-person examinations and a detailed interview on diet and lifestyle factors at baseline. Follow-up information was obtained from annual health examinations and periodical re-interviews. This cohort study was approved by the Institutional Review Board of Tianjin Medical University (approval number: TMUhMEC 201430), and written informed consent was obtained from all participants prior to participation. This study was conducted according to the guidelines laid down in the Declaration of Helsinki and all procedures involving human subjects/patients.

The current study included participants who were recruited from May 2013 to December 2018 (baseline) and followed up until 31 December 2019. Of 28 445 individuals eligible for participation, we excluded participants with CVD, cancer or diabetes at baseline. Participants were defined as having diabetes if they had a fasting blood glucose ≥ 7·0 mmol/l or self-reported history of diabetes. In addition, we excluded those with alcoholic fatty liver disease, NAFLD or other liver diseases at baseline. Ultimately, a total of 15 538 participants were included in this analysis (Fig. 1).

Fig. 1. Flow chart for analysis cohort study creation.

Ascertainment of outcome

At both baseline and subsequent follow-up examinations, hepatic steatosis was tested by abdominal ultrasonography, which was carried out by experienced sonographers using a TOSHIBA SSA-660A ultrasound machine (Toshiba), with a 2–5 MHz curved array probe. Fatty liver disease was defined as having two or more of the following abnormal findings of liver ultrasonography: ‘bright liver’, vascular blurring and deep attenuation of the ultrasound signal(Reference Farrell, Chitturi and Lau16). NAFLD was defined as the presence of fatty liver disease without significant alcohol consumption (≥ 210 g/week for men and ≥ 140 g/week for women) and/or any other causes for secondary hepatic fat accumulation(Reference Fan, Wei and Zhuang17). The outcome was a first-time NAFLD that occurred during the follow-up period, as described previously(Reference Zhang, Meng and Zhang18).

Assessments of dietary intake and added sugar intake

At the baseline visit, dietary intake was collected using a validated 100-item self-administered FFQ(Reference Zhang, Gu and Bian12). Participants were inquired how often, on average, they consumed a standard portion of foods in the previous month. There were seven response categories ranged from ‘never eating’ to ‘more than two times per day’ for foods and eight choices ranged from ‘never drinking’ to ‘more than four times per day’ for beverages. Daily intakes of total energy and nutrients were calculated based on the 2009 Chinese Food Composition Table. Since added sugar value was not available in the Chinese Food Composition Table database, estimates of the added sugar content of each sugary food and beverage in the FFQ were derived from the US Department of Agriculture nutrient database(19). The reproducibility and validity of the FFQ have been described in detail elsewhere(Reference Zhang, Wu and Bian20). Briefly, Spearman rank correlation coefficients between the two FFQ collected approximately 3 months apart were 0·68 for energy intake and 0·62–0·79 for added sugars, fruits, vegetables and beverages. Spearman’s rank correlations between the FFQ and four non-consecutive 4-d weighed dietary records, also known as the FFQ validity correlation coefficients, were 0·49 for energy intake and 0·35–0·54 for nutrients. We created a healthy diet score by combining five common elements of healthy dietary patterns, which included vegetables, fruits, fish, unprocessed red meat and processed meat(Reference Zhang, Gu and Bian12,Reference English, Ard and Bailey21) . The score ranged from 0 to 5, with a higher score indicating a healthier diet. To characterise overall dietary patterns, we employed exploratory factor analysis to derive three main dietary patterns: sugar rich dietary pattern, vegetable rich dietary pattern and animal food dietary pattern, as described in our recent study(Reference Zhang, Gu and Bian22).

In the present study, exposure variables included total added sugars, liquid added sugar, solid added sugar and added sugars from different foods. We classified sugary foods and beverages on the FFQ into seven subgroups: bread, yogurt, desserts (including cakes, cookies, pastries, ice cream and other Chinese desserts), candies/jam, soft drinks, fruit/vegetable drinks and coffee. The relative contribution of each food source to added sugar intake is presented in online Supplementary Fig. 1.

Assessment of covariates

Information on age, sex, educational level, occupation, monthly family income, smoking status and alcohol drinking status was obtained from structured questionnaires completed by participants upon enrolment in the study. Personal and family medical history was determined by self-reported survey responses at baseline. Height and waist circumference were measured to the nearest 0·1 cm, while body weight was measured to the nearest 0·1 kg. BMI was calculated as weight (kg) divided by height (m) squared. Physical activity in the most recent week and sedentary time per day were measured using the International Physical Activity Questionnaire short version(Reference Craig, Marshall and Sjostrom23). Physical activity then was calculated as the weekly metabolic equivalent hours.

Fasting blood glucose and lipid profiles were analysed using standardised laboratory procedures. Participants were defined as hyperlipidaemia if they met any of the following conditions: total cholesterol ≥ 5·17 mmol/l, TAG ≥ 1·7 mmol/l, LDL-cholesterol ≥ 3·37 mmol/l or taking lipid-lowering drugs. Blood pressure was measured at least twice using an electronic sphygmomanometer (TM-2655, A&D), and the mean of the last two measurements was calculated. Hypertension was defined as systolic blood pressure ≥ 140 mmHg and/or diastolic blood pressure ≥ 90 mmHg and/or having a history of hypertension. Depressive symptoms were determined by the Chinese version of the Zung Self-Rating Depression Scale(Reference Lee, Chiu and Wing24), and the scale scores ≥ 45 were considered to have depressive symptoms.

Statistical analysis

The quantile–quantile plot was used to examine the normal distribution of continuous variables. Baseline characteristics were presented as medians (interquartile ranges) for all continuous variables due to their non-normal distributions and as percentages for categorical variables. Differences in the distribution of baseline covariates according to NAFLD status were assessed using Wilcoxon rank sum test for continuous variables and χ 2 test for categorical variables. To assess bias from loss to follow-up, we compared the difference between the final analytic sample and those who were lost to follow-up.

Person time for follow-up was computed from the baseline survey to the date of the first occurrence of NAFLD or the end of follow-up (31 December 2019), whichever occurred first. The crude NAFLD incidence rates were calculated by taking the number of NAFLD cases and that by the sum of the person-years and were expressed as incidence per 1000 person-years. Cox proportional hazards regression models with follow-up as time scale were performed to estimate hazard ratios (HR) and their 95 % CI for the association between quartile of added sugar intake and risk of NAFLD. The proportional hazards assumption was checked by including time-by-covariate interaction terms in the Cox proportional hazards model, and all variables satisfied the conditions (all P > 0·05). Model adjustments were done in three sets. Model 1 was adjusted for age, sex and BMI. Model 2 was additionally adjusted for education, occupation, monthly family income, smoking status, alcohol drinking status, family medical history (including CVD, hypertension, hyperlipidaemia and diabetes), hypertension, hyperlipidaemia, depressive symptoms, physical activity, sedentary time, total energy intake and healthy diet score. Model 3 was identical to model 2, except that it was adjusted for BMI change from baseline to follow-up to correct for the possible effect of weight change on the risk of developing NAFLD.

In sensitivity analyses, we adjusted for sugar rich dietary pattern, vegetable rich dietary pattern and animal food dietary pattern instead of the self-calculated healthy diet score. Furthermore, to reduce the possibility of reverse causality, we excluded incident NAFLD cases that occurred within the first 2 years of follow-up.

All statistical analyses were performed using SAS statistical software, version 9.4 (SAS Institute Inc.). The significance level was set at a two-sided 0·05.

Results

The mean added sugar intake was 20·8 g/d for all participants. Among the added sugar sources, desserts (48 %) and yogurt (19 %) contributed the largest share, followed by soft drinks (12 %) and fruit/vegetable drinks (12 %).

Table 1 shows the baseline characteristics of the study population. Participants with incident NAFLD were older, were more likely to be men and have metabolic risk factors, such as higher baseline BMI, waist circumference and fasting blood glucose, than those without incident NAFLD. Furthermore, those with incident NAFLD were more likely to drink alcohol and smoke and had a higher proportion of personal and family history of disease. Moreover, participants included in this study were more likely to be younger, tended to be men, had a better metabolic profile and were more likely to have healthier lifestyles and higher socio-economic status than those who were lost to follow-up (online Supplementary Table 1).

Table 1. Baseline characteristics of the study participants by incident NAFLD status (n 15 538)*

(Percentages; medians and interquartile ranges)

ALT, alanine aminotransferase; DBP, diastolic blood pressure; FPG, fasting plasma glucose; MET, metabolic equivalent; NAFLD, non-alcoholic fatty liver disease; PA, physical activity; SBP, systolic blood pressure; TC, total cholesterol; WC, waist circumference.

* Continuous variables are presented as medians (interquartile ranges) and categorical variables as percentages.

Wilcoxon rank sum tests for continuous variables or χ 2 tests for categorical variables.

During the 54 081 person-years of follow-up (median follow-up time: 4·2 years), 3476 participants received a diagnosis of NAFLD (incident rate 64 per 1000 person-years). Table 2 displays the HR (95 % CI) of the association between added sugar intake and risk of NAFLD. Total added sugar intake and liquid added sugar intake were significantly and positively associated with the risk of NAFLD. After adjusting for age, sex and baseline BMI, the multivariable HR of NAFLD risk for the highest compared with the lowest quartile of intakes were 1·15 (95 % CI 1·05, 1·27) for total added sugar intake and 1·19 (95 % CI 1·09, 1·31) for liquid added sugar intake, respectively. By contrast, solid added sugar intake was not associated with the risk of NAFLD (Q4 v. Q1 HR = 0·98; 95 % CI 0·89, 1·08). Additional adjustment for socio-economic status, lifestyle factors, medical history, diet quality and BMI change did not appreciably change the association estimates. In the fully adjusted models, comparing the upper quartiles to the lower quartiles of exposure, the increased relative risks of NAFLD were 18 % for total added sugar intake (HR = 1·18; 95 % CI 1·06, 1·32) and 20 % for liquid added sugar intake (HR = 1·20; 95 % CI 1·08, 1·33), respectively. In contrast, there was no significant association between solid added sugar intake and the risk of NAFLD (HR = 0·96; 95 % CI 0·86, 1·07).

Table 2. Association between added sugar intake and risk of non-alcoholic fatty liver disease (n 15 538)*

(Hazard ratios and 95 % confidence intervals)

Model 1: adjusted for age (continuous; years), sex (categorical; men or women) and baseline BMI (continuous; kg/m2).

Model 2: additionally adjusted for smoking status (categorical; current smoker, ex-smoker or non-smoker), alcohol drinking status (categorical; everyday drinker, sometime drinker, ex-drinker or non-drinker), educational level (categorical: < or ≥ college graduate), occupation (categorical; managers, professionals and other), annual family income (categorical: < or ≥ 10 000 Yuan), physical activity (continuous; MET-h/week), family history of disease (including CVD, hypertension, hyperlipidaemia and diabetes (each yes or no)), hypertension (yes or no), hyperlipidaemia (yes or no), depressive symptoms (score < or ≥ 45), sedentary time (hour/d), total energy intake (kcal/d) and healthy diet score (0, 1, 2, 3, 4 or 5).

Model 3: additionally adjusted for BMI change from baseline to follow-up (continuous; kg/m2). For liquid and solid added sugars, mutual adjustment was conducted in model 3.

* Values are hazard ratios (95 % CI) unless otherwise indicated.

Test for trend based on variable containing median value for each quartile.

Median (range) intake.

Figure 2 presents the associations between added sugar intake from food sources and risk of NAFLD. Only added sugar from soft drinks was significantly associated with the risk of NAFLD, with an HR of 1·05 (95 % CI 1·03, 1·08) for each additional 5 g daily, P < 0·0001. Added sugars from bread, yogurt, desserts, candies, fruit/vegetable drinks and coffee were not significantly associated with the risk of NAFLD (P ≥ 0·09).

Fig. 2. Associations between added sugars from different food sources and the risk of non-alcoholic fatty liver disease. The multivariable Cox model included age, sex, baseline BMI, smoking status, alcohol drinking status, educational level, occupation, annual family income, physical activity, family history of disease (including CVD, hypertension, hyperlipidaemia and diabetes), hypertension, hyperlipidaemia, depressive symptoms, sedentary time, total energy intake, healthy diet score, BMI change from baseline to follow-up and intake of the other added sugar sources.

In sensitivity analyses, adjustment for sugar rich dietary pattern, vegetable rich dietary pattern and animal food dietary pattern produced comparable results with those obtained from the primary analyses (data not shown). In addition, the exclusion of NAFLD cases that occurred within the first 2 years of follow-up did not materially change the associations of added sugars with risk of NAFLD (online Supplementary Table 2).

Discussion

In this large-scale prospective cohort study, we found that intakes of total added sugars and liquid added sugars were significantly associated with an increased risk of NAFLD. There was no association observed between solid added sugar intake and the risk of NAFLD. In addition, only added sugar from soft drinks was associated with an increased risk of NAFLD. No association was observed for other sources of added sugars.

In our study population, the mean intake of added sugars was 20·8 g/d, which was similar to the estimate of 18·8 g/d from the Chinese Nutrition and Health Surveillance in 2012(Reference Su25). The 2016 Chinese Dietary Guidelines recommended that added sugar intake should be less than 50 g/d and preferably less than 25 g/d. Our findings indicated that added sugar intake in this population was below the Chinese Dietary Guidelines. Further, the added sugar intake level in our study was much lower than those of the Western countries. For example, the mean intake of added sugars in the US population was 67·8 g/d in 2017–2018(Reference Ricciuto, Fulgoni and Gaine26), while the mean added sugar intake in the Australian population was 60·3 g/d in 2011–2012(Reference Lei, Rangan and Flood27).

Regarding sources of added sugars, we found that the top sources of added sugars were desserts and yogurt, and followed by soft drinks and fruit/vegetable drinks, consistent with China National survey data(Reference Liu, Yu and Guo28). In contrast, added sugars in the USA(Reference Ricciuto, Fulgoni and Gaine26) and other Western countries(Reference Lei, Rangan and Flood27) primarily came from soft drinks. Collectively, these findings indicate that our study population (adult Chinese population) and Western populations have different food sources for added sugar intake.

Our study is the first to investigate the association between added sugar intake per se and the risk of NAFLD. The results showed a positive association between total added sugar intake and the risk of NAFLD. This finding was supported by data from the US National Health and Nutrition Examination Survey that indicated the parallel rise in NAFLD prevalence and added sugar consumption in the periods from 1988 to 2012(Reference Jensen, Abdelmalek and Sullivan8,Reference Ruhl and Everhart29) . However, the consumption of added sugars was very low in China when compared with the US population(Reference Ricciuto, Fulgoni and Gaine26). Indeed, only 7·7 % of individuals consumed more than 50 g/d of added sugars in our study population. Thus, our study provides novel evidence that total added sugar intake is positively associated with incident NAFLD among adults with low consumption of added sugars.

Furthermore, added sugars in the liquid form were associated with an increased risk of NAFLD, whereas added sugars in the solid form were not. This discrepancy between physical forms of added sugars has been found in previous studies on sugar and health outcomes. For example, a cross-sectional study of adults in the East of England showed that added sugar intake in a liquid form was associated with higher levels of chronic inflammation, one of the potential mechanisms associated with NAFLD, but sugars from solids were not associated(Reference O’Connor, Imamura and Brage30). In addition, a cohort study found that liquid sugar intake was associated with an increased risk of overall cancer incidence and mortality and all-cause mortality, but intake of sugars in solid form was unrelated to these outcomes(Reference Laguna, Alegret and Cofan31). Likewise, results from two Swedish population-based prospective cohorts also indicated that liquid added sugars in sweetened beverages were positively associated with mortality, whereas solid added sugars in treats were inversely associated(Reference Ramne, Alves Dias and Gonzalez-Padilla32). Taken together, these findings highlight that added sugars in liquid v. solid forms may impact metabolic diseases differently(Reference Pan and Hu33). One possible explanation for this difference may be that added sugars in liquid form are digested more rapidly than added sugars in solid foods; thus, quickly absorbable added sugars in liquid form increase the rate of hepatic extraction of fructose, de novo lipogenesis and production of lipids(Reference Pan and Hu33). Another possible explanation may be that liquid added sugars lead to less satiety than added sugars from solid foods(Reference DiMeglio and Mattes34). This may give rise to a positive energy balance and subsequent hepatic fat accumulation(Reference Eng and Estall14,Reference DiMeglio and Mattes34) .

Our study for the first time investigated the associations between different food sources of added sugars and risk of NAFLD. The results indicated that only added sugars from soft drinks were significantly associated with the risk of NAFLD. In line with our study, previous studies consistently showed that sugar-sweetened soft drink consumption was associated with a higher risk of NAFLD(Reference Asgari-Taee, Zerafati-Shoae and Dehghani10Reference Park, Yiannakou and Petersen13). Furthermore, we did not observe significant associations between other food sources of added sugars and risk of NAFLD. Such findings suggested that different food sources of added sugars were differently associated with the risk of NAFLD. Further, our findings support the benefits of public health campaigns to reduce soft drink intake. At the same time, this also highlights that dietary recommendations without considering specific sources of added sugars may be simplistic and insufficient.

Consistent with our previous study(Reference Zhang, Gu and Lu35), the current study showed that those with incident NAFLD were more physically active (Table 1). The reason is that men (geometric mean of physical activity was 12·0 (95 % CI 11·7, 12·4)) were more physically active than women (geometric mean of physical activity was 8·58 (95 % CI 8·35, 8·81)), and NAFLD patients were more likely to be men; thus, sex is an important factor that explains the fact that NAFLD patients are more physically active than controls. Indeed, the sex-adjusted geometric means of physical activity were 9·87 (95 % CI 9·65, 10·1) in controls and 9·94 (95 % CI 9·51, 10·4) in NAFLD patients, respectively (P value = 0·79).

The strengths of the present study include the large sample size, prospective design, objective assessment of NAFLD and careful adjustments for a wide range of lifestyle factors. In addition, the current study for the first time prospectively investigated the associations between added sugars and their forms/sources and the risk of NAFLD.

There are also several limitations to our study. First, the added sugar intake in this study was estimated using the US Department of Agriculture database due to the lack of sugar data in the Chinese Food Composition Tables. However, the estimated intakes were similar to those reported in the Chinese Nutrition and Health Survey(Reference Su25,Reference Liu, Yu and Guo28) . Second, dietary intake was measured only at baseline. However, we had excluded participants with CVD, cancer and diabetes at baseline from the current study because these diseases could result in significant dietary changes(Reference Hu, Stampfer and Rimm36). Thus, the vast majority of participants would not have changed their diet intake in such a relatively short follow-up period (median 4·2 years). Third, the diagnosis of NAFLD was conducted by abdominal ultrasound instead of liver biopsy, which is the gold diagnostic standard for NAFLD. However, abdominal ultrasound is not invasive and easy to accept by participants. Moreover, this non-invasive method is widely used in large-scale population-based studies due to its high sensitivity and specificity(Reference Hernaez, Lazo and Bonekamp37). Fourth, although we adjusted for a long list of risk factors including sedentary time, depressive symptoms and BMI change, unmeasured or residual confounding cannot be entirely ruled out. Finally, as our cohort is comprised of relatively healthy Chinese adults with low added sugar consumption, our results may not be generalisable to other ethnic populations. In addition, more than 10 % of participants lost to follow-up might bias our results.

Conclusions

In conclusion, this large prospective cohort study indicates that total added sugars and liquid added sugars, especially added sugars from soft drinks, were associated with an increased risk of NAFLD. Our findings suggest that reducing the consumption of added sugars, mainly from soft drinks, could be an effective preventive measure to prevent NAFLD.

Supplementary material

For supplementary material accompanying this paper visit https://doi.org/10.1017/S000711452200277X

Acknowledgements

The authors thank all the participants that have made this study.

This study was supported by grants from the National Natural Science Foundation of China (no. 81941024 and 81872611),  Tianjin Major Public Health Science and Technology Project (No. 21ZXGWSY00090), National Health Commission of China (No. SPSYYC 2020015), SZ got support from the China Scholarship Council (no. 202006940030), 2014 and 2016 Chinese Nutrition Society (CNS) Nutrition Research Foundation—DSM Research Fund (Nos. 2016-046, 2014-071 and 2016-023), China.

S. Z., L. Q. and K. N. conceptualised and designed the study. S. Z. performed the data analysis and drafted the manuscript. S. Z., G. M., Q. Z., L. H., H. W., Y. G., T. Z., X. W., J. Z., J. D., X. Z., Z. C., X. Z., X. D., S. S., X. W., M. Z., Q. J., K. S., Y. B., E. S., L. Q. and K. N. contributed to analysis and interpretation of data. L. Q. and K. N. contributed to the revision of the manuscript and approved the final draft. K. N. and S. Z obtained funding for the study. L. Q. and K. N. were involved in study supervision. All authors contributed to the intellectual content, critical revisions to the drafts of the paper and approved the final version.

The authors declared they have no conflicts of interest.

References

Powell, EE, Wong, VW & Rinella, M (2021) Non-alcoholic fatty liver disease. Lancet 397, 22122224.CrossRefGoogle ScholarPubMed
Loomba, R, Friedman, SL & Shulman, GI (2021) Mechanisms and disease consequences of nonalcoholic fatty liver disease. Cell 184, 25372564.CrossRefGoogle ScholarPubMed
Lee, HW & Wong, VW (2019) Changing NAFLD epidemiology in China. Hepatology 70, 10951098.CrossRefGoogle ScholarPubMed
Targher, G, Tilg, H & Byrne, CD (2021) Non-alcoholic fatty liver disease: a multisystem disease requiring a multidisciplinary and holistic approach. Lancet Gastroenterol Hepatol 6, 578588.CrossRefGoogle ScholarPubMed
Petroni, ML, Brodosi, L, Bugianesi, E, et al. (2021) Management of non-alcoholic fatty liver disease. BMJ 372, m4747.CrossRefGoogle ScholarPubMed
Tybor, DJ, Beauchesne, AR, Niu, R, et al. (2018) An evidence map of research linking dietary sugars to potentially related health outcomes. Curr Dev Nutr 2, nzy059.CrossRefGoogle ScholarPubMed
Bowman, SA (2017) Added sugars: definition and estimation in the USDA food patterns equivalents databases. J Food Compos Anal 64, 6467.CrossRefGoogle Scholar
Jensen, T, Abdelmalek, MF, Sullivan, S, et al. (2018) Fructose and sugar: a major mediator of non-alcoholic fatty liver disease. J Hepatol 68, 10631075.CrossRefGoogle Scholar
Ma, J, Karlsen, MC, Chung, M, et al. (2016) Potential link between excess added sugar intake and ectopic fat: a systematic review of randomized controlled trials. Nutr Rev 74, 1832.CrossRefGoogle ScholarPubMed
Asgari-Taee, F, Zerafati-Shoae, N, Dehghani, M, et al. (2019) Association of sugar sweetened beverages consumption with non-alcoholic fatty liver disease: a systematic review and meta-analysis. Eur J Nutr 58, 17591769.CrossRefGoogle ScholarPubMed
Chen, H, Wang, J, Li, Z, et al. (2019) Consumption of sugar-sweetened beverages has a dose-dependent effect on the risk of non-alcoholic fatty liver disease: an updated systematic review and dose-response meta-analysis. Int J Environ Res Public Health 16, 2192.CrossRefGoogle Scholar
Zhang, S, Gu, Y, Bian, S, et al. (2021) Soft drink consumption and risk of nonalcoholic fatty liver disease: results from the Tianjin Chronic Low-Grade Systemic Inflammation and Health (TCLSIH) cohort study. Am J Clin Nutr 113, 12651274.CrossRefGoogle ScholarPubMed
Park, WY, Yiannakou, I, Petersen, JM, et al. (2021) Sugar-sweetened beverage, diet soda, and non-alcoholic fatty liver disease over 6 years: the Framingham heart study. Clin Gastroenterol Hepatol (In Press)Google ScholarPubMed
Eng, JM & Estall, JL (2021) Diet-induced models of non-alcoholic fatty liver disease: food for thought on sugar, fat, and cholesterol. Cells 10, 1805.CrossRefGoogle ScholarPubMed
Zhang, S, Gan, S, Zhang, Q, et al. (2022) Ultra-processed food consumption and the risk of non-alcoholic fatty liver disease in the Tianjin Chronic Low-grade Systemic Inflammation and Health Cohort Study. Int J Epidemiol 51, 237249.CrossRefGoogle ScholarPubMed
Farrell, GC, Chitturi, S, Lau, GK, et al. (2007) Guidelines for the assessment and management of non-alcoholic fatty liver disease in the Asia-Pacific region: executive summary. J Gastroenterol Hepatol 22, 775777.CrossRefGoogle ScholarPubMed
Fan, JG, Wei, L, Zhuang, H, et al. (2019) Guidelines of prevention and treatment of nonalcoholic fatty liver disease (2018, China). J Dig Dis 20, 163173.CrossRefGoogle ScholarPubMed
Zhang, S, Meng, G, Zhang, Q, et al. (2021) Consumption of preserved egg is associated with modestly increased risk of nonalcoholic fatty liver disease in Chinese adults. J Nutr 151, 27412748.CrossRefGoogle ScholarPubMed
( US) NDL (2006) USDA Database for the Added Sugars Content of Selected Foods (Release 1.) https://permanent.fdlp.gov/lps86777/addsug01.pdf.https://permanent.fdlp.gov/lps86777/addsug01.pdf (accessed November 2021).Google Scholar
Zhang, S, Wu, X, Bian, S, et al. (2021) Association between consumption frequency of honey and non-alcoholic fatty liver disease: results from a cross-sectional analysis based on the Tianjin Chronic Low-grade Systemic Inflammation and Health (TCLSIH) Cohort Study. Br J Nutr 125, 712720.CrossRefGoogle ScholarPubMed
English, LK, Ard, JD, Bailey, RL, et al. (2021) Evaluation of dietary patterns and all-cause mortality: a systematic review. JAMA Netw Open 4, e2122277.CrossRefGoogle ScholarPubMed
Zhang, S, Gu, Y, Bian, S, et al. (2021) Dietary patterns and risk of non-alcoholic fatty liver disease in adults: a prospective cohort study. Clin Nutr 40, 53735382.CrossRefGoogle ScholarPubMed
Craig, CL, Marshall, AL, Sjostrom, M, et al. (2003) International physical activity questionnaire: 12-country reliability and validity. Med Sci Sports Exerc 35, 13811395.CrossRefGoogle ScholarPubMed
Lee, HC, Chiu, HF, Wing, YK, et al. (1994) The Zung self-rating Depression Scale: screening for depression among the Hong Kong Chinese elderly. J Geriatr Psychiatr Neurol 7, 216220.CrossRefGoogle ScholarPubMed
Su, L (2016) The Consumption Status and Changes of Sugar-Sweetened Food among Chinese Residents from 2002 to 2012 and its Relationship with Overweight and Obesity. Beijing, China: Chinese Center for Disease Control and Prevention.Google Scholar
Ricciuto, L, Fulgoni, VL III, Gaine, PC, et al. (2021) Sources of added sugars intake among the U.S. population: analysis by selected sociodemographic factors using the National Health and Nutrition Examination Survey 2011–2018. Front Nutr 8, 687643.CrossRefGoogle Scholar
Lei, L, Rangan, A, Flood, VM, et al. (2016) Dietary intake and food sources of added sugar in the Australian population. Br J Nutr 115, 868877.CrossRefGoogle ScholarPubMed
Liu, S, Yu, D, Guo, Q, et al. (2016) Consumption status and trend of added sugar containing food among Chinese from 2002 to 2012. Wei Sheng Yan Jiu 45, 398401.Google ScholarPubMed
Ruhl, CE & Everhart, JE (2015) Fatty liver indices in the multiethnic United States National Health and Nutrition Examination Survey. Aliment Pharmacol Ther 41, 6576.CrossRefGoogle ScholarPubMed
O’Connor, L, Imamura, F, Brage, S, et al. (2018) Intakes and sources of dietary sugars and their association with metabolic and inflammatory markers. Clin Nutr 37, 13131322.CrossRefGoogle ScholarPubMed
Laguna, JC, Alegret, M, Cofan, M, et al. (2021) Simple sugar intake and cancer incidence, cancer mortality and all-cause mortality: a cohort study from the PREDIMED trial. Clin Nutr 40, 52695277.CrossRefGoogle ScholarPubMed
Ramne, S, Alves Dias, J, Gonzalez-Padilla, E, et al. (2019) Association between added sugar intake and mortality is nonlinear and dependent on sugar source in 2 Swedish population-based prospective cohorts. Am J Clin Nutr 109, 411423.CrossRefGoogle ScholarPubMed
Pan, A & Hu, FB (2011) Effects of carbohydrates on satiety: differences between liquid and solid food. Curr Opin Clin Nutr Metab Care 14, 385390.CrossRefGoogle ScholarPubMed
DiMeglio, DP & Mattes, RD (2000) Liquid v. solid carbohydrate: effects on food intake and body weight. Int J Obes Relat Metab Disord 24, 794800.CrossRefGoogle Scholar
Zhang, S, Gu, Y, Lu, M, et al. (2020) Association between edible mushroom intake and the prevalence of newly diagnosed non-alcoholic fatty liver disease: results from the Tianjin Chronic Low-Grade Systemic Inflammation and Health Cohort Study in China. Br J Nutr 123, 104112.CrossRefGoogle ScholarPubMed
Hu, FB, Stampfer, MJ, Rimm, E, et al. (1999) Dietary fat and coronary heart disease: a comparison of approaches for adjusting for total energy intake and modeling repeated dietary measurements. Am J Epidemiol 149, 531540.CrossRefGoogle ScholarPubMed
Hernaez, R, Lazo, M, Bonekamp, S, et al. (2011) Diagnostic accuracy and reliability of ultrasonography for the detection of fatty liver: a meta-analysis. Hepatology 54, 10821090.CrossRefGoogle ScholarPubMed
Figure 0

Fig. 1. Flow chart for analysis cohort study creation.

Figure 1

Table 1. Baseline characteristics of the study participants by incident NAFLD status (n 15 538)*(Percentages; medians and interquartile ranges)

Figure 2

Table 2. Association between added sugar intake and risk of non-alcoholic fatty liver disease (n 15 538)*(Hazard ratios and 95 % confidence intervals)

Figure 3

Fig. 2. Associations between added sugars from different food sources and the risk of non-alcoholic fatty liver disease. The multivariable Cox model included age, sex, baseline BMI, smoking status, alcohol drinking status, educational level, occupation, annual family income, physical activity, family history of disease (including CVD, hypertension, hyperlipidaemia and diabetes), hypertension, hyperlipidaemia, depressive symptoms, sedentary time, total energy intake, healthy diet score, BMI change from baseline to follow-up and intake of the other added sugar sources.

Supplementary material: File

Zhang et al. supplementary material

Zhang et al. supplementary material

Download Zhang et al. supplementary material(File)
File 139.5 KB