bs_bs_banner
doi:10.1111/jgh.12861
H E PAT O L O G Y
Association of weight gain since age 20 with non-alcoholic fatty liver disease in normal weight individuals Takeshi Kimura,*,† Gautam A Deshpande,† Kevin Y Urayama,†,‡ Katsunori Masuda,* Tsuguya Fukui† and Yutaka Matsuyama§ *Center for Preventive Medicine, St. Luke’s International University, †Center for Clinical Epidemiology, St. Luke’s Life Science Institute, ‡ Department of Human Genetics and Disease Diversity, Graduate School of Medical and Dental Sciences, Tokyo Medical and Dental University, and §Department of Biostatistics, School of Public Health, The University of Tokyo, Tokyo, Japan
Key words body weight gain, non-alcoholic steatohepatitis, prevention and control. Accepted for publication 22 November 2014. Correspondence Dr Takeshi Kimura, Center for Clinical Epidemiology, St. Luke’s Life Science Institute, 10-1 Akashi-cho, Chuo-ku, Tokyo 104-0044, Japan. Email:[email protected] Conflict of interest: None. Financial grants: None.
Abstract Background and Aims: Interventions for lifestyle diseases including non-alcoholic fatty liver disease (NAFLD) have focused on overweight and obese populations. The impact of adult weight gain on NAFLD development among normal weight individuals remains unclear. Methods: In this cross-sectional study, we collected data from participants presenting to a health check-up program. Ultrasound-diagnosed NAFLD prevalence was examined over 1-kg increments of weight change since age 20. Relative risks were calculated in men and women stratified by current weight (normal, overweight, and obese). Logistic regression was used to estimate odds ratios (ORs) and 95% confidence intervals (CIs) adjusting for potential confounders. Results: Among 21 496 participants, 3498 cases of NAFLD (16.3%) were observed. Prevalence of NAFLD increased with weight gain since age 20; among the 10.1–11.0 kg weight gain group, 41.6% of men and 24.8% of women had NAFLD. Multivariate analysis by quartiles showed that weight change was significantly associated with NAFLD risk in men and women. Risk of NAFLD associated with weight change (10-kg increments) was significantly higher in normal weight individuals (men: OR 7.53, 95% CI: 4.99–11.36, women: OR 12.20, 95% CI: 7.45–19.98) than overweight (men: OR 1.61, 95% CI: 0.91–2.85, women: OR 2.90, 95% CI: 0.99–8.54) and obese (men: OR 4.0, 95% CI: 2.97–5.39, women: OR 2.68, 95% CI: 2.00–3.60). Conclusions: NAFLD is robustly associated with weight change since age 20. This effect appears particularly strong in individuals at normal weight, suggesting an important role for early and longitudinal weight monitoring, even among healthy individuals at normal weight.
Introduction Because of the worldwide obesity pandemic,1,2 obesity-related diseases are anticipated to become an increasingly heavy health-care burden for many countries; non-alcoholic fatty liver disease (NAFLD) has emerged as an important global problem.3 The worldwide prevalence of NAFLD, estimated at 10–30% in the general population depending on population and diagnostic methods,4–11 is expected to rise in the coming years.12 NAFLD causes significant morbidity over time. Progression to non-alcoholic steatohepatitis portends increased rates of advanced fibrosis, cirrhosis, and hepatocellular carcinoma, carrying with them further morbidity and mortality burdens.13,14 Additionally, previous studies suggest that NAFLD contributes to an increased incidence of cardiovascular disease (CVD),12 metabolic syndrome
(MS), and type 2 diabetes (T2DM).15 Therefore, a better understanding of the epidemiology of NAFLD, including identification of modifiable risk factors and prevention strategies, has the potential to substantially improve public health outcomes on a large scale. With strong evidence linking obesity to NAFLD and other lifestyle-related disease, much of the primary prevention performed in clinical practice remains focused on overweight and obese populations.16 However, adult weight gain, as opposed to absolute body weight, has also been reported to be a risk factor for various lifestyle-related disease including T2DM,17 CVD,18,19 stroke,20 and chronic kidney disease.21 It remains unclear whether the dynamics of weight gain over time, regardless of body mass index (BMI), adversely influences progression to NAFLD, thus warranting attention for primary prevention strategies.
Journal of Gastroenterology and Hepatology 30 (2015) 909–917 © 2014 Journal of Gastroenterology and Hepatology Foundation and Wiley Publishing Asia Pty Ltd
909
Fatty liver in normal weight individuals
T Kimura et al.
Given established associations between lifestyle diseases and NAFLD, we hypothesized a similar relationship between weight gain during adulthood and development of NAFLD; there are no prior reports evaluating this association. In this study, we examined the relationship between weight gain since age 20 and NAFLD, utilizing comprehensive data from a large cohort of adults presenting for an annual health check-up. Further hypothesizing that normal weight individual may similarly be at risk for NAFLD in the setting of weight gain, data were stratified by current BMI to clarify this relationship.
Methods Study population. We conducted a cross-sectional study of apparently healthy participants undergoing a health screening exam in 2008–2009 at the Center for Preventive Medicine at St. Luke’s International Hospital in Tokyo, Japan. In Japan, mandatory annual health screenings are required for all employed persons; 80% of this sample was referred to our center via employer-sponsored programs, with 20% consisting of community self-referrals. Eligible participants were selected among a total of 21 710 men and 23 732 women. Per clinically relevant criteria which could potentially bias outcomes, individuals with any of the following were excluded: (i) missing clinical data; (ii) age < 30 years; (iii) > 2 health check-ups in 2008–2009; (iv) daily alcohol intake ≥ 20 g/day for men or ≥ 10 g/day for women; (v) positive or unknown hepatitis B or C serology; (vi) pregnancy; (vii) on medication for a hepatic condition, menopausal disorder, or metabolic disease including hypertension, hyperlipidemia, and diabetes mellitus; (viii) fasting blood glucose (FBG) ≥ 126 mg or hemoglobin A1c (HbA1c) ≥ 6.5%; (ix) history of surgery or medication for malignancy; and (x) currently or previously treated CVD. Comparison of excluded participants showed similar characteristics to those included in the study. The exclusion of 23 945 individuals resulted in a study population of 21 496 individuals (Fig. 1). The study was approved by the St. Luke’s International Hospital Institutional Review Board and was exempted from individual participants’ written informed consent. Data collection. Anthropometric measurements included height and weight, measured via digital weight scale, and waist circumference measured at the midpoint between the lowest rib margin and the iliac crest. Body weight at age 20 was collected via self-administered questionnaires. Weight change, the primary exposure variable, was defined as the difference between weight (kilograms) at 2008–2009 screening and at age 20. Clinical and demographic data were collected from all participants during the health screening process. Systolic (SBP) and diastolic (DBP) blood pressure, routine labs including FBG, HbA1c, total cholesterol, low-density lipoprotein-cholesterol (LDL-C), high-density lipoprotein-cholesterol (HDL-C), triglycerides (TG), uric acid (UA), aspartate aminotransferase (AST), alanine aminotransferase (ALT), and γ-glutamyltransferase (GGT) were collected. Serum samples were obtained after a 12-h fast. All participants completed self-administered questionnaires regarding health history and lifestyle habits including smoking (non-smoker/ ex-smoker/current smoker), alcohol intake (grams per week), 910
exercise habits (no exercise/activity 1–2 times per week/3–5 times per week/6–7 times per week), present illnesses and treatments, and past medical history. Responses were verified by a trained team of nurses via face-to-face interviews conducted with all participants as a standard part of the health check-up process. To preserve confidentiality, direct patient identifiers were not collected when creating the dataset. Diagnosis of NAFLD. All participants received abdominal ultrasound as a routine part of this health screening program. Diagnosis of fatty liver was performed by ultrasound, assessing four well-established criteria: enhanced liver echogenicity, echogenicity greater in liver than kidneys, deep attenuation, and vascular blurring.22 Participants were diagnosed with NAFLD by observing, at minimum, enhanced liver echogenicity and hepatorenal contrast. Examinations were performed by trained ultrasound technicians, with findings reviewed by experienced radiologists. Definition of overweight, obesity and MS. Proposed standard definitions of overweight and obesity vary by race and ethnicity in previous reports.23 For Asians, specific definitions of obesity have been proposed, although uniformly accepted criteria have yet to be determined.24 Consistent with a previous report,25 we used Asian-specific BMI cut-offs as recommended by the World Health Organization24,26 as follows: underweight, < 18.5; normal weight, 18.5–22.9; overweight, 23–24.9; and obese, ≥ 25. To define MS, we adopted International Diabetes Federation Task Force criteria27: (i) elevated waist circumference; (ii) elevated triglyceride (≥ 150 mg/mL); (iii) reduced HDL-C (male: < 40 mg/ dL, female: < 50 mg/dL); (iv) SBP > 130 or DBP > 85 mmHg; and (v) elevated fasting glucose (≥ 100 mg/dL). Diagnosis required individuals to meet at least three criteria. Statistical analysis. Variables were reported as mean (SD), median (interquartile range), or proportion (%) based on variable characteristics. To explore differences between groups, Student’s t-test, Mann–Whitney U-test, or χ2 test was performed based on appropriate distribution of data. The primary outcome was prevalence of NAFLD. In order to facilitate clinical interpretation, weight change since age 20 was evaluated in 1-kg increments and categories, with one weight loss category and four gender-specific weight gain quartiles. NAFLD prevalence was plotted over the range of 1-kg weight change increments. Relative risk (RR) of NAFLD by weight change (compared with no gain or loss group [−0.9 to 0 kg]) was calculated, stratified by current BMI and BMI at age 20. The difference in log (RR) between BMI categories was evaluated by Wald test. Adjusting for potential confounders, multivariate logistic regression of individuals with BMI > 18.5 was used to evaluate the association between NAFLD and weight change, calculating odds ratios (ORs) with 95% confidence intervals (CIs). We considered several potential confounders and present results for two models, the first of which was adjusted for basic demographic factors and lifestyle behaviors (model 1: age, BMI at age 20, light alcohol consumption, smoking status, and exercise). In a second model, we evaluated the influence of additional clinically relevant
Journal of Gastroenterology and Hepatology 30 (2015) 909–917 © 2014 Journal of Gastroenterology and Hepatology Foundation and Wiley Publishing Asia Pty Ltd
T Kimura et al.
Fatty liver in normal weight individuals
Participants of Annual Health Check-up of 2008–2009 Men: 21 710 Women: 23 731 • Data Missing Men: 2052 Women: 1413
Men: 19 658 Women: 22 318 • Recieving Health Check-up More than Twice in 2008 • Less than 30 years old Men: 789 Women: 731
Men: 18 869 Women: 21 587 • Pregnancy (Women) • Excluding Alcohol Intake Men ≥ 20 g/day, Women ≥ 10 g/day • HBsAg positive, HCVAb positive • Medical Treatment for Liver Diseases Men: 8167 Women: 5169 Men: 10 702 Women: 16 418
• Medical Treatment for Metabolic Abnormalities (Hypertension, Lipid Disorder, Diabetes) • FBS ≥ 126 mg/dL or HbA1c ≥ 6.5 Men: 2302 Women: 2265 Men: 8400 Women: 14 153
• Medical Treatment for Malignancies Men: 267 Women: 415
Men: 8133 Women: 13 738
• Medical Treatment for Menopausal Disorder (Women) • Present Medical Treatment or Past History for Cardiovascular Disease Men: 123 Women: 252
Figure 1 Flow chart describing the study exclusions performed.
Eligible Participants Men: 8010 Women: 13 486
covariates included in previous studies (model 2: model 1 variables plus FBG, TG, UA, and ALT). We performed stratified analyses to examine potential heterogeneity in effect based on participants’ current BMI category. Statistical tests for interaction were performed using logistic regression that included a multiplicative term representing the product of 1-kg incremental weight change variable and three-category BMI variable. Two-sided P-values of < 0.05 were considered statistically significant. All statistical analyses were performed via SAS software, version 9.1 (SAS Institute Inc., Cary, NC, USA).
Results Baseline characteristic of participants. After exclusions, 21 496 individuals were included in this study, of whom 13 486 (62.7%) were women. Participants’ mean age was 47.4 years (SD, 11.0 years); all were Asian, primarily of Japanese ethnicity. NAFLD prevalence was 16.3% (men, 30.4%; women, 7.9%). Baseline data are summarized in Table 1. Comparing baseline clinical variables by presence of NAFLD, differences were seen for numerous variables. Current body
Journal of Gastroenterology and Hepatology 30 (2015) 909–917 © 2014 Journal of Gastroenterology and Hepatology Foundation and Wiley Publishing Asia Pty Ltd
911
Fatty liver in normal weight individuals
Table 1
T Kimura et al.
Baseline characteristics of participants
Variable
Number of participants (%) Age, years Height (cm) Body weight in 2008 (kg) Body weight at age 20 (kg) 2008 BW-20y BW (kg) BMI categories in 2008 (%) Normal weight in 2008 Overweight in 2008 Obesity in 2008 BMI in 2008 (kg/m2) BMI categories at age 20 (%) Underweight at age 20 Normal weight at age 20 Overweight at age 20 Obesity at age 20 BMI at age 20 (kg/m2) Alcohol consumption: light alcohol consumption (%) Smoking status (%) Non-smoker Ex-smoker Current smoker Exercise (%) 0 times/week 1 to 2 times/week 3 to 5 times/week 6 to 7 times/week Systolic blood pressure (mmHg) Diastolic blood pressure (mmHg) FBG (mmol/l) HbA1c (%) Total cholesterol (mmol/l) LDL cholesterol (mmol/L) HDL cholesterol (mmol/L) Triglyceride (mmol/L)† Uric acid (μmol/L)† AST (IU/L)† ALT (IU/L)† GGT (IU/L)† Metabolic syndrome (%)
NAFLD−
Men NAFLD+
P-value
NAFLD−
5578 (69.6) 47.7 (11.9) 171 (6) 65.2 (8) 60.4 (7.2) 4.8 (6.3)
2432 (30.4) 47.2 (10.4) 170.8 (5.9) 74(9.9) 62.3 (8.4) 11.7 (7.3)
< 0.001 < 0.001 < 0.001 < 0.001 < 0.001
12420 (92.1) 46.9 (10.7) 158.3 (5.5) 51.4 (6.9) 49.7 (5.7) 1.7 (5.7)
1066 (7.9) 52.9(10.6) 157(5.5) 62.3(10) 51.1 (7.3) 11.2(7.5)
< 0.001 < 0.001 < 0.001 < 0.001 < 0.001
58.9 25.5 12.1 22.3 (2.3)
18.6 32.1 49.3 25.3 (2.8)
< 0.001 < 0.001 < 0.001 < 0.001
66.3 9.3 5.1 20.5 (2.5)
26.9 25.0 47.6 25.2 (3.5)
< 0.001 < 0.001 < 0.001 < 0.001
12.9 75.8 7.5 3.5 20.6 (2.1) 63
8.6 71.4 12.7 7.1 21.3 (2.4) 55.3
< 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001
24.5 69.0 4.8 1.6 19.8 (2) 30.3
16.8 66.3 11.4 5.5 20.7 (2.5) 22.3
< 0.001 0.062 < 0.001 < 0.001 < 0.001 < 0.001
50.7 31.5 17.8
45.5 33.9 20.6
< 0.001 0.030 0.002
85.5 9.3 5.2
84 10.8 5.2
29.6 40.5 17.5 12.4 116.8 (14.6) 72.8 (9.5) 5.5 (0.4) 5.4 (0.3) 5.1 (0.8) 3.1 (0.7) 1.5 (0.3) 0.9 (0.7–1.3) 351.0 (65.4) 20 (18–23) 20 (16–25) 23 (17–34) 8.4
37 41.3 12.8 8.9 123.7 (15) 77.5 (10.1) 5.7 (0.4) 5.6 (0.3) 5.4 (0.8) 3.4 (0.7) 1.2 (0.3) 1.5 (1.1–2.1) 386.7 (71.4) 23 (20–29) 31 (23–44) 35 (25–54) 40.2
< 0.001 0.500 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001
36.7 35.2 18.6 9.5 109.9 (15.4) 67.8 (9.8) 5.2 (0.4) 5.4 (0.3) 5.2 (0.9) 3.0 (0.8) 1.8 (0.4) 0.7 (0.5–0.9) 255.8 (53.5) 19 (17–22) 15 (12–19) 14 (11–19) 1.3
Women NAFLD+
37.6 30.7 21.5 10.2 124.4 (16.9) 76.2 (10.5) 5.6 (0.4) 5.7 (0.3) 5.8 (0.9) 3.6 (0.8) 1.5 (0.3) 1.2 (0.9–1.7) 303.4 (59.5) 22 (19–26) 22 (17–31) 22 (16–33) 26.6
P-value
0.200 0.100 0.900 0.490 0.003 0.020 0.500 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001
Light alcohol consumption: persons who drink alcohol daily; < 20 g/day for men, or < 10 g/day for women. Values are expressed as mean (SD) unless indicated otherwise. † Median (interquartile range). ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; BW, body weight; FBG, fasting blood glucose; GGT, γ-glutamyltransferase; HDL, high-density lipoprotein; LDL, low-density lipoprotein; NAFLD, non-alcoholic fatty liver disease.
weight, weight at age 20, SBP, DBP, FBG, LDL-C, TG, UA, AST, ALT, GGT, and prevalence of MS were all higher in participants with NAFLD (P < 0.001 for all). HDL-C and frequency of exercise were significantly lower in those with NAFLD (P < 0.001).
Weight change since age 20 and prevalence of NAFLD. In both genders, prevalence of NAFLD demonstrated a 912
consistent and proportional increase with weight change since age 20 (Fig. 2). In men and women, prevalence of NAFLD with a weight change of −0.9 to 0.0 kg was 6.8% (95% CI: 4.0–9.7) and 1.1% (95% CI: 0.4–1.7), respectively. Among those experiencing a change of 9.1–10.0 kg, the prevalence of NAFLD in men and women increased dramatically to 39.2% (95% CI: 34.4–44.1) and 19.2% (95% CI: 15.0–23.5), respectively. The prevalence of NAFLD in men remained higher compared with women across all weight change increments.
Journal of Gastroenterology and Hepatology 30 (2015) 909–917 © 2014 Journal of Gastroenterology and Hepatology Foundation and Wiley Publishing Asia Pty Ltd
T Kimura et al.
Fatty liver in normal weight individuals
100 90
Table 2
70 60 50 40 30 20 10 0
–10
0 0 .1 0 . ≥2 to 2 .0 9 .1 19 o 1 .0 t .1 1 8 18 1 to 7.0 . 1 17 to 6.0 .1 1 16 to .0 .1 1 5 15 to .0 .1 1 4 14 to .0 .1 1 3 13 to 2.0 .1 1 12 1 to 1.0 . 1 1 1 to 0 . .1 10 10 o t .0 9 .1 to 9 .0 8 .1 to 8 .0 7 .1 o 7 t .0 6 .1 6 o t 0 5 .1 to 5 . .0 4 .1 o 4 t .0 3 .1 3 o t .0 2 .1 to 2 .0 1.1 to 1 0.0 – 0.1 to .0 .9 – 1 –0 to 2.0 .9 – –1 to 3.0 .9 – –2 9 to 4.0 . – –3 to .0 .9 – 5 –4 9 to 6.0 . – –5 to .0 .9 – 7 –6 to .0 .9 – 8 –7 to .0 9 .9 – 8 to – .9 –9 0 1 ≤–
Figure 2 Prevalence of non-alcoholic fatty liver disease (NAFLD) and weight change since age 20 by gender in overall population. Prevalence is plotted for each 1-kg increment weight change in men (square) and women (diamond). Participants experiencing ≥ 20-kg increase in weight change are categorized into a single group. Corresponding 95% confidence intervals are shown (dotted lines).
Prevalence of NAFLD (%)
80
Body weight change since age 20 (kg)
Logistic regression analysis for weight change (overall population)
(BW in 2008)—(BW at age 20)
Model 1 OR (95% CI)
Men Continuous variable Quartile −0.1 kg ≥ :BW loss group Q1 (0–4.1 kg) Q2 (4.1–7.7 kg) Q3 (7.7–12.1 kg) Q4 (≥ 12.1 kg) Women Continuous variable Quartile −0.1 kg ≥ :BW loss group Q1 (0–2.1 kg) Q2 (2.1–4.5 kg) Q3 (4.5–8.2 kg) Q4 (≥ 8.2 kg)
Model 2 P-value
OR (95% CI)
P-value
1.23 (1.22–1.24)
< 0.001
1.16 (1.14–1.17)
< 0.001
0.26 (0.20–0.35) 1 [reference] 2.34 (1.94–2.83) 5.36 (4.46–6.43) 18.24 (15.10–22.03)
< 0.001
< 0.001
< 0.001 < 0.001 < 0.001
0.35 (0.26–0.48) 1 [reference] 1.85 (1.51–2.26) 3.44 (2.82–4.19) 7.60 (6.18–9.34)
< 0.001 < 0.001 < 0.001
1.27 (1.25–1.28)
< 0.001
1.20 (1.18–1.21)
< 0.001
0.20 (0.13–0.3) 1 [reference] 1.59 (1.13–2.26) 4.02 (2.95–5.47) 18.69 (14.0–24.93)
< 0.001
0.25 (0.16–0.38) 1 [reference] 1.33 (0.91–1.93) 2.83 (2.03–3.95) 8.64 (6.32–11.81)
< 0.001
0.009 < 0.001 < 0.001
0.14 < 0.001 < 0.001
Model 1: adjusted by age, body mass index (BMI) at age 20, alcohol consumption, smoking status, and exercise. Model 2: adjusted by covariates of model 1 plus fasting blood glucose (FBG), triglycerides (TG), uric acid (UA), and alanine aminotransferase (ALT). BW, body weight; CI, confidence interval.
Logistic regression analysis for weight change in overall population. Because of strong evidence of interaction between weight change and gender (P interaction: P < 0.001) on NAFLD risk, men and women were considered separately in all subsequent analyses. Multivariate analysis evaluating body weight change in 1-kg increments and adjusting for age, BMI at age 20, and lifestyle factors (model 1) showed a statistically significant increased risk of NAFLD in both men (OR 1.23, 95% CI: 1.22– 1.24) and women (OR 1.27, 95% CI: 1.25–1.28) (Table 2). Additionally, weight change since age 20 was evaluated as five categories (weight loss vs four gender-specific weight gain quartiles). Compared with quartile 1, the weight loss group demonstrated a protective effect against NAFLD in both men (OR 0.26, 95% CI: 0.20–0.35) and women (OR 0.20, 95% CI: 0.13–
0.30), whereas the weight gain groups showed significant increased risks and a strong trend toward greater risks with each increasing quartile of weight gain. Comparing the highest with lowest quartile adjusted per model 1, an approximately 18-fold increased risk of NAFLD was found for both men (weight gain of ≥ 12.1 kg: OR 18.24, 95% CI: 15.10–22.03) and women (weight gain of ≥ 8.2 kg: OR 18.69, 95% CI: 14.0–24.93). When adjusting for additional laboratory covariates (model 2), risk estimates were attenuated but still demonstrated a sevenfold to eightfold increased risk of NAFLD in highest versus lowest quartile. Risk of NAFLD associated with weight gain in currently normal weight population. To evaluate the role of current body weight status on weight change and NAFLD,
Journal of Gastroenterology and Hepatology 30 (2015) 909–917 © 2014 Journal of Gastroenterology and Hepatology Foundation and Wiley Publishing Asia Pty Ltd
913
Fatty liver in normal weight individuals
T Kimura et al.
100
Prevalence of NAFLD (%)
90 80 70 60 50 40 30 20 10 0 0 7.0 7. ≥1 to 1 .0 6 .1 16 to 1 .0 5 .1 15 o 1 .0 t 4 .1 14 to 1 .0 3 .1 13 to 1 .0 2 .1 12 to 1 .0 .1 11 11 to .1 .0 10 10 to 0 9.1 o 9. t 0 8.1 8. to 0 7.1 to 7. 0 6.1 o 6. t 0 5.1 to 5. 4.1 4.0 to 0 3.1 o 3. t .0 2.1 to 2 1.1 1.0 to 0.0 0.1 to – .0 1 .9 –0 to – .0 2 .9 –1 to – .0 3 .9 –2 to – 4.0 .9 –3 to – 0 5. .9 –4 to – 6.0 .9 –5 to – 0 7. .9 –6 to – .0 .9 –8 –7 to 9.0 .9 – –8 to .9 –9 10 ≤–
–10
Body weight change since age 20 (kg)
analyses stratified by current BMI were performed. Although the currently normal weight group showed an overall lower prevalence than currently overweight and obese groups (Fig. 3), a strong increased NAFLD risk with increasing weight change increments was still observed in both genders. Results of multivariate logistic regression analyses among the currently normal weight group showed associations with weight change since age 20 similar to those of the overall population (data not shown). The crude RR of NAFLD was calculated for each weight change increment compared with weight change of −0.9 kg to 0.0 kg (reference group) stratified by current BMI categories (normal, overweight, and obese). As weight change increased, RR remained significantly higher among normal weight individuals than those in heavier BMI groups (Fig. 4a,b) in both genders. At a weight gain of 10.0–11.0 kg, RRs for currently normal weight, overweight, and obese men were 7.41, 2.05, and 1.20, respectively (normal weight vs overweight, P = 0.017; normal weight vs obese, P < 0.001). In women, corresponding RRs were 40.23, 1.98, and 1.46 respectively (normal weight vs overweight, P < 0.001; normal weight vs obese, P < 0.001). In multivariate analyses evaluating weight change in 10-kg increments, the strongest risk of NAFLD was observed for those at currently normal BMI among both men (OR 7.53, 95% CI: 4.99–11.36) and women (OR 12.20, 95% CI: 7.45–19.98) (Table 3). A test for interaction showed that the effect of weight change on NAFLD risk significantly differed by current BMI (P < 0.001).
Discussion According to our data, the prevalence of NAFLD demonstrated a proportional rise with increasing weight gain since age 20. This association was seen in both men and women, although NAFLD prevalence was consistently higher in men. After adjustment for confounders, the risk of NAFLD associated with weight gain since age 20, treated as both an increasing incremental variable and in quartile categories, remained significant in both genders. Surprisingly, individuals at currently normal weight demonstrated this 914
Figure 3 Prevalence of non-alcoholic fatty liver disease (NAFLD) according to weight change since age 20 among currently normal body weight individuals. Prevalence of NAFLD is plotted for each 1-kg increment change of body weight since age 20 years separately in men (black square) and in women (black diamond). Participants experiencing 17-kg or more increase in men and 12.1-kg or more increase in women in body weight change are categorized into a single group. Corresponding 95% confidence intervals are also shown (dotted lines).
association more robustly than heavier groups; the RR of developing NAFLD among men and women currently at normal weight was significantly stronger than those who were currently overweight and obese. The effect of weight gain since age 20 and risk of various lifestyle diseases has been previously reported,17–21 but few have examined this relationship pertaining to NAFLD. While obesity, and its association with MS, are known risk factors for NAFLD,4,28 the strong influence of weight gain in adulthood, seen even among currently normal weight individuals, is not well-understood. In one prior report involving recent weight gain among normal weight individuals, an approximately 2-kg weight gain over a 4-year interval contributed to NAFLD.29 Our study examining weight gain over an extended period spanning adulthood demonstrated that the impact of adult weight gain was especially strong among currently normal weight persons, a patient demographic conventionally considered to be at low risk for lifestyle-related diseases. This suggests a potentially useful role for longitudinal weight monitoring of apparently healthy individuals for primary prevention of NAFLD and possibly other lifestyle diseases. While further research is necessary to clarify the pathophysiology underlying the increased risk of NAFLD associated with weight gain among normal weight individuals, a possible mechanism may involve the function of subcutaneous adipose tissue (SAT). SAT has been shown to have less facilitative or even protective effects on metabolic abnormalities compared with visceral adipose tissue (VAT).30 Higher VAT/SAT ratios have been associated with increased cardiometabolic risk.31 Previous data suggest that the total number of adipocytes in both SAT- and VAT-related body fat, determined during childhood or adolescence, remains relatively static during adulthood while retaining the capacity to accommodate excess free fatty acids.32,33 Differing in expandability, exhausted SAT expandability may be compensated for by maintained VAT expansion, resulting in increasing VAT/SAT ratios.34 In effect, increased VAT/SAT ratios in adult weight gain may be more common among normal weight versus heavier individuals.32 This mechanism is further supported by our
Journal of Gastroenterology and Hepatology 30 (2015) 909–917 © 2014 Journal of Gastroenterology and Hepatology Foundation and Wiley Publishing Asia Pty Ltd
T Kimura et al.
Fatty liver in normal weight individuals
(a) 20.0 18.0 16.0 Relative risk
14.0 12.0 10.0 8.0 6.0 4.0 2.0 0.0 0 7.0 7. ≥1 to 1 .0 .1 16 16 to .0 .1 15 15 to .0 .1 14 14 to .0 3 .1 13 to 1 .0 .1 12 12 to .0 .1 11 11 to .0 .1 10 10 to . 0 9.1 o 9 t 0 8.1 8. o t 0 7.1 to 7. .0 6.1 6 o t 0 5.1 to 5. 0 4.1 4. o t 0 3.1 to 3. 0 2.1 2. o t .0 1.1 to 1 0.0 – 0.19 to 1.0 . –0 o – .0 t 2 .9 –1 to – .0 3 .9 –2 o – .0 t .9 –4 –3 to .0 5 .9 –4 o – t 6.0 .9 –5 to – .0 7 .9 –6 to – .0 .9 –8 –7 to 9.0 .9 –8 to – .9 –9 0 1 ≤–
Body weight change since age 20 (kg)
(b)
45.0 40.0
Table 3
30.0 25.0 20.0 15.0 10.0 5.0 0.0 1.0 .0 ≥1 11 to .1 0 10 10. to 9.1 9.0 to 8.1 8.0 to 7.1 7.0 to 6.1 6.0 to 5.1 5.0 to 4.1 4.0 to 3.1 3.0 to 2.1 2.0 to 1.1 1.0 to .0 0.1 o –0 t .9 1.0 –0 to – .9 2.0 –1 to – .0 .9 –2 o –3 t .9 4.0 –3 to – .0 .9 –4 o –5 t .9 6.0 –5 to – .9 .0 –6 –7 to 0 . .9 –7 o –8 t 9.0 .9 –8 to – .9 –9 10 ≤–
Figure 4 Relative risk of non-alcoholic fatty liver disease (NAFLD) by weight change since age 20 stratified by current body mass index (BMI). Relative risks in (a) men and (b) women for each 1-kg weight change compared with −0.9 to 0.0 weight change group (†reference) were calculated for those currently at normal weight (18.5 ≤ BMI < 23, triangle), overweight (23 ≤ BMI < 25, square), and obese (BMI ≥ 25,circle).
Relative risk
35.0
Body weight change since age 20 (kg)
Adjusted odds ratio of body weight change stratified by current BMI groups n
Men Normal body weight Overweight Obese Women Normal body weight Overweight Obese
OR† (95% CI)
P-value
P interaction
3736 2145 1874
7.53 (4.99–11.36) 1.61 (0.91–2.85) 4.00 (2.97–5.39)
< 0.001 0.1 < 0.001
< 0.001
8515 1425 1137
12.20 (7.45–19.98) 2.90 (0.99–8.54) 2.68 (2.00–3.6)
< 0.001 0.053 < 0.001
< 0.001
† Adjusted covariates: age, BMI at 20 years, alcohol consumption, smoking, exercise, fasting blood glucose (FBG), triglycerides (TG), uric acid (UA), and alanine aminotransferase (ALT). Odds ratio (OR) represents the risk associated with each 10-kg unit increase in weight change. BMI, body mass index; CI, confidence interval.
data when examining the effect of weight gain on NAFLD risk stratified by BMI at age 20 (data not shown). Consistent with the interpretation of our current analysis, we observed that normal and underweight individuals at age 20 had significantly increased risk
of NAFLD as weight gain increased over time; risk was significantly lower among individuals who were overweight and obese at age 20. Prospective cohort studies with baseline anthropometric measures in adolescence, tracked over time, will help clarify this
Journal of Gastroenterology and Hepatology 30 (2015) 909–917 © 2014 Journal of Gastroenterology and Hepatology Foundation and Wiley Publishing Asia Pty Ltd
915
Fatty liver in normal weight individuals
T Kimura et al.
relationship. Finally, the phenomenon of “metabolically obese but normal weight” 35,36 or “normal weight obesity,”37 defined as excessive body fat deposition in individuals with normal BMI, have been previously associated with various lifestyle-related diseases, including MS, T2DM, and CVD.38–40 Similar mechanisms involving excessive and ectopic visceral fat deposition may occur in NAFLD.41 This study has some limitations. First, this was a single-center study of one ethnic group. However, given a large sample size capturing a wide study base, as well as prior evidence linking adult weight gain to lifestyle diseases, these findings are likely applicable to broader populations; additional studies are warranted in other groups. Second, as weight at age 20 was self-reported, recall bias may have affected the results. However, previous studies have shown very high correlations between self-reported long-term recall and measured weight.42 Information on past weight was obtained before diagnostic examination. Thus, we would expect any potential misclassification of weight at age 20 to be unaffected by NAFLD diagnosis, although there is a possibility of differential misreporting of past weight in normal versus overweight/obese individuals. Our analysis, stratified by current BMI, showed associations between weight change and NAFLD consistent across all strata. While we cannot exclude the possibility that the smaller effect sizes in overweight and obese groups may have been influenced by recall bias, it is unlikely that the observed magnitude of difference is entirely due to misclassification, particularly in the context of a biologically plausible explanation. From a clinical perspective, the possibility of poor weight recall, especially among apparently healthy individuals, reinforces the need for health-care providers to play an early and active role in weight monitoring for improved long-term health, even in normal weight individuals. Finally, the sensitivity and specificity of ultrasound for diagnosing NAFLD is an inherent limitation.43 The gold standard remains liver biopsy; proton magnetic resonance spectroscopy may also offer higher sensitivity and specificity for evaluating fat deposits. However, these procedures are difficult in the primary care setting in terms of safety, accessibility, and implementation costs.43 While underdiagnosis may exist in our data, it is likely to be minimal and random in nature because of the rigorous ultrasonographic evaluation implemented in the clinical setting. Despite these limitations, our findings have a potentially large impact on clinical practice and suggest that health-care providers, uniquely poised to monitor weight over months, years, or decades, should remain vigilant for NAFLD even in normal weight patients. Similarly, these data highlight the importance of preventive healthcare models that include continuity of care in apparently healthy persons, suggesting that following trends through a dynamic adulthood has a clinically important prognostic role in preventable disease. As achieving sustained weight loss among those already overweight or obese at the time of intervention continues to pose substantial clinical challenges, our data imply that interventions aimed at halting or reversing even smaller weight gains earlier in adulthood may not only be more manageable for patients, but also effectively protect against the development lifestyle-related disease. Considering the potential importance of NAFLD on public health, weight monitoring from early adulthood, even among normal weight individuals, is an “office-friendly” parameter to evaluate and predict risk for NAFLD in daily clinical practice. 916
Acknowledgments Contributors: We would like to acknowledge Tomohiro Shinozaki, M.P.H. (Department of Biostatistics, School of Public Health, The University of Tokyo, Tokyo, Japan) for his advice on data analysis and Osamu Takahashi, M.D., M.P.H., Ph.D. (Center for Clinical Epidemiology, St. Luke’s Life Science Institute, Tokyo, Japan) for his support during manuscript development and submission.
References 1 World Health Organization. WHO global infobase: BMI/overweight/obesity. Cited 21 Mar 2014. Available from URL: https:/www0.nih.go.jp/eiken/english/research/project_nhns.htm 2 Ministry of Health, Labour and Welfare, Japan. Outline of the national health and nutrition survey Japan, 2007. Cited 21 Mar 2014. Available from URL: http://www0.nih.go.jp/eiken/english/research/ project_nhns.htm 3 International Diabetes Federation. The global burden. IDF diabetes atlas. Cited 21 Mar 2014. Available from URL: http://www.idf.org/diabetesatlas/ 4 Lazo M, Clark JM. The epidemiology of nonalcoholic fatty liver disease: a global perspective. Semin. Liver Dis. 2008; 28: 339–50. 5 Browning JD, Szczepaniak LS, Dobbins R et al. Prevalence of hepatic steatosis in an urban population in the United States: impact of ethnicity. Hepatology 2004; 40: 1387–95. 6 Church TS, Kuk JL, Ross R et al. Association of cardiorespiratory fitness, body mass index, and waist circumference to nonalcoholic fatty liver disease. Gastroenterology 2006; 130: 2023–30. 7 Bedogni G, Miglioli L, Masutti F et al. Prevalence of and risk factors for nonalcoholic fatty liver disease: the Dionysos nutrition and liver study. Hepatology 2005; 42: 44–52. 8 Lonardo A, Bellini M, Tartoni P et al. The bright liver syndrome. Prevalence and determinants of a “bright” liver echo pattern. Ital. J. Gastroenterol. Hepatol. 1997; 29: 351–6. 9 Hamaguchi M, Kojima T, Takeda N et al. The metabolic syndrome as a predictor of nonalcoholic fatty liver disease. Ann. Intern. Med. 2005; 143: 722–8. 10 Omagari K, Kadokawa Y, Masuda J et al. Fatty liver in non-alcoholic non-overweight Japanese adults: Incidence and clinical characteristics. J. Gastroenterol. Hepatol. 2002; 17: 1098–105. 11 Eguchi Y, Hyogo H, Ono M et al. Prevalence and associated metabolic factors of nonalcoholic fatty liver disease in the general population from 2009 to 2010 in Japan: a multicenter large retrospective study. J. Gastroenterol. 2012; 47: 586–95. 12 Chalasani N, Younossi Z, Lavine JE et al. The diagnosis and management of non-alcoholic fatty liver disease: practice guideline by the American Association for the Study of Liver Diseases, American College of Gastroenterology, and the American Gastroenterological Association. Hepatology 2012; 55: 2005–23. 13 Adams LA, Lymp JF, St. Sauver J et al. The natural history of nonalcoholic fatty liver disease: a population-based cohort study. Gastroenterology 2005; 129: 113–21. 14 Musso G, Gambino R, Cassader M et al. Meta-analysis: natural history of non-alcoholic fatty liver disease (NAFLD) and diagnostic accuracy of non-invasive tests for liver disease severity. Ann. Med. 2011; 43: 617–49. 15 Adams LA, Waters OR, Knuiman MW et al. NAFLD as a risk factor for the development of diabetes and the metabolic syndrome: an eleven-year follow-up study. Am. J. Gastroenterol. 2009; 104: 861–7. 16 Yoong SL, Carey ML, Sanson-Fisher RW, D’Este CA, Mackenzie L, Boyes A. A cross-sectional study examining Australian general
Journal of Gastroenterology and Hepatology 30 (2015) 909–917 © 2014 Journal of Gastroenterology and Hepatology Foundation and Wiley Publishing Asia Pty Ltd
T Kimura et al.
17
18
19 20
21
22
23
24
25
26
27
28
practitioners’ identification of overweight and obese patients. J. Gen. Intern. Med. 2014; 29: 328–34. Colditz GA, Willett WC, Rotnitzky A et al. Weight gain as a risk factor for clinical diabetes mellitus in women. Ann. Intern. Med. 1995; 122: 481–6. Chei CL, Iso H, Yamagishi K et al. Body mass index and weight change since 20 years of age and risk of coronary heart disease among Japanese: the Japan Public Health Center-Based Study. Int. J. Obes. 2008; 32: 144–51. Willett WC, Manson JE, Stampfer MJ et al. Weight, weight change, heart disease in women coronary. JAMA 1990; 273: 461–5. Rexrode KM, Hennekens CH, Willett WC et al. A prospective study of body mass index, weight change, and risk of stroke in women. JAMA 1997; 277: 1539–45. Wakasugi M, Narita I, Iseki K et al. Weight gain after 20 years of age is associated with prevalence of chronic kidney disease. Clin. Exp. Nephrol. 2012; 16: 259–68. Kojima S-I, Watanabe N, Numata M et al. Increase in the prevalence of fatty liver in Japan over the past 12 years: analysis of clinical background. J. Gastroenterol. 2003; 38: 954–61. Razak F, Anand SS, Shannon H et al. Defining obesity cut points in a multiethnic population. Circulation 2007; 115: 2111–18. World Health Organization Expert Consultation. Appropriate body-mass index for Asian populations and its implications for policy and intervention strategies. Lancet 2004; 363: 157–63. Sinn DH, Gwak GY, Park HN et al. Ultrasonographically detected non-alcoholic fatty liver disease is an independent predictor for identifying patients with insulin resistance in non-obese, non-diabetic middle-aged Asian adults. Am. J. Gastroenterol. 2012; 107: 561–7. World Health Organization/International Association for the Study of Obesity/International Obesity Task Force. The Asia-Pacific Perspective: Redefining Obesity and Its Treatment. Geneva: WHO, 2000. Alberti KGMM, Eckel RH, Grundy SM et al. Harmonizing the metabolic syndrome: a joint interim statement of the International Diabetes Federation Task Force on Epidemiology and Prevention; National Heart, Lung, and Blood Institute; American Heart Association; World Heart Federation; International Atherosclerosis Society; and International Association for the Study of Obesity. Circulation 2009; 120: 1640–5. Vernon G, Baranova A, Younossi ZM. Systematic review: the epidemiology and natural history of non-alcoholic fatty liver disease and non-alcoholic steatohepatitis in adults. Aliment. Pharmacol. Ther. 2011; 34: 274–85.
Fatty liver in normal weight individuals
29 Chang Y, Ryu S, Sung E et al. Weight gain within the normal weight range predicts ultrasonographically detected fatty liver in healthy Korean men. Gut 2009; 58: 1419–25. 30 Hocking S, Samocha-Bonet D, Milner K-L et al. Adiposity and insulin resistance in humans: the role of the different tissue and cellular lipid depots. Endocr. Rev. 2013; 34: 463–500. 31 Kaess BM, Pedley A, Massaro JM et al. The ratio of visceral to subcutaneous fat, a metric of body fat distribution, is a unique correlate of cardiometabolic risk. Diabetologia 2012; 55: 2622–30. 32 Spalding KL, Arner E, Westermark PO et al. Dynamics of fat cell turnover in humans. Nature 2008; 453: 783–7. 33 Arner P, Bernard S, Salehpour M et al. Dynamics of human adipose lipid turnover in health and metabolic disease. Nature 2011; 478: 110–13. 34 Virtue S, Vidal-Puig A. It’s not how fat you are, it’s what you do with it that counts. PLoS Biol. 2008; 6: e237. 35 Conus F, Rabasa-Lhoret R, Péronnet F. Characteristics of metabolically obese normal-weight (MONW) subjects. Appl. Physiol. Nutr. Metab. 2007; 32: 4–12. 36 Lee S-H, Ha H-S, Park Y-J et al. Identifying metabolically obese but normal-weight (MONW) individuals in a nondiabetic Korean population: the Chungju Metabolic disease Cohort (CMC) study. Clin. Endocrinol. 2011; 75: 475–81. 37 Romero-Corral A, Somers VK, Sierra-Johnson J et al. Accuracy of body mass index in diagnosing obesity in the adult general population. Int. J. Obes. 2008; 32: 959–66. 38 Romero-corral A, Somers VK, Sierra-Johnson J et al. Normal weight obesity: a risk factor for cardiometabolic dysregulation and cardiovascular mortality. Eur. Heart J. 2010; 31: 737–46. 39 Madeira F, Silva A, Veloso H et al. Normal weight obesity is associated with metabolic syndrome and insulin resistance in young adults from a middle-income country. PLoS ONE 2013; 8: e60673. 40 Wildman RP, Muntner P, Reynolds K et al. The obese without cardiometabolic risk factor clustering and the normal weight with cardiometabolic risk factor clustering: prevalence and correlates of 2 phenotypes among the US population (NHANES 1999–2004). Arch. Intern. Med. 2008; 168: 1617–24. 41 Thomas EL, Frost G, Taylor-Robinson SD et al. Excess body fat in obese and normal-weight subjects. Nutr. Res. Rev. 2012; 25: 150–61. 42 Casey VA, Dwyer JT, Berkey CS et al. Long-term memory of body weight and past weight satisfaction: a longitudinal follow-up study. Am. J. Clin. Nutr. 1991; 53: 1493–8. 43 Hernaez R, Lazo M, Bonekamp S et al. Diagnostic accuracy and reliability of ultrasonography for the detection of fatty liver: a meta-analysis. Hepatology 2011; 54: 1082–90.
Journal of Gastroenterology and Hepatology 30 (2015) 909–917 © 2014 Journal of Gastroenterology and Hepatology Foundation and Wiley Publishing Asia Pty Ltd
917
doi:10.1111/jgh.12861
H E PAT O L O G Y
Association of weight gain since age 20 with non-alcoholic fatty liver disease in normal weight individuals Takeshi Kimura,*,† Gautam A Deshpande,† Kevin Y Urayama,†,‡ Katsunori Masuda,* Tsuguya Fukui† and Yutaka Matsuyama§ *Center for Preventive Medicine, St. Luke’s International University, †Center for Clinical Epidemiology, St. Luke’s Life Science Institute, ‡ Department of Human Genetics and Disease Diversity, Graduate School of Medical and Dental Sciences, Tokyo Medical and Dental University, and §Department of Biostatistics, School of Public Health, The University of Tokyo, Tokyo, Japan
Key words body weight gain, non-alcoholic steatohepatitis, prevention and control. Accepted for publication 22 November 2014. Correspondence Dr Takeshi Kimura, Center for Clinical Epidemiology, St. Luke’s Life Science Institute, 10-1 Akashi-cho, Chuo-ku, Tokyo 104-0044, Japan. Email:[email protected] Conflict of interest: None. Financial grants: None.
Abstract Background and Aims: Interventions for lifestyle diseases including non-alcoholic fatty liver disease (NAFLD) have focused on overweight and obese populations. The impact of adult weight gain on NAFLD development among normal weight individuals remains unclear. Methods: In this cross-sectional study, we collected data from participants presenting to a health check-up program. Ultrasound-diagnosed NAFLD prevalence was examined over 1-kg increments of weight change since age 20. Relative risks were calculated in men and women stratified by current weight (normal, overweight, and obese). Logistic regression was used to estimate odds ratios (ORs) and 95% confidence intervals (CIs) adjusting for potential confounders. Results: Among 21 496 participants, 3498 cases of NAFLD (16.3%) were observed. Prevalence of NAFLD increased with weight gain since age 20; among the 10.1–11.0 kg weight gain group, 41.6% of men and 24.8% of women had NAFLD. Multivariate analysis by quartiles showed that weight change was significantly associated with NAFLD risk in men and women. Risk of NAFLD associated with weight change (10-kg increments) was significantly higher in normal weight individuals (men: OR 7.53, 95% CI: 4.99–11.36, women: OR 12.20, 95% CI: 7.45–19.98) than overweight (men: OR 1.61, 95% CI: 0.91–2.85, women: OR 2.90, 95% CI: 0.99–8.54) and obese (men: OR 4.0, 95% CI: 2.97–5.39, women: OR 2.68, 95% CI: 2.00–3.60). Conclusions: NAFLD is robustly associated with weight change since age 20. This effect appears particularly strong in individuals at normal weight, suggesting an important role for early and longitudinal weight monitoring, even among healthy individuals at normal weight.
Introduction Because of the worldwide obesity pandemic,1,2 obesity-related diseases are anticipated to become an increasingly heavy health-care burden for many countries; non-alcoholic fatty liver disease (NAFLD) has emerged as an important global problem.3 The worldwide prevalence of NAFLD, estimated at 10–30% in the general population depending on population and diagnostic methods,4–11 is expected to rise in the coming years.12 NAFLD causes significant morbidity over time. Progression to non-alcoholic steatohepatitis portends increased rates of advanced fibrosis, cirrhosis, and hepatocellular carcinoma, carrying with them further morbidity and mortality burdens.13,14 Additionally, previous studies suggest that NAFLD contributes to an increased incidence of cardiovascular disease (CVD),12 metabolic syndrome
(MS), and type 2 diabetes (T2DM).15 Therefore, a better understanding of the epidemiology of NAFLD, including identification of modifiable risk factors and prevention strategies, has the potential to substantially improve public health outcomes on a large scale. With strong evidence linking obesity to NAFLD and other lifestyle-related disease, much of the primary prevention performed in clinical practice remains focused on overweight and obese populations.16 However, adult weight gain, as opposed to absolute body weight, has also been reported to be a risk factor for various lifestyle-related disease including T2DM,17 CVD,18,19 stroke,20 and chronic kidney disease.21 It remains unclear whether the dynamics of weight gain over time, regardless of body mass index (BMI), adversely influences progression to NAFLD, thus warranting attention for primary prevention strategies.
Journal of Gastroenterology and Hepatology 30 (2015) 909–917 © 2014 Journal of Gastroenterology and Hepatology Foundation and Wiley Publishing Asia Pty Ltd
909
Fatty liver in normal weight individuals
T Kimura et al.
Given established associations between lifestyle diseases and NAFLD, we hypothesized a similar relationship between weight gain during adulthood and development of NAFLD; there are no prior reports evaluating this association. In this study, we examined the relationship between weight gain since age 20 and NAFLD, utilizing comprehensive data from a large cohort of adults presenting for an annual health check-up. Further hypothesizing that normal weight individual may similarly be at risk for NAFLD in the setting of weight gain, data were stratified by current BMI to clarify this relationship.
Methods Study population. We conducted a cross-sectional study of apparently healthy participants undergoing a health screening exam in 2008–2009 at the Center for Preventive Medicine at St. Luke’s International Hospital in Tokyo, Japan. In Japan, mandatory annual health screenings are required for all employed persons; 80% of this sample was referred to our center via employer-sponsored programs, with 20% consisting of community self-referrals. Eligible participants were selected among a total of 21 710 men and 23 732 women. Per clinically relevant criteria which could potentially bias outcomes, individuals with any of the following were excluded: (i) missing clinical data; (ii) age < 30 years; (iii) > 2 health check-ups in 2008–2009; (iv) daily alcohol intake ≥ 20 g/day for men or ≥ 10 g/day for women; (v) positive or unknown hepatitis B or C serology; (vi) pregnancy; (vii) on medication for a hepatic condition, menopausal disorder, or metabolic disease including hypertension, hyperlipidemia, and diabetes mellitus; (viii) fasting blood glucose (FBG) ≥ 126 mg or hemoglobin A1c (HbA1c) ≥ 6.5%; (ix) history of surgery or medication for malignancy; and (x) currently or previously treated CVD. Comparison of excluded participants showed similar characteristics to those included in the study. The exclusion of 23 945 individuals resulted in a study population of 21 496 individuals (Fig. 1). The study was approved by the St. Luke’s International Hospital Institutional Review Board and was exempted from individual participants’ written informed consent. Data collection. Anthropometric measurements included height and weight, measured via digital weight scale, and waist circumference measured at the midpoint between the lowest rib margin and the iliac crest. Body weight at age 20 was collected via self-administered questionnaires. Weight change, the primary exposure variable, was defined as the difference between weight (kilograms) at 2008–2009 screening and at age 20. Clinical and demographic data were collected from all participants during the health screening process. Systolic (SBP) and diastolic (DBP) blood pressure, routine labs including FBG, HbA1c, total cholesterol, low-density lipoprotein-cholesterol (LDL-C), high-density lipoprotein-cholesterol (HDL-C), triglycerides (TG), uric acid (UA), aspartate aminotransferase (AST), alanine aminotransferase (ALT), and γ-glutamyltransferase (GGT) were collected. Serum samples were obtained after a 12-h fast. All participants completed self-administered questionnaires regarding health history and lifestyle habits including smoking (non-smoker/ ex-smoker/current smoker), alcohol intake (grams per week), 910
exercise habits (no exercise/activity 1–2 times per week/3–5 times per week/6–7 times per week), present illnesses and treatments, and past medical history. Responses were verified by a trained team of nurses via face-to-face interviews conducted with all participants as a standard part of the health check-up process. To preserve confidentiality, direct patient identifiers were not collected when creating the dataset. Diagnosis of NAFLD. All participants received abdominal ultrasound as a routine part of this health screening program. Diagnosis of fatty liver was performed by ultrasound, assessing four well-established criteria: enhanced liver echogenicity, echogenicity greater in liver than kidneys, deep attenuation, and vascular blurring.22 Participants were diagnosed with NAFLD by observing, at minimum, enhanced liver echogenicity and hepatorenal contrast. Examinations were performed by trained ultrasound technicians, with findings reviewed by experienced radiologists. Definition of overweight, obesity and MS. Proposed standard definitions of overweight and obesity vary by race and ethnicity in previous reports.23 For Asians, specific definitions of obesity have been proposed, although uniformly accepted criteria have yet to be determined.24 Consistent with a previous report,25 we used Asian-specific BMI cut-offs as recommended by the World Health Organization24,26 as follows: underweight, < 18.5; normal weight, 18.5–22.9; overweight, 23–24.9; and obese, ≥ 25. To define MS, we adopted International Diabetes Federation Task Force criteria27: (i) elevated waist circumference; (ii) elevated triglyceride (≥ 150 mg/mL); (iii) reduced HDL-C (male: < 40 mg/ dL, female: < 50 mg/dL); (iv) SBP > 130 or DBP > 85 mmHg; and (v) elevated fasting glucose (≥ 100 mg/dL). Diagnosis required individuals to meet at least three criteria. Statistical analysis. Variables were reported as mean (SD), median (interquartile range), or proportion (%) based on variable characteristics. To explore differences between groups, Student’s t-test, Mann–Whitney U-test, or χ2 test was performed based on appropriate distribution of data. The primary outcome was prevalence of NAFLD. In order to facilitate clinical interpretation, weight change since age 20 was evaluated in 1-kg increments and categories, with one weight loss category and four gender-specific weight gain quartiles. NAFLD prevalence was plotted over the range of 1-kg weight change increments. Relative risk (RR) of NAFLD by weight change (compared with no gain or loss group [−0.9 to 0 kg]) was calculated, stratified by current BMI and BMI at age 20. The difference in log (RR) between BMI categories was evaluated by Wald test. Adjusting for potential confounders, multivariate logistic regression of individuals with BMI > 18.5 was used to evaluate the association between NAFLD and weight change, calculating odds ratios (ORs) with 95% confidence intervals (CIs). We considered several potential confounders and present results for two models, the first of which was adjusted for basic demographic factors and lifestyle behaviors (model 1: age, BMI at age 20, light alcohol consumption, smoking status, and exercise). In a second model, we evaluated the influence of additional clinically relevant
Journal of Gastroenterology and Hepatology 30 (2015) 909–917 © 2014 Journal of Gastroenterology and Hepatology Foundation and Wiley Publishing Asia Pty Ltd
T Kimura et al.
Fatty liver in normal weight individuals
Participants of Annual Health Check-up of 2008–2009 Men: 21 710 Women: 23 731 • Data Missing Men: 2052 Women: 1413
Men: 19 658 Women: 22 318 • Recieving Health Check-up More than Twice in 2008 • Less than 30 years old Men: 789 Women: 731
Men: 18 869 Women: 21 587 • Pregnancy (Women) • Excluding Alcohol Intake Men ≥ 20 g/day, Women ≥ 10 g/day • HBsAg positive, HCVAb positive • Medical Treatment for Liver Diseases Men: 8167 Women: 5169 Men: 10 702 Women: 16 418
• Medical Treatment for Metabolic Abnormalities (Hypertension, Lipid Disorder, Diabetes) • FBS ≥ 126 mg/dL or HbA1c ≥ 6.5 Men: 2302 Women: 2265 Men: 8400 Women: 14 153
• Medical Treatment for Malignancies Men: 267 Women: 415
Men: 8133 Women: 13 738
• Medical Treatment for Menopausal Disorder (Women) • Present Medical Treatment or Past History for Cardiovascular Disease Men: 123 Women: 252
Figure 1 Flow chart describing the study exclusions performed.
Eligible Participants Men: 8010 Women: 13 486
covariates included in previous studies (model 2: model 1 variables plus FBG, TG, UA, and ALT). We performed stratified analyses to examine potential heterogeneity in effect based on participants’ current BMI category. Statistical tests for interaction were performed using logistic regression that included a multiplicative term representing the product of 1-kg incremental weight change variable and three-category BMI variable. Two-sided P-values of < 0.05 were considered statistically significant. All statistical analyses were performed via SAS software, version 9.1 (SAS Institute Inc., Cary, NC, USA).
Results Baseline characteristic of participants. After exclusions, 21 496 individuals were included in this study, of whom 13 486 (62.7%) were women. Participants’ mean age was 47.4 years (SD, 11.0 years); all were Asian, primarily of Japanese ethnicity. NAFLD prevalence was 16.3% (men, 30.4%; women, 7.9%). Baseline data are summarized in Table 1. Comparing baseline clinical variables by presence of NAFLD, differences were seen for numerous variables. Current body
Journal of Gastroenterology and Hepatology 30 (2015) 909–917 © 2014 Journal of Gastroenterology and Hepatology Foundation and Wiley Publishing Asia Pty Ltd
911
Fatty liver in normal weight individuals
Table 1
T Kimura et al.
Baseline characteristics of participants
Variable
Number of participants (%) Age, years Height (cm) Body weight in 2008 (kg) Body weight at age 20 (kg) 2008 BW-20y BW (kg) BMI categories in 2008 (%) Normal weight in 2008 Overweight in 2008 Obesity in 2008 BMI in 2008 (kg/m2) BMI categories at age 20 (%) Underweight at age 20 Normal weight at age 20 Overweight at age 20 Obesity at age 20 BMI at age 20 (kg/m2) Alcohol consumption: light alcohol consumption (%) Smoking status (%) Non-smoker Ex-smoker Current smoker Exercise (%) 0 times/week 1 to 2 times/week 3 to 5 times/week 6 to 7 times/week Systolic blood pressure (mmHg) Diastolic blood pressure (mmHg) FBG (mmol/l) HbA1c (%) Total cholesterol (mmol/l) LDL cholesterol (mmol/L) HDL cholesterol (mmol/L) Triglyceride (mmol/L)† Uric acid (μmol/L)† AST (IU/L)† ALT (IU/L)† GGT (IU/L)† Metabolic syndrome (%)
NAFLD−
Men NAFLD+
P-value
NAFLD−
5578 (69.6) 47.7 (11.9) 171 (6) 65.2 (8) 60.4 (7.2) 4.8 (6.3)
2432 (30.4) 47.2 (10.4) 170.8 (5.9) 74(9.9) 62.3 (8.4) 11.7 (7.3)
< 0.001 < 0.001 < 0.001 < 0.001 < 0.001
12420 (92.1) 46.9 (10.7) 158.3 (5.5) 51.4 (6.9) 49.7 (5.7) 1.7 (5.7)
1066 (7.9) 52.9(10.6) 157(5.5) 62.3(10) 51.1 (7.3) 11.2(7.5)
< 0.001 < 0.001 < 0.001 < 0.001 < 0.001
58.9 25.5 12.1 22.3 (2.3)
18.6 32.1 49.3 25.3 (2.8)
< 0.001 < 0.001 < 0.001 < 0.001
66.3 9.3 5.1 20.5 (2.5)
26.9 25.0 47.6 25.2 (3.5)
< 0.001 < 0.001 < 0.001 < 0.001
12.9 75.8 7.5 3.5 20.6 (2.1) 63
8.6 71.4 12.7 7.1 21.3 (2.4) 55.3
< 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001
24.5 69.0 4.8 1.6 19.8 (2) 30.3
16.8 66.3 11.4 5.5 20.7 (2.5) 22.3
< 0.001 0.062 < 0.001 < 0.001 < 0.001 < 0.001
50.7 31.5 17.8
45.5 33.9 20.6
< 0.001 0.030 0.002
85.5 9.3 5.2
84 10.8 5.2
29.6 40.5 17.5 12.4 116.8 (14.6) 72.8 (9.5) 5.5 (0.4) 5.4 (0.3) 5.1 (0.8) 3.1 (0.7) 1.5 (0.3) 0.9 (0.7–1.3) 351.0 (65.4) 20 (18–23) 20 (16–25) 23 (17–34) 8.4
37 41.3 12.8 8.9 123.7 (15) 77.5 (10.1) 5.7 (0.4) 5.6 (0.3) 5.4 (0.8) 3.4 (0.7) 1.2 (0.3) 1.5 (1.1–2.1) 386.7 (71.4) 23 (20–29) 31 (23–44) 35 (25–54) 40.2
< 0.001 0.500 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001
36.7 35.2 18.6 9.5 109.9 (15.4) 67.8 (9.8) 5.2 (0.4) 5.4 (0.3) 5.2 (0.9) 3.0 (0.8) 1.8 (0.4) 0.7 (0.5–0.9) 255.8 (53.5) 19 (17–22) 15 (12–19) 14 (11–19) 1.3
Women NAFLD+
37.6 30.7 21.5 10.2 124.4 (16.9) 76.2 (10.5) 5.6 (0.4) 5.7 (0.3) 5.8 (0.9) 3.6 (0.8) 1.5 (0.3) 1.2 (0.9–1.7) 303.4 (59.5) 22 (19–26) 22 (17–31) 22 (16–33) 26.6
P-value
0.200 0.100 0.900 0.490 0.003 0.020 0.500 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001
Light alcohol consumption: persons who drink alcohol daily; < 20 g/day for men, or < 10 g/day for women. Values are expressed as mean (SD) unless indicated otherwise. † Median (interquartile range). ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; BW, body weight; FBG, fasting blood glucose; GGT, γ-glutamyltransferase; HDL, high-density lipoprotein; LDL, low-density lipoprotein; NAFLD, non-alcoholic fatty liver disease.
weight, weight at age 20, SBP, DBP, FBG, LDL-C, TG, UA, AST, ALT, GGT, and prevalence of MS were all higher in participants with NAFLD (P < 0.001 for all). HDL-C and frequency of exercise were significantly lower in those with NAFLD (P < 0.001).
Weight change since age 20 and prevalence of NAFLD. In both genders, prevalence of NAFLD demonstrated a 912
consistent and proportional increase with weight change since age 20 (Fig. 2). In men and women, prevalence of NAFLD with a weight change of −0.9 to 0.0 kg was 6.8% (95% CI: 4.0–9.7) and 1.1% (95% CI: 0.4–1.7), respectively. Among those experiencing a change of 9.1–10.0 kg, the prevalence of NAFLD in men and women increased dramatically to 39.2% (95% CI: 34.4–44.1) and 19.2% (95% CI: 15.0–23.5), respectively. The prevalence of NAFLD in men remained higher compared with women across all weight change increments.
Journal of Gastroenterology and Hepatology 30 (2015) 909–917 © 2014 Journal of Gastroenterology and Hepatology Foundation and Wiley Publishing Asia Pty Ltd
T Kimura et al.
Fatty liver in normal weight individuals
100 90
Table 2
70 60 50 40 30 20 10 0
–10
0 0 .1 0 . ≥2 to 2 .0 9 .1 19 o 1 .0 t .1 1 8 18 1 to 7.0 . 1 17 to 6.0 .1 1 16 to .0 .1 1 5 15 to .0 .1 1 4 14 to .0 .1 1 3 13 to 2.0 .1 1 12 1 to 1.0 . 1 1 1 to 0 . .1 10 10 o t .0 9 .1 to 9 .0 8 .1 to 8 .0 7 .1 o 7 t .0 6 .1 6 o t 0 5 .1 to 5 . .0 4 .1 o 4 t .0 3 .1 3 o t .0 2 .1 to 2 .0 1.1 to 1 0.0 – 0.1 to .0 .9 – 1 –0 to 2.0 .9 – –1 to 3.0 .9 – –2 9 to 4.0 . – –3 to .0 .9 – 5 –4 9 to 6.0 . – –5 to .0 .9 – 7 –6 to .0 .9 – 8 –7 to .0 9 .9 – 8 to – .9 –9 0 1 ≤–
Figure 2 Prevalence of non-alcoholic fatty liver disease (NAFLD) and weight change since age 20 by gender in overall population. Prevalence is plotted for each 1-kg increment weight change in men (square) and women (diamond). Participants experiencing ≥ 20-kg increase in weight change are categorized into a single group. Corresponding 95% confidence intervals are shown (dotted lines).
Prevalence of NAFLD (%)
80
Body weight change since age 20 (kg)
Logistic regression analysis for weight change (overall population)
(BW in 2008)—(BW at age 20)
Model 1 OR (95% CI)
Men Continuous variable Quartile −0.1 kg ≥ :BW loss group Q1 (0–4.1 kg) Q2 (4.1–7.7 kg) Q3 (7.7–12.1 kg) Q4 (≥ 12.1 kg) Women Continuous variable Quartile −0.1 kg ≥ :BW loss group Q1 (0–2.1 kg) Q2 (2.1–4.5 kg) Q3 (4.5–8.2 kg) Q4 (≥ 8.2 kg)
Model 2 P-value
OR (95% CI)
P-value
1.23 (1.22–1.24)
< 0.001
1.16 (1.14–1.17)
< 0.001
0.26 (0.20–0.35) 1 [reference] 2.34 (1.94–2.83) 5.36 (4.46–6.43) 18.24 (15.10–22.03)
< 0.001
< 0.001
< 0.001 < 0.001 < 0.001
0.35 (0.26–0.48) 1 [reference] 1.85 (1.51–2.26) 3.44 (2.82–4.19) 7.60 (6.18–9.34)
< 0.001 < 0.001 < 0.001
1.27 (1.25–1.28)
< 0.001
1.20 (1.18–1.21)
< 0.001
0.20 (0.13–0.3) 1 [reference] 1.59 (1.13–2.26) 4.02 (2.95–5.47) 18.69 (14.0–24.93)
< 0.001
0.25 (0.16–0.38) 1 [reference] 1.33 (0.91–1.93) 2.83 (2.03–3.95) 8.64 (6.32–11.81)
< 0.001
0.009 < 0.001 < 0.001
0.14 < 0.001 < 0.001
Model 1: adjusted by age, body mass index (BMI) at age 20, alcohol consumption, smoking status, and exercise. Model 2: adjusted by covariates of model 1 plus fasting blood glucose (FBG), triglycerides (TG), uric acid (UA), and alanine aminotransferase (ALT). BW, body weight; CI, confidence interval.
Logistic regression analysis for weight change in overall population. Because of strong evidence of interaction between weight change and gender (P interaction: P < 0.001) on NAFLD risk, men and women were considered separately in all subsequent analyses. Multivariate analysis evaluating body weight change in 1-kg increments and adjusting for age, BMI at age 20, and lifestyle factors (model 1) showed a statistically significant increased risk of NAFLD in both men (OR 1.23, 95% CI: 1.22– 1.24) and women (OR 1.27, 95% CI: 1.25–1.28) (Table 2). Additionally, weight change since age 20 was evaluated as five categories (weight loss vs four gender-specific weight gain quartiles). Compared with quartile 1, the weight loss group demonstrated a protective effect against NAFLD in both men (OR 0.26, 95% CI: 0.20–0.35) and women (OR 0.20, 95% CI: 0.13–
0.30), whereas the weight gain groups showed significant increased risks and a strong trend toward greater risks with each increasing quartile of weight gain. Comparing the highest with lowest quartile adjusted per model 1, an approximately 18-fold increased risk of NAFLD was found for both men (weight gain of ≥ 12.1 kg: OR 18.24, 95% CI: 15.10–22.03) and women (weight gain of ≥ 8.2 kg: OR 18.69, 95% CI: 14.0–24.93). When adjusting for additional laboratory covariates (model 2), risk estimates were attenuated but still demonstrated a sevenfold to eightfold increased risk of NAFLD in highest versus lowest quartile. Risk of NAFLD associated with weight gain in currently normal weight population. To evaluate the role of current body weight status on weight change and NAFLD,
Journal of Gastroenterology and Hepatology 30 (2015) 909–917 © 2014 Journal of Gastroenterology and Hepatology Foundation and Wiley Publishing Asia Pty Ltd
913
Fatty liver in normal weight individuals
T Kimura et al.
100
Prevalence of NAFLD (%)
90 80 70 60 50 40 30 20 10 0 0 7.0 7. ≥1 to 1 .0 6 .1 16 to 1 .0 5 .1 15 o 1 .0 t 4 .1 14 to 1 .0 3 .1 13 to 1 .0 2 .1 12 to 1 .0 .1 11 11 to .1 .0 10 10 to 0 9.1 o 9. t 0 8.1 8. to 0 7.1 to 7. 0 6.1 o 6. t 0 5.1 to 5. 4.1 4.0 to 0 3.1 o 3. t .0 2.1 to 2 1.1 1.0 to 0.0 0.1 to – .0 1 .9 –0 to – .0 2 .9 –1 to – .0 3 .9 –2 to – 4.0 .9 –3 to – 0 5. .9 –4 to – 6.0 .9 –5 to – 0 7. .9 –6 to – .0 .9 –8 –7 to 9.0 .9 – –8 to .9 –9 10 ≤–
–10
Body weight change since age 20 (kg)
analyses stratified by current BMI were performed. Although the currently normal weight group showed an overall lower prevalence than currently overweight and obese groups (Fig. 3), a strong increased NAFLD risk with increasing weight change increments was still observed in both genders. Results of multivariate logistic regression analyses among the currently normal weight group showed associations with weight change since age 20 similar to those of the overall population (data not shown). The crude RR of NAFLD was calculated for each weight change increment compared with weight change of −0.9 kg to 0.0 kg (reference group) stratified by current BMI categories (normal, overweight, and obese). As weight change increased, RR remained significantly higher among normal weight individuals than those in heavier BMI groups (Fig. 4a,b) in both genders. At a weight gain of 10.0–11.0 kg, RRs for currently normal weight, overweight, and obese men were 7.41, 2.05, and 1.20, respectively (normal weight vs overweight, P = 0.017; normal weight vs obese, P < 0.001). In women, corresponding RRs were 40.23, 1.98, and 1.46 respectively (normal weight vs overweight, P < 0.001; normal weight vs obese, P < 0.001). In multivariate analyses evaluating weight change in 10-kg increments, the strongest risk of NAFLD was observed for those at currently normal BMI among both men (OR 7.53, 95% CI: 4.99–11.36) and women (OR 12.20, 95% CI: 7.45–19.98) (Table 3). A test for interaction showed that the effect of weight change on NAFLD risk significantly differed by current BMI (P < 0.001).
Discussion According to our data, the prevalence of NAFLD demonstrated a proportional rise with increasing weight gain since age 20. This association was seen in both men and women, although NAFLD prevalence was consistently higher in men. After adjustment for confounders, the risk of NAFLD associated with weight gain since age 20, treated as both an increasing incremental variable and in quartile categories, remained significant in both genders. Surprisingly, individuals at currently normal weight demonstrated this 914
Figure 3 Prevalence of non-alcoholic fatty liver disease (NAFLD) according to weight change since age 20 among currently normal body weight individuals. Prevalence of NAFLD is plotted for each 1-kg increment change of body weight since age 20 years separately in men (black square) and in women (black diamond). Participants experiencing 17-kg or more increase in men and 12.1-kg or more increase in women in body weight change are categorized into a single group. Corresponding 95% confidence intervals are also shown (dotted lines).
association more robustly than heavier groups; the RR of developing NAFLD among men and women currently at normal weight was significantly stronger than those who were currently overweight and obese. The effect of weight gain since age 20 and risk of various lifestyle diseases has been previously reported,17–21 but few have examined this relationship pertaining to NAFLD. While obesity, and its association with MS, are known risk factors for NAFLD,4,28 the strong influence of weight gain in adulthood, seen even among currently normal weight individuals, is not well-understood. In one prior report involving recent weight gain among normal weight individuals, an approximately 2-kg weight gain over a 4-year interval contributed to NAFLD.29 Our study examining weight gain over an extended period spanning adulthood demonstrated that the impact of adult weight gain was especially strong among currently normal weight persons, a patient demographic conventionally considered to be at low risk for lifestyle-related diseases. This suggests a potentially useful role for longitudinal weight monitoring of apparently healthy individuals for primary prevention of NAFLD and possibly other lifestyle diseases. While further research is necessary to clarify the pathophysiology underlying the increased risk of NAFLD associated with weight gain among normal weight individuals, a possible mechanism may involve the function of subcutaneous adipose tissue (SAT). SAT has been shown to have less facilitative or even protective effects on metabolic abnormalities compared with visceral adipose tissue (VAT).30 Higher VAT/SAT ratios have been associated with increased cardiometabolic risk.31 Previous data suggest that the total number of adipocytes in both SAT- and VAT-related body fat, determined during childhood or adolescence, remains relatively static during adulthood while retaining the capacity to accommodate excess free fatty acids.32,33 Differing in expandability, exhausted SAT expandability may be compensated for by maintained VAT expansion, resulting in increasing VAT/SAT ratios.34 In effect, increased VAT/SAT ratios in adult weight gain may be more common among normal weight versus heavier individuals.32 This mechanism is further supported by our
Journal of Gastroenterology and Hepatology 30 (2015) 909–917 © 2014 Journal of Gastroenterology and Hepatology Foundation and Wiley Publishing Asia Pty Ltd
T Kimura et al.
Fatty liver in normal weight individuals
(a) 20.0 18.0 16.0 Relative risk
14.0 12.0 10.0 8.0 6.0 4.0 2.0 0.0 0 7.0 7. ≥1 to 1 .0 .1 16 16 to .0 .1 15 15 to .0 .1 14 14 to .0 3 .1 13 to 1 .0 .1 12 12 to .0 .1 11 11 to .0 .1 10 10 to . 0 9.1 o 9 t 0 8.1 8. o t 0 7.1 to 7. .0 6.1 6 o t 0 5.1 to 5. 0 4.1 4. o t 0 3.1 to 3. 0 2.1 2. o t .0 1.1 to 1 0.0 – 0.19 to 1.0 . –0 o – .0 t 2 .9 –1 to – .0 3 .9 –2 o – .0 t .9 –4 –3 to .0 5 .9 –4 o – t 6.0 .9 –5 to – .0 7 .9 –6 to – .0 .9 –8 –7 to 9.0 .9 –8 to – .9 –9 0 1 ≤–
Body weight change since age 20 (kg)
(b)
45.0 40.0
Table 3
30.0 25.0 20.0 15.0 10.0 5.0 0.0 1.0 .0 ≥1 11 to .1 0 10 10. to 9.1 9.0 to 8.1 8.0 to 7.1 7.0 to 6.1 6.0 to 5.1 5.0 to 4.1 4.0 to 3.1 3.0 to 2.1 2.0 to 1.1 1.0 to .0 0.1 o –0 t .9 1.0 –0 to – .9 2.0 –1 to – .0 .9 –2 o –3 t .9 4.0 –3 to – .0 .9 –4 o –5 t .9 6.0 –5 to – .9 .0 –6 –7 to 0 . .9 –7 o –8 t 9.0 .9 –8 to – .9 –9 10 ≤–
Figure 4 Relative risk of non-alcoholic fatty liver disease (NAFLD) by weight change since age 20 stratified by current body mass index (BMI). Relative risks in (a) men and (b) women for each 1-kg weight change compared with −0.9 to 0.0 weight change group (†reference) were calculated for those currently at normal weight (18.5 ≤ BMI < 23, triangle), overweight (23 ≤ BMI < 25, square), and obese (BMI ≥ 25,circle).
Relative risk
35.0
Body weight change since age 20 (kg)
Adjusted odds ratio of body weight change stratified by current BMI groups n
Men Normal body weight Overweight Obese Women Normal body weight Overweight Obese
OR† (95% CI)
P-value
P interaction
3736 2145 1874
7.53 (4.99–11.36) 1.61 (0.91–2.85) 4.00 (2.97–5.39)
< 0.001 0.1 < 0.001
< 0.001
8515 1425 1137
12.20 (7.45–19.98) 2.90 (0.99–8.54) 2.68 (2.00–3.6)
< 0.001 0.053 < 0.001
< 0.001
† Adjusted covariates: age, BMI at 20 years, alcohol consumption, smoking, exercise, fasting blood glucose (FBG), triglycerides (TG), uric acid (UA), and alanine aminotransferase (ALT). Odds ratio (OR) represents the risk associated with each 10-kg unit increase in weight change. BMI, body mass index; CI, confidence interval.
data when examining the effect of weight gain on NAFLD risk stratified by BMI at age 20 (data not shown). Consistent with the interpretation of our current analysis, we observed that normal and underweight individuals at age 20 had significantly increased risk
of NAFLD as weight gain increased over time; risk was significantly lower among individuals who were overweight and obese at age 20. Prospective cohort studies with baseline anthropometric measures in adolescence, tracked over time, will help clarify this
Journal of Gastroenterology and Hepatology 30 (2015) 909–917 © 2014 Journal of Gastroenterology and Hepatology Foundation and Wiley Publishing Asia Pty Ltd
915
Fatty liver in normal weight individuals
T Kimura et al.
relationship. Finally, the phenomenon of “metabolically obese but normal weight” 35,36 or “normal weight obesity,”37 defined as excessive body fat deposition in individuals with normal BMI, have been previously associated with various lifestyle-related diseases, including MS, T2DM, and CVD.38–40 Similar mechanisms involving excessive and ectopic visceral fat deposition may occur in NAFLD.41 This study has some limitations. First, this was a single-center study of one ethnic group. However, given a large sample size capturing a wide study base, as well as prior evidence linking adult weight gain to lifestyle diseases, these findings are likely applicable to broader populations; additional studies are warranted in other groups. Second, as weight at age 20 was self-reported, recall bias may have affected the results. However, previous studies have shown very high correlations between self-reported long-term recall and measured weight.42 Information on past weight was obtained before diagnostic examination. Thus, we would expect any potential misclassification of weight at age 20 to be unaffected by NAFLD diagnosis, although there is a possibility of differential misreporting of past weight in normal versus overweight/obese individuals. Our analysis, stratified by current BMI, showed associations between weight change and NAFLD consistent across all strata. While we cannot exclude the possibility that the smaller effect sizes in overweight and obese groups may have been influenced by recall bias, it is unlikely that the observed magnitude of difference is entirely due to misclassification, particularly in the context of a biologically plausible explanation. From a clinical perspective, the possibility of poor weight recall, especially among apparently healthy individuals, reinforces the need for health-care providers to play an early and active role in weight monitoring for improved long-term health, even in normal weight individuals. Finally, the sensitivity and specificity of ultrasound for diagnosing NAFLD is an inherent limitation.43 The gold standard remains liver biopsy; proton magnetic resonance spectroscopy may also offer higher sensitivity and specificity for evaluating fat deposits. However, these procedures are difficult in the primary care setting in terms of safety, accessibility, and implementation costs.43 While underdiagnosis may exist in our data, it is likely to be minimal and random in nature because of the rigorous ultrasonographic evaluation implemented in the clinical setting. Despite these limitations, our findings have a potentially large impact on clinical practice and suggest that health-care providers, uniquely poised to monitor weight over months, years, or decades, should remain vigilant for NAFLD even in normal weight patients. Similarly, these data highlight the importance of preventive healthcare models that include continuity of care in apparently healthy persons, suggesting that following trends through a dynamic adulthood has a clinically important prognostic role in preventable disease. As achieving sustained weight loss among those already overweight or obese at the time of intervention continues to pose substantial clinical challenges, our data imply that interventions aimed at halting or reversing even smaller weight gains earlier in adulthood may not only be more manageable for patients, but also effectively protect against the development lifestyle-related disease. Considering the potential importance of NAFLD on public health, weight monitoring from early adulthood, even among normal weight individuals, is an “office-friendly” parameter to evaluate and predict risk for NAFLD in daily clinical practice. 916
Acknowledgments Contributors: We would like to acknowledge Tomohiro Shinozaki, M.P.H. (Department of Biostatistics, School of Public Health, The University of Tokyo, Tokyo, Japan) for his advice on data analysis and Osamu Takahashi, M.D., M.P.H., Ph.D. (Center for Clinical Epidemiology, St. Luke’s Life Science Institute, Tokyo, Japan) for his support during manuscript development and submission.
References 1 World Health Organization. WHO global infobase: BMI/overweight/obesity. Cited 21 Mar 2014. Available from URL: https:/www0.nih.go.jp/eiken/english/research/project_nhns.htm 2 Ministry of Health, Labour and Welfare, Japan. Outline of the national health and nutrition survey Japan, 2007. Cited 21 Mar 2014. Available from URL: http://www0.nih.go.jp/eiken/english/research/ project_nhns.htm 3 International Diabetes Federation. The global burden. IDF diabetes atlas. Cited 21 Mar 2014. Available from URL: http://www.idf.org/diabetesatlas/ 4 Lazo M, Clark JM. The epidemiology of nonalcoholic fatty liver disease: a global perspective. Semin. Liver Dis. 2008; 28: 339–50. 5 Browning JD, Szczepaniak LS, Dobbins R et al. Prevalence of hepatic steatosis in an urban population in the United States: impact of ethnicity. Hepatology 2004; 40: 1387–95. 6 Church TS, Kuk JL, Ross R et al. Association of cardiorespiratory fitness, body mass index, and waist circumference to nonalcoholic fatty liver disease. Gastroenterology 2006; 130: 2023–30. 7 Bedogni G, Miglioli L, Masutti F et al. Prevalence of and risk factors for nonalcoholic fatty liver disease: the Dionysos nutrition and liver study. Hepatology 2005; 42: 44–52. 8 Lonardo A, Bellini M, Tartoni P et al. The bright liver syndrome. Prevalence and determinants of a “bright” liver echo pattern. Ital. J. Gastroenterol. Hepatol. 1997; 29: 351–6. 9 Hamaguchi M, Kojima T, Takeda N et al. The metabolic syndrome as a predictor of nonalcoholic fatty liver disease. Ann. Intern. Med. 2005; 143: 722–8. 10 Omagari K, Kadokawa Y, Masuda J et al. Fatty liver in non-alcoholic non-overweight Japanese adults: Incidence and clinical characteristics. J. Gastroenterol. Hepatol. 2002; 17: 1098–105. 11 Eguchi Y, Hyogo H, Ono M et al. Prevalence and associated metabolic factors of nonalcoholic fatty liver disease in the general population from 2009 to 2010 in Japan: a multicenter large retrospective study. J. Gastroenterol. 2012; 47: 586–95. 12 Chalasani N, Younossi Z, Lavine JE et al. The diagnosis and management of non-alcoholic fatty liver disease: practice guideline by the American Association for the Study of Liver Diseases, American College of Gastroenterology, and the American Gastroenterological Association. Hepatology 2012; 55: 2005–23. 13 Adams LA, Lymp JF, St. Sauver J et al. The natural history of nonalcoholic fatty liver disease: a population-based cohort study. Gastroenterology 2005; 129: 113–21. 14 Musso G, Gambino R, Cassader M et al. Meta-analysis: natural history of non-alcoholic fatty liver disease (NAFLD) and diagnostic accuracy of non-invasive tests for liver disease severity. Ann. Med. 2011; 43: 617–49. 15 Adams LA, Waters OR, Knuiman MW et al. NAFLD as a risk factor for the development of diabetes and the metabolic syndrome: an eleven-year follow-up study. Am. J. Gastroenterol. 2009; 104: 861–7. 16 Yoong SL, Carey ML, Sanson-Fisher RW, D’Este CA, Mackenzie L, Boyes A. A cross-sectional study examining Australian general
Journal of Gastroenterology and Hepatology 30 (2015) 909–917 © 2014 Journal of Gastroenterology and Hepatology Foundation and Wiley Publishing Asia Pty Ltd
T Kimura et al.
17
18
19 20
21
22
23
24
25
26
27
28
practitioners’ identification of overweight and obese patients. J. Gen. Intern. Med. 2014; 29: 328–34. Colditz GA, Willett WC, Rotnitzky A et al. Weight gain as a risk factor for clinical diabetes mellitus in women. Ann. Intern. Med. 1995; 122: 481–6. Chei CL, Iso H, Yamagishi K et al. Body mass index and weight change since 20 years of age and risk of coronary heart disease among Japanese: the Japan Public Health Center-Based Study. Int. J. Obes. 2008; 32: 144–51. Willett WC, Manson JE, Stampfer MJ et al. Weight, weight change, heart disease in women coronary. JAMA 1990; 273: 461–5. Rexrode KM, Hennekens CH, Willett WC et al. A prospective study of body mass index, weight change, and risk of stroke in women. JAMA 1997; 277: 1539–45. Wakasugi M, Narita I, Iseki K et al. Weight gain after 20 years of age is associated with prevalence of chronic kidney disease. Clin. Exp. Nephrol. 2012; 16: 259–68. Kojima S-I, Watanabe N, Numata M et al. Increase in the prevalence of fatty liver in Japan over the past 12 years: analysis of clinical background. J. Gastroenterol. 2003; 38: 954–61. Razak F, Anand SS, Shannon H et al. Defining obesity cut points in a multiethnic population. Circulation 2007; 115: 2111–18. World Health Organization Expert Consultation. Appropriate body-mass index for Asian populations and its implications for policy and intervention strategies. Lancet 2004; 363: 157–63. Sinn DH, Gwak GY, Park HN et al. Ultrasonographically detected non-alcoholic fatty liver disease is an independent predictor for identifying patients with insulin resistance in non-obese, non-diabetic middle-aged Asian adults. Am. J. Gastroenterol. 2012; 107: 561–7. World Health Organization/International Association for the Study of Obesity/International Obesity Task Force. The Asia-Pacific Perspective: Redefining Obesity and Its Treatment. Geneva: WHO, 2000. Alberti KGMM, Eckel RH, Grundy SM et al. Harmonizing the metabolic syndrome: a joint interim statement of the International Diabetes Federation Task Force on Epidemiology and Prevention; National Heart, Lung, and Blood Institute; American Heart Association; World Heart Federation; International Atherosclerosis Society; and International Association for the Study of Obesity. Circulation 2009; 120: 1640–5. Vernon G, Baranova A, Younossi ZM. Systematic review: the epidemiology and natural history of non-alcoholic fatty liver disease and non-alcoholic steatohepatitis in adults. Aliment. Pharmacol. Ther. 2011; 34: 274–85.
Fatty liver in normal weight individuals
29 Chang Y, Ryu S, Sung E et al. Weight gain within the normal weight range predicts ultrasonographically detected fatty liver in healthy Korean men. Gut 2009; 58: 1419–25. 30 Hocking S, Samocha-Bonet D, Milner K-L et al. Adiposity and insulin resistance in humans: the role of the different tissue and cellular lipid depots. Endocr. Rev. 2013; 34: 463–500. 31 Kaess BM, Pedley A, Massaro JM et al. The ratio of visceral to subcutaneous fat, a metric of body fat distribution, is a unique correlate of cardiometabolic risk. Diabetologia 2012; 55: 2622–30. 32 Spalding KL, Arner E, Westermark PO et al. Dynamics of fat cell turnover in humans. Nature 2008; 453: 783–7. 33 Arner P, Bernard S, Salehpour M et al. Dynamics of human adipose lipid turnover in health and metabolic disease. Nature 2011; 478: 110–13. 34 Virtue S, Vidal-Puig A. It’s not how fat you are, it’s what you do with it that counts. PLoS Biol. 2008; 6: e237. 35 Conus F, Rabasa-Lhoret R, Péronnet F. Characteristics of metabolically obese normal-weight (MONW) subjects. Appl. Physiol. Nutr. Metab. 2007; 32: 4–12. 36 Lee S-H, Ha H-S, Park Y-J et al. Identifying metabolically obese but normal-weight (MONW) individuals in a nondiabetic Korean population: the Chungju Metabolic disease Cohort (CMC) study. Clin. Endocrinol. 2011; 75: 475–81. 37 Romero-Corral A, Somers VK, Sierra-Johnson J et al. Accuracy of body mass index in diagnosing obesity in the adult general population. Int. J. Obes. 2008; 32: 959–66. 38 Romero-corral A, Somers VK, Sierra-Johnson J et al. Normal weight obesity: a risk factor for cardiometabolic dysregulation and cardiovascular mortality. Eur. Heart J. 2010; 31: 737–46. 39 Madeira F, Silva A, Veloso H et al. Normal weight obesity is associated with metabolic syndrome and insulin resistance in young adults from a middle-income country. PLoS ONE 2013; 8: e60673. 40 Wildman RP, Muntner P, Reynolds K et al. The obese without cardiometabolic risk factor clustering and the normal weight with cardiometabolic risk factor clustering: prevalence and correlates of 2 phenotypes among the US population (NHANES 1999–2004). Arch. Intern. Med. 2008; 168: 1617–24. 41 Thomas EL, Frost G, Taylor-Robinson SD et al. Excess body fat in obese and normal-weight subjects. Nutr. Res. Rev. 2012; 25: 150–61. 42 Casey VA, Dwyer JT, Berkey CS et al. Long-term memory of body weight and past weight satisfaction: a longitudinal follow-up study. Am. J. Clin. Nutr. 1991; 53: 1493–8. 43 Hernaez R, Lazo M, Bonekamp S et al. Diagnostic accuracy and reliability of ultrasonography for the detection of fatty liver: a meta-analysis. Hepatology 2011; 54: 1082–90.
Journal of Gastroenterology and Hepatology 30 (2015) 909–917 © 2014 Journal of Gastroenterology and Hepatology Foundation and Wiley Publishing Asia Pty Ltd
917
