Persistent Nonalcoholic Fatty Liver Disease Increases Risk for Carotid Atherosclerosis

Accepted Manuscript Persistent Nonalcoholic Fatty Liver Disease Increases Risk for Carotid Atherosclerosis Dong Hyun Sinn, Soo Jin Cho, Seonhye Gu, Donghyeong Seong, Danbee Kang, Hyunkyoung Kim, Byoung-Kee Yi, Seung Woon Paik, Eliseo Guallar, Juhee Cho, Geum-Youn Gwak PII: DOI: Reference:

S0016-5085(16)34613-3 10.1053/j.gastro.2016.06.001 YGAST 60504

To appear in: Gastroenterology Accepted Date: 1 June 2016 Please cite this article as: Sinn DH, Cho SJ, Gu S, Seong D, Kang D, Kim H, Yi B-K, Paik SW, Guallar E, Cho J, Gwak G-Y, Persistent Nonalcoholic Fatty Liver Disease Increases Risk for Carotid Atherosclerosis, Gastroenterology (2016), doi: 10.1053/j.gastro.2016.06.001. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. All studies published in Gastroenterology are embargoed until 3PM ET of the day they are published as corrected proofs on-line. Studies cannot be publicized as accepted manuscripts or uncorrected proofs.

1

ACCEPTED MANUSCRIPT

Persistent Nonalcoholic Fatty Liver Disease Increases Risk for Carotid Atherosclerosis

Dong Hyun Sinn,a* Soo Jin Cho,b* Seonhye Gu,c Donghyeong Seong,d Danbee Kang,d Hyunkyoung Kim,e Byoung-Kee Yi,d,f Seung Woon Paik,a Eliseo Guallar,d,g Juhee Cho,d,g

a

RI PT

Geum-Youn Gwaka

Department of Medicine, Samsung Medical Center, Sungkyunkwan University School of

Medicine, Seoul, Korea.

Center for Health Promotion, Samsung Medical Center, Seoul, Korea.

c

Biostatistics and Clinical Epidemiology Center, Samsung Medical Center, Seoul, Korea.

d

Department of Health Science and Technology, Samsung Advanced Institute for Health

M AN U

SC

b

Science and Technology, Sungkyunkwan University, Seoul, Korea. Department of Otorhinolaryngology, Samsung Medical Center, Seoul, Korea.

f

Department of Medical Informatics, Samsung Medical Center, Seoul, Korea.

g

Departments of Health, Behavior and Society and Epidemiology, Johns Hopkins Bloomberg

TE D

e

Drs. Dong Hyun Sinn and Soo Jin Cho contributed equally as first authors of this study.

AC C

*

EP

School of Public Health, Baltimore, USA.

Running Title: NAFLD and carotid atherosclerosis

Correspondence: Geum-Youn Gwak, MD, PhD, Department of Medicine, Samsung Medical Center, Sungkyunkwan University School of Medicine, 81 Irwon-ro, Gangnam-Gu, 135-710, Seoul, South Korea. Tel: +82-2-3410-3409. Fax: +82-2-3410-6983. E-mail: [email protected].

2

ACCEPTED MANUSCRIPT Specific author contributions: Sinn DH: Study design, statistical analysis, and writing of the draft manuscript. Cho SJ: Data collection, critical revision of the manuscript.

Seong D: Data collection, critical revision of the manuscript.

RI PT

Gu S: Data collection, statistical analysis, critical revision of the manuscript.

SC

Kang D: Data collection, statistical analysis, critical revision of the manuscript.

M AN U

Kim H: Data collection, statistical analysis, critical revision of the manuscript. Yi BK: Critical revision of the manuscript, study supervision.

Paik SW: Critical revision of the manuscript, study supervision.

Guallar E: Statistical analysis, critical revision of the manuscript, study supervision.

TE D

Cho J: Statistical analysis, critical revision of the manuscript, study supervision. Gwak GY: Study design, critical revision of the manuscript, study supervision.

EP

All authors approved the final submission.

AC C

Conflicts of interest

Guarantor of the article: Geum-Youn Gwak, MD, PhD. Financial support: None.

Potential competing interests: None.

3

ACCEPTED MANUSCRIPT Abstract Background & Aims: Nonalcoholic fatty liver disease (NAFLD) has been associated with subclinical atherosclerosis in cross-sectional studies. We investigated the longitudinal

RI PT

association of NAFLD with the development of subclinical carotid atherosclerosis.

Methods: We performed a retrospective cohort study of 8,020 adult men (average age, 49.2 years) without carotid atherosclerosis at baseline who underwent repeated health check-up

SC

examinations from January 1, 2005 through December 31, 2013. NAFLD status was

diagnosed by ultrasonography and classified into 4 groups based on baseline and follow-up

M AN U

findings: none, developed, regressed, or persistent NAFLD. Subclinical carotid atherosclerosis was measured by ultrasound.

Results: The age-adjusted hazard ratio for subclinical carotid atherosclerosis development

TE D

comparing participants with persistent NAFLD to those without NAFLD was 1.23 (95% confidence interval [CI], 1.13–1.35; P<.001). The association persisted after adjustment for smoking, alcohol, body mass index, and weight change (hazard ratio, 1.13; 95% CI 1.03–

EP

1.25; P=.014), but disappeared after adjustment for metabolic variables. The hazard ratio,

AC C

comparing subjects with regression of NAFLD vs those with persistent NAFLD, was 0.82 (95% CI, 0.69–0.96; P=.013). The risk of subclinical carotid atherosclerosis development was also higher among participants with high NAFLD fibrosis score, FIB-4 scores, or levels of gamma-glutamyl transferase at baseline.

Conclusion: In a large cohort study, persistent NAFLD was associated with an increased risk of subclinical carotid atherosclerosis development. This association was explained by metabolic factors that could be potential mediators of the effect of NAFLD. Markers of liver

4

ACCEPTED MANUSCRIPT fibrosis were also associated with subclinical carotid atherosclerosis development. Prospective studies are needed to determine whether treatment of NAFLD can reduce this risk.

AC C

EP

TE D

M AN U

SC

RI PT

KEY WORDS: GGT; chronic liver disease; intima-media thickness; liver steatosis

5

ACCEPTED MANUSCRIPT Introduction Nonalcoholic fatty liver disease (NAFLD) is a rapidly increasing chronic liver pathology.1, 2 Although the primary abnormality in NAFLD affects liver structure and function and may result in cirrhosis, liver failure and hepatocellular carcinoma,3 the clinical

RI PT

burden of NAFLD is not confined to liver-related morbidity and mortality. Indeed, NAFLD is associated with the metabolic syndrome, diabetes, and cardiovascular disease (CVD)

morbidity and mortality,4 indicating that NAFLD is a multisystem disease closely linked to

SC

metabolic disease and atherosclerotic CVD.

Several studies have reported that fatty liver is associated with markers of subclinical

M AN U

atherosclerosis such as coronary artery calcium,5 carotid intima-media thickness (CIMT),6 and carotid plaque,7 and a recent meta-analysis reported a strong association between NAFLD and several markers of subclinical atherosclerosis independent of traditional risk factors.8 A major limitation of the literature in this area, however, is that all studies conducted

TE D

so far have been cross-sectional, with inherent limitations in assessing causality. Therefore, we conducted a longitudinal study to assess the independent association of NAFLD with the development of subclinical carotid atherosclerosis identified by ultrasound. Subclinical

EP

carotid atherosclerosis was defined as the development of an abnormally increased CIMT or of carotid plaque, two well-established markers of subclinical atherosclerosis9 that predict

AC C

increased CVD risk.10, 11

Methods

Study Population We conducted a retrospective cohort study based on the population of men aged 20 years or older who underwent a comprehensive health check-up examination including abdominal and carotid ultrasound at Samsung Medical Center’s Health Screening Center in

6

ACCEPTED MANUSCRIPT Seoul, South Korea, between January 1, 2005 and December 31, 2013 (Figure 1). Since our objective was to evaluate the association between NAFLD status and the risk of developing subclinical carotid atherosclerosis, the analysis was restricted to men who underwent at least two screening exams including both abdominal and carotid ultrasound taken at least 1 year

RI PT

apart (n = 21,467). We excluded subjects who had any of the following exclusion criteria: history of CVD or cancer (n = 1,143); use of aspirin or other antithrombotic drugs (n = 2,385); presence of subclinical carotid atherosclerosis on ultrasound imaging at baseline (n =8,507);

SC

history of cirrhosis (n = 111); heavy alcohol drinker (≥ 30 g/day; n = 3,897); positive

serology for hepatitis B virus surface antigen or hepatitis C virus (n = 990). Since study

M AN U

participants could have more than one exclusion criteria, the number of eligible participants was 9,080. We then excluded 1,060 participants with missing information on study covariates (478 missing smoking status, 858 missing alcohol consumption, and 21 missing lipid levels). The final sample size was 8,020 men.

TE D

The study was approved by the Institutional Review Board of the Samsung Medical Center, Seoul, Korea, that waived the requirement for informed consent because the study was based on deidentified existing administrative and clinical data routinely collected for

AC C

EP

health screening purposes.

Data collection

Study participants routinely completed a self-administered health questionnaire and a

detailed physical examination in each screening exam. The questionnaire included questions on the history of diabetes, hypertension, malignancy, stroke, and CVD, as well as smoking, alcohol use and medication use. Height and weight were measured and body mass index was calculated as weight in kg divided by height in m squared. Waist circumference was measured to the nearest half-cm at the mid-point between the lower border of the rib cage and

7

ACCEPTED MANUSCRIPT the iliac crest. Blood pressure was measured using a mercury sphygmomanometer after the subject had been seated for at least 10 minutes. Hypertension was defined as a systolic blood pressure ≥ 140 mm Hg, a diastolic blood pressure ≥ 90 mm Hg, or current use of antihypertensive medications.

RI PT

Blood specimens were sampled after at least a 12-hour fast. Serum levels of total cholesterol, low-density lipoprotein (LDL) cholesterol, high-density lipoprotein (HDL) cholesterol, triglycerides, glucose, alanine aminotransferase (ALT), aspartate

SC

aminotransferase (AST), gamma-glutamyl transferase (GGT), and albumin were measured as part of the health check-up exam at Samsung Medical Center’s Department of Laboratory

M AN U

Medicine. Diabetes was defined as a fasting serum glucose ≥ 126 mg/dL or self-reported use of insulin or antidiabetic medications. Dyslipidemia was defined according to Adult Treatment Panel III criteria12 as triglyceride levels ≥ 150 mg/dl, HDL-cholesterol levels < 40 mg/dl or use of medication for dyslipidemia.

TE D

We considered ALT levels elevated if ALT was > 34 IU/L based on our previous study of a nationally representative sample of healthy Koreans from the Fourth Korea National Health and Nutrition Examination Survey.13 As there is no universally accepted cutoff of

EP

GGT levels in Korean men, we considered GGT levels elevated if GGT was > 32 U/L, the

AC C

median level in the study sample. We further classified participants ALT and GGT trajectories into 4 groups depending on the levels at baseline and at the end of follow-up (persistently low, persistently high, increasing, and decreasing). We also calculated two indices of liver fibrosis. The NAFLD fibrosis score was

calculated as -1.675 + 0.037 × age (years) + 0.094 × BMI (kg/m2) + 1.13 × impaired fasting glucose/diabetes (yes = 1, no = 0) + 0.99 × AST/ALT ratio – 0.013 × platelet count (×109/l) – 0.66 × albumin (g/dl)).14 A low NAFLD fibrosis score (< -1.455) is a strong predictor of the absence of liver fibrosis.14 The FIB-4 score was calculated as age (years) × AST

8

ACCEPTED MANUSCRIPT (U/L)/[platelet count (109/L) × ALT (U/L)1/2].15 A low FIB-4 score (< 1.45) is a strong predictor of the absence of liver fibrosis.15

Liver ultrasound

RI PT

Abdominal ultrasonography was performed by experienced radiologists at Samsung Medical Center’s Health Promotion Center. The diagnosis of fatty liver was based on

standard criteria, including parenchymal brightness, liver-to-kidney contrast, deep beam

SC

attenuation and bright vessel walls.16 Since we had excluded participants with excessive

alcohol use (≥ 30 g/d) as well as other chronic liver disease as described in the exclusion

M AN U

criteria, cases of fatty liver were considered NAFLD.

To further classify study participants based on the progression or regression of fatty liver over follow-up, we evaluated the presence of fatty liver at the baseline visit and at the last visit (defined as the first visit in which subclinical carotid atherosclerosis was identified

TE D

or the last available visit for participants who did not develop subclinical carotid atherosclerosis over follow-up). We then classified participants in 4 groups with respect to fatty liver status: 1) none, those without fatty liver both at baseline and follow-up; 2)

EP

developed, those without fatty liver at baseline but with fatty liver at follow-up; 3) regressed,

AC C

those with fatty liver at baseline but without fatty liver at follow-up; 4) persistent, those with fatty liver both at baseline and follow-up.

Carotid ultrasound

Carotid artery ultrasonography was performed using B-mode ultrasound system (Logiq 7, GE Medical System, Milwaukee, WI, USA) with a 9-MHz linear array transducer. The examination included bilateral scans of the common, internal and external carotid arteries including the bifurcation site. Abnormal CIMT was defined as a CIMT > 1.2 mm.10,

9

ACCEPTED MANUSCRIPT 17

Carotid plaque was defined as a focal wall thickening that is at least 0.5 mm or 50% greater

than surrounding CIMT, or a focal region with CIMT > 1.5 mm that protrudes into the lumen in any carotid segment.11 Subclinical carotid atherosclerosis was defined when abnormal

RI PT

CIMT or carotid plaque was found.

Statistical Analyses

The primary outcome was the development of incident subclinical carotid

SC

atherosclerosis during follow-up. Participants contributed follow-up time from the baseline visit to the visit in which subclinical carotid atherosclerosis was newly identified or to the last

M AN U

available follow-up visit among those who did not develop subclinical carotid atherosclerosis. We used Cox regression models to estimate hazard ratios (HR) with 95% confidence intervals (CI) for the development of subclinical carotid atherosclerosis associated with the presence of NAFLD adjusted for other risk factors. We fitted 3 models with progressive

TE D

degree of adjustment for potential confounders. Model 1 was adjusted for age (continuous). Model 2 was further adjusted for smoking (current, past, never), alcohol consumption (modest, none), body mass index (continuous) and weight change (continuous). Finally,

EP

model 3 was further adjusted for metabolic factors that could also be potential mediators

AC C

including systolic blood pressure (continuous), use of antihypertensive medications (yes, no), fasting blood glucose (continuous), use of hypoglycemic medications (yes, no), LDL cholesterol (continuous), HDL cholesterol (continuous), triglycerides (continuous), and use of lipid lowering medications (yes, no). P values of less than 0.05 were considered significant. All analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC) and R 3.2.1 (Vienna, Austria; http://www.R-project.org/).

Results

10

ACCEPTED MANUSCRIPT The mean (SD) age of study participants was 49.2 (7.1) years (Table 1). The prevalence of diabetes, hypertension, dyslipidemia, and smoking at baseline were 5.8, 21.8,

41.6, and 56.9%, respectively. The prevalence of NAFLD at baseline was 39.7% (n = 3,185), and of those 17.6% (562/3,185) showed regression of NAFLD by the end of follow-up.

RI PT

Among 4,835 participants without NAFLD at baseline, 23.1% (1,118/4,835) developed NAFLD by the end of follow-up.

The median follow-up period was 3.3 years (range: 1.0 – 8.9 years). During follow-up,

SC

subclinical carotid atherosclerosis developed in 2,523 participants, with a 3-year cumulative incidence of 14.3%. Those who developed subclinical carotid atherosclerosis were older,

M AN U

more likely to be current or past smokers, and metabolically unhealthy (higher body mass index, more diabetes, hypertension and dyslipidemia) than those who did not (Table 1). The 3-year cumulative incidences of subclinical carotid atherosclerosis for those with none, developed, regressed and persistent NAFLD were 13.6, 12.2, 11.4 and 16.8% ,

TE D

respectively (p < 0.001). The age-adjusted HR (95% CI) for subclinical carotid atherosclerosis development comparing participants with persistent NAFLD to those without NAFLD was 1.23 (1.13-1.35; p < 0.001; Table 2). The association persisted after further

EP

adjustment for smoking, alcohol, body mass index, and weight change (Model 2), but

AC C

disappeared after adjustment for metabolic variables (Model 3), suggesting that these factors mediate the association between NAFLD and the development of subclinical carotid atherosclerosis. The HR (95% CI) for subclinical carotid atherosclerosis development comparing participants with regression of NAFLD to those with persistent NAFLD was 0.82 (0.69 – 0.96; p = 0.013; Supplementary Table 1). This association persisted after adjustment for potential confounders and metabolic variables. In multivariable-adjusted analysis, the increased risk of subclinical carotid atherosclerosis development in participants with persistent NAFLD compared to those

11

ACCEPTED MANUSCRIPT without NAFLD was more evident in participants without diabetes, hypertension, or

dyslipidemia, but the P values for the interaction between NAFLD status and metabolic risk factor were not statistically significant (Table 3 and Supplementary Table 2). Among participants with NAFLD at baseline (n = 3,185), the risk of subclinical

RI PT

carotid atherosclerosis development was higher among participants with high levels of

NAFLD fibrosis score, FIB-4 score, ALT or GGT (Table 4, model 2). The differences for NAFLD fibrosis score, FIB-4 score, and GGT were statistically significant.

SC

Finally, the risk of subclinical carotid atherosclerosis development was higher for participants with persistently high ALT or GGT levels (Table 5). Also, the risk was

M AN U

significantly lower among participants with decreased GGT levels compared to those with persistently high GGT levels (p < 0.05 for all models). Serum GGT levels at baseline and follow-up showed a close association with NAFLD status (Supplementary Table 3).

TE D

Discussion

In this large cohort study with longitudinal assessment of the association between NAFLD status and the risk of carotid atherosclerosis, we found that participants with

EP

persistent NAFLD were at higher risk of subclinical carotid atherosclerosis development

AC C

compared to those without NAFLD. This association was mediated by metabolic risk factors. Furthermore, among participants with NAFLD at baseline, regression of NAFLD over follow-up was associated with a reduced risk of subclinical carotid atherosclerosis development. The risk of subclinical carotid atherosclerosis development was also higher among participants with high NAFLD fibrosis score, FIB-4 score, and GGT levels at baseline. As for NAFLD status, participants with persistently high GGT levels had a higher risk of subclinical carotid atherosclerosis development compared to those with persistently low GGT levels, and those with decreased GGT levels had a lower risk compared to those with

12

ACCEPTED MANUSCRIPT

persistently high GGT levels. Our findings thus indicate that persistent NAFLD may be a risk factor for the development of carotid atherosclerosis and that its effects are mediated by metabolic factors. Several previous cross-sectional studies as well as meta-analyses have established a

RI PT

cross-sectional association between NAFLD and subclinical atherosclerosis.8, 18-21 NAFLD is considered an early hepatic manifestation of the metabolic syndrome, with insulin resistance as a common pathophysiological mechanism.3, 22 NAFLD may further promote insulin

SC

resistance possibly leading to accelerated atherosclerosis.23 Indeed, previous studies have established an association between NAFLD and endothelial dysfunction, a key mechanism in

M AN U

the early development of atherosclerosis.24, 25

In our study, the association of persistent NAFLD with the development of carotid atherosclerosis was attenuated after adjusting for metabolic risk factors. Diabetes, hypertension, and dyslipidemia are well-established risk factors for the development of CVD

TE D

including carotid atherosclerosis.26 Furthermore, among participants with NAFLD at baseline in our study, the risk of developing carotid atherosclerosis was significantly reduced in those in whom NAFLD regressed over follow-up compared to those with persistent NAFLD. This

EP

is consistent with data from Sung et al. showing that resolution of NAFLD attenuated the risk

AC C

of incident hypertension over 5 years of follow-up.27 Our results suggest that the effects of NALFD may be mediated by diabetes, hypertension or dyslipidemia. These findings also highlight the need of cohort studies to understand the role of NAFLD in early cardiovascular disease.

Some studies have suggested that NAFLD uncomplicated by steatohepatitis or fibrosis is insufficient to increase CVD risk.28 Indeed, while insulin resistance is strongly associated with hepatic steatosis,29 hepatic steatosis is often self-limited and does not always accompany insulin resistance.3 In some studies, NAFLD with advanced fibrosis was more

13

ACCEPTED MANUSCRIPT strongly associated with atherosclerosis than NAFLD without fibrosis.30, 31 In our study, participants with high NAFLD fibrosis or FIB-4 scores had a higher risk of subclinical carotid atherosclerosis development. Participants with NAFLD and elevated baseline GGT, however, showed an increased risk of subclinical carotid atherosclerosis development, and

RI PT

persistently elevated GGT levels were associated with the risk of subclinical carotid

atherosclerosis. Serum GGT is a predictor of insulin resistance,32 and has been suggested as a link between fatty liver and the development of early atherosclerosis.7, 33 Further research is

SC

needed to understand the risk implications of elevated NAFLD fibrosis score, FIB-4 score, and GGT levels and their role in atherogenesis in NAFLD patients.

M AN U

In our study, participants with NAFLD at baseline who showed regression of NAFLD over follow-up showed a risk of subclinical carotid atherosclerosis comparable to that of participants without NAFLD at baseline. This observation highlights the importance of persistent NAFLD as a risk factor and suggests that resolution of NAFLD may reduce the

TE D

risk of atherosclerotic CVD. Our study, however, was not an intervention trial, so caution is needed in interpreting the data. Currently, there is no effective treatment for NAFLD.2 Most treatment strategies involve lifestyle modification such as weight reduction, hypocaloric diet,

EP

increased physical activity, smoking cessation, and treatment of any associated metabolic risk

AC C

factors (diabetes, hypertension and dyslipidemia). Since lifestyle changes reduce CVD risk,34 it is possible that the reduced risk of CVD among participants with resolved NAFLD in the present study may be the consequence of lifestyle changes and not the direct consequence of NAFLD resolution.

Other limitations also need to be considered in the interpretation of our findings. Ultrasound is a practical and safe method for assessing fatty liver and carotid plaque in clinical and population settings, but it is associated with measurement error.35 Abdominal ultrasonography can lead to an incorrect diagnosis of NAFLD in 10 – 30% of cases and

14

ACCEPTED MANUSCRIPT cannot differentiate simple steatosis to steatohepatitis.36 Also, because of the long study duration, different radiologists and vascular ultrasound technicians were involved in

performing abdominal and carotid ultrasounds over time. Study personnel collecting the data, however, were unaware of the study aims, and measurement errors in liver and carotid

RI PT

ultrasound testing were independent and non-differential. As a consequence, measurement error in the exposure and outcome variables in this study likely resulted in an underestimation of the association between NAFLD and subclinical carotid atherosclerosis development. Also,

SC

we lacked detailed data on lifestyle factors that may have changed over follow-up and may explain both changes in NALFD and risk of carotid atherosclerosis. Finally, our study

apply to other populations.

M AN U

subjects were all Korean men who visited a health check-up center, and our findings may not

Our study also had several strengths, including the longitudinal design, the large sample size, the availability of data on multiple cardiovascular, metabolic, and liver risk

TE D

factors, and the use of high quality clinical and laboratory methods that added to the strength of our findings.

In conclusion, we showed that persistent NAFLD was associated with an increased

EP

risk of subclinical carotid atherosclerosis development compared to participants who did

AC C

never developed NAFLD, while regression of NAFLD status associated with a reduced risk of carotid atherosclerosis compared to participants with persistent NAFLD. The association was mediated by metabolic risk factors such as diabetes, hypertension or dyslipidemia. The risk of subclinical carotid atherosclerosis development was also higher among participants with high NAFLD fibrosis score, FIB-4 score, and GGT levels at baseline. Prospective interventional trials focused on whether managing NAFLD can lead to the reduction of CVD risk are warranted.

ACCEPTED MANUSCRIPT 15 Figure legends

AC C

EP

TE D

M AN U

SC

RI PT

Figure 1. Flow diagram of study participants.

ACCEPTED MANUSCRIPT 17 References 1.

Chalasani N, Younossi Z, Lavine JE, et al. The diagnosis and management of nonalcoholic fatty liver disease: practice guideline by the American Gastroenterological

RI PT

Association, American Association for the Study of Liver Diseases, and American College of Gastroenterology. Gastroenterology 2012;142:1592-609. 2.

Korean Association for the Study of the L. KASL clinical practice guidelines:

3.

SC

management of nonalcoholic fatty liver disease. Clin Mol Hepatol 2013;19:325-48.

Cohen JC, Horton JD, Hobbs HH. Human fatty liver disease: old questions and new

M AN U

insights. Science 2011;332:1519-23. 4.

Byrne CD, Targher G. NAFLD: A multisystem disease. J Hepatol 2015;62:S47-S64.

5.

Sung KC, Wild SH, Kwag HJ, et al. Fatty liver, insulin resistance, and features of metabolic syndrome: relationships with coronary artery calcium in 10,153 people.

6.

TE D

Diabetes Care 2012;35:2359-64.

Targher G, Bertolini L, Padovani R, et al. Relation of nonalcoholic hepatic steatosis to early carotid atherosclerosis in healthy men: role of visceral fat accumulation. Diabetes

7.

EP

Care 2004;27:2498-500.

Kozakova M, Palombo C, Eng MP, et al. Fatty liver index, gamma-glutamyltransferase,

8.

AC C

and early carotid plaques. Hepatology 2012;55:1406-15.

Oni ET, Agatston AS, Blaha MJ, et al. A systematic review: burden and severity of subclinical cardiovascular disease among those with nonalcoholic fatty liver; should we care? Atherosclerosis 2013;230:258-67.

9.

Cicorella N, Zanolla L, Franceschini L, et al. Usefulness of ultrasonographic markers of carotid atherosclerosis (intima-media thickness, unstable carotid plaques and severe

ACCEPTED MANUSCRIPT 18 carotid stenosis) for predicting presence and extent of coronary artery disease. J Cardiovasc Med (Hagerstown) 2009;10:906-12. 10.

Stein JH, Korcarz CE, Hurst RT, et al. Use of carotid ultrasound to identify subclinical

RI PT

vascular disease and evaluate cardiovascular disease risk: a consensus statement from the American Society of Echocardiography Carotid Intima-Media Thickness Task Force. Endorsed by the Society for Vascular Medicine. J Am Soc Echocardiogr 2008;21:93-111;

11.

SC

quiz 189-90.

Touboul PJ, Hennerici MG, Meairs S, et al. Mannheim carotid intima-media thickness

M AN U

and plaque consensus (2004-2006-2011). An update on behalf of the advisory board of the 3rd, 4th and 5th watching the risk symposia, at the 13th, 15th and 20th European Stroke Conferences, Mannheim, Germany, 2004, Brussels, Belgium, 2006, and Hamburg, Germany, 2011. Cerebrovasc Dis 2012;34:290-6.

Grundy SM, Cleeman JI, Daniels SR, et al. Diagnosis and management of the metabolic

TE D

12.

syndrome: an American Heart Association/National Heart, Lung, and Blood Institute Scientific Statement. Circulation 2005;112:2735-52. Park HN, Sinn DH, Gwak GY, et al. Upper normal threshold of serum alanine

EP

13.

aminotransferase in identifying individuals at risk for chronic liver disease. Liver Int

14.

AC C

2012;32:937-44.

Angulo P, Hui JM, Marchesini G, et al. The NAFLD fibrosis score: a noninvasive system that identifies liver fibrosis in patients with NAFLD. Hepatology 2007;45:846-54.

15.

Sterling RK, Lissen E, Clumeck N, et al. Development of a simple noninvasive index to predict significant fibrosis in patients with HIV/HCV coinfection. Hepatology 2006;43:1317-25.

ACCEPTED MANUSCRIPT 19 16.

Saverymuttu SH, Joseph AE, Maxwell JD. Ultrasound scanning in the detection of hepatic fibrosis and steatosis. Br Med J (Clin Res Ed) 1986;292:13-5.

17.

Howard G, Sharrett AR, Heiss G, et al. Carotid artery intimal-medial thickness

Investigators. Stroke 1993;24:1297-304. 18.

RI PT

distribution in general populations as evaluated by B-mode ultrasound. ARIC

Mellinger JL, Pencina KM, Massaro JM, et al. Hepatic steatosis and cardiovascular

19.

SC

disease outcomes: An analysis of the Framingham Heart Study. J Hepatol 2015.

Ozturk K, Uygun A, Guler AK, et al. Nonalcoholic fatty liver disease is an independent

20.

M AN U

risk factor for atherosclerosis in young adult men. Atherosclerosis 2015;240:380-6. Kim NH, Park J, Kim SH, et al. Non-alcoholic fatty liver disease, metabolic syndrome and subclinical cardiovascular changes in the general population. Heart 2014;100:938-43. 21.

Sookoian S, Pirola CJ. Non-alcoholic fatty liver disease is strongly associated with

22.

TE D

carotid atherosclerosis: a systematic review. J Hepatol 2008;49:600-7. Than NN, Newsome PN. A concise review of non-alcoholic fatty liver disease. Atherosclerosis 2015;239:192-202.

Luo J, Xu L, Li J, et al. Nonalcoholic fatty liver disease as a potential risk factor of

EP

23.

cardiovascular disease. Eur J Gastroenterol Hepatol 2015;27:193-9. Villanova N, Moscatiello S, Ramilli S, et al. Endothelial dysfunction and cardiovascular

AC C

24.

risk profile in nonalcoholic fatty liver disease. Hepatology 2005;42:473-80.

25.

VanWagner LB, Wilcox JE, Colangelo LA, et al. Association of nonalcoholic fatty liver disease with subclinical myocardial remodeling and dysfunction: A population-based study. Hepatology 2015.

ACCEPTED MANUSCRIPT 20 26.

Bonora E, Kiechl S, Willeit J, et al. Carotid atherosclerosis and coronary heart disease in the metabolic syndrome: prospective data from the Bruneck study. Diabetes Care 2003;26:1251-7. Sung KC, Wild SH, Byrne CD. Development of new fatty liver, or resolution of existing

RI PT

27.

fatty liver, over five years of follow-up, and risk of incident hypertension. J Hepatol 2014;60:1040-5.

Ghouri N, Preiss D, Sattar N. Liver enzymes, nonalcoholic fatty liver disease, and

SC

28.

incident cardiovascular disease: a narrative review and clinical perspective of prospective

29.

M AN U

data. Hepatology 2010;52:1156-61.

Kawano Y, Cohen DE. Mechanisms of hepatic triglyceride accumulation in nonalcoholic fatty liver disease. J Gastroenterol 2013;48:434-41.

30.

Chen Y, Xu M, Wang T, et al. Advanced fibrosis associates with atherosclerosis in

31.

TE D

subjects with nonalcoholic fatty liver disease. Atherosclerosis 2015;241:145-50. Sunbul M, Agirbasli M, Durmus E, et al. Arterial stiffness in patients with non-alcoholic fatty liver disease is related to fibrosis stage and epicardial adipose tissue thickness.

32.

EP

Atherosclerosis 2014;237:490-3.

Ryoo JH, Oh CM, Kim HS, et al. Clinical association between serum gamma-

AC C

glutamyltransferase levels and the development of insulin resistance in Korean men: a 5year follow-up study. Diabet Med 2014;31:455-61.

33.

Fraser A, Harris R, Sattar N, et al. Gamma-glutamyltransferase is associated with incident vascular events independently of alcohol intake: analysis of the British Women's Heart and Health Study and Meta-Analysis. Arterioscler Thromb Vasc Biol 2007;27:2729-35.

ACCEPTED MANUSCRIPT 21 34.

Lloyd-Jones DM, Hong Y, Labarthe D, et al. Defining and setting national goals for cardiovascular health promotion and disease reduction: the American Heart Association's strategic Impact Goal through 2020 and beyond. Circulation 2010;121:586-613. Hernaez R, Lazo M, Bonekamp S, et al. Diagnostic accuracy and reliability of

RI PT

35.

ultrasonography for the detection of fatty liver: A meta-analysis. Hepatology 2011. Dasarathy S, Dasarathy J, Khiyami A, et al. Validity of real time ultrasound in the

EP

TE D

M AN U

SC

diagnosis of hepatic steatosis: a prospective study. J Hepatol 2009;51:1061-7.

AC C

36.

ACCEPTED MANUSCRIPT Table 1. Characteristics of Study Population at Baseline.

49.2±7.1 24.6±2.6

1,294 (23.5) 1,087 (19.8) 3,116 (56.7)

1,232 (15.4) 6,788 (84.6) 95.8±17.3

828 (15.1) 4,669 (84.9) 95.0±16.0

404 (16.0) 2,119 (84.0) 97.4±19.7

<0.001

156 (2.8)

113 (4.5)

<0.001

278 (5.1) 117.0±14.8 74.9±10.7

185 (7.3) 118.1±14.9 75.0±10.4

<0.001 0.004 0.55

603 (11.0)

332 (13.2)

0.005

1,752 (21.8) 195.2±32.1 125.6±28.5 52.2±12.6 146.7±85.2

1,159 (21.1) 193.6±31.8 123.7±28.0 52.4±12.7 144.1±84.4

593 (23.5) 198.6±32.3 129.7±29.1 51.9±12.3 152.1±86.6

0.015 <0.001 <0.001 0.21 <0.001

86 (1.1)

69 (1.3)

17 (0.7)

0.019

3,335 (41.6) 28.0±18.8 45.8±44.6

2,224 (40.5) 28.0±19.3 44.5±43.4

1,111 (44.0) 27.9±17.8 48.5±47.0

0.003 0.10 <0.001 0.006

3,717 (46.4) 1,118 (13.9) 562 (7.0) 2,623 (32.7)

2,615 (47.6) 759 (13.8) 386 (7.0) 1,737 (31.6)

1,102 (43.7) 359 (14.2) 176 (7.0) 886 (35.1)

463 (5.8) 117.3±14.8 74.9±10.6

TE D

935 (11.7)

M AN U

269 (3.4)

AC C

512 (20.3) 567 (22.5) 1,444 (57.2)

SC

1,806 (22.5) 1,654 (20.6) 4,560 (56.9)

EP

Age (years) BMI (kg/m2) Smoking Never Past Current Alcohol consumption None Modest FBG (mg/dl) Hypoglycemic medication Diabetes* SBP (mmHg) DBP (mmHg) Antihypertensive medication Hypertension* TC (mg/dl) LDL-C (mg/dl) HDL-C (mg/dl) Triglyceride (mg/dl) Dyslipidemia medication Dyslipidemia* ALT (U/L) GGT (U/L) NAFLD None Developed Regressed Persistent

RI PT

All (n= 8,020)

Developed subclinical carotid atherosclerosis over follow-up No Yes P value (n = 5,497) (n = 2,523) 48.4±7.0 51.0±7.0 <0.001 24.5±2.6 24.8±2.5 <0.001 0.001

0.27

BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure; TC, total cholesterol, LDLC, low density lipoprotein cholesterol; HDL-C, high density lipoprotein cholesterol; FBG, fasting blood glucose; ALT, alanine aminotransferase; GGT, gamma-glutamyl transferase; NAFLD, non-alcoholic fatty liver disease. Values are expressed as mean ± standard deviation or number (%).

1

ACCEPTED MANUSCRIPT *

AC C

EP

TE D

M AN U

SC

RI PT

Definitions: Hypertension, SBP ≥ 140 mmHg or DBP ≥ 90 mmHg or use of antihypertensive medication; Dyslipidemia, HDL-C < 40 mg/dl or triglyceride ≥ 150 mg/dl or use of dyslipidemia medication; Diabetes, FBS ≥ 126 mg/dl or use of hypoglycemic medication.

2

ACCEPTED MANUSCRIPT

Table 2. Risk of Subclinical Carotid Atherosclerosis Development According to Non-Alcoholic Fatty Liver Disease (NAFLD) Status. Model 2 HR (95% CI) P value

RI PT

None (n = 3,717) Developed (n = 1,118) Regressed (n = 562)† Persistent (n = 2,623)†

Model 1 HR (95% CI) P value

1.00 (Reference) 1.00 (Reference) 1.00 (Reference) 1.00 (0.88-1.12) 0.94 1.05 (0.93-1.18) 0.45 0.96 (0.85-1.09) 0.97 (0.83-1.14) 0.73 0.97 (0.83-1.14) 0.73 0.98 (0.83-1.15) 1.20 (1.10-1.31) <0.001 1.23 (1.13-1.35) <0.001 1.13 (1.03-1.25)

SC

NAFLD status

Unadjusted HR (95% CI) P value

0.52 0.77 0.014

Model 3 HR (95% CI) P value 1.00 (Reference) 0.91 (0.80-1.03) 0.89 (0.75-1.05) 1.00 (0.90-1.10)

0.12 0.15 0.92

AC C

EP

TE D

M AN U

NAFLD, non-alcoholic fatty liver disease; HR, hazard ratio; CI, confidence interval. Model 1. Adjusted for age. Model 2. Further adjusted for smoking, alcohol, body mass index (kg/m2) and weight change (kg/year). Model 3. Further adjusted for systolic blood pressure, use of antihypertensive medications, fasting blood glucose, use of hypoglycemic medications, LDL cholesterol, HDL cholesterol, triglycerides and use of lipid lowering medications. † P values for regressed versus persistent NAFLD were < 0.05 for all models (Supplementary Table 1).

3

ACCEPTED MANUSCRIPT

None (n = 3,717) HR (95% CI)

NAFLD Status Developed (n = 1,118) Regressed† (n = 562) HR (95% CI) P value HR (95% CI) P value 0.95 (0.84-1.08) 1.18 (0.67-2.07)

0.42 0.56

1.00 (Reference) 1.00 (Reference)

0.95 (0.83-1.10) 0.94 (0.73-1.22)

0.49 0.65

1.00 (Reference) 1.00 (Reference)

1.04 (0.88-1.22) 0.82 (0.67-0.99)

0.66 0.043

1.01 (0.85-1.19) 0.57 (0.27-1.21)

Persistent† (n = 2,623) HR (95% CI) P value

0.95 0.14

1.10 (0.99-1.22) 1.14 (0.79-1.64)

0.067 0.49

1.00 (0.83-1.20) 0.89 (0.63-1.25)

0.98 0.48

1.14 (1.02-1.27) 1.06 (0.86-1.30)

0.027 0.58

0.92 (0.73-1.16) 0.97 (0.76-1.23)

0.46 0.79

1.12 (0.97-1.29) 1.05 (0.90-1.22)

0.12 0.54

SC

1.00 (Reference) 1.00 (Reference)

M AN U

Diabetes* No (n = 7,557) Yes (n = 463) Hypertension* No (n = 6,268) Yes (n = 1,752) Dyslipidemia* No (n = 4,685) Yes (n = 3,335)

RI PT

Table 3. Risk of Subclinical Carotid Atherosclerosis Development According to Non-Alcoholic Fatty Liver Disease (NAFLD) Status in Clinically Relevant Subgroups.

AC C

EP

TE D

HR, hazard ratio; CI confidence interval. Definitions: Diabetes, FBS ≥ 126 mg/dl or use of hypoglycemic medication; Hypertension, SBP ≥ 140 mmHg or DBP ≥ 90 mmHg or use of antihypertensive medication; Dyslipidemia, HDL-C < 40 mg/dl or triglyceride ≥ 150 mg/dl or use of dyslipidemia medication. Adjusted for age, smoking, alcohol, body mass index (kg/m2) and weight change (kg/year). * P values for interaction between NAFLD status and metabolic abnormality were 0.40 for diabetes, 0.32 for hypertension, and 0.33 for dyslipidemia. † P values for regressed versus persistent were 0.021, 0.050, 0.013, 0.13, 0.024 and 0.088 for non-diabetes, diabetes, non-hypertension, hypertension, non-dyslipidemia and dyslipidemia, respectively (Supplementary Table 2).

4

ACCEPTED MANUSCRIPT

Table 4. Severity of Non-Alcoholic Fatty Liver Disease (NAFLD) at Baseline and Risk of Subclinical Carotid Atherosclerosis Development.

HR (95% CI)

Model 1 P value

HR (95% CI)

Model 2 P value

NAFLD fibrosis score 1.00 (Reference)

High (≥-1.455) (n = 720)

1.35 (1.17-1.55)

NA <0.001

1.00 (Reference)

High (≥1.45) (n = 341)

1.43 (1.20-1.71)

NA <0.001

NA

ALT levels (U/L) 1.00 (Reference)

High (>34) (n = 1,184)

0.99 (0.87-1.12)

1.00 (Reference) 0.86

GGT levels (U/L) 1.00 (Reference)

High (>32) (n = 2,088)

1.10 (0.97-1.26)

1.07 (0.94-1.21)

0.31

1.00 (Reference)

EP

Low (≤32) (n = 1,097)

--

TE D

Low (≤34) (n = 2,001)

--

M AN U

Low (<1.45) (n = 2,844)

0.13

Model 3 P value

1.00 (Reference)

NA

FIB-4 score

HR (95% CI)

SC

Low (<-1.455) (n = 2,463)

RI PT

Unadjusted

1.20 (1.05-1.36)

1.31 (1.20-1.44)

<0.001

<0.001

1.17 (1.03-1.34)

<0.001

1.43 (1.19-1.70)

<0.001

1.00 (Reference) 0.25

1.00 (Reference) 0.007

1.29 (1.17-1.41)

1.00 (Reference)

1.00 (Reference) 1.08 (0.95-1.23)

P value

1.00 (Reference)

1.00 (Reference) 1.45 (1.21-1.74)

HR (95% CI)

1.04 (0.91-1.19)

0.54

1.00 (Reference) 0.020

1.12 (0.97-1.28)

0.12

AC C

Abbreviation: NAFLD, nonalcoholic fatty liver disease; FIB-4, fibrosis-4; ALT, alanine aminotransferase; GGT, gamma glutamyl transferase; HR, hazard ratio; CI, confidence interval; NA, not applicable. Model 1. Adjusted for age. Age-adjusted models were not calculated for NAFLD fibrosis score or for FIB-4 score because age is used for calculating the scores. Model 2. Further adjusted for smoking, alcohol, BMI (kg/m2) and weight change (kg/year). For NAFLD fibrosis score, the model was not adjusted age or BMI. For FIB-4 score the model was not adjusted for age. Model 3. Further adjusted for systolic blood pressure, use of antihypertensive medications, fasting blood glucose, use of hypoglycemic medications, LDL cholesterol, HDL cholesterol, triglycerides and use of lipid lowering medications. For NAFLD fibrosis score, the model was not adjusted age, BMI or fasting blood glucose. For FIB-4 score the model was not adjusted for age.

5

ACCEPTED MANUSCRIPT

Table 5. Risk of Subclinical Carotid Atherosclerosis Development According to Serum Alanine Aminotransferase or Gamma Glutamyl Transferase Levels.

0.77 0.70 0.10

1.00 (Reference) 1.00 (0.85-1.18) 0.96 0.91 (0.79-1.03) 0.14 1.23 (1.13-1.34) <0.001

Model 2 HR (95% CI) P value

Model 3 HR (95% CI) P value

1.00 (Reference) 1.11 (0.97-1.27) 0.14 1.09 (0.96-1.24) 0.19 1.28 (1.12-1.45) <0.001

1.00 (Reference) 1.06 (0.93-1.22) 0.38 1.06 (0.93-1.21) 0.35 1.18 (1.03-1.35) 0.017

1.00 (Reference) 1.12 (0.95-1.32) 0.17 0.93 (0.81-1.06) 0.26 1.37 (1.26-1.50) <0.001

1.00 (Reference) 1.00 (Reference) 1.06 (0.90-1.24) 0.52 1.05 (0.89-1.24) 0.54 0.90 (0.79-1.04) 0.14 0.85 (0.74-0.97) 0.020 1.27 (1.16-1.39) <0.001 1.20 (1.09-1.32) <0.001

SC

1.00 (Reference) 1.02 (0.89-1.17) 1.03 (0.90-1.16) 1.11 (0.98-1.27)

M AN U

ALT levels during follow-up Persistently low (n = 5,540) Elevated (n = 780) Decreased (n = 872) Persistently high (n = 828) GGT levels during follow-up Persistently low (n = 3,382) Elevated (n = 589) Decreased (n = 842) Persistently high (n = 3,207)

Model 1 HR (95% CI) P value

RI PT

Unadjusted HR (95% CI) P value

1.00 (Reference) 0.99 (0.87-1.14) 1.01 (0.88-1.15) 1.09 (0.95-1.25)

0.91 0.93 0.24

AC C

EP

TE D

HR, hazard ratio; CI confidence interval. Cutoffs to define elevated levels were 34 U/L for alanine aminotransferase and 32 U/L gamma-glutamyl transferase. Model 1. Adjusted for age. Model 2. Further adjusted for smoking, alcohol, BMI (kg/m2) and weight change (kg/year). Model 3. Further adjusted for systolic blood pressure, use of antihypertensive medications, fasting blood glucose, use of hypoglycemic medications, LDL cholesterol, HDL cholesterol, triglycerides and use of lipid lowering medications.

6

ACCEPTED MANUSCRIPT

Supplementary Table 1. Resolution of Non-Alcoholic Fatty Liver Disease (NAFLD) and Risk of Subclinical Carotid Atherosclerosis Development.

Regressed (n = 562) Persistent (n = 2,623)

Unadjusted HR (95% CI) P value 0.82 (0.69-0.96) 0.013 1.00 (Reference)

Model 1 HR (95% CI) P value 0.79 (0.67-0.93) 0.004 1.00 (Reference)

Model 2 HR (95% CI) P value 0.77 (0.65-0.91) 0.003 1.00 (Reference)

RI PT

NAFLD status

Model 3 HR (95% CI) P value 0.79 (0.67-0.94) 0.009 1.00 (Reference)

AC C

EP

TE D

M AN U

SC

Model 1. Adjusted for age. Model 2. Further adjusted for smoking, alcohol, BMI (kg/m2) and weight change (kg/year). Model 3. Further adjusted for systolic blood pressure, use of antihypertensive medications, fasting blood glucose, use of hypoglycemic medications, LDL cholesterol, HDL cholesterol, triglycerides and use of lipid lowering medications.

7

ACCEPTED MANUSCRIPT Supplementary Table 2. Resolution of Non-Alcoholic Fatty Liver Disease (NAFLD) and Risk of Subclinical Carotid Atherosclerosis Development in Clinically-Relevant Subgroups. NAFLD Regressed (n = 562) HR (95% CI) P value 0.021 0.050

1.00 (Reference) 1.00 (Reference)

0.78 (0.64-0.95) 0.76 (0.53-1.09)

0.013 0.13

0.75 (0.85-0.96) 0.82 (0.64-1.03)

RI PT

0.81 (0.68-0.97) 0.48 (0.23-1.00)

SC

Diabetes* No (n = 7,557) Yes (n = 463) Hypertension* No (n = 6,268) Yes (n = 1,752) Dyslipidemia* No (n = 4,685) Yes (n = 3,335)

Persistent (n = 2,623)

0.024 0.088

1.00 (Reference) 1.00 (Reference)

1.00 (Reference) 1.00 (Reference)

AC C

EP

TE D

M AN U

Adjusted for age, smoking, alcohol, body mass index (kg/m2) and weight change (kg/year). * P values for interaction between NAFLD status and metabolic abnormality were 0.21 for diabetes, 0.95 for hypertension, and 0.44 for dyslipidemia.

8

ACCEPTED MANUSCRIPT Supplementary Table 3. Gamma Glutamyl Transferase (GGT) Levels by Non-Alcoholic Fatty Liver Disease (NAFLD) Status.

AC C

EP

TE D

M AN U

SC

None (n = 3,717) Developed (n = 1,118) Regressed (n = 562) Persistent (n = 2,623) P value

GGT levels at end of follow-up (U/L) 37.5±48.6 52.2±61.0 38.2±45.0 53.0±46.5 <0.001

RI PT

GGT levels at baseline (U/L) 37.5±38.8 47.0±43.9 48.6±42.4 56.4±50.4 <0.001

NAFLD status

9

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT