The Prevalence of Non-alcoholic Fatty Liver Disease and Diabetes Mellitus in an Iranian Population.

BACKGROUND Type II diabetes mellitus (T2DM) and non-alcoholic fatty liver disease (NAFLD) are important causes of morbidity and mortality worldwide. We aimed to estimate the prevalence of DM in the context of NAFLD. METHODS In this cross-sectional study, we studied 5052 participants, aged 18 years and older, of a baseline population-based cohort in northern Iran (N=6143). The prevalence of DM was estimated in individuals with and without NAFLD. The association between NAFLD and T2DM was evaluated using logistic regression with the adjustment of confounding effects of age, sex, body mass index, lipid profiles, and fasting insulin. RESULTS In men, the prevalence (95% confidence interval) of T2DM was 5.34% (4.35%-6.34%) and 15.06% (13.12%-17.00%) in individuals without and with NAFLD, respectively (p <0.001). In women without NAFLD, the prevalence was 8.27% (6.83%-9.71%) while in the presence of NAFLD, the prevalence was 27.21% (24.59%-29.83%), (p <0.001). In univariate analysis, the chance of having T2DM was 3.700 (3.130-4.380) times more in patients with NAFLD compared with subjects without NAFLD (p<0.001). This chance was reduced (Odds Ratio=1.976, 95% CI: 1.593-2.451, p <0.001) after removing the effects of other variables. CONCLUSION The prevalence of T2DM is increased in the context of NAFLD. This condition may be considered as an independent predictor of T2DM.

INTRODUCTION

Non-alcoholic fatty liver disease (NAFLD) is defined by cytoplasmic accumulation of fat in hepatocytes when the other causes of steatosis, such as excessive alcohol consumption, drugs, hepatitis C and B or genetic disorders are ruled out.[1] The prevalence of NAFLD in general population of USA was estimated to be 27-34%.[2] The prevalence of NAFLD was 60-70% in American patients with diabetes and 42.6% and 69.5% in European patients with diabetes.[3,4] Based on one meta-analysis, the prevalence of diabetes mellitus (DM) in adults>40 years was 24% and it was estimated that the prevalence increases by 0.4% with one year increase in age after 20 years of age.[5]

NAFLD is considered as an emerging healthcare problem in Asian countries,[6,7] with a prevalence of 15-45%, which is rising.[7] The rising trend in the prevalence of obesity plays a critical role in the increasing trend of NAFLD prevalence.[8-11] According to the results of a meta-analysis in Iran, the prevalence of NAFLD was estimated 33.9%.[12]

On the other hand, the prevalence of type II diabetes mellitus (T2DM) and other metabolic disorders is dramatically increasing worldwide.[13,14] Previous studies showed an increase in the incidence of NAFLD in case of T2DM.[15] Consequently, the increasing trend of NAFLD seems likely to continue in upcoming years due to the increasing prevalence of obesity and T2DM.[15-17]

Iran, as one of the low and middle income countries (LMIC), is experiencing a high prevalence of obesity, which is predicted to keep rising in near future.[18] Regarding the growing trend in the prevalence of both NAFLD and T2DM and the association between these two conditions, the present study was conducted to estimate and compare the prevalence of T2DM in individuals with and without NAFLD.

MATERIALS AND METHODS

In compliance with the principles of the Declaration of Helsinki, we studied 5052 of 6143 participants of a baseline cohort study of individuals aged 10 to 90 years.[19-24], Our study subjects were 18 years and older among whom 2859 (56.6%) were male and 2193 (43.4%) were female. Enrolment process is shown in figure 1. Since September 2008, the baseline cohort study has been conducted in Amol, a populated city in northern Iran, by the Gastrointestinal and Liver Disease Research Centre (GILDRC) of Iran University of Medical Sciences (IUMS). Research Ethics Committee of IUMS approved the study protocol (reference number: IR.IUMS.rec.1394.19753) and all participants gave informed consent prior to their inclusion in the study. The written informed consent (in Persian) contained statements pertaining the rationale and benefits of participation in the cohort, the procedures such as blood sampling, the right to withdraw from the study at any time without giving reasons and without any penalty, and confidentiality of any disclosing data for each participant. The sampling frame was obtained from local health centers, where almost all the study participants had health records. Further explanation of sampling of cohort study is stated elsewhere.[25,26]

Fig.1

Schematic flowchart of enrolment and exclusions (HBsAg: hepatitis B virus surface antigen, anti-HCV Ab: anti hepatitis C virus antibody

Data collection: Anthropometric measures (weight, height) and blood pressure were measured by trained healthcare staff. Height was measured at standing position with heels and buttocks pressed up against a wall. Blood pressure was measured using a fitted sphygmomanometer cuff after resting 5 minutes in a quiet room. We used Korotkoff sounds to determine systolic blood pressure (SBP) and diastolic blood pressure (DBP).

After 12-hour fasting, venous blood sample was taken from each participant to assess fasting blood sugar (FBS) and lipid profiles. All laboratory tests, including FBS, triglycerides (TGs), high-density lipoprotein (HDL), low-density lipoprotein (LDL), cholesterol, and aminotransferases were assessed enzymatically using BS200 Auto-analyzer (Mindray, Shenzhen, China). 10% of the blood samples were re-evaluated by the Iranian National Reference Laboratory. The coefficients of variation ranged from 1.7% to 3.8% for all laboratory values.

NAFLD was defined by sonographic evidence of hepatic steatosis and ruling out other causes of acute or chronic hepatitis as well as secondary hepatic fat accumulation due to alcohol consumption, taking steatogenic medications, or hereditary disorders. A single expert radiologist did all ultrasound examinations using a 3-5 MHz transducer. Sagittal, longitudinal, lateral, and intercostal views of the liver parenchyma were obtained. Steatosis was confirmed when there was a markedly increased hepatic echogenicity. T2DM was defined by history of type II DM or taking anti-hyperglycemic drugs (oral agents, insulin), or a FBS≥125 mg/dL without any history of type I DM.

Data analyses: We used descriptive statistics (frequency distribution, mean and standard deviation) to report the findings. Two group proportion test was used to compare the prevalence of T2DM in patients with and without NAFLD. The population was divided into 16 strata based on sex and age groups with 10-year intervals and comparisons were made in subgroups. Independent t test was used to compare normally distributed variables, such as age, body mass index (BMI), SBP, DBP, mean arterial pressure (MAP), FBS, TG, total cholesterol, HDL, LDL, and aminotransferases. We used Mann-Whitney test to compare fasting insulin levels and insulin resistance indices (HOMA-IR) as they were not normally distributed in the two groups of patients (with and without NAFLD). We evaluated the associations between T2DM and age, sex, BMI, MAP, TG, LDL, HDL, fasting insulin, and NAFLD by univariate and multivariate logistic regression models and reported them in terms of odds ratio (OR) and confidence intervals (CI) and P values. The significance level for all analyses was considered 0.05. The data were inferred statistically using IBM SPSS software version 22 (IBM Corp., Armonk, USA) and STATA software version 12 (StataCorp, Texas, USA).

RESULTS

Our study subjects were 18 years and older among whom 2860 (56.6%) were male and 2192 (43.4%) were female with a mean age (±standard deviation) of 44.77 (±16.77) and 43.78 (±15.43) years, respectively. Table 1 shows the basic characteristics of the study population in “NAFLD” and “without NAFLD” groups. In both sexes, all characteristics were significantly higher in “NAFLD” group compared with “without NAFLD” group, except HDL that was significantly lower in “NAFLD” group. Table 2 demonstrates the prevalence of T2DM in people with and without NAFLD stratified by sex and age groups. There was a significantly higher prevalence of T2DM in NAFLD group in all subgroups except for women aged 65 years and older. We defined T2DM as the outcome variable to perform univariate and multivariate logistic regression analyses. According to univariate analysis, the relative risk of having T2DM in patients with NAFLD, compared with subjects without NAFLD, was 3.700 (95% CI: 3.130-4.380, p<0.001). After adjustment of confounding variables including age, sex, BMI, TG, LDL, HDL, and insulin, the risk was reduced to 1.976 (95% CI: 1.593-2.451, p <0.001). Among variables entered into multivariate model, age (p <0.001), sex (p <0.001), TG (p <0.001), HDL (p =0.001), and fasting insulin (p =0.028) had a significant association with T2DM in addition to NAFLD (p <0.001). More details are represented in table 3.

Table 1

Basic characteristics of the study subjects with and without non-alcoholic fatty liver disease

Sex	Variables	Mean±Standard Deviation		Mean difference (95% CI)
		Without NAFLD	With NAFLD
Male	Age (years)	42.21±17.79	48.58±14.29	6.37 (5.22-7.52)
	BMI (Kg/m2)	24.33±3.66	29.66±3.98	5.32 (5.06-5.59)
	SBP (mmHg)	114.27±14.84	122.14±15.94	7.87 (6.81-8.94)
	DBP (mmHg)	73.74±12.19	80.93±12.21	7.19 (6.33-8.04)
	MAP (mmHg)	87.25±12.19	94.66±12.66	7.41 (6.55-8.28)
	FBS (mg/dL)	94.73±26.89	104.42±32.99	9.69 (7.63-11.75)
	TG (mg/dL)	123.90±78.40	175.46±113.84	51.56 (44.75-58.38)
	Total Cholesterol (mg/dL)	171.13±40.70	189.48±41.39	18.35 (15.48-21.21)
	HDL (mg/dL)	45.46±11.41	40.21±10.84	-5.25 (-6.05- -4.45)
	LDL (mg/dL)	99.82±30.40	112.16±30.30	12.34 (10.17-14.50)
	Fasting insulin (mIU/L)	7.40±5.37	11.08±7.89	3.68 (3.23-4.13)
	HOMA-IR	1.74±1.39	2.86±2.32	1.12 (0.99-1.25)
	AST (IU/L)	22.74±10.75	25.84±14.86	3.10 (2.22-3.97)
	ALT (IU/L)	22.44±15.57	32.46±21.95	10.02 (8.73-11.30)
Female	Age (years)	38.28±15.28	50.72±12.54	12.44 (11.32-13.55)
	BMI (Kg/m2)	26.96±4.85	33.07±4.73	6.11 (5.73-6.49)
	SBP (mmHg)	110.36±15.89	121.82±17.91	11.47 (10.16-12.78)
	DBP (mmHg)	72.28±12.24	80.27±12.68	7.99 (7.01-8.97)
	MAP (mmHg)	84.97±12.28	94.11±13.51	9.14 (8.12-10.15)
	FBS (mg/dL)	95.29±31.17	114.97±49.19	19.68 (16.52-22.84)
	TG (mg/dL)	115.14±66.10	171.83±114.92	56.70 (49.25-64.14)
	Total Cholesterol (mg/dL)	180.71±40.66	200.51±43.81	19.80 (16.48-23.11)
	HDL (mg/dL)	49.04±11.69	43.01±11.72	-6.03 (-6.98- -5.08)
	LDL (mg/dL)	104.87±30.53	116.56±31.74	11.69 (9.18-14.19)
	Fasting insulin (mIU/L)	9.42±6.50	11.55±7.15	2.13 (1.59-2.66)
	HOMA-IR	2.22±1.73	3.35±3.02	1.12 (0.93-1.31)†
	AST (IU/L)	18.79±6.72	21.47±12.35	2.67 (1.92-3.43)
	ALT (IU/L)	17.34±12.29	22.42±15.51	5.08 (3.99-6.16)

ALT: alanine aminotransferase, AST: aspartate aminotransferase, BMI: body mass index, CI: confidence interval, DBP: diastolic blood pressure, FBS: fasting blood sugar, HDL: high-density lipoprotein, HOMA-IR: homeostatic model assessment-insulin resistance, LDL: low-density lipoprotein, MAP: mean arterial pressure, NAFLD: non-alcoholic fatty liver disease, SBP: systolic blood pressure, TG: triglycerides

Table 2

Prevalence of type 2 diabetes mellitus in subjects with and without non-alcoholic fatty liver disease according to sex and age groups

Sex	Age group (years)	Percentage		OR (95%CI)
		Without NAFLD	With NAFLD
Male	18-39	1.27	4.49	3.667 (1.717-7.830)
	40-64	6.67	15.59	2.590 (1.832-3.662)
	≥65	16.10	33.34	2.605 (1.664-4.078)
	Total≥18	5.34	15.06	3.140 (2.450-4.024)
Female	18-39	1.58	9.95	6.887 (3.387-14.00)
	40-64	16.39	30.36	2.224 (1.669-2.963)
	≥65	24.49	35.81	1.720 (0.973-3.042)
	Total ≥18	8.27	27.21	4.147 (3.290-5.227)

NAFLD: non-alcoholic fatty liver disease, OR: odd ratio

Table 3

Results of univariate and multivariate logistic regression analyses with type 2 diabetes mellitus as outcome variable

Analysis	Variables	𝛽 coefficient (SE)	Wald test	Odd ratio (95% CI)
Univariate	Age	0.058 (0.003)	423.51	1.059 (1.053-1.065)
	Sex	0.673 (0.081)	69.64	1.960 (1.674-2.296)
	BMI	0.081(0.007)	124.13	1.084 (1.069-1.100)
	MAP	0.032 (0.003)	127.70	1.033 (1.027-1.039)
	TG	0.004 (0.0003)	105.31	1.004 (1.003-1.005)
	LDL	0.007 (0.001)	33.96	1.007 (1.005-1.010)
	HDL	-0.026 (0.004)	49.69	0.974 (0.967-0.981)
	Fasting Insulin	0.018 (0.005)	12.75	1.019 (1.008-1.029)
	NAFLD	1.308 (0.086)	232.38	3.700 (3.130-4.380)
Multivariate	Age	0.061(0.003)	303.351	1.063 (1.055-1.070)
	Sex	0.846 (0.099)	72.287	2.329 (1.917-2.831)
	BMI	-0.008 (0.011)	0.528	0.992 (0.970-1.014)
	MAP	0.005 (0.003)	2.017	1.005 (0.998-1.012)
	TG	0.003 (0.0005)	39.174	1.003 (1.002-1.004)
	LDL	-0.002 (0.001)	1.744	0.998 (0.995-1.001)
	HDL	-0.014 (0.004)	10.660	0.986 (0.978-0.995)
	Fasting Insulin	0.015 (0.007)	4.847	1.015 (1.002-1.029)
	NAFLD	0.681 (0.110)	38.411	1.976 (1.593-2.451)

BMI: body mass index, CI: confidence interval, HDL: high-density lipoprotein, LDL: low-density lipoprotein, MAP: mean arterial pressure, NAFLD: non-alcoholic fatty liver disease, SE: standard error, TG: triglycerides.

DISCUSSION

Our results showed that in both men and women, the prevalence of T2DM dramatically increases in patients with NAFLD compared with individuals without NAFLD. Based on our results, 15% of men and 27% of women aged 18 years and older with NAFLD had comorbid T2DM, while the prevalence was 5% and 8% in men and women without NAFLD, respectively. Similar results were reproduced after the data were analyzed according to age groups. In an agreement with our findings, Ortiz-Lopez and colleagues reported a dramatically higher prevalence of pre-diabetes and T2DM in patients with NAFLD compared with ones without NAFLD.[27] Prior population-based studies showed an 11-18% prevalence of T2DM in patients with NAFLD.[28,29] NAFLD was suggested as a better predictor for T2DM than BMI in a Chinese population.[30] Another observation estimated that 66% of patients with diabetes would have NAFLD and significant hepatic fibrosis.[31]

Patients with NAFLD have a 3.700 (95% CI: 3.130-4.380) times more chance of having T2DM compared with subjects without NAFLD, we observed. A previous study calculated a relative risk of 4.492 (95% CI: 3.640-5.542) for T2DM in patients with NAFLD compared with individuals without NAFLD.30 We found a significant association between NAFLD and T2DM after adjusting confounding effects of other variables, such as BMI, MAP, lipid profiles, fasting insulin, age, and sex (Odds Ratio=1.976; 95% CI: 1.593-2.451, P<0.001). In line with our results, Li and co-workers declared NAFLD as a risk factor for T2DM independent of obesity or overweight.[30]

As mentioned before, we evaluated the prevalence of T2DM in the patients with NAFLD in different age groups (18-39, 40-64, over 64 years) of both sexes. The prevalence of T2DM was significantly higher in patients with NAFLD than patients without this condition in all age groups and the difference was markedly greater in 40-64 years age group of both sexes and in men over 64 years. Note worthily, NAFLD is thought to be a hepatic manifestation of metabolic syndrome (MetS) and a strong risk factor of MetS and also an important predictor of development of T2DM.[32-34]

The excessive liver fat is associated with hepatic, adipose tissue, and muscle insulin resistance. Consequently, individuals harboring fatty liver, either with or without obesity, may be predisposed to develop MetS and T2DM. On the other hand, T2DM may be considered as a risk factor for developing moderate and severe fibrosis and also a predisposing risk factor for the development of hepatocellular carcinoma in the context of NAFLD.[32-36] A reciprocal relationship is conceivable between NAFLD and metabolic disorders such as T2DM.[32-34] Furthermore, NAFLD can lead to hepatocellular carcinoma even without any evidence of underlying cirrhosis.[37,38] The results show an integrated approach for prevention and control of these two conditions can lead to proper management of non-communicable diseases (NCD) at the community level. Regarding its high prevalence in population, NAFLD may have a high population attributable fraction (PAF) for associated conditions including T2DM, MetS, cardiovascular diseases (CVD), and malignancies. In the other words, the incidence of population diseases, including conditions mentioned above, may decrease if exposure to NAFLD is reduced. On the other hand, almost all conditions mentioned above have common underlying causes and risk factors. As a result, an integrating and comprehensive community-based intervention should be recommended to control the related modifiable risk factors, particularly in LMICs where the high incidence of NCDs, such as NAFLD, T2DM, CVD, and malignancies, is considered an emerging challenge of globalization in primary healthcare.

In our study, NAFLD was defined by radiological evidence investigated by a single operator, while liver biopsy is the gold standard for diagnosis of NAFLD. However, biopsy is not an optimal tool for diagnosis of NAFLD in the population-based studies due to its complications and its high false positive rate.[39,40] These limitations in addition to methodological ones, emanated from the descriptive design, affect the study psychometrics.

In conclusion, the prevalence of T2DM dramatically increases in the context of NAFLD. This condition may be considered as an independent predictor of T2DM. An integrated population-based intervention may reduce the burden of many NCDs, such as NAFLD and T2DM.