SEARCH

SEARCH BY CITATION

Keywords:

  • diabetes mellitus;
  • glucose intolerance;
  • hepatitis B virus;
  • hepatitis C virus;
  • insulin resistance

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Background and Aim:  Chronic hepatitis C virus infection has been known to increase the risk of diabetes. Whether this association holds true for chronic hepatitis B virus (HBV) infection remains unclear. We thus conducted this study to investigate the influence of asymptomatic chronic HBV infection on the incidence of diabetes in a longitudinal cohort.

Methods:  A total of 1233 adults who received health examinations in 1997–1998 and in 2000–2001 were enrolled. Among them, 483 subjects who received a third health examination in 2006–2008 were further sampled. The prevalence and incidence of diabetes between asymptomatic HBV carriers and non-HBV controls were compared using the χ2-test and logistic regression.

Results:  In 1997–1998, the prevalence rates of diabetes (9.49 and 12.0%) and glucose intolerance (28.5 and 25.4%) in HBV carriers and non-HBV controls were comparable (P > 0.05). There was no significant correlation between asymptomatic HBV infection and the presence of diabetes in subjects examined in 1997–1998, 2000–2001, or 2006–2008 when adjusted for age, gender, and body mass index (P > 0.05). In 296 non-diabetic subjects during 1997–1998, the ten-year incidence of diabetes/glucose intolerance was similar between HBV carriers and non-HBV controls (40.0 and 38.7%, P > 0.05). Moreover, no significant correlation was found between asymptomatic HBV infection and the incidence of diabetes/glucose intolerance in 2006–2008 (P = 0.775).

Conclusions:  Compared to non-HBV controls, subjects with asymptomatic chronic HBV infection do not have an increased risk of diabetes, and thus HBV itself is not pro-diabetic.


Abbreviations
HBV

hepatitis B virus

HCV

hepatitis C virus

CHC

chronic hepatitis C

ALT

alanine aminotransferase

AST

aspartate aminotransferase

FPG

fasting plasma glucose

BMI

body mass index

PPG

postprandial plasma glucose

Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Ample evidence has documented the interactions between metabolic factors and hepatitis C virus (HCV) infection. For example, cross-sectional and longitudinal studies have shown that chronic HCV infection is associated with an increased risk of developing insulin resistance and type 2 diabetes.1–3 Whether this association holds true for chronic hepatitis B virus (HBV) infection remains unclear.

Although effective vaccines have been available for more than two decades, chronic hepatitis B virus (HBV) infection remains a major global health problem. The World Health Organization estimates that there are more than 350 million HBV carriers and these carriers are at increased risk of developing cirrhosis and liver cancer in their lifetimes.4 Moreover, recent studies have also suggested that chronic HBV infection is associated with the development of diabetes mellitus,5,6 and this positive association may be attributed to the presence of hepatic fibrosis or cirrhosis due to HBV infection.7,8 However, earlier reports indicated that the incidence of diabetes was higher in chronic hepatitis patients of various etiologies compared to the general population.9–11 Taking these lines of evidence together, whether HBV infection itself leads to the emergence of diabetes remains controversial and thus deserves further studies.

Taking advantage of rampant HBV infection in Taiwan, we conducted a longitudinal study to understand the impact of asymptomatic chronic HBV infection on the incidence of diabetes.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Study subjects

Study subjects were selected from 1413 apparently healthy adults of ≥ 20 years of age who received health examinations twice at the Health Management Center of National Taiwan University Hospital: in 1997–1998 and in 2000–2001 (Fig. 1). The main reason for choosing subjects who received two health examinations was to confirm the status of asymptomatic HBV carriage. This study was performed in accordance with the principles of the Declaration of Helsinki, and was approved by the Ethics Committee of the National Taiwan University Hospital.

image

Figure 1. Flow diagram of enrolling 1233 study subjects (143 asymptomatic chronic hepatitis B virus [HBV] carriers and 1090 non-HBV controls). HbA1c, plasma glycosylated hemoglobin; HBsAg, hepatitis B surface antigen; PPG, postprandial glucose.

Download figure to PowerPoint

Of 1413 adults, 180 were excluded: due to the positivity of hepatitis C virus antibody (anti-HCV) in any of the two blood tests in 51 subjects; due to different results for the hepatitis B surface antigen (HBsAg) in the two blood tests in 45 subjects (19 subjects with one positive and one negative result, 19 subjects with one weak positive and one negative result, 6 subjects with one positive and one weak positive result, and one subject with two weak positive results); due to lacking laboratory data of plasma 2-h postprandial glucose (2-h PPG, mg/dL) and/or plasma glycosylated hemoglobin (HbA1c, %) in 11 subjects; due to diagnosis of suspect or definite parenchymal liver disease or liver cirrhosis on abdominal ultrasound report in 51 subjects; and due to no report of abdominal ultrasound in 22 subjects. Finally, 1233 adults were enrolled in this study, including 143 asymptomatic HBV carriers and 1090 non-HBV controls.

Of the 1233 enrolled subjects, 493 received a third health examination in 2006–2008 and were further studied. Ten of 493 subjects were excluded due to the following reasons. Five asymptomatic HBV carriers became HBsAg-negative, another carrier had weakly positive HBsAg, one control had a positive anti-HCV antibody, one control lacked data of plasma 2-h PPG level, one control had a diagnosis of parenchymal liver disease on abdominal ultrasound report, and another control had a diagnosis of intra-hepatic malignancy on ultrasound report. Finally, 483 subjects (55 asymptomatic HBV carriers and 428 non-HBV controls) were included to evaluate a ten-year effect of asymptomatic HBV infection on the incidence of diabetes. The age (mean ± standard error [SE]) in 1997–1998, gender ratio (male/female), and case number ratio of HBV carriers/non-HBV controls of these 483 subjects were not statistically different from those of the total 1233 subjects enrolled in 1997–1998 (54.0 ± 0.44 vs 54.8 ± 0.27 years, 370/113 vs 903/330, and 55/428 vs 143/1090; P = 0.124, 0.170, and 0.969, respectively). The follow-up period of the 483 subjects was 9.86 ± 0.04 (mean ± SE) years, and that of the asymptomatic HBV carriers group (9.75 ± 0.14 years, n = 55) was similar to that of the non-HBV control group (9.87 ± 0.05 years, n = 428, P > 0.05).

Laboratory data

Laboratory results of health examinations were retrospectively collected from either databank or medical chart records. Age, gender, bodyweight (kg), body length (meters), fasting plasma glucose (FPG, mg/dL), 2-h PPG (mg/dL), plasma glycosylated hemoglobin (HbA1c, %), serum levels of aspartate aminotransferase (AST, U/L) and alanine aminotransferase (ALT, U/L), results of serum HBsAg, anti-hepatitis C virus antibody, and reports of ultrasound examination were collected. In our hospital, the normal upper limits of serum AST and ALT levels for men are ≤ 37 U/L and ≤ 41 U/L, respectively, and for women both AST and ALT normal levels are ≤ 31 U/L.

Definition of asymptomatic chronic HBV carrier

On the basis of laboratory data from two health examinations (one in 1997–1998 and one in 2000–2001), asymptomatic chronic HBV infection was defined as follows: (i) persistent positive results for HBsAg; (ii) persistent negative results for anti-HCV; and (iii) no evidence of parenchymal liver disease or cirrhosis on abdominal ultrasound report.

Definition of diabetes mellitus

Type 2 diabetes mellitus was diagnosed on the basis of both positive history and the laboratory results of plasma glucose measurements [10]. Without a positive history, diabetes mellitus was diagnosed with either an increased FPG ≥ 126 mg/dL or an increased 2-h PPG ≥ 200 mg/dL. In subjects without a positive history of diabetes, glucose intolerance was determined if either one of the following three conditions was met: (i) FPG between 100 and 125 mg/dL with a normal 2-h PPG level (< 140 mg/dL); (ii) 2-h PPG between 140 and 199 mg/dL with a normal FPG level (< 100 mg/dL); or (iii) FPG between 100 and 125 mg/dL and 2-h PPG between 140 and 199 mg/dL. In subjects without a positive history of diabetes, normal FPG and 2-h PPG levels with an elevated plasma HbA1c level ≥ 6.0 % were classified as undetermined cases.

Statistical analysis

The enrolled 1233 subjects were divided into the asymptomatic HBV carrier group (n = 143) and the non-HBV control group (n = 1090). Between these two groups, statistical analyses using the sas program were performed as follows: (i) comparisons of demographic and clinical data (Table 1); (ii) comparisons of the prevalence of diabetes in 1997–1998, 2000–2001, and 2006–2008 separately using the χ2-test (Table 2); (iii) logistic regression analysis of the presence of diabetes on the asymptomatic HBV carrier status with adjustments for age, gender, and body mass index (BMI, equal to bodyweight/body length2, kg/m2, Table 3); (iv) comparisons of the ten-year incidence of diabetes in 483 subjects who underwent a third health examination in 2006–2008 using the χ2-test (Table 4); and (v) logistic regression analysis of the incidence of diabetes on the asymptomatic HBV carrier status in these 483 subjects (Table 5). In estimating the ten-year incidence of diabetes, 296 non-diabetics in 1997–1998 were analyzed first. Then, another 125 subjects with glucose intolerance in 1997–1998 were added and analyzed once more (Table 4 and 5).There were 36 cases, three cases, and one case of undetermined diagnosis for diabetes in 1997–1998, 2000–2001, and 2006–2008, respectively. These undetermined cases of diabetes were excluded from the above statistical analyses.

Table 1.  Demographic and clinical data of study subjects in 1997–1998 (n = 1233)
 Study group
HBV carrier (n = 143)Control (n = 1090)P-value*
  • Values are mean ± standard error.

  • *

    Probability values were estimated using the Student's t-test except for the male/female ratio, which was estimated using Fisher's exact test.

  • ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; FPG, fasting plasma glucose; HBV, hepatitis B virus; PPG, postprandial glucose; HbA1c, plasma glycosylated hemoglobin.

Age (year)52.7 ± 0.7355.1 ± 0.290.004
Male/female ratio113/30790/3000.108
BMI (kg/m2)24.4 ± 0.2524.6 ± 0.090.450
Serum AST (U/L)25.9 ± 0.9621.3 ± 0.22< 0.0001
Serum ALT (U/L)34.7 ± 2.3224.3 ± 0.41< 0.0001
Plasma FPG (mg/dL)96.0 ± 1.4197.4 ± 0.650.453
Plasma 2-h PPG (mg/dL)121.5 ± 3.39128.5 ± 1.480.101
Plasma HbA1c (%)5.21 ± 0.075.38 ± 0.030.050
Table 2.  Comparison of the prevalence of diabetes between asymptomatic HBV carrier group and control group (n = 1233)
 Case number
DiabetesGINormalP-value*
  • *

    Probability values were estimated using the χ2-test with three levels of diagnosis for diabetes, e.g. diabetes, glucose intolerance, and normal.

  • Thirty-six cases, three cases, and one case of undetermined diagnosis for diabetes in 1997–1998, 2000–2001, and 2006–2008, respectively, were excluded from the analyses. See Methods section for the definition of an undetermined case of diabetes.

  • GI, glucose intolerance; HBV, hepatitis B virus.

Subjects in 1997–1998 (n = 1197)    
 HBV carrier group133985 
 Control group1272696640.577
Subjects in 2000–2001 (n = 1230)    
 HBV carrier group1626101 
 Control group1412317150.514
Subjects in 2006–2008 (n = 482)    
 HBV carrier group131625 
 Control group701611970.288
Table 3.  Logistic regression analyses for the presence of diabetes on the presence of asymptomatic HBV carrier status with adjustments for age, gender, and BMI
 Presence of diabetes*
P value*OR95% CI
  • *

    Logistical regression analyses were performed with an ordinal response for the diagnosis of diabetes. In the analysis, diabetes, glucose intolerance, and normal were represented digitally by 3, 2, and 1, respectively. Female and male were represented by 2 and 1, respectively.

  • Thirty-six cases, three cases, and one case of undetermined diagnosis for diabetes in 1997–1998, 2000–2001, and 2006–2008, respectively, were excluded from the analyses.

  • BMI, body mass index; CI, confidence interval; HBV, hepatitis B virus; OR, odds ratio.

Subjects in 1997–1998 (n = 1197)   
 Age (years)< 0.00011.0471.033–1.060
 Gender*0.0610.7670.580–1.013
 BMI (kg/m2)< 0.00011.1811.131–1.232
 HBV carrier or control group0.6501.0900.750–1.585
Subjects in 2000–2001 (n = 1230)   
 Age (years)< 0.00011.0481.034–1.061
 Gender*0.1871.2380.902–1.701
 BMI (kg/m2)< 0.00011.0921.057–1.128
 HBV carrier or control group0.5840.8980.612–1.319
Subjects in 2006–2008 (n = 482)   
 Age (years)< 0.00011.0541.035–1.073
 Gender*0.7670.9400.623–1.417
 BMI (kg/m2)0.0851.0570.992–1.127
 HBV carrier or control group0.4571.2290.714–2.116
Table 4.  Comparison of the ten-year incidence of diabetes between asymptomatic HBV carrier group and control group in 483 subjects who underwent a third health examination in 2006–2008
 Case number
DiabetesGINormalP-value*
  • *

    Probability values were estimated using the χ2-test with three levels of diagnosis for diabetes, e.g. diabetes, glucose intolerance, and normal.

  • The follow-up period of the 296 non-diabetics was 9.82 ± 0.05 years, and that of the HBV carrier group (9.56 ± 0.18 years, n = 30) was similar to that of the non-HBV control group (9.85 ± 0.06 years, n = 266, P > 0.05).

  • The analyzed 421 cases included the above 296 non-diabetics and another 125 subjects with glucose intolerance, but not definite diabetes, in 1997–1998.

  • GI, glucose intolerance; HBV, hepatitis B virus.

296 non-diabetic cases in 1997–1998    
 HBV carrier group3918 
 Control group9941630.208
421 cases without definite diabetes in 1997–1998    
 HBV carrier group61625 
 Control group261571910.282
Table 5.  Logistic regression analyses of the incidence of diabetes on the presence of asymptomatic HBV carrier status adjusted for age, gender, and BMI in subjects who underwent a third health examination in 2006–2008
 Incidence of diabetes*
P-value*OR95% CI
  • *

    Logistical regression analyses were performed with an ordinal response for the diagnosis of diabetes. In the analysis, the diabetes, the glucose intolerance, and the normal were represented digitally by 3, 2, and 1, respectively. Female and male were represented by 2 and 1, respectively.

  • The 421 analyzed cases included the above 296 non-diabetics in 1997–1998 and another 125 subjects with glucose intolerance, but not definite diabetes, in 1997–1998.

  • BMI, body mass index; CI, confidence interval; HBV, hepatitis B virus; OR, odds ratio.

296 non-diabetic cases in 1997–1998
 Age (years)< 0.00011.0581.032–1.085
 Gender*0.7291.1040.632–1.928
 BMI (kg/m2)0.5661.0270.938–1.123
 HBV carrier or control group0.7751.1230.508–2.481
421 cases without definite diabetes in 1997–1998
 Age (years)< 0.00011.0461.026–1.068
 Gender*0.8400.9550.609–1.496
 BMI (kg/m2)0.0941.0620.990–1.140
 HBV carrier or control group0.8381.0650.581–1.952

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

The demographic and clinical data of 1233 study subjects in 1997–1998 are shown in Table 1. The mean age (± SE) of the asymptomatic HBV carrier group was slightly but significantly younger than that of the control group (52.7 ± 0.73 and 55.1 ± 0.29 years, respectively, P < 0.005). Asymptomatic HBV carriers had significantly higher serum AST and ALT levels than controls (both P < 0.0001). No significant difference was found between the asymptomatic HBV carrier group and the control group in terms of gender ratio, BMI, plasma FPG level, 2-h PPG level, or HbA1c level.

Table 2 shows the prevalence rates of diabetes between asymptomatic HBV carriers and controls analyzed using the χ2-test. The rates of diabetes between these two groups were comparable in 1997–1998, 2000–2001, and 2006–2008, separately.

By using logistic regression analyses, the influence of asymptomatic HBV infection on the presence of diabetes after adjustment for age, gender, and BMI is shown in Table 3. The regression analyses were performed with an ordinal response for the diagnoses of diabetes, e.g. the diabetes, the glucose intolerance, and the normal were represented digitally by 3, 2, and 1, respectively. In 1997–1998, the presence of diabetes had a positive and significant association with age and BMI (both P < 0.0001, Table 3), but was not associated with asymptomatic HBV infection (P = 0.650). Same analyses for 2000–2001 and 2006–2008 separately also revealed no significant association between the presence of diabetes and the asymptomatic HBV infection (P = 0.584 and 0.457, respectively). When 220 subjects with an abnormally high serum AST level and/or ALT level measured in either 1997–1998 or 2000–2001 were excluded, repeated above regression analyses revealed similar results (all P > 0.500).

The results of the ten-year incidence of diabetes from 1997–1998 to 2006–2008 are shown in Table 4. The follow-up period of 296 non-diabetics in 1997–1998 was 9.82 ± 0.05 years, and that of the asymptomatic HBV group (9.56 ± 0.18 years, n = 30) was similar to that of the non-HBV control group (9.85 ± 0.06 years, n = 266, P > 0.05). The incidence rates of diabetes between these two groups were not statistically different using the χ2-test (P = 0.208). After adding another 125 subjects with glucose intolerance in 1997–1998, the results of analyses were similar (P = 0.282, n = 421).

By logistic regression adjusted with age, gender, and BMI, no significant relation was found between asymptomatic HBV infection and the incidence of diabetes in the ten-year observation of the 296 non-diabetics in 1997–1998 (P = 0.775, n = 296, Table 5). The results of the same analyses after adding another 125 subjects with glucose intolerance in 1997–1998 were similar (P = 0.838, n = 421, Table 5).

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

After a ten-year period of clinical observation, we found that, compared to non-HBV controls, asymptomatic HBV infection doesn't increase the risk of developing type 2 diabetes, suggesting chronic HBV infection is not pro-diabetic. These data provide indirect evidence for the current concept that the positive association between chronic HBV infection and increased incidence of diabetes may be attributable to HBV-related parenchymal liver diseases rather than the virus itself.7–11

Although chronic HCV infection has been linked to the emergence of insulin resistance and subsequent diabetes,1 little is known about the relationship between chronic HBV infection and the development of insulin resistance as well as diabetes.12,13 Several recent studies reported a higher incidence of diabetes in patients with chronic HBV infection.5,6,14 However, the study populations were heterogeneous with different stages of liver disease. Therefore, the data with special emphasis on asymptomatic HBV carriers are still lacking. A previous report from Hong Kong indicated that HBsAg-carrier mothers had a significantly higher incidence of gestational diabetes than control mothers;15 however, serum aminotransferase levels were not mentioned. Accordingly, using the stringent definition, the study subjects might not be asymptomatic HBV carriers without any liver damage.

The severity of liver disease in patients with chronic HBV infection has been shown to correlate with the incidence of diabetes. Papatheodoridis et al. indicated that the presence of diabetes was independently associated with a higher serum gamma-glutamyl-transpeptidase level, a more severe fibrosis, or the presence of cirrhosis.7 Similarly, Qureshi et al. showed that diabetes was more frequent in patients having liver cirrhosis than in those having chronic hepatitis.8 These findings further support the notion that the severity of liver disease is the key to developing diabetes, regardless of the etiologies.

The positive association between liver disease severity and diabetes could be reasoned as follows. First, the liver plays an important role in glucose homeostasis and thus may affect blood glucose levels.16–19 Thus, impaired glucose homeostasis and diabetic status are prone to occur in the scenario of moderate to severe liver disease. Second, persistent inflammatory activities in the liver may also cause defective glucose homeostasis.20–22

Of particular note in this study is that asymptomatic HBV carriers had significantly higher serum AST and ALT levels than non-HBV controls, implying that even in clinically asymptomatic infection, HBV may still produce adverse effects on liver cells. Although the differences in the absolute values of means of AST and ALT levels were modest, their longer-term effects on liver cells as well as metabolic profiles await further examinations.

A limitation of our study is that many of the 1233 subjects enrolled in 1997–1998 did not receive a third health examination in 2006–2008. This reduced the case number of a ten-year observation down to only 483 (55 asymptomatic chronic HBV carriers and 428 non-HBV controls) and thus might reduce the value of the paper. But because we had used very strict conditions to enroll both the asymptomatic chronic HBV carriers and the non-HBV controls, we think the case numbers of 55 and 428, respectively, should be able to provide reliable results. In addition, the age, gender ratio, and case number ratio of HBV carriers/non-HBV controls of these 483 subjects in 2006–2008 were not statistically different from those of original sample in 1997–1998 (see data in Study subjects of the Methods section). Another limitation of our study is the lack of data on hepatitis B e antigen status and HBV DNA titer, which were not measured during the health examination program. Studies with these two data would provide more substantial information and understanding about the correlation between chronic HBV infection and the risk of diabetes.

On the basis of a longitudinal study, our findings suggest that asymptomatic chronic HBV infection does not increase the risk of diabetes over time. In combination with our previous study, which found that chronic HBV infection is not associated with an increased insulin resistance,12 and earlier reports of an increased diabetic incidence in chronic hepatitis patients of various etiologies,9–11 these observations strongly suggest that the current concept regarding a positive association between chronic HBV infection and diabetes may be secondary to HBV-induced liver disease rather than HBV itself.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

This work was supported by grants from the National Taiwan University Hospital, the Liver Disease Prevention and Treatment Research Foundation, the Department of Health, and the National Science Council, Executive Yuan, Taiwan. Z. S. Huang planned and conducted this project, analyzed data, and wrote this paper. T. S. Huang helped in planning the project and interpreted study results. T. H. Wu helped in preparation and analysis of data. M. F. Chen helped in collecting and analysis of data. C. S. Hsu helped in analysis of data, and J. H. Kao planned this project, interpreted results and wrote the paper. None of the authors have any conflicts of interest or financial disclosures to declare.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References