- Top of page
- Materials and methods
Non-alcoholic fatty liver disease (NAFLD) is emerging as the most common chronic liver disease in developed countries. It is associated with cardiovascular morbidity,1, 2 and may progress to liver cirrhosis and hepatocellular carcinoma.3, 4 Owing to the westernized dietary habits in Asian countries, there may also be an epidemic of metabolic syndrome and NAFLD in the near future. In a recent population survey of 15 540 Chinese adults, 9.8% of men and 17.8% of women had metabolic syndrome, while around 30% were overweight.5 A screening of 3175 Chinese showed that 15% of the population had NAFLD diagnosed by ultrasonography.6 Among Chinese patients with biopsy-proven NAFLD, necroinflammation and fibrosis were found in 86% and 26%, respectively.7 According to various prospective studies, a quarter to half of these patients have progression in liver fibrosis with time.8–10
Non-alcoholic fatty liver disease is associated with components of the metabolic syndrome, including diabetes mellitus, hypertension, dyslipidaemia and central obesity.7, 11–13 According to the 2006 American Diabetes Association (ADA) criteria, the diagnosis of diabetes is made if the fasting glucose exceeds 7.0 mm. The cut-off was lowered from 7.8 mm in 1997 because the new value correlated better with postchallenge glucose levels. Alternatively, using the 75 g oral glucose tolerance test (OGTT), a 2-h plasma glucose above 11.1 mm also indicates diabetes. However, ADA discouraged clinicians to use OGTT because of the greater cost and inconvenience.14
Nevertheless, there have been suggestions that fasting and postchallenge glucose criteria might diagnose two different groups of diabetic patients. According to 11 population-based studies in Asia, only 37% of the diabetic patients fulfilled both the fasting and 2-h plasma glucose criteria.15 In a subsequent review of 13 European studies, 2-h plasma glucose increased with age, while fasting glucose did not.16 Moreover, NAFLD patients have hepatic insulin resistance together with peripheral insulin resistance.17 Whether this will affect postload glucose level is unknown. With this background, fasting glucose alone may not be adequate in the assessment of NAFLD patients.
In this study, we performed OGTT for Chinese patients with biopsy-proven NAFLD and determined the prevalence of undiagnosed diabetes and postchallenge hyperglycaemia. We also aimed to assess factors associated with diabetes in these individuals and define the role of OGTT among them.
- Top of page
- Materials and methods
We set out to assess the prevalence of undiagnosed diabetes in Chinese NAFLD patients. The response to oral glucose challenge was abnormal in a significant proportion of this cohort. Thirty-three percentage of the patients had diabetes and another 29% had IGT. Notably, isolated postchallenge hyperglycaemia was common in this cohort. Twelve percentage of the patients had isolated IGT but normal fasting glucose, and 25% of the diabetic NAFLD patients had normal fasting glucose.
Abnormal glucose metabolism has been found in different NAFLD series. For example, among 19 Italian patients with biopsy-proven NAFLD, five had IGT and one had IFG.20 The higher prevalence of IGT than IFG echoed our findings. The overall prevalence of abnormal glucose metabolism is lower than that of our series probably because the Italian cohort contained patients with less severe disease (only three patients had grade 3 necroinflammation and no patient had stage 4 fibrosis). In 114 Turkish patients with elevated serum aminotransferase levels and bright liver on ultrasonography, 50 had IGT or diabetes according to the postload glucose level, suggesting the need to perform OGTT in NAFLD patients.21 However, the investigators combined IGT and diabetes in the analysis and did not take account of the fasting glucose level, making it difficult to assess the prevalence of isolated postchallenge hyperglycaemia. In another large population screening in Japan, the prevalence of ultrasonography-diagnosed NAFLD rose from 27% in people with normal fasting glucose, 43% in people with IFG, to 62% in people with newly diagnosed diabetes.22
Isolated postchallenge hyperglycaemia also exists but is much less common in the general population. Summarizing 13 European series, around 2–6% of people with fasting glucose below 7.0 mm had 2-h plasma glucose (11.1 mm).16 The corresponding percentage in the Asian series was 3%.15 As our series represent a group of patients with high risk of having metabolic syndrome, the prevalence of postchallenge hyperglycaemia was considerably higher. Our findings suggest strong insulin resistance among NAFLD patients. Using insulin clamp technique, it has been demonstrated that glucose disposal is reduced almost by half even among non-diabetic NAFLD patients.20, 23, 24 The defect is more severe among NASH patients than those with simple steatosis.25 On the other hand, hepatic glucose output is less suppressed after administration of insulin, indicating that hepatic sensitivity to insulin is reduced in NAFLD patients.23, 25
Although fasting glucose was an independent predictor of diabetes, using fasting glucose alone to screen NAFLD patients was limited by its low sensitivity. When normal fasting glucose was set at 5.6 mm according to the latest ADA criteria, the sensitivity in detecting diabetes was only 79%. Using 6.1 mm, the old definition of normal fasting glucose, the sensitivity for fasting glucose alone in detecting diabetes further decreases to 63%. In other words, OGTT can be considered a routine assessment for NAFLD patients in view of the high prevalence of postchallenge hyperglycaemia and poor sensitivity of fasting glucose alone. To the least, OGTT should be performed in NAFLD patients with IFG and/or low HDL cholesterol level, both of which being independent factors associated with diabetes in our cohort.
Does it matter to miss postchallenge hyperglycaemia? In our study, IGT and diabetes patients were more likely to have significant necroinflammation or liver fibrosis. While the majority of patients with normal glucose regulation had no or minimal liver fibrosis, stage 3 or 4 fibrosis was present in 5% and 17% of patients with IGT and diabetes, respectively. Similarly, 92% of patients with normal glucose regulation had no or minimal necroinflammation. By contrast, grade 2 necroinflammation was present in 33% and 25% of patients with IGT and diabetes, respectively. Besides having more severe NAFLD, patients with isolated postchallenge hyperglycaemia are also at risk of developing cardiovascular complications. Among 181 Swedish patients with acute myocardial infarction, previous undiagnosed IGT and diabetes occurred in 40% and 25%, respectively.26 Moreover, isolated postchallenge hyperglycaemia doubles the risk of cardiovascular events and mortality.27–29 In a prospective study on more than 2500 Japanese, IGT, but not IFG, increased the risk of dying from cardiovascular disease by 2.2-fold.30 These data suggest that isolated postchallenge hyperglycaemia not only is common in high-risk individuals, but also predicts poor outcome.
The association among dysglycaemia, cardiovascular disease and NASH probably reflects that the diseases are part of the metabolic syndrome and share common aetiologies. Diabetes is a chronic low-grade inflammatory state. For example, proinflammatory cytokines-like tumour necrosis factor-α, interleukin-6 and interleukin-18 are hypersecreted, whereas anti-inflammatory cytokines-like adiponectin are suppressed. In fact, some of these cytokines are also involved in the development of NASH. Among different populations, hypoadiponectinemia is associated with NAFLD and even NASH.31, 32 In addition, nuclear factor-κB activation and upregulation of tumour necrosis factor-α, interleukin-6 and intercellular adhesion molecule-1 are involved in animal models of steatohepatitis.33
Our study has several limitations. First, the number of patients is relatively small. However, our cohort has comprehensive metabolic profile and histological data. This is superior to radiological studies that are limited by their accuracy and inability to assess necroinflammation and fibrosis.34 Secondly, our patients were recruited from secondary and tertiary referral centres and our data may not reflect the true prevalence of IGT and diabetes among NAFLD patients in the community. Nevertheless, the metabolic profiles of NAFLD patients in our cohort are similar to those identified in the community by ultrasound screening.6, 35 As NAFLD is strongly associated with cardiovascular events, our data further support the need for comprehensive workup including OGTT in these patients.1, 2 Thirdly, our cohort contains Chinese patients only. These findings need to be confirmed in other ethnic groups. Fourthly, the control group in this study represents healthy individuals from the community. Although this provides data regarding the prevalence of postchallenge hyperglycaemia and diabetes in the general population in this age- and gender-group, our data cannot discern whether the high prevalence of dysglycaemia in the NAFLD patients was due to the liver disease or other metabolic factors. In particular, the control group was not matched for BMI with the NAFLD cases. Obesity per se is associated with a higher prevalence of dysglycaemia.
In conclusion, isolated postchallenge hyperglycaemia is common among Chinese NAFLD patients without history of diabetes. It is associated with histological severe disease, and cannot be accurately predicted by any fasting glucose cut-off.