Should all patients with nonalcoholic fatty liver disease undergo oral glucose tolerance test?

Abstract

Nonalcoholic Fatty Liver Disease (NAFLD) has become a global epidemic, affecting 20–40% of the general adult population. In some patients, the disease runs a progressive course, resulting in cirrhosis, hepatocellular carcinoma and liver-related mortality. Since NAFLD was first described, its association with metabolic syndrome and insulin resistance has been well recognized. Incident diabetes is also commonly diagnosed in NAFLD patients. With this background, the Asia-Pacific Working Party on NAFLD recommends anthropometric measurements and metabolic screening for all NAFLD patients. The question that remains is what tests clinicians should arrange. Should oral glucose tolerance test (OGTT) be part of the routine workup? The issue of OGTT is not new. Although fasting glucose is commonly used to screen for or diagnose diabetes, it is well known that the correlation between fasting and postprandial glucose is not perfect. In particular, isolated post-challenge hyperglycemia after OGTT is more often found in male and elderly subjects. Using fasting glucose alone, a significant proportion of patients with diabetes or impaired glucose tolerance would be missed. On the other hand, OGTT involves blood taking at two or more time points after oral glucose challenge. This may be inconvenient and costly to patients and clinicians. The intake of concentrated glucose drinks may also cause gastrointestinal upset and vomiting. Therefore, before one recommends OGTT as a routine test, a few questions need to be answered. How common do NAFLD patients have abnormal OGTT (i.e. number needed to screen)? Does it matter to have subclinical diabetes or impaired glucose tolerance not identified by fasting blood tests alone? Would the finding alter clinical management? In this issue of the Journal, two studies provided important information in this area. Kimura and colleagues performed 75-g OGTT in 173 Japanese biopsy-proven NAFLD patients without prior diagnosis of type 2 diabetes. Overall, 60% of the subjects had abnormal OGTT. Thirty-seven percent had impaired glucose tolerance and 23% had diabetes. Although patients with different degree of liver fibrosis had similar glucose levels, those with advanced fibrosis had significantly higher plasma insulin level throughout a 3-h period. After adjusting for age and aspartate aminotransferase level, plasma insulin level at 2 h remained significantly associated with advanced fibrosis. The similar glucose levels among patients with different fibrosis stages deviates somewhat from our usual understanding. In patients with NAFLD or viral hepatitis, cirrhosis or advanced fibrosis has consistently been associated with increased risk of diabetes. In fact, cirrhosis itself is a cause of insulin resistance. However, the lack of cirrhotic patients in Kimura’s cohort might partly explain the observation. Besides, the mean plasma glucose level among patients with different fibrosis stages was compared using Mann–Whitney U-test. It might be helpful to report the number of patients with impaired glucose tolerance and diabetes in each fibrosis subgroup. In the second article by Manchanayake and colleagues, OGTT was performed in 76 Australian NAFLD patients without prior diagnosis of diabetes. One-third of the subjects had abnormal OGTT, with 22% having impaired glucose tolerance and 9% having diabetes. Impaired fasting glucose only had 25% sensitivity in predicting abnormal OGTT. Similarly, based on plasma insulin levels at 0, 60 and 120 min, hyperinsulinemia and/or insulin resistance was almost universally found, including even patients with normal glucose tolerance. Although few patients in this cohort had liver biopsy, cirrhosis was found in 12% of patients with abnormal OGTT and 4% of those with normal glucose tolerance. These two studies concur with previous studies showing that one-third to one-half of NAFLD patients had undiagnosed diabetes and impaired glucose tolerance. Thus, the number needed to screen for OGTT to detect abnormal glucose regulation is only 2 to 3. Even in NAFLD patients with normal fasting glucose, the number needed to screen is less than 5. In cross-sectional and longitudinal studies, post-challenge hyperglycemia appears to be associated with adverse clinical outcomes. In a study of 1040 patients undergoing coronary angiogram, impaired glucose tolerance or diabetes by OGTT was associated with cardiovascular events (defined as a composite of vascular deaths, non-fatal myocardial infarctions, non-fatal strokes, percutaneous coronary interventions, bypass graftings, and revascularizations of non-coronary arteries) at a mean follow-up of 3.8 years. In another histological cohort, impaired glucose tolerance and diabetes were also associated with advanced liver fibrosis or cirrhosis. While it is unclear if post-challenge hyperglycemia itself contributes to the pathogenesis of these clinical complications, OGTT may help clinicians to identify high-risk patients for more intensive monitoring and treatment. At present, lifestyle changes and insulin sensitizers have been shown to be useful in the treatment of impaired glucose tolerance or pre-diabetes. In the landmark Diabetes Prevention Program Study, 3234 patients with elevated fasting or post-challenge plasma glucose were randomized to placebo, metformin, or a lifestylemodification program. While both treatments prevented the progression to diabetes, it was remarkable that lifestyle modification achieved even lower incidence of diabetes than metformin treatment. This highlights the importance of lifestyle intervention in patients with metabolic disorders. Recently, randomized controlled trials have confirmed that lifestyle modification is effective in reducing hepatic fat in NAFLD patients. The full extent of benefits will become clearer when larger clinical trials are completed. Do we have new markers to replace OGTT? Although clinicians commonly use glycosylated hemoglobin levels to monitor Accepted for publication 8 November 2010.