icon-folder.gif   Conference Reports for NATAP  
  9th Conference on Retroviruses and Opportunistic Infections
Seattle, Washington, February, 2002
Back grey_arrow_rt.gif
Insulin Resistance, Glucose Intolerance: What does it all mean?
Written for NATAP by Judith Aberg, Washington University, St Louis
  Will you get diabetes on HAART? This is similar to the question regarding the risk of hyperlipidemia and heart disease. No one knows for sure, but what we did see more posters this year addressing the risks and prevalence of glucose intolerance. The problem is that we do not know the clinical significance of elevated insulin levels. Will individuals with insulin resistance go on to develop Diabetes mellitus (DM). DM is becoming a major medical problem in the world. Long-term consequences of uncontrolled diabetes include neuropathy, nephropathy (kidney disease), retinopathy (leading to blindness) and heart disease. It is also associated with vascular disease and is one of the major reasons for amputations in the USA. Dr Currier reports below that HIV-infected patients may be at a higher risk for diabetes than HIV-negative patients. Dr Yoon from NYC reported that a number of traditional risk factors play significant roles in contributing to the development of diabetes in HIV-infected patients including the condition of the liver may play a role. Dr Mehta's report suggests factors other than protease inhibitors may contribute, and that nevirapine appeared to be less likely to lead to diabetes. Dr Falutz reported findings suggesting that the oral glucose tolerance test may be the most reliable screening test.
It has been estimated that 1-6% of patients with HIV who receive PI therapy may be at risk for diabetes. Metabolic disorders in individuals infected with HIV are a topic of current concern and discussion. Metabolic complications (hyperlipidemia, insulin resistance, and fat redistribution) have been associated with HIV infection and antiretroviral (ARV) treatment. Several studies have demonstrated a direct link between exposure to protease inhibitors (PIs) and the development of insulin resistance both in vitro (in the test tube and animals) and in vivo. However, there is limited knowledge of the epidemiology (the percent of HIV-infected patients with diabetes) and pathogenesis (the cause) of DM in this population. We are not yet certain as to whether PI therapy directly leads to diabetes, but research is looking at this. The results of studies reported below suggest uncertainty about the preliminary research suggesting a PI may lead to diabetes. The studies suggest there are many contributing factors. As people with HIV experience increased survival, an understanding of the incidence, risk factors (both treatment related and non-treatment related) and management of chronic diseases, such as DM, is essential.
Dr. Currier and colleagues (Abstract 677-T) examined Medi-Cal (California Medicaid) claims from July 1994 to June 2000 for DM incidence rates in men and women over age 18 with and without HIV diagnoses. A total of 7219 (61% male) individuals with HIV and 2,792,971 (30% male) individuals without HIV were studied. A total of 7,101,180 person-years were examined. To assess only incident cases of DM diagnoses, individuals had to be free of DM related diagnoses and oral anti-diabetic pharmacy claims for at least one year prior to study inclusion.
HIV-infected persons were 3.32 times more likely to have diabetes in this study. Overall incidence of DM per 100 person-years in HIV infected individuals is 10.68 (95% CI 10.15-11.23) and 2.91 (95% CI 2.89-2.92) in non-HIV infected individuals, odds ratio 3.32 (p<0.0001). Age-specific relative risk for DM in individuals with HIV compared to those without HIV range from 7.74 (95% CI: 5.03, 11.91, p<0.001) in individuals age 18-24 to 2.16 (95% CI: 1.44, 3.25, p< 0.001) in individuals age 65+. The rates for DM were higher in all age groups for HIV+ vs HIV-negative persons. But the greatest increased risk in HIV+ persons vs HIV-negative persons was in individuals 25-44 years of age, where the risk appeared 5-6 times greater. The risk for diabetes was equal for men and women. These data suggest an association between increased incidence of DM in HIV-infected individuals compared to non-infected individuals. Given the constraints of the database, the investigators were unable to examine HIV medication effects. But further analysis of this dataset will attempt to explore the relationship between exposure to HIV therapies and the risk of diabetes.
Risk Factors For Diabetes. Although glucose abnormalities have been associated with use of protease inhibitors (PIs), other factors may contribute to the development of DM in HIV-infected patients. To delineate risk factors for DM, Dr.Yoon and colleagues (Abstract 678-T) conducted a case-control study in an urban HIV clinic at NY Presbyterian Hospital-Cornell, which has provided care for 5400 HIV-infected patients from May 1991- December 2000. 2 controls (HIV infected non-diabetic) per case (HIV infected diabetic) were matched on age +/- 5 years, race, gender, and length of clinic follow-up. HIV risk factors, family history of DM, HBVinfection, HCV infection, antiretroviral use, body mass index (BMI), lipodystrophy (physician diagnosis), CD4 cell counts (nadir and at time of DM diagnosis), HIV viral load (PCR), mean serum alanine aminotransferase levels (ALT), and mean cholesterol were evaluated at or prior to onset of diabetes in cases or during the equivalent period of follow-up in controls. From May 1991 to December 2000, they identified 50 cases for DM. 1 case was excluded as no matched controls could be found, leaving 49 cases and 98 controls. Mean time from start of PI to DM was 22 months. Mean time from initial visit to DM diagnosis was 27 months, mean age was 45 years, 63% were male, 39% African American, 35% Hispanic, and 27% Caucasian. Compared with controls, cases had a higher mean BMI (30.0 vs 25.3 kg/m2; matched odds ratio [OR] 1.20; p < 0.001), higher mean ALT (66.3 vs 43.7 U/L; OR 1.01; p= 0.013), stronger family history of DM (50% vs 29.1%; OR 3.30; p= 0.009), and higher prevalence of fat accumulation (30% vs 12.8%; OR 3.40; p= 0.025). HCV coinfection was more common in cases than controls (51.1% vs 36.5%; OR 2.10; p= 0.066), and cases had greater prior/current PI use (71% vs 58.2%; OR 2.30; p= 0.072). In multivariate analyses, only BMI (OR 1.18 /kg/m2; p= 0.010), family history (OR 9.41; p= 0.034), and ALT (OR 1.02 /U/L; p= 0.047) were associated with DM. Although PI use and HCV co-infection may contribute to risk of DM, as shown by others, traditional risk factors (i.e. obesity, family history) account for much of the risk of DM in HIV-infected persons. The association of serum ALT level with DM may reflect liver injury or steatosis, which may predispose to hyperglycemia. These findings suggest that there are complex interrelationships between genetic factors, treatment-induced metabolic changes, and liver injury in the pathogenesis of DM in HIV-infected population.
Incidence of Diabetes Similar for All Specific Protease Inhibitors and Efavirenz, But Not For Nevirapine. Dr. Mehta and colleagues from Johns Hopkins University (Abstract 679-T. The Effect of HAART and HCV Infection on the Development of Diabetes Mellitus) reported similar results as Yoon and colleagues. They reported that DM is marginally more common in HCV/HIV co-infected patients, but HCV did not substantially increase the risk of incident DM in persons receiving HAART. In addition, although incident DM appears to be common during HAART, particularly among patients without traditional risk factors for diabetes (e.g., older age and obesity), the incidence rate of DM was similar for all PIs and efavirenz, suggesting that factors other than PI use may contribute to the pathogenesis of DM in HIV-infected persons. Nevirapine had a lower incidence of diabetes per 100 person-years. In a multivariate statistical analysis, older age, African-American race, baseline glucose levels and failure to gain weight during therapy were independent predictors of developing diabetes.
Mehta looked at patients receiving HAART from April 1995 to August 2001 at Johns Hopkins Hospital HIV Clinic. They looked at 513 patients, out of 1002 receiving HAART, who were on their first HAART regimen for at least 6 months and had at least 2 non-fasting glucose measurements. Of the 513 persons studied, 244 (48%) were HCV negative and 269 (52%) were HCV positive. Persons with HCV infection were older and more often African-American than those without HCV infection. Persons with HCV infection were more likely to report injection drug use (IDU) than those without HCV infection. There were no differences in prevalence of hepatitis B virus infection, baseline CD4 cell counts, HIV RNA viral load or nonfasting glucose levels.
Dr. Saves and colleagues (Abstract 682-T), in a study designed to look at lipodystrophy, noted a 10% incidence of DM at 36 months in a cohort of HIV-infected patients starting a PI-containing regimen. Traditional risk factors such as older age and a higher BMI were significantly associated with a higher risk of diabetes. HCV status nor any antiretroviral treatment were associated with the onset of diabetes
Various Screening Tests for Diabetes. Dr. Falutz and Gardiner (Abstract 676-T) addressed the issue of what is the best screening tool for insulin resistance (IR). Given we still do not the clinical meaning of elevated insulin levels, routine screening tests for IR remain controversial. Tests under consideration include the oral glucose tolerance test (OGTT), the homeostasis model assessment (HOMA), elevated fasting insulin (FI), and fasting glucose (FBG). The investigators compared the usefulness of several screening tests for IR to the standard OGTT. 32 HIV-positive males (n=29) and females (n=3) underwent a standard OGTT and FI determination. The HOMA was calculated (HOMA = [FI x FBG]/22.5). Subjects were characterized as having either normal glucose tolerance (NGT) or impaired glucose tolerance (IGT), consistent with IR. A HOMA of >4.0, an elevated FI, and a FBG 6.1-7.0 mmol/L, represents IR. Of the 32 patients evaluated, 22 (69%) had NGT on the basis of the OGTT, 9 (28%) had IGT and 1 had DM. Elevated FBG levels consistent with a diagnosis of impaired fasting glucose (IFG) were found in 4/32 (12.5%); elevated FI occurred in 7/32 (22%), and a calculated HOMA >4.0 occurred in 8/32 (25%). There was no difference in the BMI (Ht/m2) or age between the NGT group (22.95.9) and the IGT group (24.93.7). Of the 9 identified with IGT, only 6 were also identified on the basis of either an elevated FI and/or an elevated HOMA. Of 8 patients with a HOMA >4.0, 7/8 had elevated FI, 4/8 had IFG, and 6/8 had IGT. Elevated FBG alone did not identify most patients who had abnormal parameters on other screening tests. Based on these preliminary results, the use of either elevated FBG, elevated FI, or an elevated HOMA cannot currently be recommended for IR screening. The investigators concluded that the standard OGTT remains the most reliable screening test available in this population.
There remains a lot to be done in deciphering the clinical meaning of insulin resistance. How does insulin resistance fit into the equation of the metabolic syndrome or lipodystrophy? Dr. Wohl discusses some glucose disturbances in his review on lipodystrophy for NATAP. There were no presentations dealing with the treatment of insulin resistance outside the context of lipodystrophy and/or lipoatrophy. Frank diabetes needs to be treated as one does in diabetics without HIV infection. Unlike dyslipidemia, which rarely causes an immediate medical consequence, elevated or extremely low blood glucose levels may be associated with life-threatening conditions. Expect much more to be presented during next year's meeting.