Insulin Resistance and Diabetes

Introduction

Metabolic complications associated with HIV disease and its treatment -- including insulin resistance and diabetes, abnormal cholesterol and triglyceride levels (dyslipidemia), and body fat gain or loss -- remain a medical mystery and a topic of intense interest for AIDS researchers and people with HIV alike. While these complications sometimes have been collectively referred to as "lipodystrophy syndrome," it remains unclear whether or how they are related and what causes them (see "HAART Attack: Metabolic Disorders during Long-Term Antiretroviral Therapy," BETA, April 1999).

Scientists are urgently trying to better understand these conditions, which may have a negative impact on quality of life, interfere with adherence to antiretroviral therapy, and lead to long-term health problems. High blood glucose levels (hyperglycemia) and dyslipidemia are a particular concern because in the population at large they have been linked with increased risk of heart disease (see "Cardiovascular Disease in People With HIV," BETA, Summer/Autumn 2002). Much research is underway and new clues are steadily emerging, but Daniel Kuritzkes, M.D., of Boston's Brigham and Women's Hospital predicts, "We'll need several more years of follow-up to get a better perspective."

Blood Glucose Abnormalities: The Basics

The body requires sugar, or glucose, to provide energy for all its functions. A hormone called insulin allows glucose to enter individual cells. Normally, after eating, beta cells in the pancreas (an abdominal organ) produce more insulin to process the incoming sugar. Glucose and insulin levels in the blood also regulate the breakdown of glycogen (a stored form of energy) and the production of new glucose (gluconeogenesis) by the liver, as well as the release of fatty acids from fat cells (adipocytes). When normal glucose metabolism goes awry, several disorders may develop.

Insulin resistance is a condition in which the body's cells do not respond properly to the hormone and cannot take up glucose, which then builds up in the bloodstream. This causes the beta cells to release extra insulin, leading to high blood insulin levels (hyperinsulinemia). Over time, the beta cells can fail to secrete enough insulin. When the body cannot produce sufficient insulin, or the cells do not respond to it efficiently, the result is hyperglycemia -- impaired fasting glucose (IFG) and impaired glucose tolerance (IGT). Eventually, this process can lead to diabetes mellitus (sugar diabetes), a condition characterized by persistent hyperglycemia (see "Progression of Insulin Resistance to Type 2 Diabetes" below).

There are two primary forms of diabetes mellitus: type 1 and type 2. (Diabetes insipidus is an uncommon condition characterized by excess urine production unrelated to blood sugar abnormalities. Pregnant women may also develop a transient condition known as gestational diabetes. This article is limited to diabetes mellitus).

Type 1 diabetes (also called juvenile onset or insulin-dependent diabetes mellitus [IDDM]) typically occurs at a young age and is believed to result from the destruction of insulin-producing beta cells by the immune system. People with type 1 diabetes produce little or no insulin and usually must receive daily insulin injections.

Type 2 diabetes (also called adult onset, insulin-resistant, or non-insulin-dependent diabetes mellitus [NIDDM]) typically develops later in life -- though it is now being seen in children -- and commonly occurs in people who are overweight. (See "Risk Factors for Blood Glucose Abnormalities" below for more diabetes risk factors.) Type 2 diabetes is a progressive illness that involves a gradual decline in insulin sensitivity and production. It can take years or decades for mild insulin resistance to progress to full-blown diabetes, and many people with impaired insulin sensitivity never develop frank (clinically apparent) diabetes. Those with type 2 diabetes can often be treated with diet modification, increased exercise, weight loss, and/or oral medications, and usually do not require insulin injections. The blood glucose abnormalities that develop in people with HIV resemble type 2, not type 1, diabetes.

Insulin resistance and diabetes are a concern because untreated high blood sugar can lead to a wide range of long-term health problems, including kidney dysfunction, retina damage leading to blindness, nerve damage, erectile dysfunction in men, and pregnancy complications in women. In fact, diabetes is the sixth leading cause of death in the U.S. Yet these complications can occur even in people who never progress from impaired glucose tolerance to frank diabetes.

Hyperglycemia can also contribute to blood vessel abnormalities and cardiovascular disease, including heart attacks and strokes. This process is not well understood -- especially in people with HIV -- but it is thought that excess sugar in the blood may promote blood clotting and make cholesterol more likely to adhere to blood vessel walls. Both the Caerphilly Heart Study, which followed more than 2,500 men in a Welsh town between 1979 and 1983, and the Prospective Cardiovascular Munster (PROCAM) study, which followed 2,754 men, found that diabetes was associated with about a 2.5-fold increased risk of heart disease.

Insulin Resistance and HIV

Before the availability of highly active antiretroviral therapy (HAART), blood glucose abnormalities were infrequently seen in people with HIV. But in June 1997, soon after protease inhibitors (PIs) came into widespread clinical use, the U.S. Food and Drug Administration (FDA) issued a health advisory warning of an association between PIs and hyperglycemia and diabetes mellitus. Since then, there have been continued reports of insulin resistance in people using anti-HIV therapy.

Different studies have yielded widely varying estimates of the prevalence of impaired glucose metabolism in people on HAART, in part because they have used different tests and inconsistent definitions of the condition. The prevalence of frank diabetes mellitus in people with HIV is relatively low, with studies reporting rates from 0.5% to 15%. But, says Michael Dubé, M.D., of Indiana University School of Medicine, diabetes is "only the tip of the iceberg." Impaired glucose tolerance is considerably more common, affecting an estimated 15-25%, and research suggests that some degree of insulin resistance may occur in one-half of people taking PIs.

Research also indicates that coinfection with the hepatitis C virus (HCV) -- which affects as many as 40% of people with HIV in the U.S. -- increases the risk of blood glucose abnormalities. Studies have shown that people with chronic hepatitis C are more likely to develop insulin resistance and type 2 diabetes. For example, Shruti Mehta, M.P.H., and colleagues from Johns Hopkins University in Baltimore found that people with HCV were four times more likely to develop type 2 diabetes than HCV negative people; however, they found no association between hepatitis B and diabetes. Mehta's team also found that HIV/HCV-coinfected people were five times more likely to develop hyperglycemia than those with HIV alone. Similarly, Michel Duong, M.D., and colleagues from Dijon studied 29 HIV/HCV-coinfected individuals receiving HAART, 76 people with HIV alone, and 121 with HCV alone. Both the coinfected subjects and those with HCV alone had a significantly higher rate of insulin resistance than those with only HIV.

Although it is not clear how chronic hepatitis promotes blood sugar abnormalities, it is believed that liver damage affects the metabolism of glycogen and the production of glucose. In a letter published in the December 15, 2003 issue of the Journal of Acquired Immune Deficiency Syndromes (JAIDS), Raymond Chung, M.D., and colleagues from Harvard Medical School and Massachusetts General Hospital in Boston reported that elevated alanine transaminase (ALT) liver enzyme levels -- an indication of liver inflammation -- predicted insulin resistance in HIV-positive individuals with lipodystrophy whether or not they were coinfected with hepatitis B or C.

It is not yet known whether blood glucose abnormalities in people with HIV will have the same negative health consequences as they do in the population at large, though there is little reason to expect otherwise.

According to Dr. Dubé, the high prevalence of insulin resistance in people with HIV taking HAART "raises concern about the eventual development of increased cardiovascular morbidity in this population." And, say Oluwatoyin Falusi, M.D., and Judith Aberg, M.D., of the Adult AIDS Clinical Trials Group (AACTG) Cardiovascular Disease Focus Group, "Even if [HIV-positive] patients are not at increased risk for cardiovascular disease, they are at least at the same risk as HIV negative, age-matched persons with similar risk factors."

Indeed, the prevalence of blood glucose abnormalities is substantial in the general population. An estimated 6-9% of Americans have diabetes, although rates are considerably higher among older people (20% of those over age 65) and among African Americans (13%), Latinos (10%), Native Americans (15%), and Asian Americans and Pacific Islanders. As many as one-third of Americans have some degree of insulin resistance. A recent report by the National Center for Health Statistics stated that the rate of diabetes had increased 27% between 1997 and 2002, which many attribute to the rising incidence of obesity.

Some experts have suggested that the increasing incidence of blood glucose abnormalities in people with HIV is due to the fact that, thanks to effective treatment, such individuals are now living long enough to experience the normal problems associated with aging. (Interestingly, HIV-positive children on HAART rarely develop insulin resistance, although they do develop elevated blood fat levels.) But many others believe that HAART -- especially PIs -- or HIV itself share much of the blame.

Causes of Blood Glucose Disorders

Blood sugar abnormalities are due to too much glucose (e.g., reduced uptake of glucose by cells, increased production of glucose by the liver), too little insulin (decreased insulin release by beta cells), or some combination of both. A number of different theories have been put forth to explain the increased occurrence of insulin resistance and diabetes in people with HIV. While the bulk of research implicates PIs, other explanations cannot be discounted, and it is likely that multiple factors are at play simultaneously.

Protease Inhibitors

As noted previously, insulin resistance and diabetes were not common in people with HIV before the advent of HAART, and many studies have found an association between blood glucose abnormalities and PI therapy. In fact, PI use appears to be more directly related to disorders of glucose metabolism than to other metabolic complications such as body fat gain or loss.

Kathleen Mulligan, M.D., of the University of California at San Francisco (UCSF) and colleagues analyzed data from 20 HIV-positive individuals who started treatment with a PI, nine who received only nucleoside reverse transcriptase inhibitors (NRTIs), and 12 who received no antiretroviral therapy. Those who started a PI-based regimen had elevated fasting insulin and blood glucose levels, which suggest increased insulin resistance, as well as increased triglyceride and LDL ("bad") cholesterol levels. Signs of insulin resistance were apparent after an average of 3.4 months. However, no body shape changes were seen during this period. The group treated with only NRTIs did not experience similar glucose and lipid abnormalities.

Georg Behrens, M.D., of Hannover Medical School in Germany and colleagues reported that 46% of 38 PI recipients in their study had impaired glucose tolerance and 13% had diabetes, compared with 24% and none, respectively, among PI-naive subjects. Ravi Walli, M.D., and colleagues from Ludwig-Maximilians-Universität in Munich reported that 61% of 67 PI-treated subjects had reduced insulin sensitivity, which was seen in none of the 13 HIV-positive, treatment-naive controls. Frank Goebel, M.D., and colleagues, also from Munich, detected some degree of insulin resistance in 55% of people receiving PIs, compared with 27% of those receiving NRTIs, but Thierry Saint-Marc, M.D., and associates from Hôpital Edouard Herriot in Lyon found that only PIs -- not NRTIs -- were associated with glucose abnormalities.

However, some physicians believe these insulin resistance rates are too high. "We simply do not see insulin resistance in 50% of patients taking PIs," says George Beatty, M.D., M.P.H., of UCSF.

In the large Women's Interagency HIV Study (WIHS), women receiving PIs were significantly more likely to report that they had diabetes than HIV-negative women (2.8 cases per 100 person-years vs. 1.4 cases per 100 person-years, respectively). Interestingly, in this study HIV-positive women who received no antiretroviral therapy or received only NRTIs were even less likely to report diabetes than HIV-negative women (1.2 cases per 100 person-years).

Among the PIs, indinavir (Crixivan) has been most strongly associated with impaired glucose metabolism. Dr. Dubé and colleagues detected signs of insulin resistance in people with HIV within eight weeks of starting indinavir. Another study showed reduced insulin sensitivity as soon as two weeks after starting the drug. Mustafa Noor, M.D., from the Veterans Affairs Medical Center in San Francisco and colleagues found that insulin resistance (but not elevated lipid levels or visceral fat accumulation) developed within four weeks after starting indinavir in HIV-negative volunteers. In one of their studies, glucose disposal (uptake of glucose by cells) was reduced after a single dose of the drug. Because Dr. Noor's study subjects were neither HIV positive nor taking other classes of antiretroviral drugs, these results offer evidence that indinavir itself directly triggers insulin resistance.

Some other PIs appear less likely than indinavir to cause blood glucose abnormalities. For example, Dr. Walli and colleagues found that indinavir led to greater insulin resistance than either saquinavir (Fortovase, Invirase) or amprenavir (Agenerase). At the 3rd International Workshop on Adverse Drug Reactions and Lipodystrophy in HIV in October 2001, Jacqueline Capeau, M.D., and associates from INSERM in Paris reported that in laboratory studies, indinavir had the greatest inhibitory effect on a regulatory protein that helps control insulin resistance (discussed below), followed by nelfinavir (Viracept) and to a lesser extent amprenavir. Dr. Dubé's team saw a trend toward decreased insulin sensitivity in people treated with amprenavir, but only at the end of the 48-week study period, after increases in triglyceride and LDL cholesterol levels and abdominal fat accumulation had already occurred. Most trial data suggest that the newer PI atazanavir (Reyataz) has a minimal effect on glucose and lipid metabolism, but the drug requires further long-term study.

It is unclear exactly how PIs affect glucose metabolism, but research points to a variety of possible mechanisms, including reduced glucose uptake by peripheral cells, decreased insulin production by beta cells in the pancreas, and increased glucose production by the liver.

Several laboratory, animal, and clinical studies suggest that PIs may directly interfere with the transport of glucose into cells. An insulin-sensitive protein called Glut-4 plays a key role in transporting glucose into fat and muscle cells after eating. In laboratory studies using 3T3-L1 adipocytes (a type of fat cell), Haruhiko Murata, Ph.D., and colleagues from Washington University in St. Louis found that indinavir and other PIs reduced glucose uptake by inhibiting Glut-4 activity. At 100 micrometers (µm), indinavir reduced glucose uptake by 63%, while a 10 µm dose (closer to the concentrations used in humans) caused a 26% decrease. This inhibition occurred within minutes, and was reversed when indinavir was removed. In another study in frog egg cells, indinavir, amprenavir, and ritonavir (Norvir) reduced glucose uptake by 45%, 42%, and 54%, respectively. Noting that mutant mice lacking Glut-4 have almost no subcutaneous (under the skin) fat, the authors suggested that peripheral lipoatrophy (fat loss in the limbs and face) in people with HIV may be mediated by PIs' effect on the transport protein.

A related protein, Glut-2, allows beta cells in the pancreas to take up glucose to monitor blood sugar levels and regulate insulin release. Joseph Koster, Ph.D., and colleagues, also from Washington University, found that indinavir in doses similar to those used in humans -- and other PIs at higher concentrations -- inhibit the activity of Glut-2, thus reducing glucose uptake by beta cells. If the beta cells cannot detect elevated glucose levels, the authors suggest, they will not produce extra insulin to compensate, leading to hyperglycemia. These laboratory findings may help explain data from Dr. Dubé's team showing that reduced insulin sensitivity and increased fasting glucose did not trigger beta cells to release more insulin in people taking indinavir.

Proposing yet another mechanism, Dr. Capeau and colleagues found that in laboratory tests, PIs (indinavir, nelfinavir, and amprenavir) inhibit the production and activity of sterol regulatory element binding protein (SREBP), a key fat cell messenger that triggers stem cells to differentiate into adipocytes. SREBP also stimulates increased production of peroxisome proliferating activation factor gamma (PPAR-gamma), which promotes cellular glucose uptake in the presence of insulin.

Marc van der Valk, M.D., from the University of Amsterdam and colleagues reported that in addition to decreased cell sensitivity to insulin, glucose production by the liver is increased in people taking PIs. Using the hyperinsulinemic euglycemic clamp technique (discussed below in the "Diagnosis and Monitoring" section), the researchers found that hepatic glucose production was 47% higher in the PI recipients than in HIV-negative control subjects. In addition, insulin suppressed glucose production less in the PI group than in controls. Similarly, Dr. Noor's team found that hepatic glucose production (both gluconeogenesis and glycogenolysis, or breakdown of stored sugar) increased within four weeks of starting indinavir.

In an article in the December 2000 issue of Clinical Infectious Diseases, Dr. Dubé summarized the evidence for a direct effect of PIs in inducing blood glucose abnormalities. This includes the rapid development of glucose abnormalities soon after starting PI therapy, the reversal of glucose abnormalities when PIs are halted, the onset of insulin resistance before changes in body fat distribution occur, and plausible biological mechanisms. People with other risk factors for diabetes (for example, family history or obesity) may be especially susceptible to the effect of PIs on blood glucose.

Blood Glucose and Lipodystrophy

Much remains to be learned about the relationship between blood glucose abnormalities and other metabolic manifestations in people with HIV. As Dr. Dubé notes, while blood fat abnormalities, abdominal obesity, and loss of peripheral fat frequently coexist with insulin resistance, "It is not clear whether all of these result from a common pathogenic mechanism."

In the HIV-negative population, glucose abnormalities, elevated triglyceride levels, decreased HDL ("good") cholesterol, high blood pressure, and visceral abdominal obesity often occur together -- a syndrome variously known as insulin resistance syndrome, metabolic syndrome, or syndrome X. Here, too, it is unclear how these conditions are linked, but they occur together often enough to suggest they are interrelated.

Research indicates that altered glucose metabolism, dyslipidemia, and fat gain or loss are linked in people with HIV as well, and some researchers suggest that glucose abnormalities may in fact be attributable to body fat changes. Studies have shown, for example, that the accumulation of visceral abdominal fat can promote insulin resistance. Reduced insulin sensitivity may also result when fat cells are broken down -- a process called lipolysis -- as occurs during peripheral lipoatrophy. Further, research in mice suggests that loss of subcutaneous fat is associated with fat accumulation in insulin-sensitive tissues such as the liver and skeletal muscle, again contributing to insulin resistance.

Andrew Carr, M.D., and colleagues from St. Vincent's Hospital in Sydney found that among people taking PIs, insulin resistance was more common in those with body shape changes -- either abdominal obesity or peripheral fat loss. Dr. Carr's group also reported that people with "buffalo hump" (accumulation of fat at the back of the neck) were at higher risk for insulin resistance and diabetes, although other studies have yielded conflicting results.

Similarly, Colleen Hadigan, M.D., from Massachusetts General Hospital and colleagues found that among 101 HIV-positive people in the Framingham Offspring Study (a large study of cardiovascular risk), those with body fat changes were more likely to have impaired glucose tolerance (evidenced by elevated insulin levels and increased glucose levels after a glucose tolerance test) and frank diabetes. In another study, Dr. Hadigan's team found that insulin levels were most elevated in HIV-positive women with abdominal fat accumulation, independent of PI use. The same research group also reported insulin resistance in men with AIDS-related wasting syndrome who were treated with NRTIs but not PIs, and noted that reduced lean body mass and increased abdominal fat were the primary predictors of hyperinsulinemia. In addition, they found that when 52 HIV-positive hypogonadal (low testosterone level) men with AIDS-related wasting were given supplemental testosterone therapy, their insulin sensitivity improved as their lean body mass increased. (It should be noted, however, that supplemental testosterone may provide no additional benefit in men who already have normal levels.)

Dr. Corinne Vigouroux of INSERM and colleagues found that among study participants receiving PIs, 11 out of 14 (79%) with severe facial wasting had either insulin resistance or diabetes, compared with just four out of 20 (20%) without facial fat loss. In this study, elevated triglycerides were also more common in the group with facial wasting (79% vs. 35%). Dennis Mynarcik, M.D., from the State University of New York at Stony Brook and colleagues reported that among 12 HIV-negative study subjects, 15 HIV-positive participants with lipodystrophy, and 14 HIV-positive individuals without lipodystrophy, insulin resistance was greatest in those with fat loss. Likewise, Ove Andersen, M.D., and colleagues from Hvidovre University in Copenhagen reported that loss of limb fat was the strongest predictor of insulin resistance and decreased insulin production, independent of the type of antiretroviral therapy used.

Free Fatty Acids

Some research suggests that the relationship between body fat changes and glucose abnormalities may be mediated by free fatty acids. Normally, fatty acids are released when blood sugar levels are low to provide the liver with "raw material" for gluconeogenesis. High blood levels of free fatty acids -- related to both visceral fat accumulation and peripheral fat loss -- may interfere with normal glucose regulation and are associated with greater insulin resistance.

Dr. Hadigan's group found that HIV-positive individuals receiving antiretroviral therapy (58% on PIs, 74% on NRTIs) experienced heightened fasting lipolysis that increased further after consuming glucose, an indication of reduced insulin sensitivity (normally, lipolysis decreases after glucose consumption). Those with the greatest rates of lipolysis had the most severe insulin resistance. When an agent called acipimox was used to lower free fatty acid levels, insulin sensitivity improved, but it did not return to normal, suggesting that other factors were also involved.

If the hypothesis that body fat changes promote blood glucose abnormalities holds true, PIs may play an indirect role in glucose disorders by altering lipid metabolism and causing body fat abnormalities, perhaps in addition to their more direct effect. NRTIs (especially d4T [stavudine, Zerit]) also may indirectly contribute to glucose abnormalities by causing peripheral fat loss. As noted above, Dr. Goebel and colleagues detected evidence of insulin resistance in 27% of HIV-positive people treated with NRTIs (although the rate in those receiving PIs was twice as high). Dr. Mulligan, too, found that insulin sensitivity was reduced by 10% in people taking antiretroviral regimens that excluded PIs. Further, Dr. Andersen's team reported that glucose disposal was reduced in people with elevated lactic acid levels, a possible indication of NRTI-induced damage to the mitochondria, energy-producing organelles in cells that are involved in glucose metabolism.

Interestingly, research has not implicated non-nucleoside reverse transcriptase inhibitors (NNRTIs) in blood glucose abnormalities, although they have been linked with other metabolic manifestations in some studies.

Yet other data indicate that blood glucose abnormalities are not directly caused by body fat changes or dyslipidemia. Dr. Mulligan and colleagues, for example, found that blood glucose abnormalities developed just a few months after people began taking PIs, well before body shape changes occurred. Dr. Saint-Marc's team reported that when individuals with peripheral fat loss switched from d4T (which is strongly associated with lipoatrophy) to either abacavir (Ziagen) or AZT (zidovudine, Retrovir), they experienced increased subcutaneous fat but no improvement in insulin resistance.

Dr. van der Valk and colleagues demonstrated that when PIs were stopped for 96 weeks in eight HIV-positive men with lipodystrophy, glucose production decreased and lipolysis was reduced, although no body fat changes were seen. In a study of diet and exercise in obese HIV-positive women, Ellen Engelson, Ed.D., of Columbia University in New York City and associates found that after completing a 12-week weight loss program, the women experienced reducti