Metabolic and Skeletal Complications of HIV Infection: Title and subTitle BreakThe Price of Success

A 57-year-old man with human immunodeficiency virus (HIV) infection complicated by type 2 diabetes mellitus, hyperlipidemia, hypertension, and osteonecrosis was seen in routine follow-up.

The patient was diagnosed as having HIV infection in 1983, at which time his CD4 cell count was 500/μL. He began antiretroviral therapy in 1989 with zidovudine monotherapy and received sequential monotherapy followed by dual therapy, primarily with zidovudine and didanosine, until 1996, when he began combination antiretroviral therapy with stavudine, lamivudine, nevirapine, and saquinavir; the latter 2 drugs were replaced by nelfinavir shortly thereafter (Figure 1). In 1997, he was diagnosed as having Pneumocystis pneumonia and improved following treatment with trimethoprim-sulfamethoxazole and prednisone.

He was first seen at the National Institutes of Health in 1999, at which time he was still receiving stavudine, lamivudine, and nelfinavir and was also taking trimethoprim-sulfamethoxazole for Pneumocystis prophylaxis. Previous antiretroviral regimens failed to suppress the HIV viral load to below detection limits; however, after a change to stavudine, didanosine, indinavir, and ritonavir in 2000, the HIV viral load dropped to less than 50 copies/mL. He interrupted treatment for 5 months in late 2001 because of fatigue and concern about antiretroviral toxicity. Therapy was restarted in March 2002 with his current regimen of didanosine, tenofovir, abacavir, lamivudine, and lopinavir/ritonavir.

His course has been complicated by the development of body composition changes. He first noted fat loss in the face and extremities in 1996, followed by increases in abdominal and dorsocervical fat beginning in 1997. These changes persisted following the change in his antiretroviral therapy, and he received poly-l-lactic acid injections for the cosmetic treatment of facial fat loss with good results.

Intermittent elevations in fasting blood glucose levels were noted in the late 1990s, and the patient was formally diagnosed as having type 2 diabetes mellitus in 2003. Despite efforts at diet control, blood glucose elevations persisted and the patient began treatment with rosiglitazone in 2005.

Mild elevations in triglycerides, total cholesterol, and low-density lipoprotein cholesterol were noted prior to the initiation of highly active antiretroviral therapy (HAART). Lipid abnormalities increased following initiation of HAART, and he began atorvastatin in 1999, which was changed to rosuvastatin in 2004. Gemfibrozil was added in 2004 for persistent hypertriglyceridemia. Despite these medications, mild lipid abnormalities persist.

As a participant in a screening study of avascular necrosis in asymptomatic HIV-infected adults, he was diagnosed as having bilateral osteonecrosis of the hips in 2001. Despite the use of a cane and minimized weight bearing, he required bilateral total hip replacements in 2003. In 2005, he fractured 2 ribs after falling from a curb. Chest and rib radiographs suggested osteoporosis.

The patient now returns to the clinic for routine follow-up of his immune and viral measurements, as well as for ongoing management of diabetes mellitus and dyslipidemia.

This patient presents a panorama of the HIV/AIDS epidemic in the United States from its beginnings to the present day. Early in the epidemic, his management focused on treatment and prevention of opportunistic infections, the major causes of morbidity and mortality associated with AIDS during that time. After the introduction of protease inhibitors in 1996, his viral load was eventually suppressed, his CD4 cell count increased, and the course of his HIV-related immunodeficiency was reversed. Then complications began to emerge that, in many ways, are the face of HIV infection in the United States today: body composition changes progressed and he developed worsening lipid abnormalities, glucose intolerance, and osteonecrosis. Management of his HIV infection no longer depended simply on control of the viral load but also required an understanding of these chronic complications, the medications needed to control them, and the potential interactions between these medications and antiretroviral drugs.

Combination antiretroviral therapy is now recommended as the standard of care for patients with progressive HIV infection¹ and is likely to be continued for such patients' entire lives, with the exception of short periods of drug interruption due to toxic effects or patient desire. In the decade since the introduction of HAART regimens, which include a combination of antiretroviral agents that usually contains 2 nucleoside reverse transcriptase inhibitors (NRTIs) plus a protease inhibitor or nonnucleoside reverse transcriptase inhibitor (NNRTI), a number of metabolic disorders and bone diseases have been described in association with HIV infection and antiretroviral therapy that were not identified during the initial clinical trials of these agents ( Article ). Often, a patient will develop more than 1 of these complications, as was the case of the patient presented herein. Recognition of the association of metabolic complications with increased cardiovascular disease risk has increased the urgency of efforts to understand and combat these disorders.

This discussion focuses on 5 major complications of HIV disease and treatment: fat redistribution (lipodystrophy), dyslipidemia, insulin resistance, cardiovascular disease, and bone disorders. The incidence of these complications depends on a variety of factors, including specific antiretroviral drugs and drug combinations, lifestyle choices, and genetic determinants, as well as undefined factors presumably in part related to HIV infection.

Shortly after the approval of the first protease inhibitor in late 1995, the initial enthusiasm about the effectiveness of these drugs was tempered by reports of unexpected metabolic complications apparently associated with their use. These reports described abnormalities in fat distribution (lipodystrophy), often in association with insulin resistance and dyslipidemia^{2 - 4} ; together, these have been termed lipodystrophy syndrome. For simplicity, each aspect is discussed separately herein, but it is important to remember that they often occur together in an individual patient and can all potentially contribute to increased cardiovascular disease risk, the most significant long-term consequence of lipodystrophy syndrome.

The fat redistribution of lipodystrophy is characterized by 2 primary patterns: accumulation of central fat and loss of subcutaneous fat, which can occur separately but are commonly seen together. Central fat accumulation is manifested by increased abdominal girth resulting from increased deposition of visceral fat (Figure 2); enlargement of the dorsocervical fat pad (“buffalo hump”), breast enlargement, and lipomatosis have also been described. Peripheral fat wasting, or lipoatrophy, is characterized by the selective loss of adipose tissue in the face, extremities, and buttocks (Figure 2).

The body composition changes often significantly affect patient quality of life, as they have been associated with social stigmatization, increased psychological stress, and reduced self-esteem,⁵ and can lead to reluctance to initiate antiretroviral therapy as well as to antiretroviral medication noncompliance.⁶

The clinical diagnosis of lipodystrophy was initially based on patient-reported fat loss in the face and extremities, with or without accompanying central fat gain, with confirmation by physical examination.⁷ Such a definition is difficult to codify since it is subject to observer (patient or physician) interpretation. Case definitions derived from models that combine clinical and laboratory measurements as well as measures of body composition by dual-energy x-ray absorptiometry and computed tomography scanning have been validated in a large clinical trial⁸ ; however, such definitions have not been broadly adapted.

The prevalence of lipodystrophy among HIV-infected persons receiving antiretroviral therapy has been estimated to range from 20% to 84%.^{7 ,9 - 10} Prospective cohort studies have suggested that 20% to 50% of patients will develop at least 1 sign of lipodystrophy within 2 years of initiating HAART.^{9 - 10}

Distinct mechanisms appear to be responsible for peripheral fat loss and central fat accumulation. Cross-sectional studies that have included HIV-negative controls have consistently demonstrated lipoatrophy in HIV-infected patients, especially those receiving antiretroviral therapy, but have only inconsistently demonstrated fat accumulation.^{11 - 13} Longitudinal studies in patients initiating antiretroviral therapy have demonstrated both lipoatrophy and fat accumulation.^{14 - 16} Thus, lipoatrophy appears to be directly related to antiretroviral therapy; however, it is currently unclear how often fat accumulation represents physiological improvement in the setting of suppressed viral replication vs a pathological response to drug therapy.

The risk of developing lipoatrophy has been linked repeatedly to the use of NRTIs, especially stavudine.^{17 - 18} Newer NRTI agents, abacavir and tenofovir, have not been associated with lipoatrophy. Most NRTI-specific adverse effects are thought to be manifestations of mitochondrial toxicity, resulting from inhibition of mitochondria-specific DNA polymerase γ, ultimately leading to impaired production of adenosine triphosphate. Mitochondrial depletion and dysfunction have been demonstrated in adipose tissue from HIV-infected adults with lipodystrophy.^{19 - 20}

Protease inhibitor use appears to accelerate the rate of development of NRTI-associated lipoatrophy.²¹ This may not be a class effect, as it was not seen in a 48-week trial of atazanavir.²² In vitro, protease inhibitors have been shown to impair adipose cell differentiation and insulin sensitivity. The postulated mechanism is interference with the transcription factor sterol regulatory element-binding protein–1 and decreased activation of downstream pathways.²³

The causes of visceral fat accumulation are poorly understood. While the pattern of fat distribution has some similarities to Cushing disease, no abnormalities in the hypothalamic-pituitary-adrenal axis have been identified.^{4 ,24}

Given the substantial impact on self-image resulting from body composition changes, a number of strategies attempting to reverse these changes have been evaluated. One logical approach has been to stop the offending antiretroviral agents or to switch to regimens that are not associated with such changes, but such approaches have been generally unsuccessful.^{25 - 26} The only such intervention demonstrated to be effective in randomized trials is switching patients with lipoatrophy from a thymidine-based nucleoside analogue (eg, stavudine) to abacavir,^{25 - 26} but improvement is only moderate. As with any antiretroviral regimen change, the risk that the new regimen will fail to adequately control HIV replication must be weighed against potential benefits.

Though many other potential pharmacologic treatments are under investigation for these body composition changes, none has shown consistent efficacy in reversing these changes, and many of the interventions are limited by toxic effects. These treatments include insulin-sensitizing agents; hormone therapies such as growth hormone, growth hormone–releasing hormone, and testosterone; statins; and the dietary supplement uridine.

Thiazolidinediones, which are used to treat type 2 diabetes mellitus, can reverse lipoatrophy in non–HIV-infected patients. Two randomized trials in patients with lipoatrophy found no difference in fat gain between HIV-infected patients receiving rosiglitazone and those receiving placebo.^{27 - 28} However, another randomized trial of patients with lipoatrophy and hyperinsulinemia demonstrated small but statistically significant gains in leg fat in patients receiving rosiglitazone.²⁹ Of note, rosiglitazone has been shown to increase total cholesterol, low-density lipoprotein cholesterol, and triglyceride levels.^{27 - 29} Pioglitazone, another thiazolidinedione with a distinct lipid profile, similarly led to small but statistically significant gains in limb fat in a study of patients with self-reported lipoatrophy.³⁰ Thus, these agents may help treat lipoatrophy in some patients, but additional studies are needed.

In 1 randomized trial of patients with hyperinsulinemia or impaired glucose tolerance as well as increased waist-hip ratio and clinical evidence of fat redistribution, metformin improved insulin sensitivity and reduced both subcutaneous and visceral abdominal fat.³¹ Subsequent studies have failed to show benefit in patients without evidence of insulin resistance or glucose intolerance. Metformin has been associated with reduced limb fat in lipoatrophic patients and should be used with caution in patients with lipoatrophy.

Growth hormone, a lipolytic drug, has been shown to reduce HIV-associated fat accumulation but is associated with significant adverse effects, including impaired glucose metabolism, fluid retention, and joint pain.³² In preliminary studies, growth hormone–releasing factor appears to have similar beneficial effects on fat accumulation but without the same adverse effect profile.³³

As a result of the uncertain efficacy and potential adverse effects of many pharmacologic therapies, many patients and clinicians have looked to nonpharmacologic approaches, including diet, exercise, and, for treatment of facial fat loss, soft-tissue augmentation. A variety of fillers and implants have been utilized for the latter, including autologous fat transfer and injection of silicone, collagen, hyaluronic acid, poly-L-lactic acid, and polyacrylamide. These have been shown to improve body image satisfaction, as occurred in the current patient, and depression indices, but are limited by expense and, for many of the agents, the short duration of correction.^{34 - 35} Poly-L-lactic acid (Sculptra, Dermik Laboratories, Berwyn, Pa [previously marketed in Europe as New-Fill, Biotech Industry SA, Luxembourg]) is the only injectable filler approved by the US Food and Drug Administration for use in correcting facial lipoatrophy in HIV-infected adults, although the cost can be prohibitive.

As the case presentation illustrates, lipid disturbances are very frequent in patients receiving combination antiretroviral therapy. After starting protease inhibitor therapy, the patient's triglyceride and cholesterol elevations worsened; both normalized rapidly during treatment interruption, but triglyceride levels increased immediately after resuming HAART (Figure 1). Abnormalities in lipid metabolism were first described in HIV-infected patients prior to the development of HAART.³⁶ Following the introduction of protease inhibitors, the prevalence of dyslipidemia increased, with lipid disturbances affecting as many as 60% of patients receiving these agents.³⁷ Among protease inhibitors, however, impact on the lipid profile differs, with newer agents such as atazanavir and fosamprenavir having a more favorable lipid profile compared with older agents.^{38 - 40} Certain protease inhibitors have been shown to have a direct effect on lipid metabolism in the absence of HIV infection or other antiretroviral agents,⁴¹ in part by increasing hepatic triglyceride synthesis and reducing the clearance of triglyceride-rich very low-density lipoprotein particles.^{42 - 43}

Though lipid perturbations appear to be most significant with the protease inhibitor class of medications, in a large, cross-sectional study, lipid disturbances were found in HIV-infected patients regardless of treatment regimen (Figure 3).^{44 - 45} Among NRTIs, tenofovir has been shown to result in a more favorable lipid profile compared with stavudine.⁴⁶ In the NNRTI class, nevirapine is associated with gains in high-density lipoprotein cholesterol and significantly less increase in total and low-density lipoprotein cholesterol and triglycerides compared with efavirenz.⁴⁷

Lipid profiles may be improved with discontinuation of certain antiretroviral agents. Replacement of a protease inhibitor with nevirapine, efavirenz, or abacavir reduces triglycerides, total cholesterol, and low-density lipoprotein cholesterol.^{48 - 50} However, in a randomized trial comparing substitution of nevirapine or efavirenz for a protease inhibitor vs statin or fibrate therapy, lipid-lowering therapy was more effective.⁵¹ Limited data exist regarding the effects of treatment discontinuation on dyslipidemia.⁵² In light of recent studies suggesting an increased risk of infectious, cardiovascular, and other complications in patients with treatment interruption,⁵³ interruption strategies for the management of hyperlipidemia should be used only after a careful weighing of risks and benefits.

Management of hyperlipidemia in HIV-infected adults should follow the National Cholesterol Education Program guidelines established for the non–HIV-infected population.⁵⁴ Lipid-lowering therapy should be considered for isolated hypertriglyceridemia and elevated low-density lipoprotein and total cholesterol levels that do not respond to lifestyle modification or, if used, antiretroviral regimen change. Although the most common lipid abnormality among HIV patients is hypertriglyceridemia, for which fibrate therapy is indicated, the class of antilipid agents most commonly used in HIV-infected populations is the statins.⁵⁵

Statins, with the exception of pravastatin, rosuvastatin, and fluvastatin, are metabolized by the cytochrome p450 (CYP) 3A4 isoenzyme. Because protease inhibitors inhibit CYP3A4, the use of statins metabolized through the p450 system in combination with protease inhibitors should be done cautiously. Based on potency, pharmacokinetics, and safety, pravastatin or rosuvastatin are used preferentially in HIV-infected patients.^{56 - 57} Atorvastatin has a lower affinity for CYP3A4 and has also been used successfully in patients receiving HAART.⁵⁸

A fibric acid derivative, either gemfibrozil or fenofibrate, is recommended for the treatment of primary hypertriglyceridemia.⁵⁷ Niacin may be used as an alternative agent for the treatment of mixed dyslipidemia. Niacin is associated with insulin resistance and hepatotoxicity but appears to be well tolerated in HIV patients receiving antiretroviral therapy.⁵⁹ Antilipid agents may be combined, as was done in the current case, but patients should be closely monitored for liver and muscle toxicity. In non–HIV-infected patients, fenofibrate is associated with less myopathy when combined with a statin than is gemfibrozil.⁶⁰

Insulin resistance is characterized by decreased sensitivity to insulin in insulin-sensitive tissue, including muscle, liver, adipose tissue, and endothelium, and predicts the development of type 2 diabetes mellitus.⁶¹ Insulin resistance and impaired glucose tolerance have been observed since the introduction of protease inhibitors.^{3 ,62} In a cohort of HIV-infected patients receiving combination antiretroviral therapy,⁶³ 35% had impaired glucose tolerance, and in a cross-sectional study of patients taking protease inhibitors,⁷ 7% had diabetes mellitus while impaired glucose tolerance was found in an additional 16%. Insulin resistance has been demonstrated in patients with fat redistribution, even in patients not receiving protease inhibitors.⁶³

In HIV-uninfected populations, insulin resistance has been associated with an increased risk of cardiovascular disease.^{64 - 65} Although an association between insulin resistance in HIV-infected persons and increased cardiovascular risk has not been definitively established, insulin resistance in HIV patients with lipodystrophy, especially those with visceral fat accumulation, has been associated with multiple markers of increased cardiac risk.^{63 ,66}

The pathophysiology of insulin resistance in HIV-infected adults receiving antiretroviral therapy remains unknown. Studies of protease inhibitors in HIV-negative adults have shown that indinavir, ritonavir, and lopinavir can induce insulin resistance over a short treatment course, suggesting that it is not exclusively a secondary effect of changes in fat distribution.^{67 - 68} Insulin resistance is drug-specific and not a class effect: in similar studies of amprenavir^{69 - 70} and atazanavir,⁶⁸ insulin resistance was not induced. One proposed mechanism is protease inhibitor interference with glucose transporter-4–mediated glucose transport.⁷¹

There is no universally accepted measure of insulin resistance. The majority of research studies use the euglycemic, hyperinsulinemic clamp for measurement of insulin sensitivity; however, this approach is not practical in clinical settings or in large clinical trials. In HIV-infected adults, the combination of an oral glucose tolerance test with measurement of insulin levels may offer the best compromise between simplicity and accuracy, though additional study is needed.⁷² However, caution should be taken to ensure that a standardized insulin assay is used.

Because protease inhibitor therapy may increase insulin resistance, some experts have suggested that assessment of glucose tolerance be integrated into the management of patients starting these agents. In patients with known disorders of glucose metabolism, avoidance of protease inhibitor medications most associated with worsening insulin resistance is recommended when possible.⁵⁷ In patients already receiving protease inhibitor therapy, insulin resistance has been shown to improve with the substitution of abacavir or an NNRTI for the protease inhibitor^{73 - 74} ; however, 3 NRTIs alone are currently considered inadequate anti-HIV therapy.

Management should follow the standards established for the non–HIV-infected population.⁷⁵ Lifestyle modification (weight loss, diet modification and increased exercise) should be recommended for HIV patients with evidence of impaired glucose tolerance, given their benefits in other populations.⁷⁶ In patients with diabetes mellitus unresponsive to lifestyle modification, treatment with insulin-sensitizing agents such as metformin or thiazolidinediones should be initiated, as was done in the current case. Metformin has been shown to improve insulin sensitivity and improve markers of increased cardiovascular disease risk in HIV-infected adults with insulin resistance.^{18 ,31} Metformin use should be reserved for patients with normal hepatic and renal function. Thiazolidinediones (rosiglitazone and pioglitazone) also improve insulin sensitivity in HIV-infected adults and appear to be safe, though limited longitudinal data exist.^{27 - 29} Insulin therapy may be required in patients who are unresponsive to oral agents.

Although early case reports suggested a possible association of HIV infection with premature coronary artery disease, recent retrospective and prospective studies of cardiovascular disease risk in HIV patients, especially as related to antiretroviral therapy exposure, have produced conflicting results.^{77 - 80} The most convincing data have been collected by the DAD (Data Collection on Adverse Events of HIV Drugs) study, which is the largest cohort trial to date to prospectively investigate the association of HIV infection and antiretroviral therapy with risk of cardiovascular disease.⁸¹

Through February 2005, the DAD study included a total of 94 469 patient-years of follow-up; 345 study participants had a cardiac event, for an overall incidence rate of 3.65 per 1000 patient-years of treatment.⁸² After adjustment for known cardiac risk factors, including age, sex, tobacco use, and elevated blood pressure, risk increased 16% per year of antiretroviral exposure (95% confidence interval, 9%-23%).^{81 - 82} In analysis of antiretroviral classes, protease inhibitors were found to increase risk by 16% per year of exposure, while NNRTIs were not significantly associated with increased risk. The effect of protease inhibitors was partially but not completely explained by effects on lipids. Use of NRTIs did not appear to increase cardiovascular risk. The absolute risk of a cardiac event in the cohort was low, approximately 0.4% per year of follow-up.

The mechanisms underlying atherosclerosis in HIV disease remain poorly elucidated. It has been postulated that HIV itself promotes atherosclerosis through proinflammatory effects on endothelial cells. Antiretroviral effects on lipids and insulin sensitivity may contribute to premature atherosclerosis. Certain protease inhibitors have been shown to impair endothelial function (a precursor of atherosclerosis) in HIV-negative volunteers, independent of alterations in lipid profile or blood pressure.^{83 - 84}

In the absence of data specific to HIV-infected patients, guidelines that were developed for risk reduction in the general population should be followed. All HIV patients should be evaluated for cardiac disease risk, including assessment of traditional risk factors, such as tobacco use, hypertension, dyslipidemia, and family history, as well as emerging risk factors, such as weight and antiretroviral medication use. Lifestyle modification, especially smoking cessation, should be initiated, along with treatment of hypertension, hyperlipidemia, and diabetes mellitus. When possible, antiretroviral therapy should be selected that includes agents with the least association with cardiovascular risk.

The patient presented here has multiple cardiovascular disease risk factors, including diabetes mellitus, hypertension, and hyperlipidemia, as well as possible additional risk related to protease inhibitor use and underlying HIV infection. In line with the approach outlined above, his medical management has been optimized and includes daily aspirin, an angiotensin-converting enzyme inhibitor, and antilipid agents.

A high prevalence of osteopenia and osteoporosis has been described in HIV-infected patients, including those who have not received antiretroviral drugs. In treatment-naive participants, one study reported a prevalence of osteopenia of 23% to 28%, a significant increase over rates seen in HIV-negative controls.⁴⁶ Another study found osteoporosis rates of 50% in patients receiving protease inhibitors compared with 23% in HIV patients not receiving protease inhibitors and 29% in matched controls.⁸⁵ It is noteworthy, however, that an increased rate of fragility fractures has not been documented to date.

Factors associated with accelerated bone loss in HIV-infected cohorts include low body weight, low body mass index, smoking history, and duration of HIV infection.^{86 - 87} Traditional risk factors, including increasing age, heavy alcohol consumption, and steroid exposure, also confer additional risk.⁸⁸ Though there are no formal guidelines for the management of osteoporosis in HIV-infected adults, patients at high risk of osteoporosis, based on multiple traditional risk factors or a history of fracture, may be evaluated using dual-energy x-ray absorptiometry scanning. Patients should be counseled to maintain an adequate intake of calcium and vitamin D and to incorporate weight-bearing exercise into their fitness routines. Alendronate, combined with calcium and vitamin D supplementation, has been shown to be safe and effective for increasing bone mineral density in HIV-infected patients.⁸⁹

Osteonecrosis, also know as avascular necrosis, was first described in HIV-infected patients in 1990; subsequently, numerous case reports and retrospective case series have been published. Our group has demonstrated an unexpectedly high prevalence (4.4%) of magnetic resonance imaging–documented osteonecrosis of the hip in a cohort of 339 asymptomatic patients,⁹⁰ as well as an incidence rate of symptomatic disease approximately 100 times higher than that of the general population.^{91 - 92} Bilateral hip involvement and involvement of other bones is common (Figure 4). Symptomatic patients demonstrated a higher rate of progression to total hip replacement than asymptomatic patients.

Most reports of osteonecrosis in HIV-infected patients have identified the presence of 1 or more recognized risk factors, especially the use of corticosteroid drugs, although in a portion of these patients, no risk factor could be identified.^{93 - 94} Based on available data, neither HAART nor HIV itself acts as an independent risk factor for osteonecrosis.

Currently, there are no generally accepted guidelines for the prevention or early detection of osteonecrosis, and therapy following diagnosis is aimed at providing symptomatic relief. Recent studies suggesting a beneficial effect of bisphosphonate therapy in osteonecrosis need to be validated in other populations, including those with HIV infection.⁹⁵

As highlighted by the patient presented here, therapy for metabolic complications is based primarily on approaches and algorithms developed in non–HIV-infected populations, with modifications to minimize drug interactions. This patient will need ongoing, presumably lifelong monitoring to maintain lipids and glucose at acceptable levels and to minimize his risk of cardiovascular disease while ensuring that his HIV status remains stable. The transformation of HIV infection from an ultimately fatal disease to a chronic disease has been a remarkable achievement of medicine but has come at a cost. The challenge now is to develop treatment strategies and drug regimens that minimize the risks of these complications and optimize their management.

A PHYSICIAN: You have reviewed a large number of complications resulting from HAART therapy—the dark side of HAART. Can you comment on new developments in understanding the mechanistic causes of these complications and is there any hope that treatments can be developed to target these mechanisms and prevent these complications?

DR MORSE: The mechanisms underlying antiretroviral-associated metabolic complications are an active area of research. While progress has been made, the causes are still incompletely understood. The good news is that drug developers are aware of these adverse effects and are now working to develop antiretroviral agents that minimize these complications. For example, the newer protease inhibitor atazanavir appears to impact lipids and markers of insulin sensitivity much less than older protease inhibitors.

A PHYSICIAN: How effective are insulin-sensitizing agents in HIV patients who continue to receive protease inhibitors? Would you expect these agents to be as effective when compared with use in patients who do not have an extrinsic cause of insulin resistance?

DR MORSE: You raise an interesting point: in the setting of drug-induced metabolic abnormalities, will therapies designed for patients in other clinical settings be effective? No one knows for sure if the approaches used for non-HIV patients with hyperlipidemia, insulin resistance, or cardiovascular disease will be as effective for HIV patients receiving antiretroviral therapy. Studies have shown similar short-term benefits, but the long-term efficacy is unknown. In my patients, insulin sensitizers, combined with lifestyle modification, have been effective at the usual doses.

A PHYSICIAN: Your findings of high rates of osteopenia and osteonecrosis of the hips in your patients are very interesting. In light of the recent descriptions of osteonecrosis of the jaw in patients receiving bisphosphonates, can you comment on the use of bisphosphonates in your patients? Additionally, have you noted any association between osteonecrosis and the use of bisphosphonates?

DR MORSE: We did not specifically collect information about bisphosphonate use in osteonecrosis patients. The majority of our patients had exposure to corticosteroids as their major risk factor, though no patient received long-term steroids and, if I recall correctly, none received bisphosphonate therapy. We are not aware of any data suggesting that bisphosphonate use is associated with osteonecrosis of the jaw in HIV-infected patients; studies of bisphosphonates for osteoporosis in this population have suggested that it is safe.