A 39-Year-Old Man With HIV-Associated Lipodystrophy

DR LIBMAN: Mr B was diagnosed as having human immunodeficiency virus (HIV) infection in 1994. His baseline CD4 cell count was 80/μL, and he has had no opportunistic diseases. He developed severe anemia while taking zidovudine and peripheral neuropathy while taking didanosine and stavudine. Other drugs previously received include lamivudine, indinavir, and ritonavir. In 2003, he self-discontinued all of his medications because of the aforementioned toxic effects and the development of fat loss in his cheeks, fat gain in his neck, chest, and abdomen, and hyperlipidemia. His CD4 cell count at that time was 249/μL, and his HIV RNA viral load was greater than 100 000 copies/mL.

Based on the prior drug toxic effects that Mr B experienced, including lipodystrophy, and his HIV genotype test results (which showed multiple nucleoside reverse transcriptase inhibitor [NRTI] mutations), he started a combination of tenofovir, lamivudine, efavirenz, and ritonavir-boosted atazanavir. Soon thereafter, he developed a diffuse skin rash from efavirenz and subsequently from nevirapine, and antiretroviral therapy was interrupted. In February 2004, Mr B started a combination of tenofovir, lamivudine, lopinavir/ritonavir, and fosamprenavir, and he has continued this regimen to the present. While receiving treatment, his CD4 cell counts have ranged from 367/μL to 631/μL and his viral loads have ranged from 111 to 826 copies/mL, with no particular trends noted. Mr B states that he has missed only “a few” medication doses per month. Over time, he has become increasingly concerned about his worsening body morphology changes. His weight ranged between 179 and 189 lb (80.5-85.1 kg) over the past 4 years.

Mr B's medical history is otherwise noteworthy for a methicillin-resistant Staphylococcus aureus finger abscess, oral herpes simplex infection, molluscum contagiosum, common warts, allergic rhinitis, and acute pyelonephritis. Mr B also has a remote history of a positive skin test result for tuberculosis, for which he received isoniazid prophylaxis, and hepatitis A and B infections. His family medical history is noteworthy for a father and a paternal grandfather with hyperlipidemia and coronary artery disease.

Mr B is currently taking fosamprenavir, 700 mg, 2 tablets twice daily; lamivudine, 300 mg once daily; lopinavir/ritonavir, 200 mg/50 mg, 3 tablets twice daily; tenofovir, 300 mg once daily; and fexofenadine, 60 mg twice daily as needed. In addition to the adverse drug reactions listed above, he has a history of hives due to penicillin.

Mr B is employed as a software engineer and is active in local nonprofit advocacy organizations for HIV-infected persons. He smokes a quarter-pack of cigarettes per day and has 3 to 4 alcoholic drinks per week. He has no history of drug use. Mr B's diet is mainly vegetarian.

On physical examination, Mr B's blood pressure was 118/84 mm Hg, his pulse was 72/min and regular, and his weight was 197 lb (88.7 kg), with a height of 70 in (178 cm). His waistline measured 33 in (84 cm). He had bilateral facial wasting and increased fat deposition on the back of his neck and in his chest and abdominal regions. The remainder of his examination results were normal.

Laboratory studies showed a normal complete blood cell count, and results of renal and hepatic function tests were normal. His fasting glucose level was 85 mg/dL (4.7 mmol/L), and his fasting lipid profile showed a total cholesterol level of 291 mg/dL (7.53 mmol/L), with a directly measured low-density lipoprotein (LDL) component of 77 mg/dL (1.99 mmol/L) and a high-density lipoprotein (HDL) component of 31 mg/dL (0.80 mmol/L), and a triglyceride level of 1130 mg/dL (12.7 mmol/L). His CD4 cell count was 530/μL and his HIV RNA viral load was 826 copies/mL.

I was diagnosed with HIV in October of 1994. At that point, I was told I was not only HIV-positive but I also had AIDS. My T-cell count was 80. Prior to that, I had been traveling over the summer and I started getting sick. I suspected that I did have HIV when I got sick, but I still couldn't tell my parents. So when the doctor told me that not only was I positive, but that I had AIDS, I was in shock.

When I was diagnosed, the doctor told me that I may have 12 to 18 months to live, so I sought medical care right away. The first medication I was on was [zidovudine]. Perhaps a month or 2 months after I started the [zidovudine], I developed severe anemia. So they stopped that, and I went through blood transfusion. Subsequently, I was a lot better. After that, I was put on some other medications, which caused neuropathy in my feet. When I started [efavirenz], my entire body broke out in hives. It took several attempts for me to get to a regimen that suited me.

My doctors told me that a new drug, [indinavir], had been released. But I had to apply for that through a mail lottery system because it was not yet available on the market. I was accepted into the trial, and that is when my history of lipodystrophy started. It was a gradual change, and by the second year of being on [indinavir], I began to see the effects. I noticed that under my chin, I started getting fat retention, and my cheeks were becoming a bit sunken.

My doctor told me that my cholesterol and triglyceride levels were very high. I have a history of heart disease in my family, so I have that in my genetic makeup, but I never thought it would be this high. Because my diet is primarily vegetarian, I pretty much established that it was related to the drugs.

After several years of being on the same regimen, in my mind I felt like I had found the silver bullet. Although I had high triglycerides and body fat changes, it seemed to be working. My T cells were good. All of my counts were good, so I wanted to stick with this as long as I could. But it came to a point where I realized I couldn't go this route anymore, because of the way I’m looking, because of the lipodystrophy. It's very difficult being a part of society and noticing changes in your body. People say, “Oh, you’ve gained weight, but it's only in the face.” I became afraid of meeting people that I hadn't seen in a long time.

I’ve done all the research; I’ve gone to a plastic surgeon. I’m in the process of getting insurance to cover this. They still think it's cosmetic, and I beg to differ. It is something that is absolutely caused by the medications that I’ve been taking.

I think that for someone from an Asian background, there is far more stigma associated with this disease. It's a very difficult process. You have let your family down, you're sort of ostracized in the community, your elders are looking at you differently. And it's interesting how the HIV and sexuality almost go hand in hand. As a gay Asian male, I’m told by my parents, “Well, once HIV is cured, you’ll get married.”

It's been a tough road. I think that in many ways one has to accept oneself. And it takes a very long time to be able to accept who you are.

What are the long-term treatment complications associated with HIV disease? What are the clinical characteristics, epidemiology, and pathophysiology of lipodystrophy syndrome? What is known about their effect on self-image and medication adherence? How should lipoatrophy and lipohypertrophy be managed? How should hyperlipidemia and glucose intolerance be managed in this context? Is switching antiretroviral drugs helpful for any of these manifestations? Are diet and exercise helpful for any of these manifestations? What would you recommend for Mr B?

DR FULLER: Mr B has been living with a diagnosis of AIDS (based on his first CD4 cell count of <200/μL) for nearly 15 years, has been treated with the 3 main classes of antiretroviral therapy, and has experienced complications from his treatment that have included transfusion-requiring anemia, peripheral neuropathy, hypertriglyceridemia, and lipodystrophy, which includes simultaneous lipoatrophy of the face and lipohypertrophy of the chin (“horse collar”). Although there was a time when he thought that he could tolerate any complications as long as his “numbers were good,” the body changes he has developed have now become so bothersome to him that he is actively seeking remediation of his facial lipohypertrophy through liposuction.

Numerous complications have been associated with long-term antiretroviral therapy, some of which have been experienced by Mr B. In some cases, clear links have been established between particular adverse effects and specific drugs or classes of drugs, while in others these relationships are not so clearly defined. As discussed herein, one form of HIV lipodystrophy, HIV lipoatrophy, has been associated with the thymidine analogs stavudine and zidovudine, while HIV lipohypertrophy has not been attributed to individual drugs.

Lipodystrophy is defined as “abnormal metabolism of fat”¹ and, most recently, as “abnormal depositions or wastings of adipose tissue, or combinations of these changes, seen in consequence of some protease inhibitor (PI) use in AIDS patients.”¹ Quiz Ref IDHuman immunodeficiency virus–associated lipodystrophy refers to fat accumulation (HIV lipohypertrophy) and to fat wasting (HIV lipoatrophy). Human immunodeficiency virus lipohypertrophy can manifest as an accentuated dorsocervical fat pad or “buffalo hump” (especially seen in men² ), increased chin fat or “horse collar,” increased breast tissue (especially seen in women² ), and/or central adiposity. Because HIV-associated central adiposity often involves increased visceral adipose tissue rather than increased abdominal subcutaneous adipose tissue,³ these affected patients have increased abdominal girth with thin skin folds (Figure 1A), compared with HIV-seronegative (or HIV-seropositive) obese patients, who may have the same girth but with much thicker skin folds (Figure 1B). Quiz Ref IDIncreased visceral adipose tissue can have important consequences because it is associated with an increased risk of diabetes mellitus and coronary artery disease,⁴ contributes to atherogenesis by increasing levels of plasminogen activator inhibitor 1,^{5 - 6} and impairs endothelial function.⁷

Human immunodeficiency virus lipoatrophy can manifest as loss of subcutaneous adipose tissue from the face, limbs, buttocks and/or abdomen, with visual prominence of extremity blood vessels known as pseudovenomegaly⁸ (Figure 2). Small series and case reports of HIV-associated lipodystrophy document supraclavicular fat pads,^{9 - 10} spinal epidural lipomatosis,¹¹ intramuscular fat accumulation,¹² and angiolipomas (tender or painful subcutaneous nodules composed of mature adipocytes and small blood vessels).¹³

Lipodystrophy syndrome has been associated with as much as a 2- to 5-fold increased risk of hypertension,^{14 - 15} reduced bone density (in 1 study, lumbar spine density was significantly reduced in the 21 patients with HIV infection and lipodystrophy compared with the 20 HIV-positive men without lipodystrophy and 18 age- and body mass index–matched controls),¹⁶ osteonecrosis of the hip in up to 2% of treated patients,¹⁷ and various metabolic disturbances (hypercholesterolemia in 10% to 27% of HIV-treated patients, hypertriglyceridemia in 23% to 40% of HIV-treated patients, HDL cholesterol less than 35 mg/dL [<0.91 mmol/L] in 19% to 27% of HIV-treated patients, diabetes mellitus in 7% of HIV-infected patients with lipoatrophy, and impaired glucose tolerance in 35% of HIV-infected patients³ ).

Body habitus may be assessed via subjective reporting by both patients and examiners; bioelectric impedance analysis, which can measure overall body composition but not regional body composition data; dual-energy x-ray absorptiometry (DEXA), which can provide regional body composition data (eg, on extremity and truncal fat); computed tomography (CT) scanning, which can provide quantifiable data on subcutaneous adipose tissue and visceral adipose tissue; and magnetic resonance imaging, which has been used in some studies to measure facial fat.¹⁸ It is important to note that DEXA and CT scans are costly, that CT scans involve significant x-ray exposure, and that neither has standardized means of interpretation for these purposes.^{3 ,19 - 20} Although some research studies define significant loss of limb subcutaneous adipose tissue as a 20% decrease from baseline, at least 1 study has suggested that a 30% change in limb subcutaneous adipose tissue is required to be recognized by patients or clinicians.²¹ The International AIDS Society-USA (IAS-USA) and the Infectious Diseases Society of America (IDSA) guidelines for management of the metabolic complications associated with HIV therapy^{22 - 23} note that at the present time, “no technique has demonstrated sufficient sensitivity, specificity, or predictive value to be recommended for routine clinical assessment and monitoring of body fat distribution changes.” Mr B did not have quantitative scanning to evaluate his lipodystrophy as part of a clinical evaluation, which is appropriate in the absence of evidence of the benefit of such scanning.

The Centers for Disease Control and Prevention defines AIDS wasting syndrome as an AIDS-defining condition when it results in the involuntary loss of more than 10% of baseline body weight in association with either chronic diarrhea or weakness with fever, but no specific mention of adipose tissue is included in this definition.²⁴ Human immunodeficiency virus–associated wasting syndrome related to a decrease in total body weight and diarrhea was recognized in Uganda during the early years of the AIDS epidemic, leading some to refer to African AIDS patients as having “slim disease.”^{25 - 26} However, at least 1 early study of the wasting syndrome describes it as predominantly due to loss of nonadipose tissue, with no regional changes in fat distribution or relationship to particular treatments being identified.²⁷

The first published description of fat loss in association with hyperlipidemia and insulin resistance in persons receiving PIs was reported from a cross-sectional study of outpatients in Australia.²⁸ The authors compared 116 HIV-positive men receiving PIs with 32 HIV-positive men who were PI-naive and with 47 HIV-negative control men, using physical examination and a questionnaire to assess lipodystrophy and DEXA scanning to assess body composition. While PI-naive and uninfected controls had similar body composition, lipodystrophy was noted in 64% of PI recipients after a mean of 13.9 months, with fat loss occurring in all regions except the abdomen after a median of 10 months. Patients treated with PIs who developed lipodystrophy also had significantly higher total cholesterol, triglyceride, insulin, and C-peptide levels than did PI recipients without lipodystrophy.

Numerous subsequent reports have estimated the prevalence of lipodystrophy in HIV-infected cohorts as ranging from 6% to 69%, depending on what definition of lipodystrophy is used, whether a metabolic component such as hyperlipidemia is included, the stage of HIV infection being studied, the duration and type of antiretroviral treatment, the means used for measurement, and the characteristics of the study design (cross-sectional vs prospective; controlled vs not controlled).^{29 - 32}

Although these population-based evaluations of HIV-infected cohorts have demonstrated a significant prevalence of both HIV lipoatrophy and HIV lipohypertrophy, at least 4 large studies comparing HIV-infected individuals with uninfected control populations have shown different results. These have included the Multicenter AIDS Cohort Study (MACS), a multicenter, prospective study of the natural history of HIV disease³³ ; the Fat Redistribution and Metabolic Change in HIV Infection (FRAM) study in men and women, in which HIV-infected patients were compared with age-correlated patients who had been enrolled in CARDIA, a prospective coronary artery risk development trial^{34 - 35} ; and the Women's Interagency HIV Study (WIHS), an ongoing, multisite, prospective cohort of women with and without HIV infection.³⁶ These studies observed an increase in lipoatrophy among HIV-infected individuals compared with controls but also noted that controls had consistently greater central adiposity than individuals infected with HIV. Possible explanations for the discrepancy between clinician impression and study results could be the modest sample size (the number of HIV-infected persons in these studies ranged from 350 to 825 and control populations ranged from 142 to 339) or the relatively low ratio of controls to study patients (ranging from 0.34 to 0.59). Additionally, 3 of the 4 studies were cross-sectional (both FRAM studies and MACS) rather than longitudinal, and the fourth (WIHS) involved health care practitioner measurements but not scans of body habitus. In contrast to these data, a study of 306 HIV-infected persons and 107 HIV-negative controls found that the relative amount of visceral adipose tissue for a given body mass index was higher rather than lower in HIV-infected patients.³⁷ In the context of these contradictory data on changes in body habitus for HIV-infected persons vs controls, more controlled, prospective studies will be needed to help clarify this question.

From the perspective of pathogenesis, lipodystrophy is not one unifying syndrome, and even within the categories of lipoatrophy and lipohypertrophy multiple pathophysiologic abnormalities appear to be operative. In the case of lipoatrophy, several NRTIs, especially didanosine, stavudine, and zalcitabine, have been associated with mitochondrial toxicity as a result of interference with the function of mitochondrial DNA polymerase-γ, with possibly associated symptoms including polyneuropathy, bone marrow suppression, pancreatitis, myopathy, cardiomyopathy, lactic acidosis, and lipodystrophy.^{38 - 40} There is now considerable evidence that lipoatrophy is associated with use of stavudine^{41 - 44} and (to a lesser degree) zidovudine,^{43 ,45} although the relationship to mitochondrial toxicity as a mechanism for loss of adipose tissue is less well established, since not all implicated drugs (most notably didanosine) have been associated with lipoatrophy. The association between HIV lipoatrophy and the thymidine analogs has recently been confirmed with data from AIDS Clinical Trials Group protocol 5142, which compared 3 class-sparing regimens.⁴⁶ This study found that limb fat recovered more with the NRTI-sparing regimen (lopinavir/ritonavir plus efavirenz) than with treatment with efavirenz plus 2 NRTIs or lopinavir/ritonavir plus 2 NRTIs. Two-thirds of patients in the latter 2 groups were receiving zidovudine or extended-release stavudine in addition to lamivudine, providing further evidence of the role of these drugs in lipoatrophy.⁴⁶ (Stavudine has also been associated with insulin resistance and mitochondrial dysfunction even in non–HIV-infected patients.⁴⁷ )

Mr B has a somewhat unusual combination of central facial lipoatrophy and anterior cervical lipohypertrophy. The cause of HIV lipohypertrophy (especially central lipohypertrophy) is even less well understood than is lipoatrophy. Dexamethasone suppression of plasma cortisol excluded Cushing syndrome as the cause of dorsocervical fat pad enlargement in a series of 8 patients.⁴⁸ Early cohort studies tentatively identified exposure to PIs (especially indinavir) with this clinical manifestation, leading to widespread use of the term “Crix-belly” (invoking part of indinavir's trade name, Crixivan) to describe any HIV drug–associated central adiposity, even in patients who had never been treated with PIs.⁴⁹ However, the putative association between PIs and central adiposity appears to have been more a function of the prevalent drugs being prescribed in any given patient cohort at the time of study than due to a causal relationship, as more recent studies have failed to find any particular drugs or drug categories associated with lipohypertrophy.^{48 ,50 - 51}

Other proposed mechanisms for HIV lipoatrophy and/or HIV lipohypertrophy have included impaired adipocyte regulation,^{52 - 53} unopposed production of proinflammatory cytokines,⁵⁴ dysregulation of 11-β-hydroxysteroid dehydrogenase,⁵⁵ effects of antiretroviral therapy on the autonomic nervous system,⁵⁶ polymorphisms in hemochromatosis,⁵⁷ mitochondrial polymerase-γ⁵⁸ or MDR-1⁵⁹ genes, interference with sterol regulatory enhancer binding protein 1 and interference with the insulin-regulated glucose transporter GLUT-4,¹⁹ decreased levels of lipin mRNA expression in adipose tissues,⁶⁰ inhibition of cis-9-retinoic acid,⁶¹ inhibition of peroxisome proliferator–activated receptor γ,⁶² and alterations of tumor necrosis factor T-cell homeostasis.⁶³

Numerous studies have suggested that in addition to antiretroviral drugs, certain patient characteristics can also contribute to lipodystrophy. Risk factors for lipoatrophy have included HIV treatment for longer than 2 years (odds ratio [OR], 4.4),⁴⁵ baseline viral load greater than 100 000 copies/mL (OR, 4.3),⁴⁵ age older than 40 years (OR, 3.2),⁴⁵ white race (OR, 5.4),⁴⁵ body mass index less than 24,⁶⁴ CD4 cell count less than 100/μL,⁶⁴ and female sex.^{65 - 66} Risk factors for lipohypertrophy have included duration of therapy and markers of disease severity⁵⁵ as well as increased age (adjusted OR, 2.02 for age ≥40 years compared with age <40 years),³² gain in body mass index (adjusted OR, 3.14 for change in body mass index ≥2 compared with change <2),³² and amount and duration of immune recovery.^{32 ,64}

Quiz Ref IDClinicians have become aware of the negative psychological impact of lipodystrophy, as Mr B describes, with some patients being worried that their body shape changes will disclose their HIV serostatus. Such concerns have potentially serious implications for medication nonadherence and the subsequent risk of developing viral resistance. When asked if Mr B was more concerned about the fat changes in his body or about his lipid levels, he replied, “I was more worried about the fat changes. It's very difficult being a part of society and noticing changes in your body. At that point I was not ‘out’ about my status to a lot of people, so it was far more difficult to have people notice. Just being online in the dating world, if you sent a picture of yourself, they would say, ‘Well, the body doesn't match the face.’ And then it came to a point where you had to disclose your status.”

Similar concerns about the negative psychological impact of lipodystrophy on patients have been raised in the medical literature, such as “The mere possibility that such [lipodystrophic] changes may develop could deter treatment initiation”⁶⁷ ; “These body changes give an unhealthy aspect that resembles wasting and are a reminder of disease in otherwise healthy HIV-infected patients”⁶⁸ ; and “This syndrome of lipodystrophy (often graphically depicted at medical meetings and in journals) inspired fear and led some of our patients to make unwise decisions regarding the initiation of HAART [highly active antiretroviral therapy] for fear of being irreversibly disfigured.”⁶⁹

Numerous studies have confirmed these negative effects. In a survey of 141 HIV-infected gay men in the Amsterdam Cohort Studies, the 50% who experienced lipodystrophy reported a drastic decrease in sexual activity, enjoyed sex less, felt less physically well, and were less confident in relationships.⁷⁰ In a qualitative survey of 33 HIV-infected heterosexual women and men with lipoatrophy, dominant themes that arose included erosion of self-image and self-esteem, problems in social and sexual relations, threats to locus of control, forced HIV disclosure, demoralization, depression, and clinicians' minimization of the importance of lipoatrophy.⁷¹

One prospective study of 207 patients found that adherence to HIV medication was significantly associated with adipose tissue alteration, even though patient perception of changes was not always confirmed by medical diagnosis of morphologic alterations.⁷² An analysis of 277 patients from a French cohort who were initially adherent when interviewed 4 months after initiation of treatment found that, at 20 months after initiation, 76.4% reported at least 1 lipodystrophy-related symptom and 30% were failing to maintain medication adherence. The number of self-reported lipohypertrophy (but not lipoatrophy) symptoms was independently associated with adherence failure.^{73 - 74}

Antiretroviral Drug Substitutions. Because of the known association between mitochondrial dysfunction and cell death,⁷⁵ researchers have examined substituting other drugs for NRTIs suspected of causing adipocyte loss through this mechanism. In observational studies, stavudine has been associated with loss of subcutaneous adipose tissue,⁷⁶ and several studies have observed improvement in subcutaneous adipose tissue loss by switching from stavudine to zidovudine, tenofovir, or abacavir.^{77 - 80}

Several prospective trials that randomized antiretroviral-naive patients to stavudine- or zidovudine-containing vs tenofovir-, abacavir-, or emtricitabine-containing NRTI regimens have all shown less lipoatrophy with regimens that avoided stavudine or zidovudine.^{81 - 83} Another study also showed decreased HIV lipoatrophy and visceral adipose tissue when patients received nevirapine plus lopinavir/ritonavir compared with zidovudine and lamivudine plus lopinavir/ritonavir, although total and LDL cholesterol levels were higher in the nevirapine group.⁸⁴

In 1 study, 81 patients with peripheral lipoatrophy treated with PIs were randomized to continue NRTIs but to substitute PIs with abacavir, nevirapine, adefovir, and hydroxyurea or to continue all therapy (with an option to switch at week 24). Those patients randomized to the initial switch had significantly improved lipids (total cholesterol decreased by 1.24 mmol/L [P = .001] and triglycerides decreased by 29.2 mg/dL [0.33 mmol/L] [P = .002]) and developed less total body fat (−1.6 and −0.4 kg, respectively; P = .006) and weight gain (−2.8 and −0.5 kg, respectively; P < .001). Limb subcutaneous adipose tissue was also significantly reduced (P = .049), but intra-abdominal fat did not change significantly.⁸⁵ Another study of 20 patients who switched from PIs to efavirenz did not observe a reversal of lipodystrophy, although triglyceride levels and fasting insulin resistance index declined significantly.⁸⁶

Diabetes Drugs and Human Growth Hormone . Metformin. Metformin is a biguanide insulin-sensitizing agent that decreases hepatic glucose output and stimulates peripheral glucose uptake.⁸⁷ In patients with lipohypertrophy, metformin reduced the mean (SEM) insulin area under the curve 120 minutes after oral glucose tolerance testing (−2930 [912] vs −414 [432] μIU/mL [−20349 {6334} vs −2875 {3000} pmol/L]; P = .01), weight (−1.3 [0.6] vs 1.1 [0.4] kg; P = .005), and diastolic blood pressure (−5 [4] vs 5 [2] mm Hg; P = .009) vs controls, respectively.⁸⁸ However, metformin has had mixed effects on visceral adipose tissue (ranging from a nonsignificant reduction^{88 - 89} to a mean reduction of 73 cm² [P < .01]).⁹⁰ It has also been associated with a significant decrease in appendicular fat mass and should be used with caution in the treatment of lipodystrophy, being reserved for patients with adequate peripheral fat and marked insulin resistance.⁹¹ Care must also be taken to monitor for development of lactic acidosis in patients who are taking metformin in addition to NRTIs, since metformin can increase the risk of developing this condition.⁹²

Thiazolidinediones. Thiazolidinediones such as rosiglitazone theoretically could affect HIV lipoatrophy through their inhibition of peroxisome proliferator–activated receptor γ, which influences the transcription of genes regulating adipogenesis and glucose and lipid metabolism, thus potentially leading to increased glucose transport into muscles and increased adipogenesis.⁹³ However, studies of rosiglitazone have had mixed results with respect to rosiglitazone's effects on HIV lipoatrophy and associated metabolic measures.^{89 ,94 - 95} In contrast, pioglitazone, which has not been associated with the cardiovascular risks noted with rosiglitazone,⁹⁶ was associated with an increase in limb fat of 0.45 kg vs 0.04 kg (P = .13) in 130 HIV-infected patients with self-reported lipoatrophy randomized to pioglitazone vs placebo (unless patients were also taking stavudine). Numerous questions remain about the use of this drug class for HIV lipoatrophy, with a recent review noting that the limited effect of pioglitazone on subcutaneous adipose tissue occurs only among patients not taking stavudine and suggests that thiazolidinediones should only be used for HIV lipoatrophy if patients also have diabetes mellitus.⁹⁷

Human Growth Hormone. Human growth hormone (hGH) has been approved since 1996 for treatment of HIV-associated muscle wasting conditions but is not approved for lipodystrophy.⁹⁸ Quiz Ref IDIn the context of HIV lipohypertrophy, hGH has been shown to decrease visceral adipose tissue (−32.6 vs 0.5 cm²; P < .001), to decrease limb fat (0.04 vs 0.2 kg; P < .001), to decrease non-HDL cholesterol (−13.0 vs −2.8 mg/dL; P = .02), and to improve lipid profiles,⁹⁹ but, in a study of 30 patients, it also was associated with hyperglycemia (n = 3), arthralgias (n = 15), and fluid retention (n = 4).¹⁰⁰ Human growth hormone–releasing factor (tesamorelin) has been associated with decreased visceral adipose tissue (by 27.8 cm² [−32.9%]; P < .001), an improved lipid profile (total cholesterol decreased by 10 mg/dL [0.26 mmol/L]; P < .001), and significant decreases in patients' “belly image distress” (P = .03) without significant adverse effects, suggesting that this approach may be more appropriate than the use of hGH itself.¹⁰¹ However, no trials have compared the 2 treatments head to head, and continuous treatment with tesamorelin may be required to maintain benefit.¹⁰²

Liposuction and Use of Facial Fillers . Liposuction. Liposuction has been proposed as a means of decreasing HIV lipohypertrophy, especially when it involves the anterior or posterior cervical regions (“horse collar” or “buffalo hump”). Two cases series (1 reported in abstract form only) of 25 and 10 patients found that liposuction for lipohypertrophy is often initially successful, or at least partially so.^{103 - 104} However, recurrence requiring repeat surgery has been seen in up to 50% of patients, and it is unclear whether postoperative problems (including pancreatitis, anemia, and pneumococcal pneumonia in 2 immunized patients) are related to this procedure.^{103 - 104} It should be noted that liposuction can only be used to remove subcutaneous fat and that liposuction of visceral fat in non–HIV-infected patients has not resulted in improved metabolic measures.¹⁰⁵

Facial Fillers. Numerous agents are available to treat facial lipoatrophy, with poly-L-lactic acid probably being the most studied and widely used. Poly-L-lactic acid is a synthetic, biodegradable, immunologically inert, resorbable polymer that has been used in sutures, meshes, and drug delivery devices.¹⁰⁶ Poly-L-lactic acid stimulates fibroblast multiplication and collagen production, leading to a gradual and progressive increase in volume.⁶⁸ In a prospective, open-label study of 94 patients with facial lipoatrophy, poly-L-lactic acid was injected into both cheeks every 2 weeks, with the primary end point being patients' self-perception of improvement on a visual analog scale and with objective confirmation of improvement provided by 3-dimensional photographs showing increased dermal thickness (median increase, 1.99 mm) at the end of treatment. Patients reported a benefit for up to 15 months beyond the last injection, although objective measurements of facial soft tissue volume showed no significant improvement.⁶⁸ A second open-label, prospective study enrolled 50 patients to receive 4 sets of injections at 2-week intervals. By 6 weeks, 19% of patients had a median change of total cutaneous thickness of greater than 10 mm (the study's benchmark), peaking at 48 weeks at 61% of patients and gradually declining to 43% of patients by week 96.¹⁰⁷ A controlled trial randomized 100 HIV-positive patients with moderate or severe facial lipoatrophy to receive 4 open-label poly-L-lactic acid injections at weeks 0, 2, 4, and 6 (immediate group) or to receive the same injections after a delay of 24 weeks (deferred group). Despite positive evaluations of physical improvement by patients and clinicians, and patient-perceived benefits of well-being, social functioning, and quality of life in the immediate group, objective measurements of facial thickness showed only modest improvements (2.2-mm change in soft tissue depth at maxilla and 1.0-mm change at base of nasal septum).¹⁰⁶ The authors attributed this possibly to the nonstandardized methods used to measure facial thickness but suggested that from the patient perspective (albeit nonblinded), there was still a treatment benefit.¹⁰⁶

A second filler approved by the US Food and Drug Administration in 2006 for long-term correction of facial HIV lipoatrophy associated with antiretroviral treatment is calcium hydroxylapatite. One prospective, open-label study of calcium hydroxylapatite in 100 HIV-infected patients found that all patients met the unblinded primary end point of efficacy (“an improved aesthetic result” as judged by comparison of pretreatment and posttreatment photographs at 1, 3, 6, and 12 months) and the secondary end point of increased cheek thickness (which showed an increase of 2.2-2.5 mm by 12 months; P < .001); no adverse effects were reported.¹⁰⁸

Guidelines for Management of Metabolic Complications and Body Shape Changes . The IAS-USA and the IDSA guidelines suggest that avoidance of PIs should be considered in patients with known problems with glucose metabolism or who have first-degree relatives with diabetes mellitus, with consideration given to substituting nevirapine,¹⁰⁹ efavirenz,⁸⁶ or abacavir¹¹⁰ to improve insulin resistance. Fasting glucose levels should be measured before and during treatment with a PI, and, for patients who develop hyperglycemia, established guidelines for treating this condition in non–HIV-infected individuals should be followed.¹¹¹ Oral glucose tolerance testing may be considered in addition to fasting glucose levels, since some HIV patients with lipodystrophy (but not uninfected controls) have normal fasting glucose levels but impaired glucose tolerance test results (P = .001 compared with controls¹¹² ), which are also a risk factor for cardiovascular disease.¹¹³

Some PI therapy (especially ritonavir, even if only used in “boosting” doses,¹¹⁴ but not atazanavir¹¹⁵ ) may result in hyperlipidemia, which is associated with an increased risk of myocardial infarction; the Data Collection on Adverse Events of Anti-HIV Drugs (DAD) prospective cohort study showed that each year of taking PI therapy has been associated with a 16% excess risk of myocardial infarction.¹¹⁶ Mr B's triglyceride elevation is substantially higher (1130 mg/dL [12.7 mmol/L]) than patients who experienced myocardial infarctions in that study, who had an interquartile range of 142-345 mg/dL [1.6-3.9 mmol/L].¹¹⁶ A more recent analysis of these data suggests that based on this nonrandomized study, certain NRTIs can also contribute to this risk, with abacavir associated with a 90% increase and didanosine with a 49% increase.¹¹⁷

Quiz Ref IDAs with any patient with elevated lipid levels, the guidelines of the National Cholesterol Education Program,¹¹⁸ which advocate modification of hypertension, diabetes mellitus, smoking, obesity, physical inactivity, and diet, should be adhered to, and drug therapies added when modification of these factors alone is insufficient to achieve individualized targets. Patients who are treated with both protease inhibitors and statins (especially simvastatin and lovastatin) may have a PI-mediated reduction in statin metabolism through inhibition of the cytochrome p450 system; in this context, pravastatin and atorvastatin are the preferred drugs.¹¹⁸ Because efavirenz can induce the p450 system, simvastatin, atorvastatin, or pravastatin dosages may need to be increased when used in combination with this drug.¹¹⁹

Exercise and Diet. Several studies of exercise in the context of lipodystrophy,^{120 - 121} sometimes with adjunctive meformin¹²² or rosiglitazone,¹²³ have shown improvements in central adiposity. One retrospective analysis of diet in HIV-infected patients with and without fat deposition indicated that a high-quality diet, rich in fiber and adequate in energy and protein, may help prevent fat deposition.¹²⁴ A second, prospective study of 356 HIV-infected and 162 community-derived HIV-negative controls noted that hypertriglyceridemia and metabolic abnormalities were associated with an increased intake of saturated fat (31 vs 27 g/d), suggesting that this dietary element (rather than a reduction in total fat) should be especially targeted in HIV-infected individuals with metabolic abnormalities.¹²⁵ Based on these results as well as the association of saturated fat with heart disease risk, Mr B should minimize his saturated fat intake. Supplementation with chromium nicotinate, 400 μg weekly for 16 weeks, has also been associated with decreased insulin resistance and decreased trunk fat in HIV-infected persons.¹²⁶

Mr B is obviously well-invested in his care, and I urge him to continue speaking with his physician about how important body changes are for his quality of life. His current HIV medical regimen is likely not contributing to worsening lipodystrophy, but he and his physician might consider adding an integrase inhibitor and/or CC chemokine receptor 5 (CCR5) coreceptor blocker (if his virus is only CCR5-tropic) to prevent persistent low-level viremia from progressing to overt resistance (without risking further dyslipidemic adverse effects). (Mr B should also know that even though he has experienced significant adverse effects because of antiretroviral therapy, data from the Strategies for Management of Antiretroviral Therapy trial suggest that staying on therapy actually lowers the risk of major cardiovascular, renal, or hepatic disease, even in the presence of metabolic abnormalities, either by suppressing viral replication, or possibly by improving immune function.¹²⁷ ) Mr B and his physician might also consider adding fenofibrate, niacin, or fish oil to reduce his triglycerides, especially given his family history of cardiac disease. Mr B should continue his healthy diet, avoid saturated fats, and get regular aerobic exercise. Finally, I recommend that Mr B continue to read and discuss issues related to his care, since HIV disease and its management continue to be a rapidly changing field. Since lipohypertrophy is potentially affecting his adherence and is certainly altering his psychological status, I would also support his efforts toward liposuction, although considerable challenges exist in working with insurers in this area.

I encourage clinicians to ask patients about their perceived changes in body habitus; clinicians should not presume that what appears to us to be minimal is inconsequential to patients. Ask patients about the implications of such perceived changes on their medication adherence, social confidence, mood, and sexual and social relationships. Avoid highly implicated drugs (stavudine, zidovudine) or substitute another agent for them, if possible, at the first signs of lipoatrophy. Support efforts for facial fillers, if desired and affordable, and advocate for liposuction if lipohypertrophy is causing pain, restricted range of motion, or other medical complications.

QUESTION: Has there been any investigation of the modes of disposal of these suspected drugs? Is their half-life longer in patients who have lipohypertrophy and atrophy?

DR FULLER: That is a very interesting question, especially as we’ve become aware (with efavirenz being the best example) that some populations metabolize antiretrovirals at different rates, leading to potentially different adverse effect profiles for the same drug dose, indicating a need in some cases to reduce dosing to maintain a therapeutic concentration.^{128 - 129} However, in reviewing this topic I have not come across anything that specifically relates drug levels to the development of lipodystrophy.

QUESTION: What can be done to have these “cosmetic” interventions covered by insurance so that patients can actually have access to them, especially since it is pretty clear that they are a complication of necessary medications for their infection?

DR FULLER: I am not aware that anyone has been successful in arguing for insurance companies to compensate for facial filling, although it is possible to obtain the drug for free from the manufacturing company if patients fit certain economic criteria. I have 1 such patient, but it cost him several thousand dollars per injection sequence (he had 2) to have plastic surgeons do the procedures. On the other hand, if patients with dorsocervical or anterior cervical fat pads are experiencing restricted range of motion, sleep apnea, or pain, I think one could make a medical argument for insurance coverage. It is not easy to apply for prior approval or to make appeals on denial, but I think we owe it to our patients to at least make the attempt, especially when there are resulting functional problems.

MR B: Regarding the insurance issue, I have been in the process of trying to fight my insurance and have been denied twice. With the help of my doctor letter and so forth, I’m still in the fight, but it's amazing to me that my insurance company looks at this as cosmetic, with all the side effects. Even in a roomful of HIV-positive individuals, you are still stigmatized by other people who are HIV-positive because of the way you look.

DR FULLER: Unfortunately, I do not think the trend of insurance coverage for complications of HIV infection and its treatment is moving in a helpful direction. Just in the last month, I have been dealing with an insurance company which will only pay for testosterone replacement in patients who have documented deficiency if they are also experiencing osteopenia, loss of lean muscle mass, or anemia, but not just for subphysiologic levels and associated decreased sexual functioning. So the restrictions on some of these drugs are becoming quite intense.