Important note: Information in this article was accurate in September 2000. The state of the art may have changed since the publication date.
Despite the numerous presentations on lipodystrophy (LD) in Durban, the overall picture remains fuzzy. It is clear that LD is a significant problem, that the more we look for it the more we find it, and that it is very important to patients. A GMHC study found that 30% of those surveyed changed and 7% stopped medicines because of body shape changes. Of those who hadn't made a change, 57% were considering such a change, and 46% would change medicine if symptoms worsened [Kasper, Abstract 1380].
The continued lack of a working definition hampers our efforts to study this phenomenon. Interestingly, in a French study designed to assess the cost of LD, 11 of 47 experts agreed to serve on a panel but were unable to reach consensus on a definition [Richter, Abstract 4251]. There was broad consensus among conference attendees and presenters that the term encompasses fat accumulation in the breast, dorsocervial or visceral/abdominal area or in discrete lipomas; fat loss involving the face and/or extremities; or both. Associated with LD and probably occurring earlier are lipid changes: increased cholesterol, decreased HDL, elevated cholesterol/HDL ratio, and increased triglycerides (TG). Measurement of LD remains controversial, but nearly all of the studies presented in Durban used physician or patient identification. Although patients and providers generally agree on the presence of LD, patients tend to report more severe findings [Bernasconi, Abstract 703]. We continued to learn that anthropomorphic measurements like waist-hip ratio (WHR) are not helpful because of poor specificity [Belloso, Abstract 1436; Macallan, Abstract 4246]. Our inability to agree on how to define and measure these problems helps to explain why the reported prevalence still ranges from 15 to 96%.
In interventional studies objective measurements by DEXA, MRI or CT are necessary. Based on a comparison of 60 HIV-negative controls with 61 HIV-infected patients, 16 without LD and 45 with LD, Martinez and colleagues suggested that brachial and malar sonography may be another tool, with 85-88% sensitivity and 75-84% specificity [Abstract 4280].
In clinical practice, patient report or provider observation is the most common method for detecting LD. A number of participants in Durban argued for serial body measurements (neck, breast, abdomen, buttock) or serial photographs of patients. The bottom line is that if a patient complains about LD, it is likely to be present. Clinicians should look for evidence of early LD, since in the case of fat loss, there is little that can be done to reverse the process once it's become obvious.
Several authors used a cross-sectional design to assess risk factors for LD. Multiple studies found an association between d4T use and fat loss, and between use of PIs, especially RTV and IDV, and fat accumulation. Older age and possibly sex are emerging as important cofactors. There is some evidence that men may be more likely to develop fat atrophy and women to develop fat accumulation. Other factors reported by several authors include advanced disease, duration of disease, and duration of HAART.
A large Swiss cohort of 1379 subjects found a prevalence of 43 % (578) for at least one sign of LD [Bernasconi, Abstract 703]. Twenty-eight percent had fat loss (49% of those noted facial fat loss), while 30% had fat accumulation (of those: 69% had abdominal fat accumulation, 6% had a dorsocervical fat pad, and 16%, mostly women, had breast enlargement). In this cohort, 2.4% had diabetes mellitus, 43% (582) had elevated TG levels, and 35% had a total cholesterol/HDL ratio of >5. After adjusting for age, sex, severity of disease and CD4 count, there was a strong association of fat loss with the use of d4T for over 2 years (OR 3.2). IDV use and fat accumulation were less strongly associated (OR 2.0). There was no increased risk associated with AZT use, and although ABC was associated with a slight increased risk (OR 1.2), the number of subjects on ABC was limited. In a multiple logistic regression, lipoatrophy was associated with d4T use, older age, increased lactate and total NRTI experience; it was not associated with CD4 count, viral load, or sex. In the multivariate analysis for fat accumulation, only age >41 years old and IDV use were associated.
The association of fat loss with older age and d4T use was observed in a number of abstracts. A Spanish cohort of 232 patients, approximately half of whom were women, found the odds ratio for d4T use to be 7.6 [Carmena, Abstract 4215]. In a CDC-sponsored study involving 1077 mostly gay, white men from eight U.S. cities, correlates of fat loss were d4T use, age >40, severity and duration of HIV infection, and PI use of at least 2 years [Lichtenstein, Abstract 704]. In a report based on results from an anonymous questionnaire administered to 1161 respondents from 10 European countries, older age, longer duration or severity of HIV, IDV or RTV use, and d4T use were all risk factors for LD [Schrooten, Abstract WePeB4268]. Multivariate regression analysis of results from a multi-center German cohort of 212 subjects also found an association with use of d4T for over more than one year, CD4 nadir <200 cells/mm3, and increased TG [Mauss, Abstract 1486]. Additional studies implicated advanced age; d4T or PI use. [Fabre, Abstract 4245; Wolf, Abstract 1382; Galli, Abstract 702; d'Armino Manforte, Abstract 4242].
A few studies failed to find an association with d4T. The SMART (Systematic Monitoring of Antiretroviral Therapy) cohort of 207 subjects in Amsterdam found that only duration of ART, older age and weight at initiation of HAART were significantly associated with d4T [Hillebrand, Abstract 4247]. In this study the prevalence of LD was low (15.5%), and patients changed medicines frequently. This made it difficult to implicate a specific drug; however, NRTI use in general was associated with LD. Domingo, from Spain, found that only a sedentary lifestyle was associated with LD [Abstract 705]. Taken together, the weight of evidence continues to implicate d4T as a risk factor for fat loss, especially in older patients. RTV and/or IDV were consistently associated with fat accumulation and lipid abnormalities. Only patients on PI-containing regimens developed LD (23%) in a study of 115 subjects on HAART for over 6 months with viral RNA <200 c/ml [Trape, Abstract 4281]. The authors didn't comment on the specific components of the HAART regimen or examine the effect of specific NRTIs, but they found no differences in metabolic abnormalities. Nineteen HAART-naïve patients, 46% with advanced HIV disease, began RTV/IDV (400/400). Compared to baseline measurements, cholesterol and TG levels were elevated at all follow-up time points (4, 12, 24 and 36 weeks), and those with greater adherence had the largest increase in lipids [Scerpella, Abstract 4232]. French investigators followed 1158 subjects from 47 centers after initiation of PI treatment [Raffi, Abstract 4271]. In multivariate analysis, high TG was associated with male sex (OR 2.1), age >35 years (OR 1.7), BMI >25 (OR 1.7), PI use/duration, especially RTV (OR 2.0), and prior ART use (OR 1.7). High cholesterol was associated with age >35 (OR 2.1), viral load >500 c/ml (OR 4.0), and CD4 increase >100 cells/mm3 (OR 1.7).
As noted above, it has been suggested that women may experience more fat accumulation, and men more fat loss. Crabe and colleagues reported a higher prevalence of atrophy in males (44%) than females (23%) [Abstract 4212]. Once again, older age was also associated. Marshal Glesby reviewed 526 subjects from SALSA (Self Ascertained Lipodystrophy Syndrome Assessment), using those with LD but no fat accumulation as the "control" group, and found that female sex, BMI >28 and viral load suppression to <500 c/ml were all risk factors associated with fat accumulation [Abstract 4272]. A final study found that fat accumulation was more common among women [Falutz, Abstract 4273]. Some participants commented on potential differences by ethnicity or race but there was essentially no data to support these claims.
We still do not know what causes LD. Clearly there is an interplay among host factors, medications, and HIV. Some metabolic changes occur prior to institution of ART. A CPCRA study of 132 ART naïve patients found increased cholesterol in 12.4% but only 1.2% had cholesterol >240 mg/dl [Raghavan, Abstract 4269]. Triglyceride levels of >190 mg/dl were common (19.4%). High TG was associated with lower CD4 counts (p=0.02) a history of AIDS (p=0.004), and white race (p=0.02). Hyperglycemia (3%) was associated with age >45 (p=0.02). Disease control itself has been implicated by some. However, Rodriguez, in a cohort of 88 HIV-infected injection drug users in Miami, found that hypertriglyceridemia was more common among those on ART who did not achieve an undetectable viral load [Abstract 2162]. Interestingly, he also found low selenium (<120) to be associated with increased TG, the significance of which is unclear. In a case-control study of 212 subjects, elevated cholesterol, present in all treatment arms, was more common in PI-treated subjects [Carosi, Abstract 4225]. In the PI-treated group there was a correlation between elevated cholesterol and both CD4 increase as well as viral load decrease. TG increased only in the PI group; there was a positive correlation with CD4 increase and an inverse correlation with post-treatment viral load. Duration of treatment, the specific NRTIs used, and other differences between these two groups were not reported. In an Argentine study of 115 subjects with LD, there was also no association observed with viral suppression [Trape, Abstract 4281]. In this study, however, the effect of specific antiretroviral agents was not evaluated.
There were a few intriguing studies suggesting a mechanism. In a late breaker presentation, Fessel from San Francisco noted that in contrast to centrally distributed brown fat, white fat, which is distributed peripherally (limbs, buttocks, face), has fewer mitochondria and is therefore more sensitive to mitochondrial toxicity by drugs such as d4T [Abstract LB115]. In his study of 23 subjects with dorsocervical fat pads, HDL was significantly lower (30.4 mg/dl v. 41.9, p=0.01) and the cholesterol/HDL ratio higher (7.4 v.5.6, p=0.01) in those with LD defined by various methods. Intra-abdominal fat by CT and HDL were negatively correlated. Fessel postulated that mishandling and storage of fat coupled with mitochondrial damage is responsible for LD and that treatment might include restoration of HDL. Using age, sex and BMI matched controls, Hadigan also found low HDL cholesterol (37 v.48 mg/dl) and higher TG (332 v.120 mg/dl) to be associated with LD [Abstract 762]. After adjusting for waist-hip ratio, HIV-infected patients were more likely to have impaired glucose tolerance, suggesting that central obesity alone is not responsible for metabolic disturbances. Using a standard test meal with a radioactive tracer, Morlese demonstrated abnormal lipid processing in six patients with LD on PIs, compared to six HIV-negative, age matched controls and seven HIV-infected, PI-naïve subjects [Abstract 759]. Once the naïve patients began PI treatment, there was a non-significant trend toward progressive abnormalities in lipid handling at three and six months, with decreased peripheral uptake of lipids and enhanced oxidation.
Although cross-sectional studies demonstrate associations rather then causation, it now appears that d4T use is strongly associated with fat atrophy, PI use (especially RTV and IDV) with fat accumulation, and that older age is a risk factor for both. Metabolic changes are also associated and are likely to precede body changes in many cases. The complex interplay of host factors, viral control, and HIV effects remains elusive.
Management of LD remains problematic. Medicines already exist to treat the metabolic changes associated with ART, but we need to be aware of drug interactions. For example, simvastatin levels can increase up to 2500% with RTV/SQV. In Durban, two abstracts reported on pravastatin, which is not metabolized by cytochrome P450 and has very low plasma protein binding. In an open-label, randomized trial, Moyle showed that pravastatin (40 mg a day) with dietary advice tended to be superior to dietary advice alone, with a 17.3% decrease in total cholesterol compared to 5.5% [Abstract 1438]. He found no change in TG, weight or glucose. In a smaller study of 19 subjects, Baldini found similar results, but also observed a 37% decrease in TG [Abstract 4277]. Given these findings and the low likelihood of drug interactions, pravastatin appears to be a reasonable choice of agents to treat high cholesterol in patients taking PIs.
Treatment of the body changes is proving much more difficult. For fat loss, early recognition with possible modification of ART regimen is all we have. Stopping medicines has anecdotally been reported to help. In an NIH sponsored study, 26 white men with viral loads <400 c/ml for 12 months who stopped medicine for 5.9 weeks experienced decreases in cholesterol, LDL, and TG [Hatano, Abstract 4221]. Twenty-four hour urinary free cortisol levels increased, and 24 hour urinary 17-OHCS normalized. There were no changes in insulin resistance or anthropomorphic measurements. However, it is neither realistic nor safe to stop medicines in most patients, especially since reversal of morphologic changes, if it occurred at all, would presumably take a long time.
Changing regimens seems to be a feasible option for patients with reasonable ART alternatives. Cotton and colleagues evaluated six papers or presentations from international conferences that evaluated changing from a PI to an NNRTI, with a total of 307 subjects Abstract 4197]. Taken together, mean cholesterol levels decreased 16% and TG 27.6% by six months. Changes in body shape were inconsistent. Tebas evaluated the switch to NVP in 40 subjects with VL <200 c/ml for at least six months and <40 c/ml at the time of substitution [Abstract 1485]. Parameters of glucose metabolism improved, TG decreased 44% (325 to 225), HDL increased 26%, and LDL did not significantly change. As measured by DEXA, there was a trend toward decreased fat; there was no improvement in bone density at 36 weeks. In a small French study of varying stages of disease, LD resolved in two subjects, improved in two and demonstrated no change in 12 subjects. There was a trend toward improvement in metabolic parameters [Badsi, Abstract 4206]. Subjects reporting lipoatrophy did not improve at 24 weeks in a multicenter non-randomized study of patients "intolerant" of PIs [Casado, Abstract 4171]. In the LISTOP Study 52% of 17 subjects reported subjective improvement in appearance, but there were no changes in metabolic parameter or objective measurements of fat [Lafon, Abstract 4257]. Changing from a PI to an NNRTI after LD has developed, therefore, has met with little success. There are several problems applying these studies to clinical practice: most were not controlled studies; most did not address changing both the PI and the specific NRTI; and in most patients, significant LD was already present. It is conceivable that if LD is recognized earlier, effective ART substitutions can be made to prevent further progression. At this point we must assume that once significant LD occurs, especially fat loss, reversal may not possible.
Growth hormone (rhGH) can provide temporary treatment of fat accumulation, but with significant toxicity and a very high cost. Engelson reported on use of open-label rhGH, 6 mg/day, in subjects with truncal obesity determined by both patient and physician [Abstract 1437]. Subjects were on stable ART, and multiple sophisticated measurements were performed to assess fat, muscle and metabolism (DEXA, total body K+, 2 hr GTT, lipid measurements). They demonstrated a 46% decrease in visceral adiposity; however, subcutaneous adipose tissue also continued to decrease (by 18%). There was a 3% gain in lean body mass by week 24. There were a significant number of side effects and dropouts. After stopping medicine, all changes reverted to baseline. A late breaker presentation reported on eight subjects receiving rhGH for LD in a 12-week crossover, placebo-controlled study [Furrer, Abstract LB114]. Seven of eight patients noted side effects. Mean trunk fat decreased by 2.4 kg (p= 0.02), total mean fat decreased by 2.6 kg (p=0.02). This short-term study found no differences in subcutaneous fat, or fasting lipids. Bone mineral density also increased, and there was a trend toward an increase in lean body mass. With the exception of one participant who became an extreme gym enthusiast, all the improvements reversed during the washout. Other studies reported anecdotal changes in fat using rhGH. Testosterone did not have an impact.
Growth hormone therapy has a number of problems. In addition to the high cost and significant toxicity, the benefits are temporary, and the use of growth hormone exacerbates fat atrophy while it treats fat accumulation. Exercise and diet may also be effective in treating fat accumulation, but there are no known therapies for fat loss.
The long-term cardiovascular risk of metabolic changes is unclear and unproven. Two French studies failed to demonstrate a significant association with vascular disease. There was a suggestion of abnormal carotid intima-medial thickening, a marker of cardiovascular damage, in those with LD; however, this was not significant in the final model [Mercie, Abstract 761]. The incidence of myocardial infarction was also not significantly increased [Mary-Krause, Abstract 4231]. We should remember that cardiovascular complications take a long time to develop, and in another decade we may well see a significant increase in such complications. Uncontrolled hypertriglyceri-demia or hypercholesterolemia should be aggressively managed.
LD and metabolic/lipid derangement are significant problems with a relatively high prevalence. The importance of these changes to patients cannot be underestimated. Host factors (age, sex, baseline body composition), disease factors (CD4 depletion and recovery, viral load and degree of suppression) and drug factors (d4T, RTV, IDV, duration of treatment) all seem important and the relationships complex. Mitochondrial damage and abnormal fat, glucose and lipid processing have also been implicated. Many of the metabolic changes can be treated with standard medicines, but changes in body shape are much more difficult to reverse. Fat accumulation may be partially reversible with exercise, diet, use of growth hormone, or discontinuation of protease inhibitors, but there is no evidence that we can reverse fat atrophy once it has occurred. Continued evaluation with prospective studies using objective measures and blinding are needed to help us understand and treat these frustrating complications of antiretroviral therapy.
20000901
JH20000906
©1997,1998,1999,2000. The Johns Hopkins University AIDS Service, Division of Infectious Diseases. Permission to use and reproduce portions of this newsletter is hereby granted provided that author and publication are fully credited and both copyright and permission notice appear with reprinted material. Inquiries may be directed to Sharon McAvinue, Managing Editor. Website: Johns Hopkins AIDS Service.
The original of this article can be found at http://hopkins-aids.edu/publications/report/sept00_6.html
ÆGiS is made possible through unrestricted grants from Boehringer Ingelheim, iMetrikus, Inc., the National Library of Medicine, and donations from users like you. Always watch for outdated information. This article first appeared in 2000. This material is designed to support, not replace, the relationship that exists between you and your doctor.
ÆGiS presents published material, reprinted with permission and neither endorses nor opposes any material. All information contained on this website, including information relating to health conditions, products, and treatments, is for informational purposes only. It is often presented in summary or aggregate form. It is not meant to be a substitute for the advice provided by your own physician or other medical professionals. Always discuss treatment options with a doctor who specializes in treating HIV.
Copyright ©1990, 2000. ÆGiS & the Sisters of Saint Elizabeth of Hungary. All materials appearing on ÆGiS are protected by copyright as a collective work or compilation under U.S. copyright and other laws and are the property of ÆGIS and the Sisters of Saint. Elizabeth of Hungary, or the party credited as the provider of the content.