| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
CLINICAL REVIEW 153 |
Division of Nutrition and Metabolic Diseases, Department of Internal Medicine and the Center for Human Nutrition, University of Texas Southwestern Medical Center, Dallas, Texas 75390
Address all correspondence and requests for reprints to: Abhimanyu Garg, M.D., University of Texas Southwestern Medical Center, 5323 Harry Hines Boulevard, Dallas, Texas 75390-9052. E-mail: abhimanyu.garg{at}utsouthwestern.edu.
Abstract
Human immunodeficiency virus (HIV) infection is a major global health problem. Recently, combination therapy including HIV-1 protease inhibitors (PIs) has dramatically improved the long-term survival of HIV-infected patients. However, such therapy is associated with a lipodystrophy syndrome characterized by selective loss of sc fat from the face and extremities and, in some patients, accumulation of fat around the neck, dorsocervical region, abdomen, and trunk. Lipodystrophy in HIV-infected patients (LDHIV) is associated with insulin resistance and its metabolic complications such as impaired glucose tolerance, diabetes, hypertriglyceridemia and low serum high density lipoprotein cholesterol levels. PIs appear to be the strongest link to LDHIV; however, fat loss has been reported in some patients taking non-PI antiretroviral drugs. Other factors, such as duration of HIV infection, age, and gender, may also contribute to the risk of development of LDHIV. The molecular basis of LDHIV remains unknown. There is no specific therapy for LDHIV. Avoiding weight gain by reducing energy intake and increasing physical activity may be beneficial in reducing fat accumulation as well as improving metabolic complications. Antihyperglycemic drugs may be used to treat diabetes. Management of dyslipidemia may require lipid-lowering drugs; however, the safety and efficacy of such intervention require further studies. Substitution of PIs with other antiretroviral drugs can mitigate dyslipidemia and glucose intolerance, but whether reversal of lipodystrophy occurs remains unknown. Future research is needed to discover the biochemical and molecular markers of lipodystrophy in HIV patients and develop PIs or other antiretroviral agents that are free of metabolic toxicity.
RECENTLY, GREAT PROGRESS has been made in the treatment of human immunodeficiency virus (HIV) infection. Highly active antiretroviral therapies (HAARTs) that include HIV-1 protease inhibitors (PIs) in particular result in marked HIV suppression and have dramatically improved the clinical course, prognosis, and survival of patients infected with HIV. In the last few years, however, lipodystrophy, characterized by redistribution of body fat and insulin resistance, has been reported in many HIV-infected patients, and its relationship with antiretroviral drugs and HIV infection per se has become a subject of debate and investigation. This review is intended to discuss the clinical features and possible etiology and pathogenesis of lipodystrophy in HIV-infected patients.
Clinical and metabolic characteristics of lipodystrophy in HIV-infected patients (LDHIV)
The first anecdotal report of body fat redistribution in an HIV-infected patient undergoing antiretroviral therapy including a PI (indinavir) was published in the medical literature in 1997 (1). In early 1998, Carr et al. (2) provided detailed description of a syndrome of peripheral lipodystrophy, dyslipidemia, and insulin resistance related to the use of PIs. Subsequently, many more cases have been reported. Currently, different names are being used to describe this syndrome, such as pseudo-Cushing’s syndrome, fat redistribution or maldistribution syndrome, and PI-associated or HIV-associated lipodystrophy syndrome. We prefer the term lipodystrophy syndrome in HIV-infected patients (LDHIV).
Body fat redistribution
LDHIV is recognized primarily as loss of sc adipose tissue from the facial (sunken cheeks) and peripheral regions, particularly the extremities (2, 3, 4, 5, 6) (Fig. 1
and Table 1). Affected patients appear to be muscular and have prominent superficial veins in the extremities (6). Some patients have concomitant deposition of excess adipose tissue around the neck (double chin), over the dorsocervical spine (buffalo hump) (7, 8, 9, 10, 11, 12, 13, 14), upper torso (1) and intraabdominal region (6, 13, 15, 16, 17). Breast enlargement has been observed in both women (12, 18, 19, 20) and men (21, 22, 23), but whether it is due to excess sc fat, glandular hypertrophy, or both is not clear. Compared with men, peripheral fat loss in women with LDHIV is often more subtle, whereas increased truncal adiposity is the main complaint (24, 25). Women may develop menstrual irregularities (26). Acanthosis nigricans, a feature of other lipodystrophies, has not been reported. Most patients with LDHIV are clinically well and do not have significant weight loss or opportunistic infections. Their peripheral blood CD4 cell counts are relatively high and HIV viral burden low due to effective antiretroviral therapy (3, 8). However, patients affected by severe LDHIV are often troubled by the disfiguring facial and body appearance, altered (stooping) posture, changing size of clothes, and discomfort from buffalo hump when in a supine position. Furthermore, because of disfigurement, many patients may discontinue HAART therapy (27).
|
|
Several factors may influence the accuracy of the patient’s self-report and the physician’s determination of body fat loss in affected individuals. First, because fat loss may occur gradually over months or years, early recognition may be difficult, particularly in obese individuals and in women, who have more sc fat than men. Moreover, an individual’s awareness and acceptance of body fat loss or gain may be different. For example, an obese person may not be as concerned with mild peripheral fat loss as with increasing central adiposity.
Although a 98% concordance between the patient’s self-reports and the physician’s diagnosis was claimed according to one study (32), significant discrepancy (50%) was reported more recently by the same investigators (33) and others (34). Therefore, the sensitivity and specificity of subjective criteria currently used for LDHIV are poor.
Objective measurement of body fat with dual energy x-ray absorptiometry has confirmed a significant reduction of fat in the extremities, but preservation of, or increase in, fat in the truncal region in patients with LDHIV (2, 32, 35, 36, 37, 38). Further, computerized tomography (6, 15, 17, 38, 39, 40, 41) or magnetic resonance imaging studies (13, 16) have shown increased intraabdominal fat, but not sc fat, in patients with increased abdominal girth. An increased waist to hip circumference ratio and reduced skinfold thickness in the extremities have also been noted in both men and women with LDHIV (41, 42).
Metabolic abnormalities
Patients with LDHIV often have dyslipidemia, impaired glucose tolerance, and insulin resistance (28, 43, 44) (Table 2). Metabolic abnormalities may precede changes in body fat distribution (45).
|
Hypertriglyceridemia, hypercholesterolemia, and low serum high density lipoprotein (HDL) cholesterol levels characterize dyslipidemia in LDHIV. In an early case report (46), a 5-fold increase in serum cholesterol and a 15-fold increase in serum triglyceride concentrations were noted in a patient 5 months after initiation of ritonavir therapy, and these values returned toward baseline 5 wk after discontinuing ritonavir. Severe hypertriglyceridemia after starting PI-containing therapy can lead to chylomicronemia, eruptive xanthomas, and acute pancreatitis (47, 48, 49). Carr et al. (32) reported dyslipidemia in 67% of patients receiving PI-containing therapy compared with 26% prevalence in PI-naïve patients. Others have also reported similar prevalence (50–70%) of dyslipidemia in patients treated with PI-containing regimen (14, 50, 51, 52). Ritonavir and, in particular, a combination therapy with two PIs have been reported to increase the risk of hypertriglyceridemia (7.2-fold) and hypercholesterolemia (19.6-fold) (53). Several reports also indicate an increase in the levels of lipoprotein(a) (14, 53, 54).
Dyslipidemia was noted in the pre-PI era in HIV-infected patients, but was characterized by decreased serum levels of total cholesterol, HDL cholesterol, and low density lipoprotein cholesterol and elevated level of triglycerides (55). Dyslipidemia was associated with low blood CD4 counts and increased plasma levels of interferon-
and was more common in patients with AIDS (56, 57, 58, 59, 60).
Impaired glucose tolerance and insulin resistance.
Most studies report a prevalence of hyperglycemia ranging from 0–20% (12, 14, 61, 62) during PI-containing HAART therapy. Impaired glucose tolerance was reported in 62% patients taking PIs (50), and a rise in fasting blood glucose levels was reported after initiation of PI-based therapy (63, 64). Worsening of preexisting hyperglycemia or new onset of reversible diabetes mellitus with PI therapy is relatively uncommon (65, 66, 67, 68, 69, 70). However, Vigouroux et al. (71) reported diabetes occurring in 11 of 14 (79%) patients with LDHIV, compared with 4 of 20 (20%) patients taking PI-containing HAART without LDHIV. On nucleoside reverse transcriptase inhibitors (NRTI) treatment, hyperglycemia occurred in only 14% of the patients (72). Increased truncal adiposity and visceral adipose tissue may be related to a higher prevalence of dyslipidemia and insulin resistance (12, 17, 73), but not diabetes (12), in these patients.
Many investigators reported elevation of fasting and/or postprandial serum insulin and C peptide concentrations in patients receiving PI therapies and in those with LDHIV (2, 3, 12, 32, 42, 64). In an 8-wk prospective study of 10 patients receiving indinavir-based therapy, insulin sensitivity, assessed by minimal model, decreased by 30% (63). Walli and co-workers (74) reported a higher prevalence of insulin resistance in patients taking PI-containing HAART compared with those treated with NRTIs alone (55% and 27%, respectively), and this was more prevalent in those receiving prolonged (7–13 months) treatment with PI. Furthermore, a reduction in the insulin sensitivity has been documented using hyperinsulinemic euglycemic clamp in patients with LDHIV (75, 76, 77). Mynarcik et al. (75) further showed that insulin resistance in patients with LDHIV was related to the loss of limb fat and not to the accumulation of truncal fat. Glucose intolerance, hyperinsulinemia, and insulin resistance in patients receiving PI-containing HAART therapy may be early manifestations of lipodystrophy syndrome.
Insulin resistance and glucose intolerance may help in distinguishing LDHIV from AIDS-associated wasting (Table 2
). For example, in contrast to the patients with LDHIV, HIV-infected patients (Centers for Disease Control, group IV) in the pre-PI era were noted to have increased peripheral insulin sensitivity and lower fasting serum insulin and glucose levels (78, 79) (Table 2).
Prevalence of LDHIV
Due to the differences in diagnostic criteria, selection of study populations, and duration of follow-up, considerable differences in the reported prevalence of LDHIV exist, ranging from 8–84% and averaging 42%, in patients treated with PI-containing HAART (Table 3). Average incidence varies from 7.3 (61) to 11.7 (80) per 100 patient-yr. Despite these differences, however, a generally higher prevalence was reported in patients after longer-term therapy. From pooled data, the prevalence of LDHIV is 17% in adults treated with PI-containing therapy for less than 1 yr and 43% in those treated for 1 yr or more (Fig. 2). The prevalence of LDHIV in women has been reported to range from 10.5% (35) to 37% (12), but it is probably underreported. Each additional 6 months of treatment with HAART is associated with a 45% increased risk of lipodystrophy (80). An increase in the prevalence of LDHIV is expected in the future with longer follow-up and continued use of HAART. It is interesting that LDHIV has become the most prevalent type of lipodystrophy, whereas the other types of acquired lipodystrophies as well as familial lipodystrophies are still rare diseases (29).
|
|
). Risk factors for LDHIV
The direct cause of LDHIV is unknown, but several risk factors have been associated with the occurrence of LDHIV (Table 4). PIs appear to be the strongest link to LDHIV; however, non-PI antiretroviral drugs, duration of HIV infection, age, and gender may contribute to the risk of development of LDHIV.
|
Many case reports (3, 7, 15, 24) and large observational studies have shown an association between LDHIV and PI-based HAART. During a 21-month follow-up of 39 patients receiving PI therapy, the rate of total body fat loss was 0.4 kg/month during the first year and 0.13 kg/month during the second year (32). The fat loss involved all peripheral regions, but not the abdomen, whereas total and extremity lean body masses remained unchanged (32). In contrast, body fat remained essentially unchanged in PI-naïve patients (32). Similarly, a 15% fat loss/yr from the legs was observed in patients taking PI-containing HAART (84). In addition, PI use is associated with dyslipidemia, impaired glucose tolerance, and insulin resistance (Table 5).
|
NRTIs
Fat loss and accumulation have also been reported in PI-naïve patients treated with NRTIs alone (Table 3). Lamivudine (3TC) (35) and stavudine (d4T) (41, 84, 86) as well as various combinations of NRTI (72, 87) have been implicated in causing body fat redistribution (Table 1,4). Carr et al. (86) suggested that the features of body fat loss in 14 patients treated with NRTI alone were indistinguishable from those of PI-induced lipodystrophy. However, these patients also presented with recent weight loss, fatigue, nausea, hepatomegaly, hepatic dysfunction, and high plasma concentrations of lactic acid (86).
Data are inconsistent concerning the effects of NRTIs on glucose and lipid metabolism (Table 2). Although some investigators have found dyslipidemia (88) in patients treated with NRTIs alone, no relationship between metabolic abnormalities characteristic of insulin resistance and NRTI therapy has been reported by others (86). In fact, patients with NRTI-associated body fat redistribution had either normal blood lipids, glucose, and insulin levels or lower values compared with those in PI-related LDHIV patients and values similar to those in patients not receiving any antiretroviral treatment (86, 88).
It is possible, therefore, that the fat loss and fat accumulation in patients treated with NRTIs alone represent a different disorder than the lipodystrophy syndrome in patients treated with PI-containing HAART and could be complicated by organ toxicity or failure and AIDS wasting in many of them. Whether a particular NRTI causes more fat loss than others is not yet clear. A long duration of NRTI therapy may be required to observe significant fat loss, as no significant body fat changes were noted in 151 patients participating in a randomized, controlled trial of combination NRTI therapy for a 6-month period (89). It is speculated that NRTIs may cause slow fat loss that is accelerated with the addition of PIs (84).
HIV infection
The development and severity of LDHIV have been positively associated with the duration of HIV infection, negatively associated with previous HIV viral load, and both positively and negatively associated with blood CD4 lymphocyte counts in various studies (Table 4). However, how these factors play a role in determining or modifying the development of LDHIV remains unclear.
Nutritional status, age, and adiposity
Body adiposity before receiving PI-containing HAART may also affect features of LDHIV. For example, a cross-sectional study suggested that overweight men and women (body mass index, >28 kg/m2) had a higher prevalence of buffalo hump and breast enlargement (women), but a lower prevalence of facial and gluteal fat loss compared with underweight subjects (body mass index, <20 kg/m2) (90). Older people tend to have greater body fat mass, particularly intraabdominal fat (91, 92, 93, 94), which may contribute to or modulate the body fat changes seen in LDHIV (95).
A disproportionately higher gain in fat mass relative to lean tissue was observed during refeeding after malnutrition and weight loss, although there is substantial individual variation (96). HIV-infected patients with wasting syndrome (weight loss >10% of baseline body weight) lost more fat than lean body mass (97). Thus, during recovery from wasting, body fat may accumulate disproportionately in certain areas. However, LDHIV can occur in patients without a previous history of wasting.
Pathogenesis of LDHIV
The precise mechanisms by which PIs or NRTIs cause body fat changes are not known. However, various speculations and hypotheses have been proposed.
Mechanisms of PI-induced body fat changes and metabolic abnormalities.
Some similarities in body fat distribution between patients with LDHIV and Cushing’s syndrome prompted examination of the hypothalamic-pituitary-adrenal axis in patients with LDHIV (1, 7, 8, 9, 10, 98, 99). Only a few patients with LDHIV had mildly increased serum cortisol concentration or 24-h urinary excretion of free cortisol, but they all had preserved cortisol circadian rhythm and normal response to dexamethasone. Thus, overt hypercortisolism was essentially ruled out.
However, cortisol may be locally produced in adipose tissue from conversion of biologically inactive cortisone by the enzyme 11ß-hydroxysteroid dehydrogenase type 1. The expression of this enzyme and of glucocorticoid receptors is significantly higher in omental fat than in sc fat (100, 101, 102). Therefore, a locally increased glucocorticoid concentration or action without systemic hypercortisolism may induce regional adiposity, but whether this mechanism contributes to the development of LDHIV has not been assessed.
Carr et al. (103) identified a homology between a 12-amino acid sequence of the catalytic domain of HIV-1 protease and the retinoic acid-binding domain of cytoplasmic retinoic acid-binding protein-1 (CRABP-1) and the lipid-binding domain of low density lipoprotein receptor-like protein (LRP). CRABP-1 carries retinoic acid (104), which, when isomerized to cis-9-retinoic acid, activates nuclear retinoid X receptor- peroxisome proliferator-activated receptor-
(PPAR) complex known to regulate adipocyte proliferation and differentiation (105, 106). Thus, PIs, by inhibiting CRABP-1, may inhibit adipocyte differentiation. However, preliminary results of in vitro studies do not support this hypothesis. For instance, three-dimensional crystal analyses of CRABP-1 and HIV-1 protease showed no structural similarity (107). Furthermore, none of the PIs directly binds to retinoid X receptor- or PPAR (108, 109). Inhibition of LRP by PIs may not account for hyperlipidemia, as inactivation of LRP in the livers of wild-type mice did not result in hyperlipidemia (110). Other investigators propose that nonspecific inhibition of human proteins, such as insulin-degrading enzymes or cathepsins (aspartyl proteases), by PIs can cause primary hyperinsulinemia (44, 45, 111, 112). However, this mechanism cannot explain the loss of body fat.
Limited information about the histopathology of fine needle biopsy or surgical specimens of the adipose deposition reveals nonencapsulated mature adipose tissue (7, 9), some with fibrotic changes (113), ruling out dysplastic or neoplastic pathology. Subcutaneous adipocyte apoptosis has also been reported in lipodystrophic areas of patients with LDHIV (114); however, control, PI-naive, HIV-infected subjects were not studied. The same group reported that apoptosis was not reversed on switching from indinavir to nevirapine despite improvement in the metabolic parameters (115). In vitro studies with C3H10T1/2 murine mesenchymal stem cells (108), 3T3-L1 preadipocytes (109), and human preadipocytes (116, 117) showed that several PIs inhibited adipocyte differentiation. Recently, Bastard et al. (118) showed reduced mRNA expression of many transcription factors, including sterol regulatory element-binding protein-1 (SREBP-1) in adipocytes from patients with LDHIV taking PIs compared with healthy subjects; however, SREBP-1 protein levels were increased. A similar increase in the SREBP-1 protein expression has been noted in the liver and adipose tissue of ritonavir-treated mice (119) and after nelfinavir and ritonavir exposure of 3T3-L1 preadipocytes (117, 120), but in 3T3-F442A cells, indinavir reduced SREBP-1 protein expression (121). As SREBP-1c overexpression in adipose tissue of mice causes lipodystrophy (122), PI-induced alterations of SREBP-1 expression may be a plausible mechanism for LDHIV.
Inhibition of insulin-stimulated glucose uptake in 3T3-L1 adipocytes has been produced by indinavir (123) and saquinavir (124), which affect the intrinsic transport activity of glucose transporter-4 (123, 125). These inhibitory effects of PIs on glucose transporter-4 are selective and noncompetitive and may potentially affect insulin sensitivity adversely (126).
When given alone to the healthy volunteers (54) and for the prevention of occupational exposure of the HIV infection to the health care workers, PIs caused hypertriglyceridemia (127). Several mechanisms have been proposed for PI- induced hypertriglyceridemia, including reduction of lipoprotein lipase activity by 81% (124) and protection of apolipoprotein B from degradation by proteasome (128). These mechanisms may pertain to the development of dyslipidemia in patients with LDHIV.
Recently, autosomal dominant familial partial lipodystrophies were reported to be due to defects in lamin A/C (LMNA) (31, 129, 130) and PPARG (131) genes, and autosomal recessive congenital generalized lipodystrophy (CGL) to be due to mutations in 1-acylglycerol-3-phosphate O-acyltransferase-2 (AGPAT2) (132) and Berardinelli-Seip congenital lipodystrophy-2 (BSCL2) genes (133). There are phenotypic similarities among LDHIV and familial partial lipodystrophies. Therefore, whether lipodystrophy in HIV-infected patients occurs due to PI-induced changes in the expression of these genes or other homologous genes, particularly those involved in the triglyceride or phospholipid biosynthetic pathways or adipocyte differentiation, remains to be ascertained.
Lipodystrophies, including LDHIV, present unique opportunities to study regional heterogeneity among adipose tissue and skeletal muscles, particularly their differentiation, regulation, and metabolism in relation to the insulin sensitivity. For example, an increased proportion of type II skeletal muscle fibers (134) and an increase in intramyocellular lipids (135) may contribute to insulin resistance in CGL patients. Whether such alterations occur in patients with LDHIV remains to be investigated.
Mechanisms of NRTI-induced body fat changes.
Recently, Brinkmann et al. (136) emphasized the similarity in body fat distribution in patients with LDHIV and multiple symmetric lipomatosis (MSL), which is characterized by marked accumulation of nonencapsulated adipose tissue around the neck (horse collar and buffalo hump), shoulders, and upper torso regions (137). Although heavy ethanol intake causes MSL in most patients, point mutations in mitochondrial DNA (138, 139, 140, 141) and mitochondrial dysfunction (142) have been identified in some nonalcoholic patients with familial MSL.
Interestingly, NRTIs inhibit DNA polymerase- and thus replication of mitochondrial DNA (143, 144). Mitochondrial toxicity, including reduced cytochrome c oxidase activity and impaired ß-oxidation of fatty acids, is a recognized side-effect of NRTIs and is implicated in hepatotoxicity, myopathy, and neuropathy. NRTIs increase lactate production and can cause lactic acidemia (86, 136) and, in a few cases, severe lactic acidosis (145, 146, 147, 148, 149). Some investigators speculate that, as in familial MSL, mitochondrial dysfunction by NRTIs can induce body fat changes (136, 150). However, patients with MSL do not have facial fat loss, and reduction in sc fat distal to the mid-arms and mid-thighs is relatively mild. Furthermore, most MSL patients with mitochondrial DNA mutations had multiple, discrete, and encapsulated lipomas in the neck and trunk. It should also be noted that patients with MSL usually do not develop metabolic abnormalities associated with insulin resistance, and they have high levels of serum HDL cholesterol (137, 151, 152, 153, 154, 155). Therefore, whether the mitochondrial toxicity of NRTIs is related to body fat changes remains controversial. Whether lactic acidemia is an integral component of NRTI-induced body fat changes is also unclear.
Differential diagnosis
Several disorders that cause loss or redistribution of body fat need to be differentiated from LDHIV. First of all, generalized body fat loss is commonly seen in HIV-infected patients with AIDS wasting syndrome (Tables 1 and 2). These individuals often are clinically ill due to a high viral burden. Such patients lose significant weight, including body fat and lean body mass, as opposed to LDHIV in which lean body mass is preserved. Furthermore, patients with AIDS wasting syndrome do not have impaired glucose tolerance or hyperinsulinemia, although they may have hypertriglyceridemia (47). Poorly controlled diabetes can also present with severe weight loss. MSL should be considered if there is a history of heavy ethanol intake. A history of steroid or megesterol use should be determined, as Cushing’s syndrome is associated with central obesity and dorsocervical fat deposition. Testosterone therapy can cause muscular appearance of the limbs, requiring differentiation from lipodystrophy.
Management
Observation.
Most patients with LDHIV have only mild symptoms. It is unclear how frequent and to what extent LDHIV is reversible, either spontaneously or with intervention. As PIs provide superior HIV suppression and, at present, are the most effective treatment for long-term survival of HIV-infected patients, the benefits from continued use of PIs probably outweigh the side-effects in many patients.
Diet and exercise.
Increased physical activity, particularly aerobic exercise, should be encouraged to improve dyslipidemia and insulin resistance. Resistance exercise has been shown to be beneficial in reducing truncal fat as well as total body fat in a 16-wk pilot study (156). Exercise training can lower serum triglycerides (157, 158) by as much as 25% (157) and along with dietary restriction can reduce total abdominal and visceral fat as well (159).
Severely hypertriglyceridemic patients should be advised to consume very low fat diets and avoid ethanol consumption to prevent chylomicronemia and acute pancreatitis (160). Low fat, high carbohydrate diets, however, exacerbate hypertriglyceridemia and hyperinsulinemia in nondiabetic subjects and in patients with type 2 diabetes (161, 162). Increased doses of 
-3 polyunsaturated fatty acids (5–10 g/d) from concentrated fish oil preparations may effectively lower plasma triglyceride concentrations.
Cessation and alteration of antiretroviral therapies.
In patients with severe LDHIV syndrome who were unable to continue current therapy, improvement in fat loss after cessation of PI therapy has been observed by some investigators (163), but not by others (3, 9). PI withdrawal may also improve dyslipidemia, insulin resistance, or hyperglycemia in such patients (46, 48, 66, 67). As the severity of dyslipidemia induced by different PIs alone or in combination may differ (2, 164, 165), switching PIs (such as from the ritonavir/saquinavir combination to indinavir or nelfinavir) may also result in an improvement in lipid profile in certain patients.
Although substituting PI with an NRTI may improve insulin sensitivity (74), a non-NRTI class of antiretroviral drugs may be another alternative. Replacing PI with nevirapine in combination antiretroviral therapy for 6–12 months has been reported to improve and even normalize dyslipidemia, glycemia, and insulin resistance, whereas HIV suppression was maintained (166). Quality of life and subjective perception of lipodystrophy were also improved after switching PI to nevirapine during a 3- to 6-month follow-up (23, 166, 167). However, lipodystrophy associated with nevirapine (168) and failure of HIV suppression after switching PI to nevirapine in some patients (169) has been reported. The incidence of developing lipodystrophy was reported to be lower in patients treated with efavirenz than in those given indinavir-containing regimens during a 24- to 88-wk follow-up (170). However, conflicting effects of efavirenz on dyslipidemia are reported (40, 171). Substitution of stavudine with other NRTIs for 6 months was reported to increase total body fat in patients receiving PI-containing and other regimens (172). Switching to abacavir has been reported to improve lipid levels and decrease insulin resistance (173). Recently, switching from PIs to abacavir, nevirapine, adefovir, and hydroxyurea led to an improvement in dyslipidemia, but not in loss of limb fat, over a period of 6 months. Interestingly, fasting plasma glucose and insulin remained unchanged. A reduction in the intraabdominal fat was also noted (174). Further prospective controlled studies are needed before definite recommendations can be made regarding substitution of PIs with other therapies.
Recombinant human GH (rhGH) and anabolic steroids.
Therapy with rhGH for 3–6 months caused noticeable reduction in the size of buffalo hump and truncal adiposity in a small number of patients with LDHIV, but there was no improvement in peripheral lipodystrophy and dyslipidemia. The studies, however, were not placebo controlled (175, 176, 177). In other studies rhGH therapy decreased visceral fat and fasting serum triglyceride concentrations, but worsened insulin sensitivity (77, 177). Nandrolone therapy for 8 wk had no significant effect on body fat redistribution (178). The effects of testosterone therapy on LDHIV have not been studied, but it may reduce intraabdominal fat accumulation (179).
Surgical correction.
Liposuction has been used to remove excess cervical and truncal fat (180, 181). However, excess fat deposition may recur if patients continue to take PI-containing HAART (7, 9).
Drug therapy for dyslipidemia and diabetes.
Fibric acids (gemfibrozil, benzafibrate, and fenofibrate) lower serum triglycerides by approximately 40% (182, 183, 184). Hydroxymethylglutaryl coenzyme A reductase inhibitors (statins) have also been used effectively (182). However, because many statins (simvastatin, lovastatin, and atorvastatin) are metabolized by cytochrome P4503A4, which is inhibited by PIs, simultaneous use of PIs and statins can elevate plasma levels of statins and increase the risk of myopathy. Pravastatin is not metabolized by P4503A4 and should be used with caution (185). Fibrates are also metabolized through cytochrome P450, but the predominant pathway is CYP4A (186); thus, PIs are less likely to interfere with their metabolism. The long-term safety and efficacy of lipid-lowering drugs in LDHIV patients remain to be investigated. Extreme hypertriglyceridemia leading to pancreatitis may require plasmapheresis (49).
Treatment of diabetes with common antihyperglycemic medications is reported to be effective (66, 68, 69). Metformin is beneficial (187, 188), improves insulin sensitivity, and reduces visceral fat (188). However, lactic acidosis remains a concern, particularly when metformin is used with NRTIs. Further, metformin can cause fat loss (188). In a pilot study of six patients with PI-associated diabetes, treatment with troglitazone, a PPAR activator, improved insulin sensitivity, glucose homeostasis, and dyslipidemia (189, 190). However, in a double-blind, randomized, placebo-controlled trial, rosiglitazone exacerbated hypertriglyceridemia in patients with LDHIV (191). Estrogen therapy should be avoided in women with LDHIV and hypertriglyceridemia either as oral contraceptives or as postmenopausal hormone replacement therapy because it can accentuate hypertriglyceridemia.
Future directions
The first step in studying LDHIV is to establish reliable diagnostic criteria. Large-scale epidemiological studies and, ultimately, prospective, randomized, and controlled clinical trials are needed to establish risk factors for LDHIV. The underlying molecular basis of LDHIV needs to be elucidated. This knowledge may lead to the discovery of new antiretroviral agents that are a low risk for lipodystrophy and hyperlipidemia and may provide insight into the mechanisms of insulin resistance in other adipose tissue disorders.
Acknowledgments
We acknowledge Dolores Peterson, M.D., Ph.D., and Claudia Quittner, R.N., for referring HIV-infected patients with lipodystrophy to us, and Angela Osborn for technical assistance.
Footnotes
This work was supported by NIH Grants DK-56583 and MO1-RR-00633.
Abbreviations: CRABP-1, Cytoplasmic retinoic acid-binding protein-1; HAART, highly active antiretroviral therapy; HDL, high density lipoprotein; HIV, human immunodeficiency virus; LDHIV, lipodystrophy in HIV-infected patients; LRP, low density lipoprotein receptor-like protein; MSL, multiple symmetric lipomatosis; NRTI, nucleoside reverse transcriptase inhibitor; PI, protease inhibitor; PPAR, peroxisome proliferator-activated receptor-; rhGH, recombinant human GH; SREBP-1, sterol regulatory element-binding protein-1.
Received May 23, 2002.
Accepted August 8, 2002.
References
/RXR heterodimer. J Endocrinol 164:R7–R10
