HIV protease inhibitors confer virological, immunological, clinical and survival benefits [1,2]. Protease inhibitors in combination with HIV reverse transcriptase inhibitors are now recommended as standard antiretroviral therapy [3,4]. The potency and sustained effects of combination protease inhibitor therapy have led to its widespread usage.
Known toxicities of HIV protease inhibitors include renal calculi with indinavir, nausea, diarrhoea and perioral paraesthesiae with ritonavir, and diarrhoea with nelfinavir. These adverse effects generally occur early in therapy, are not usually serious and resolve rapidly with discontinuation. Excessive bleeding in haemophiliacs, hepatitis and portal vein thromboses are relatively rare [5,6]. Long-term side effects of protease inhibitors have not been described. Generalized wasting is a common manifestation of HIV infection and is predominantly due to loss of muscle mass [7,8]. However, regional fat wasting as a consequence of any drug therapy has not been reported.
HIV protease inhibitors can cause hyperglycaemia, perhaps a result of insulin resistance [1,9,10]. Insulin resistance correlates closely with abdominal obesity and hypertriglyceridaemia and underlies type 2 (non-insulin-dependent) diabetes mellitus [11,12]. HIV protease inhibitors can also cause hyperlipidaemia . These observations led us to explore the interaction between HIV protease inhibitors, lipids and insulin sensitivity. We report a syndrome of peripheral lipodystrophy, hyperlipidaemia and insulin resistance due to HIV protease inhibitors.
The aims of this cross-sectional study were to describe and characterize a syndrome of peripheral fat wasting (peripheral lipodystrophy) associated with HIV protease inhibitor therapy. HIV-infected patients who were receiving at least one HIV protease inhibitor were compared with HIV-infected protease inhibitor-naive patients. A second comparison group comprised healthy men with comparable mean age and body mass index . Patients with recent opportunistic infection or malignancy, receiving anabolic steroids or immunomodulators were excluded. All HIV-infected patients without an exclusion criterion and seen for routine clinical care (A.C., D.A.C.) over a 4-week period in August-September 1997 were included. To avoid recruitment bias, no patient was specifically referred for the study.
Lipodystrophy was defined clinically by physical examination and by patient report of fat wasting in the face, arms or legs with or without central obesity. Because data on lipodystrophy were collected retrospectively, patients were initially asked a general question about any changes in body appearance (without reference to protease inhibitor therapy), followed by questions with specific reference to the regions mentioned above, month of onset for change in each region, and whether the changes had resolved. Patients with weight change but without peripheral fat wasting were not defined as having lipodystrophy. Lipodystrophy was attributed to one or more protease inhibitors if lipodystrophy appeared during, but not before, protease inhibitor therapy. For those receiving sequential protease inhibitors, cause was attributed to the protease inhibitor(s) prescribed at onset of lipodystrophy, as long as the duration of therapy exceeded 3 months. Patients completed a validated questionnaire detailing physical activity over the preceding 12 months .
Demographic data collected for each HIV-infected patient were as follows: duration of HIV infection and AIDS, duration and types of all antiretroviral therapies and antimicrobial prophylactic therapies, and family history of diabetes mellitus.
The following measurements were recorded for each HIV-infected patient: weight prior to protease inhibitor therapy and current weight, height, total and high density lipoprotein (HDL) cholesterol, triglyceride, glucose, insulin (Access immunoassay, Beckman, Australia), C-peptide (Linco, St Charles, Missouri, USA), free fatty acid (NEFA, Wako, Osaka, Japan), fructosamine, testosterone, sex hormone-binding globulin (SHBG), prolactin, cortisol, C3, leptin (Linco), and tumour necrosis factor (TNF)-α (Quantikine, R&D Systems, Minneapolis, Minnesota, USA) levels (all measured after a 12 h overnight fast), liver enzymes (total protein, albumin, alanine aminotransferase, aspartate aminotransferase, alkaline phosphatase), CD4+ lymphocyte counts and HIV RNA levels (also recorded prior to protease inhibitor therapy) and, in a randomly selected subgroup, total body and regional fat levels. Total and HDL cholesterol, triglyceride, glucose, fructosamine, liver enzyme, testosterone, SHBG, prolactin, cortisol and C3 levels were determined using routine methods, and insulin resistance was estimated using the homeostasis model . For healthy men, anthropometry, fasting glucose, insulin, C-peptide levels and body fat were determined.
Changes in body weight were expressed as kg per month of protease inhibitor therapy. Total body and regional fat and lean masses and percentages were measured by dual-energy X-ray absorptiometry (DEXA; Lunar DPXL, Madison, Wisconsin, USA). Central abdominal fat measured by DEXA correlates strongly with insulin sensitivity  and with measures of abdominal fat by computed tomography . DEXA was chosen in preference to computed tomography because DEXA precisely measures total body and regional fat levels, except in the face.
HIV infection was confirmed by enzyme-linked immunosorbent assay and immunoblotting. CD4 cell counts were determined by three-colour flow cytometry. Plasma HIV RNA levels were determined by the Amplicor HIV-1 Monitor assay (Roche Molecular Systems, Branchburg, New Jersey, USA); results below the assay's limit of detection were assigned a value of 2.6 log10 copies/ml plasma.
Continuous variables were analysed by analysis of variance methods. Leptin levels were adjusted for total fat levels using residuals from simple linear regressions. Time to lipodystrophy was estimated by the Kaplan–Meier method. Risk factors for development of lipodystrophy were assessed using Cox regression models.
Clinical and laboratory findings in study subjects
Three groups were studied: 116 HIV-infected subjects currently receiving at least one protease inhibitor for a mean 13.6 months (range, 1–39 months; 77 receiving indinavir, 25 receiving ritonavir plus saquinavir, nine receiving nelfinavir plus saquinavir, four receiving nelfinavir, and one receiving saquinavir); 32 HIV-infected subjects who were protease inhibitor-naive; and 47 healthy men (Table 1). There was no significant difference in age, weight, body mass index or albumin levels between the groups. HIV-infected patients had significantly higher C-peptide and lower HDL levels than healthy men. Fasting total cholesterol (P = 0.0001) and triglyceride (P = 0.003) levels were significantly greater in protease inhibitor recipients than in protease inhibitor-naive patients, and leptin (P = 0.004) and HIV RNA (P = 0.002) levels were significantly greater in the protease inhibitor-naive patients than in protease inhibitor recipients.
Body composition in study subjects
DEXA was performed in 61 HIV-infected patients and all healthy men. There was no significant difference in any biochemical or clinical parameter between those HIV-infected patients who did or did not have DEXA performed (data not shown). Patients receiving protease inhibitors had comparable body weight and fat-free mass but significantly lower fat mass overall and in each body region except the central abdomen than both protease inhibitor-naive patients and healthy men (Table 2). Although abdominal distension was reported commonly, mean central abdominal fat mass was not different in protease inhibitor recipients. HIV protease inhibitor-naive recipients had similar total and central fat mass to healthy men.
Features and prevalence of lipodystrophy and diabetes mellitus
Lipodystrophy was observed in 74 (64%) patients receiving a protease inhibitor and in one (3%) protease inhibitor-naive HIV-infected patient (χ2 test, P = 0.0001). Lipodystrophy occurred with equal frequency in all body regions, including the trunk, except the abdomen (data not shown) where patients reported relative abdominal obesity (Fig. 1). Lipodystrophy was attributed to indinavir in 41 patients and to ritonavir–saquinavir in 25 patients. Kaplan–Meier analysis estimated the median time to lipodystrophy to be 10 months (Fig. 2a). This interval was shorter in patients who received ritonavir–saquinavir than in those receiving indinavir (8 and 12 months, respectively; P = 0.013, Fig. 2b). Lipodystrophy was attributed to nelfinavir in three patients and to saquinavir in one patient, but patient numbers were too small to calculate median time to development. Lipodystrophy did not resolve in any patient but improved in three patients who switched from ritonavir–saquinavir to indinavir.
Patients with lipodystrophy had significantly lower fat in all regions except the central abdomen as well as significantly higher triglyceride, insulin and C-peptide levels, and greater insulin resistance than those without lipodystrophy (Table 3). Patients who developed lipodystrophy experienced a relative weight loss of 0.5 kg per month compared with those without lipodystrophy (P = 0.0005).
Three (2%) protease inhibitor recipients had worsening (n = 1) or new (n = 2) diabetes mellitus. For the longstanding, type 1 (insulin-dependent) diabetic patient, daily insulin requirements increased by 70%. In the two new diabetics, one required insulin for symptomatic hyperglycaemia after 4 weeks of indinavir at which time lipodystrophy was noted. The second patient had asymptomatic hyperglycaemia 4 weeks after switching from indinavir to ritonavir–saquinavir that required no therapy and had noted increased fat wasting after 9 months of indinavir.
Risk factors for lipodystrophy
Patients with protease inhibitor-induced lipodystrophy had significantly longer duration of protease inhibitor therapy than those without lipodystrophy (15.2 and 10.9 months, respectively; P = 0.0001). Lipodystrophy was more common in patients receiving ritonavir–saquinavir than in those receiving indinavir (relative risk, 1.70; P = 0.038; Table 4). Lipodystrophy was also more severe in ritonavir–saquinavir patients than in indinavir patients, as shown by less total body fat (8.4 and 14.4 kg, respectively; P = 0.008), and higher insulin (11.4 and 8.4 mIU/l, respectively; P = 0.004), triglyceride (6.7 and 2.2 mmol/l, respectively; P = 0.0001) and total cholesterol (7.0 and 5.6 mmol/l, respectively; P = 0.0008) levels.
Lipodystrophy was not more likely in those with a family history of diabetes mellitus. Other clinical variables, CD4+ lymphocyte counts and HIV RNA levels were also not independent risk factors for lipodystrophy (Table 4). Neither the use of protease inhibitors nor the presence of lipodystrophy was associated with significant differences in levels of liver function, testosterone, SHBG, prolactin, cortisol, C3 or TNF-α (data not shown). Leptin levels were significantly related to total fat mass (r = 0.35; P = 0.008), but were not related to lipodystrophy after adjustment for total fat mass.
Patients receiving HIV protease inhibitors frequently develop a syndrome of peripheral lipodystrophy, hyperlipidaemia and insulin resistance, which is common with prolonged therapy but occurrence of secondary diabetes mellitus is relatively rare. This syndrome was more frequent and profound in those receiving ritonavir–saquinavir than those receiving indinavir. This might be because protease inhibitors have differential capacities to cause this syndrome or may merely represent the use of more than one protease inhibitor. Differentiating the relative contributions of ritonavir and saquinavir was not possible, although ritonavir monotherapy frequently causes hyperlipidaemia . Whether the new formulation of saquinavir with greater bioavailability (about 12%) will cause lipodystrophy as monotherapy is not known.
The study was cross-sectional and relied on a subjective clinical definition, although the clinical data were supported by body composition data obtained by DEXA. The cross-sectional design may have excluded those who had stopped protease inhibitors and therefore a fully representative population was not studied. Prior to this study, however, patients were not ceasing protease inhibitors because of fat wasting. Prospective studies are in progress to further assess the syndrome's incidence and severity, and to determine whether any clinical or biochemical parameter predicts the syndrome. Studies of lipodystrophy in women and children receiving protease inhibitors and its reversibility upon ceasing or switching antiretroviral regimens are required.
The pathogenesis of this syndrome is unclear. Published data on the metabolic effects of HIV protease inhibitors are lacking. Peripheral fat wasting may be due to abnormal lipid release or storage perhaps via adipocyte apoptosis. Fat mobilization leads to elevated circulating fatty acids that can interfere with insulin signalling  or provide oxidative substrate competition between glucose and fatty acid cycles [18,19]. Nevertheless, this study could not determine the sequence of biochemical events or why central fat mass was unchanged. HIV protease inhibitors have high affinity for the catalytic site (Asp-Thr-Gly) of HIV protease, and may bind and alter the function of an homologous human protein(s) involved in lipid metabolism or insulin signalling. It is possible that altered cytochrome P450 metabolism of a substance involved in lipid regulation could also be involved.
Type 2 diabetes generally results from both insulin resistance and impaired insulin secretion. This may explain why lipodystrophy was common but hyperglycaemia relatively rare.
The protease inhibitor-naive patients had slightly increased C-peptide levels and insulin resistance than controls, suggesting that HIV disease may have an effect on insulin sensitivity, although a smaller study did not demonstrate this . However, lipodystrophy was rarely reported in our protease inhibitor-naive patients, and these patients did not have hyperlipidaemia. Any role of HIV infection in the pathogenesis of this syndrome remains to be determined.
Patients with advanced HIV disease clearly benefit from protease inhibitors in terms of disease progression, survival  and reversal of some opportunistic infections . However, any survival advantage in early HIV disease is unproven, although biologically plausible and widely advocated [5,6]. Alternative strategies for complete suppression of HIV replication, such as combining two nucleoside analogues with a non-nucleoside reverse transcriptase inhibitor , although less reliable, might be appropriate, particularly for those with low HIV viral load. Cessation of protease inhibitors should be considered for patients who have failed therapy if there is evidence of lipodystrophy or diabetes mellitus. Furthermore, several patients with lipodystrophy were mistakenly assumed to have HIV wasting syndrome with its psychological, social and economic consequences.
Longer term follow-up is required to assess whether vascular complications of insulin resistance and hyperlipidaemia will develop and whether there is significant morbidity associated with long-term, severe fat depletion, especially for those with HIV-associated wasting. Any role for dietary modification or lipid-lowering drugs for the treatment or prevention of lipodystrophy should be explored.
Newer HIV protease inhibitors that do not cause lipodystrophy, hyperlipidaemia and insulin resistance are required. Proteases from other pathogenic viruses including hepatitis C virus and cytomegalovirus have been proposed as targets for antiviral therapy. These proteins are serine proteases, which are more numerous than aspartyl proteases in humans. The present study highlights the need for more thorough interpretation of prelicensing data and post-marketing surveillance as part of the development of antiprotease agents.
The authors thank the patients who participated in this study, C. Satchell for performing leptin and TNF-α assays, G. Howard for recruitment of the control subjects, and J. Eisman and L. Campbell for review of the manuscript.
1. Danner SA, Carr A, Leonard J, et al.
: Safety, pharmacokinetics and preliminary efficacy of ritonavir, an inhibitor of HIV-1 protease
. N Engl J Med
2. Hammer SM, Squires KE, Hughes MD, et al.
: A controlled trial of two nucleoside analogues plus indinavir in persons with human immunodeficiency virus infection and CD4 cell counts of 200 per cubic millimeter or less
. N Engl J Med
3. Carpenter CCJ, Fischl MA, Hammer SM, et al.
: Antiretroviral therapy for HIV infection in 1997: updated recommendations of the International AIDS Society–USA panel
4. BHIVA Guidelines Co-ordinating Committee: British HIV Association guidelines for antiretroviral treatment of HIV seropositive individuals
5. Deeks SG, Smith M, Holodny M, Kahn JO: HIV-1 protease inhibitors: a review for clinicians
6. Carr A, Brown D, Cooper DA: Portal vein thrombosis in patients receiving indinavir, an HIV protease inhibitor [letter]
7. Grunfeld C, Feingold KR: Metabolic disturbances and wasting in the acquired immunodeficiency syndrome
. N Engl J Med
8. Paton NI, Macallan DC, Jebb SA, et al.
: Longitudinal changes in body composition measured with a variety of methods in patients with AIDS
. J Acquir Immune Defic Syndr Hum Retrovirol
9. Lumpkin MM: Reports of Diabetes and Hyperglycemia in Patients Receiving Protease Inhibitors for the Treatment of Human Immunodeficiency Virus Infection
. sFDA Public Health Advisory; 11 June 1997.
10. Dube MP, Johnson DL, Currier JS, Leedom JM: Protease inhibitor-associated hyperglycaemia [letter]
11. Matsuzawa Y, Shimomura I, Nakamura T, Keno Y, Tokonaga K: Pathophysiology and pathogenesis of visceral fat obesity
. Ann NY Acad Sci
12. Carey DG, Jenkins AB, Campbell LV, Freund J, Chisholm DJ: Abdominal fat and insulin resistance in normal and over- weight women: direct measurements reveal a strong relationship in subjects at both low and high risk of NIDDM
13. Nguyen T, Howard G, Kelly P, Eisman J: Bone mass, lean mass and fat mass: same gene or same environment
? Am J Epidemiol
14. Wick J: Guide to Exercise
. Canberra: National Heart Foundation of Australia; 1983.
15. Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC: Homeostasis model assessment: insulin resistance and B-cell function from fasting plasma glucose and insulin concentrations in man
16. Jensen MD, Kanaley JA, Reed JE, Sheedy PF: Measurement of abdominal and visceral fat with computed tomography and dual-energy X-ray absorptiometry
. Am J Clin Nutr
17. Schmitz-Peiffer C, Browne CL, Oakes ND, et al.
: Alterations in the expression and cellular localisation of protein kinase C isozymes
Θ are associated with insulin resistance in skeletal muscle of the high-fat-fed rat
18. Randle PJ, Garland PB, Hales CN, Newholme EA: The fatty acid cycle: its role in insulin sensitivity and the metabolic disturbances of diabetes mellitus
19. Jenkins AB, Storlien LH, Chisholm DJ, Kraegen EW: Effects of nonesterified fatty acid availability on tissue-specific glucose utilisation in ratsin vivo
. J Clin Invest
20. Heyligenberg R, Romijn JA, Hommes MJ, Endert E, Eeftinck-Schattenkerk JK, Sauerwein HP: Non-insulin-mediated glucose uptake in human immunodeficiency virus-infected men
. Clin Sci
21. Carr A, Marriott D, Field A, Vasak E, Cooper DA: Combination antiretroviral therapy of HIV-associated microsporidiosis and cryptosporidiosis
22. Montaner JG, Reiss P, Cooper DA, et al.
: A randomized, double-blind trial comparing the immunologic and virologic effects of nevirapine, didanosine and zidovudine combinations among antiretroviral naive, AIDS-free, HIV-1 infected patients with CD4 counts between 200 and 600 cells/mm3
1998 (in press).