Secondary Logo

Journal Logo


Fat distribution evaluated by computed tomography and metabolic abnormalities in patients undergoing antiretroviral therapy: preliminary results of the LIPOCO* study

Saint-Marc, Thierrya; Partisani, Mariac; Poizot-Martin, Isabelled; Rouviere, Olivierb; Bruno, F.e; Avellaneda, R.a; Lang, Jean-Marie; Gastaut, Jean-Albert; Touraine, Jean-Louisa

Author Information
  • Free



The use of HIV protease inhibitors in the treatment of HIV infection has demonstrated clinical, immunological, and survival benefits [1,2] However, several reports have documented some unexpected side-effects, including a group of conditions characterized by altered lipid and glucose metabolism and body fat distribution in patients undergoing combined antiretroviral regimens. The clinical pictures reported are the accumulation of adipose tissue in the dorsocervical region, commonly referred to as `buffalo hump' [3] fatty infiltration and bulging supraclavicular fat pads, called `Madelung's disease' [4] visceral abdominal-fat accumulation [5] benign symmetric lipomatosis [6] and subcutaneous fat wasting of the face and limbs, commonly referred to as `lipodystrophy' [7] Metabolic abnormalities which may be associated include hyperlipidaemia, hyperinsulinaemia, impaired glucose tolerance and diabetes mellitus [7] and increased 24-h urinary free cortisol levels [8] Some metabolic abnormalities, such as hypertriglyceridaemia, may precede antiretroviral therapy [9] and be exacerbated during therapy. The pathogenesis of body fat and metabolic changes occurring among HIV-infected patients treated with antiretroviral therapy remains obscure, and their interrelations are uncertain.

Carr et al. [10] have recently hypothesized that HIV protease inhibitors might induce peripheral lipodystrophy by binding and inhibiting a homologous human protein involved in lipid metabolism. However, recent published reports strongly suggest that nucleoside reverse transcriptase inhibitor (NRTI) therapy, particularly that including stavudine, may also be associated with alterations in body shape [11–13]

Furthermore, a recent study from Gervasoni et al. demonstrates that redistribution of body fat might be an unusual byproduct of effective control of the virus, particularly in the case of patients who have received antiretroviral therapy for a long time [12], as already suggested by Miller et al. [5]

Using multiple regression analysis, Kotler et al. have shown a significant association between log plasma HIV viral content and body fat distribution (estimated from anthropometric measurements) whereas neither antiretroviral and especially protease inhibitor (PI) therapy, nor CD4+ lymphocyte counts were associated with fat distribution [8]

The diagnosis of fat distribution disorders usually relies on clinical criteria. However, Pettit et al. [14] have shown a lack of correlation between patient and physician diagnoses when both were asked to diagnose `truncal obesity'. Furthermore, total body fat measured on DEXA Scanning was not useful in diagnosing lipodystrophy because of the normal variation in body fat mass in the HIV-uninfected and HIV-infected populations [15] Skinfold thickness may be useful, but one recent study suggests that tricep skinfold thickness had a sensitivity of only 60% with a specificity of 70%[16]

The purpose of this study was to compare regional body fat distribution by using quantitative computed tomography (CT) and metabolic parameters in a cohort of 154 HIV-infected male patients experiencing various fat distribution disorders and treated with different antiretroviral regimens. The relative effects of antiretroviral drugs, as well as correlations between insulin levels and viral load, CD4 cell count, lipid and endocrine profiles, and adipose tissue areas of the patients were also evaluated.

Patients and methods


The LIPOCO study is a French observational cohort study of risk factors for the development of changes in body fat distribution, carried out in unselected HIV-infected patients either therapy-naive or undergoing antiretroviral bi- or multi-therapy. This ongoing study has a prospective design with follow-up visits scheduled at 6-month intervals. Enrolment started in July 1998 and ended in June 1999. The group of patients enrolled during this period is defined as the LIPOCO-1 cohort. Participants recruited at five AIDS clinical trial units met the following unrestrictive eligibility criteria: age of more than 18 years, biologically documented HIV-1 infection (positive HIV-antibody test), and active follow-up or medical management of their disease over the previous 12 months. The study protocol was approved by the Institutional Ethics Committee of the University of Lyon, and informed consent was obtained from all patients.

All male patients participating in the LIPOCO study as of February 15, 1999, were eligible for this preliminary analysis of data collected at baseline.

History taking and clinical evaluation of fat distribution patterns

Patients were evaluated as ambulatory outpatients. Information was collected from patients' charts and during patient interviews performed at enrolment and recorded on a standardized data collection form. The demographic characteristics recorded were date of birth, race, country of origin, and transmission category. Prior and current use of antiretroviral drugs were recorded in terms of date of initiation, date of permanent discontinuation, and whether the patient was taking the drug at the time of enrolment. Dates of the last clinical visit, diagnoses of opportunistic infections and malignancies were also recorded. AIDS-defining events were determined according to criteria issued in 1993 by the Centers for Disease Control [17] All forms were checked by the scientific staff at the LIPOCO steering committee for logical error.

Investigators were asked to assess each patient's fat distribution pattern, using information obtained from history taking and 10 preselected clinical items (Table 1). On the basis of this clinical evaluation, investigators subjectively divided patients into four groups: (1) a group of purely lipoatrophic patients with symptoms of fat wasting only (LA group); (2) a group of patients with isolated symptoms of fat accumulation (obese group); (3) a group of patients with mixed symptoms of fat wasting and fat accumulation (mixed group); and (4) a group of patients without any changes in body fat distribution (control group). Lipoatrophy (fat wasting) was identified by patient self-report of relevant changes in their face (loss of Bichat and/or preauricular fat pads), arms, legs and subcutaneous abdominal tissue. This was subsequently confirmed qualitatively by physical examination. Fat accumulation was defined as a progressive increase in either abdominal girth, breast size or dorsocervical fat pads, or supraclavicular fat accumulation.

Table 1
Table 1:
Items used for clinical evaluation of lipodystrophy.

Anthropometric and body composition measurements

Patients wore light indoor clothes and no shoes when anthropometric measurements were performed, including body weight (kg) and height (cm). In addition, weight data recorded at initiation of the current antiretroviral regimen were retrospectively obtained from medical files. Body mass index (BMI) was calculated (weight in kilograms divided by the square of the height in metres). Mid-arm circumference was measured (to the nearest millimetre) using a tape measure, mid-way between the acromial process of the shoulder and the elbow tip, with the right arm hanging relaxed at the side.

Body composition was measured by tetrapolar bioelectrical impedance analysis (BIA, RJL-Akern BIA 109, Data-Input; Frankfurt, FRG) in supine position as described earlier [18] This method allows calculation of the following indexes of body composition: body fat, fat-free mass and extracellular water. Fat expressed as absolute mass and as percentage of body weight was calculated.

Body fat composition was also estimated using caliper measurements (Lange skinfold caliper; Cambridge Scientific Industries, Cambridge, Massachusetts, USA) of skinfold thickness at four sites (scapular, bicipital, tricipital and suprailiac areas) taken at the right side. All measurements were made in triplicate and averaged.

Percentage of body fat was calculated by the method of Durnin and Womersley [19] Most measurements of body composition were made by the same dietitian.

Regional fat distribution was measured by computed tomography (CT) performed at two body levels to measure abdominal and mid-thigh subcutaneous fat areas as well as intra-abdominal fat area. Patients were imaged on a SR 7000 CT-scanner (Philips Medical Systems, Best, The Netherlands). An unenhanced helical abdominal acquisition was obtained with the following parameters; Kvp 120, 1As 250, collimation 10 mm, table speed 10 mm/s, gantry rotation period 1 second (pitch I : I), 10–20 rotations. Imaging data were reconstructed every 10 mm. The slice passing through the umbilicus (fourth lumbar vertebra), reported as a valid predictor of total abdominal fat in men and women [20] and that passing through the middle of the thigh were transferred to an Easyvision (Philips Medical Systems, Best, The Netherlands) workstation. The areas of intra-abdominal, i.e. visceral adipose tissue (VAT) and abdominal subcutaneous adipose tissue (SAT), expressed in square centimetres, were calculated by summing the area of pixels in the slice with CT values from −150 to −50 hounsfiels units, according to the method described by Miller et al. (5). The ratio between the visceral and total abdominal adipose tissue (TAT) was also calculated. The radiologists who performed the CT fat measurements were kept blinded to the patient treatment details.

Laboratory analyses

All laboratory investigations were performed after a 10 h overnight fast and at least 15 min after the placement of the peripheral intravenous catheter.

Serum total cholesterol and triglycerides, measured by enzymatic methods with the same automated analyzer, serum high-density lipoprotein cholesterol, isolated by chemical precipitation, serum low-density lipoprotein cholesterol, apoliproprotein A1 and B, uric acid, and free fatty acids were measured on fasting venous blood.

Basal morning evaluation of gonadal and adrenal functions included the measurement of plasma testosterone, follicle stimulating hormone (FSH), luteinizing hormone (LH) and plasma cortisol concentrations.

CD4 cell count (cells × 106/l) and quantitative HIV-1 RNA (viral load; log10 copies/ml) were also measured on venous blood.

An oral glucose tolerance test (OGTT) was used to assess glucose tolerance and to measure the insulin response to oral glucose. Patients were given a 75 g oral glucose load in 250 ml water. The glucose solution was consumed within a 3-min period. Venous blood samples were taken 15 min before the glucose was administered, at the time of administration T0 and 30, 60, 90 and 120 min thereafter for determination of plasma glucose and insulin concentrations. The glucose tolerance status of each patient was categorized according to the criteria of the World Health Organization [21]

Nutritional intake assessment

All patients completed a self-administered questionnaire for the 3 days preceding study entry. For each meal, patients were asked to report full details of everything they had eaten or drunk, and the time of meal and drink intake. For each food item recorded, they were asked to describe the mode of preparation (i.e. raw, fried, boiled), list all major ingredients, and record amounts in ounces, tablespoons or units. The macro- and micronutitional contents of each diet were calculated by using FoodPerfect, an interactive nutritional analysis computer program based on the Canadian Nutrient File database [22] Total caloric, protein, fat, carbohydrate and cholesterol intakes were evaluated.

Statistical analysis

Results of investigations and laboratory data were compared: (1) between the four clinical categories used by investigators to classify patients according to their fat distribution clinical characteristics; and (2) between patients receiving a zidovudine-containing regimen (ZDV-experienced patients) and patients receiving a stavudine-containing regimen (d4T-experienced patients). All statistical analyses were performed with the use of programs of the SAS Institute (Statistical Analysis System, Cary, North Carolina, USA). Continuous variables were analysed by analysis of variance methods. A P-value of < 0.05 was determined to be statistically significant. Logistic regression was used to establish multivariate models to assess odds ratio (OR) of developing changes in body fat distribution, lipoatrophy and mixed syndromes according to antiretroviral drugs. Adjustment was made for gender, age, HIV-1 RNA level, CDC staging, global time of exposure to antiretroviral therapy and duration on current antiretroviral therapy. Associations between two variables were quantified using Pearson product–moment correlation coefficients.


Study population

As of February 15, 1999, a total of 214 patients (154 men eligible for the present analysis and 60 women) had been recruited in the LIPOCO cohort. Mean age (± SD) of the 154 male participants was 40.24 ± 0.67 years. Fifteen (9.7%) were therapy-naive, 39 (25.3%) on long-term NRTI therapy (NRTI group) for a mean time of exposure of 28.25 ± 3.04 months, and 100 (65%) on antiretroviral multiple combination therapy including at least one protease inhibitor (PI group) for a mean time of exposure of 21.31 ± 2.71 months. Eighty four (54.6%) patients were taking one PI and 16 (10.4%) were taking two PIs.

Table 2 shows the HIV-infection characteristics and antiretroviral treatment duration of the 154 patients included in the analysis. Homo/bisexual contact remained the most common transmission category. Forty three (27.9%) patients had been diagnosed with AIDS at the time of enrolment in the study.

Table 2
Table 2:
Diseases status and antiretroviral treatment characteristics of the study population.

Mean baseline plasma HIV-RNA levels and CD4 cell counts were 3.15 ± 0.11 log10 copies/ml and 429 ± 17.62 × 106 cells/l, respectively.

Of the 39 patients on NRTI therapy, 17 were currently receiving a zidovudine-containing regimen (ZDV-experienced patients) for a mean time of exposure of 21.17 ± 2.3 months [15 (9.7%) were taking zidovudine plus didanosine and two (1.3%) were taking zidovudine plus zalcitabine] and 22 a stavudine-containing regimen (d4T-experienced patients) for a mean of 12.81 ± 1.59 months [10 (6.5%) were taking stavudine plus didanosine and 12 (7.8%) were taking stavudine plus lamivudine]. The mean time of exposure to current treatment was significantly higher in ZDV-experienced patients as compared with d4T-experienced patients (P = 0.004). Fifteen (68.1%) patients among the 22 d4T-experienced patients had been pretreated prior to receiving stavudine and two patients among the 17 ZDV-experienced patients had been pretreated with zidovudine monotherapy. The mean duration of a preceding treatment with NRTIs (especially zidovudine) was significantly higher in d4T-experienced patients than in ZDV-experienced patients (18.86 ± 4.21 months versus 2.64 ± 1.81 months, P = 0.0028). The mean time of exposure to NRTIs was higher in d4T-experienced patients as compared with ZDV-experienced patients, but the difference did not reach statistical significance (31.68 ± 4.71 months versus 23.82 ± 3.26 months, P = 0.2).

There were 29 ZDV-experienced patients and 71 d4T-experienced patients among the 100 patients under multiple combination therapy for a mean time of exposure to current treatment of 11.32 ± 1.7 months and 10.88 ± 1.02 months, respectively [seven (4.5%) were taking zidovudine plus didanosine, two (1.3%) were taking zidovudine plus zalcitabine, 18 (11.7%) were taking zidovudine plus lamivudine and 21 (13.60%) were taking stavudine plus didanosine, one (0.65%) was taking stavudine plus zalcitabine, 40 (26%) were taking stavudine plus lamivudine]. The mean duration of a preceding treatment with NRTIs (especially ziduvodine) did not differ significantly among d4T and ZDV-experienced patients (42.94 ± 3.94 months versus 39.10 ± 6.60 months). Fifty-six (78.9%) patients had received zidovudine therapy before starting stavudine. Of the 100 patients on at least one PI, 41 were receiving indinavir, 23 receiving ritonavir, eight receiving saquinavir, 12 receiving nelfinavir, five receiving ritonavir + indinavir, five receiving ritonavir + saquinavir, six receiving saquinavir + nelfinavir. In these PI-experienced patients, the mean time of exposure to antiretroviral therapy did not differ significantly between d4T- and ZDV-experienced patients (53.83 ± 6.62 months versus 50.4 ± 6.62 months).

Although the mean duration of total antiretroviral therapy was higher in d4T-experienced patients than in ZDV-experienced patients among the 139 treated patients (NRTI and PI-experienced patients), the difference was of borderline statistical significance (49.19 ± 3.1 months versus 40.51 ± 4.41 months, P = 0.071).

Analysis of results by type of therapy

In terms of metabolic profile, although plasma glucose and fasting insulin levels were similar among the three groups, glucose-stimulated plasma insulin and fasting C-peptide levels were significantly higher in the PI group (Table 3). The PI-experienced patients were also hyperlipidaemic, with higher total cholesterol, low-density lipoprotein, triglyceride and free fatty acids (FFA) concentrations (P < 0.01) compared with the naive or NRTI groups. High-density lipoprotein concentrations were significantly lower in the PI group compared with naive and NRTI groups (P < 0.05), but did not differ between the naive and NRTI groups. Testosterone levels did not differ significantly among the groups whereas FSH levels were significantly higher in the PI group (P < 0.05) and LH levels in the therapy naive group (P < 0.01). Cortisol levels were significantly higher in the PI group compared with the NRTI and therapy-naive groups (P < 0.01), but did not differ between naive and NRTI groups.

Table 3
Table 3:
Metabolic and fat distribution parameters in patients grouped by antiretroviral therapy.

Although the three groups were similar in terms of body mass index, waist-to-hip ratio (WHR) and fat-free mass, both treated groups had, overall, a significantly lower fat mass in comparison with therapy-naive patients. This decrease was significantly more pronounced in patients treated with PIs.

With regard to regional fat distribution evaluated by CT scan, patients in both treated groups had significantly lower abdominal (SAT) and mid-thigh subcutaneous adipose tissue areas than the untreated patients. Visceral adipose tissue (VAT) was significantly increased in PI-experienced patients compared with the other groups, this resulting in abdominal symptoms such as distension, fullness, and bloating, whereas values were similar between the therapy-naive and NRTI-experienced patients. The differences in fat distribution between the treatment groups were more evident when the SAT : VAT and VAT : TAT ratios were calculated.

Analysis of results by clinical patterns of body fat distribution

A syndrome of body fat distribution alteration was observed in 82 (53.25%) patients whereas 72 (46.75%) were subjectively considered as clinically normal (control group). Investigators classified 34 (15.89%) patients in the lipoatrophy group (12 and 22 in the NRTI and PI groups, respectively), nine (4.21%) in the obesity group (PI group), 39 (18.22%) in the mixed group (three and 36 in the NRTI and PI groups, respectively).

Tables 4 and 5 compare the metabolic and endocrine parameters measured in each clinical category of fat distribution used by investigators to classify the patients.

Table 4
Table 4:
Metabolic and fat distribution parameters, by clinical fat distribution category in patients undergoing nucleoside reverse transcriptase inhibitor (NRTI) therapy (n = 39).
Table 5
Table 5:
Metabolic and fat distribution parameters, by clinical fat distribution category in patients undergoing multiple combination therapy (n = 100).

Fasting triglyceride levels were higher in the LA group than in all other groups (P < 0.005 and P < 0.01 versus NRTI and PI groups, respectively). Higher levels of insulin (T0, T120) and C-peptide (T0) were observed in obese and mixed patients, as compared to patients of the LA group and controls, but these parameters did not differ between the LA and control groups.

This difference could be clearly shown when the sums of the insulin concentrations following the OGTT were compared {sum [insulin] (Σ[insulin]) = [insulin] at T0 + [insulin] at T60 + [insulin] at T120. However, no differences in plasma glucose levels either in fasting conditions or in response to OGTT were observed between the fat distribution groups. Patients in the mixed and obese groups had significantly higher free fatty acid levels, although there was no difference between the LA and control groups. Testosterone, FSH, LH and cortisol levels did not differ between the fat distribution groups in both NRTI and PI groups.

Results of the body composition and fat distribution measurements in each category of clinical fat distribution pattern are shown in Tables 4 and 5.

Overall fat mass, mid-arm circumference, biceps, triceps and supra-iliac skinfold thickness were all lower in the lipoatrophy group compared with the mixed, obese and control groups, although the waist-to-hip ratio was similar across all groups. Purely lipoatrophic patients had lower subcutaneous abdominal and mid-thigh fat area, compared to all other groups.

Intra-abdominal fat area values were significantly higher in the mixed and obese patients as compared with the LA and control patients. Although the difference was not significant, there was also a trend in the pure LA group on PI therapy towards greater VAT levels.

Patients in the mixed group were then grouped based on whether they were receiving ZDV or d4T as part of their combination therapy (Table 6). Metabolic parameters did not differ significantly between ZDV and d4T-experienced patients although a trend was observed towards increased triglyceride levels in the d4T group. Analysis of fat distribution revealed significant decreases in percentage of body fat mass and mid-thigh subcutaneous adipose tissue for d4T-experienced patients. A trend towards decreased subcutaneous abdominal fat was also observed for d4T treatment versus ZDV, although the difference was not significant. However, although TAT was similar between the two groups, SAT : VAT was significantly lower in d4T-experienced mixed patients as compared with ZDV-experienced mixed patients (P < 0.001).

Table 6
Table 6:
Metabolic and fat distribution parameters values measured in ZDV-experienced patients and d4T-experienced patients undergoing multiple combination therapy.

Multivariate analysis

Multivariate analysis (Table 6) demonstrated that the odds ratio for developing body fat distribution abnormalities was significantly correlated with d4T-containing regimens compared with ZDV-containing regimens in both NRTI and PI groups: OR, 85.3 [95% confidence interval (CI), 3.6–999;P = 0.0058 and OR, 4.01 (95% CI, 1.2–12.7;P = 0.018) for d4T-experienced patients in NRTI and PI groups, respectively, versus OR, 0.012 (95% CI, 0.04–0.27;P = 0.0058) and OR, 0.25 (95% CI, 0.078–0.79;P = 0.019) for ZDV-experienced patients.

Alterations in body fat did not significantly correlate with lamivudine or didanosine use, in both NRTI and PI groups.

The use of PIs did not correlate with altered body fat distribution (OR, 0.34; 95% CI, 0.16–0.70;P = 0.003). When considering the use of each PI (ritonavir, indinavir, saquinavir and nelfinavir), no statistical correlations were found with body fat distribution alterations (Table 7). Differences observed in each clinical category according to antiretroviral drugs remained highly significant in both NRTI and PI groups, although the differences were less pronounced in those patients receiving a combination therapy including at least one PI.

Table 7
Table 7:
Odds ratio (95% confidence intervals) of developing changes in body fat distribution, lipoatrophy and mixed syndromes according to antiretroviral regimen.

Correlation coefficients

Correlation coefficients between insulin levels evaluated during the OGTT, and adipose tissue areas measured by computed tomography at abdominal and mid-thigh levels, BMI, WHR, triglyceride, total and high-density lipoprotein cholesterol, FFA, cortisol and HIV RNA levels, CD4 cell counts, in both NRTI and PI groups, are shown in Table 8.

Table 8
Table 8:
Correlation coefficients between Insulin levels (T0, T120, Σ) and viral load, CD4 cell count, triglyceride levels and adipose tissue areas in both nucleoside reverse transcrptase inhibitor (NRTI) and protease inhibitor (PI) groups.

No associations were found between CD4 cell count, HIV RNA level, ΔCD4 cell count and fasting, 2 h and Σ insulin concentrations in both NRTI and PI groups, whereas the ΔHIV RNA was negatively correlated with fasting insulin (r = −0.29, P = 0.041) and Σ insulin concentrations (r = − 0.24, P = 0.044) in PI group. The BMI and WHR did not show significant associations with concentrations of insulin (T0, T120, Σ) in both NRTI and PI groups. However, BMI was positively associated with fasting insulin levels (r = 0.31, P = 0.0017) in the PI group whereas WHR was positively associated with fasting insulin levels in the NRTI group (r = 0.88, P = 0.0038).

SAT, VAT, TAT, SAT : VAT ratio, and triglyceride levels were correlated with fasting, 2 h and Σ insulin concentrations in the PI group, while VAT, TAT, and triglyceride levels were correlated with the insulin levels (T0, T120, Σ) in the NRTI group.


The term lipodystrophy syndrome has been proposed to label alterations in fat distribution observed in HIV-infected patients treated with various antiretroviral combination drug regimens.

First reports of this syndrome described a combination of peripheral lipodystrophy (fat wasting), central adiposity, elevated triglyceride levels and insulin resistance in patients treated with a PI [7,23] More recently, fat distribution abnormalitites were reported in patients treated with other antiretroviral drugs (NRTIs and non-NRTIs), suggesting that factors other than PIs may contribute to the syndrome [11–13,24,25]

A difficulty in understanding the lipodystrophy syndrome has been that patients affected may not present with all of its originally described features; in particular, in patients suffering from proven limb and facial fat wasting during long-term NRTI therapy [11,24] the metabolic changes commonly seen with PI therapy, were not apparent. Moreover, all reports describing fat distribution disorders were published after the introduction of PIs in clinical practice, which resulted in an almost automatic tendency to associate them with this class of drugs.

In addition, it should be kept in mind that stavudine was available 6 months before PIs and that in September 1997, stavudine was the most commonly used drug in triple (54%) and quadruple therapy (73%) as demonstrated by Kirk et al. [26] This made it difficult to interpret reported data and to assess attributability of fat distribution disorders to the antiretroviral drugs in use.

For these reasons, a classification of the various pictures of fat distribution abnormalities observed in HIV-infected patients under active antiretroviral therapy is clearly needed. The LIPOCO study was undertaken to better characterize and differentiate such fat distribution disorders and to try to establish a case definition of lipodystrophy syndromes.

Our study demonstrates that the metabolic and physical changes previously grouped under the general term of lipodystrophy may be further divided into distinct clinical syndromes.

In contrast to the study reported by Carr et al. [7] showing a syndrome of peripheral fat wasting associated with insulin resistance, our data demonstrate that a great majority of our patients with pure lipoatrophy had fasting and glucose-stimulated plasma glucose and insulin levels as well as visceral adipose tissue areas similar to those of therapy-naive patients or non-lipodystrophic patients in both NRTI and PI groups.

However, in four patients on triple therapy initially classified as purely lipoatrophic (using subjective criteria), CT scan measurements demonstrated increases in VAT greater than 180 cm2, resulting in a SAT : VAT ratio of less than 0.5, similar to ratios seen in the mixed patients. Insulin and C-peptide levels in these four patients were also increased, again mirroring the changes seen in the mixed group. This suggests that these patients should be reclassified from the pure lipoatrophy group into the mixed group. Thus, CT analysis of VAT may distinguish those patients with lipoatrophy alone and insulin secretion within the normal range from those with both lipoatrophy and insulin resistance (mixed patients).

Furthermore, when patients on PI therapy subjectively defined as normal and obese patients were analysed according to the degree of subcutaneous fat accumulation, CT scan evaluation revealed a group of patients with high abdominal subcutaneous fat accumulation, defined by SAT : VAT ratio of more than 1.5 (normal range, 0.9–1.3) (five out of the 20 clinically normal patients and two out of the nine obese patients) [20] In these seven patients, both fasting and post-OGTT insulin levels did not differ from those observed in non-lipodystrophic patients. Our results showed that the changes in subcutaneous fat in these patients were the result of increased caloric intake (data not shown) and might simply reflect improved health because of HIV-1 suppression in patients on HAART. Therefore, we suggest that an appropriate third group (subcutaneous fat adiposity) should be defined.

The remaining of obese patients had the same clinical and biologic profile as the mixed patients (decreases in abdominal and mid-thigh SAT, increases in VAT, hyperinsulinaemia, elevated FFA and triglyceride levels) suggesting they would be best classified in the mixed group.

These examples confirm that subjective assessment by physical examination is not sufficient to identify each syndrome, and illustrate the difficulties inherent in the classification of such patients if CT scan measurements are not available. In addition, it was noted that there was wide variability in investigator opinion regarding the subjective assessment of changes in body fat.

Finally, in a study reported by Gharakhanian et al.[27] evaluating the frequency of lipodystrophy assessed subjectively and factors associated with glucose and lipid abnormalities, the high frequency of diabetes and glucose intolerance observed in patients defined as purely lipoatrophic was probably due to the mixing of patients who were really lipoatrophic and those with fat redistribution (mixed patients).

The primary differences in regional fat distribution measured by CT scan were related to the following parameters: between LA and mixed patients, visceral fat area; between mixed patients and non-lipodystrophic patients, visceral fat area and abdominal and mid-thigh subcutaneous adipose tissue – the latter being particularly decreased in patients currently undergoing stavudine therapy; and between LA and non-lipodystrophic patients, abdominal and mid-thigh subcutaneous fat areas.

From a metabolic point of view, the clinical categories of fat distribution disorders could be distinguished on the basis of plasma triglyceride levels, which were higher in the LA group, and plasma insulin and C-peptide levels as well as FFA levels, which were higher in the mixed and the subjectively named obese group. In the two latter groups, hyperinsulinaemia was not associated with a significant increase in blood glucose levels.

Although preliminary, our results suggest that the lipodystrophy syndrome should be subdivided into three distinct entities which may, or may not, coexist in the same patient.

In their pure varieties, the three conditions give distinctive clinical presentations and may be defined as follows: (i) the syndrome of fat depletion or `lipoatrophy syndrome' refers to the condition characterized by a combination of subcutaneous fat loss, especially affecting the limbs, face and buttocks, and deep fat loss visible at the level of `Bichat', preauricular or orbital fat pads, resulting in pain elicited by mastication, sunken eyeballs and retro-orbital pain; in the present analysis, this condition was associated with an increase in triglyceride levels – probably reflecting the destruction of adipocytes, perhaps via adipocyte apoptosis [24,25] – and insulin levels within the normal range; (ii) the mixed syndrome or syndrome of fat redistribution is characterized by a redistribution of fat, including peripheral fat loss (limbs, buttocks, face, but no deep fat loss) and visceral fat adiposity; in our patients, this condition was associated with the metabolic and fat distribution characteristics of incepting insulin resistance (similar to those seen in type II diabetes): raised insulin and C-peptide concentrations, and elevated free fatty acid levels; (iii) the syndrome of subcutaneous adiposity, the magnitude of which is normally less than that found in classical obesity, with no increase in visceral adipose tissue area and no insulin resistance. Fat accumulation at the level of dorsocervical region (`buffalo-hump') and breast (mainly in female patients) may occur either in patients presenting with the mixed syndrome or in patients experiencing only subcutaneous adiposity independently of metabolic changes, whereas Madelung's disease is generally associated with a syndrome of insulin resistance.

The pathogenesis of these various pictures of fat distribution abnormalities remains unexplained. In particular, whether such conditions are correlated with the type of therapy received is unclear. In the present study, the results of multiple regression analysis showed that the use of d4T significantly correlated with body fat alterations and in particular with the syndrome of fat wasting (lipoatrophy syndrome) in both NRTI- and PI-experienced patients. Our findings are consistent with those recently reported by Gervasoni et al. in HIV-infected women undergoing combined antiretroviral therapy [12] and by Polo et al.[13] In addition, among our mixed-group patients, there was a trend toward lower abdominal and mid-thigh subcutaneous areas in d4T-experienced patients in comparison with ZDV-experienced patients. Controlled studies are currently under way to address this important issue.

No associations were observed between the use of protease inhibitors and body fat distribution abnormalities, suggesting that the mechanisms responsible for these pathological states are largely independent of the action exerted by PIs, a result also noted by other investigators [8,12]

There were no significant association between insulin levels and CD4 cell count, ΔCD4 cell count or viral load. However, insulin levels were negatively associated with the magnitude of viral load decline, suggesting that the mixed syndrome or a syndrome of fat redistribution might be a manifestation of chronic HIV infection exacerbated by the successful effect of highly active antiretroviral therapy (HAART).

Before the era of HAART, insulin clearance and sensitivity of peripheral tissues to insulin were reported to be increased in HIV-infected patients [28] in particular in those with AIDS-defining symptoms.

Recently, Walli et al. [29] have demonstrated that treatment with PIs was associated with peripheral insulin resistance, although a trend towards decreased insulin sensitivity was also seen in patients treated with NRTIs alone. Thus, the mixed syndrome combining visceral fat adiposity and peripheral fat loss might be the consequence of hyperinsulinaemia, the latter being itself secondary to the viral load decline obtained in HIV-infected patients after HAART was introduced.

The recent observations of body fat distribution abnormalities including increased abdominal girth and wasting of arms and legs (i.e., a mixed syndrome) in patients naive to HIV protease inhibitors and receiving a triple combination therapy including two NRTIs and one NNRTI or three NRTIs are consistent with this hypothesis [30]

In addition, the absence of improvements in insulin levels in patients switching from PI to a NNRTI combination therapy and maintaining undetectable viral load [31] reinforces this speculation.

A few endocrine alterations associated with the various pictures of fat distribution abnormalities have been reported in the literature, but results are controversial. In particular, decreases in testosterone levels have been reported in men, but this has not yet been formally demonstrated; in the present study, no differences were seen between each category of clinical fat distribution in both NRTI and PI groups, although a great majority of patients complained of a decreased libido.

In conclusion, many critical issues related to the nature and mechanisms of fat distribution disorders in the HIV-infected population remain to be addressed before a full comprehension of such disorders can be achieved. Our preliminary results suggest a classification that needs confirmation by results of the longitudinal follow-up of the LIPOCO cohort but might have important implications for the management of the disease. Both hypertriglyceridaemia and insulin resistance are major risk factors for cardiovascular events and their associated morbidity and mortality. Such risk factors and their potential relations with antiretroviral drugs deserve consideration in the current context of long-term, highly active antiretroviral therapy for HIV-infected patients.


We thank the patients who participated in the study and R. Avellaneda for assistance with the data collection.


1. 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 of less. N Engl J Med 1997 337725–733.
2. Carpenter CQJ Fischl MA Hammer SM et al Antiretroviral therapy for HIV infection in 1997.:Updated recommendations of the international AIDS society-USA panel. JAMA 1997 2771962–1969.
3. Lo JC, Mulligan K, Tai VW, Algren H, Schambelan M. Buffalo hump in HIV-infected patients on antiretroviral therapy. 5th Conference on Retroviruses and Opportunistic Infections. Chicago–Illinois 1998 [abstract 409].
4. Ruane PJ. Atypical accumulation of fatty tissue. 37th Intersciences Conference on Antimicrobial Agents and Chemotherapy. Toronto 1997, I:85.
5. Miller K Jones E Yanovsky J Shankar R Feuerstein I Falloon J Visceral abdominal-fat accumulation associated with use of indinavir. Lancet 1998 351871–875.
6. Hengel RL Watts NB Lennox JL Benign symmetric lipomatosis associated with protease inhibitors. Lancet 1997 3501596.1596.
7. Carr A Samaras K Burton S et al A syndrome of peripheral lipodystrophy, hyperlipidaemia and insulin resistance in patients receiving HIV protease inhibitors. AIDS 1998 12F51–F58.
8. Kotler DP Rosenbaum K Wang J Pierson R Studies of body composition and fat distribution in HIV-infected and control subjects. J Acquir Immune Defic Syndr 1999 20228–237.
9. Grunfeld C Pang M Doerrler W Shigenaga JK Jensen P Feingold KR Lipids, lipoproteins, triglyceride clearance, and cytokines in human immunodeficiency virus infection and the acquired immunodeficiency syndrome. J Clin Endocrinal Metab 1992 741045–1052.
10. Carr A Samaras K Chisholm DJ et al Pathogenesis of HIV-1 protease inhibitor-associated peripheral lipodystrophy, hyperlipidaemia, and insulin resistance. Lancet 1998 3511881–1883.
11. Saint-Marc T, Poizot-Martin I, Partisani M, Fabre J, Touraine JL. A syndrome of lipodystrophy in patients receiving a stable nucleoside-analogue therapy. 6th conference on Retroviruses and Opportunistic Infections. Chicago, January 1999 [abstract 653].
12. Gervasoni C Ridolfo AL Trifiro G et al Redistribution of body fat in HIV-infected women undergoing combined antiretroviral therapy. AIDS 1999 13465–471.
13. Polo R, Verdejo J, Gomez Cano M, et al. Lipodystrophy related to NRTI inhibitors in HAART therapy.3rd Conference on Nutrition and HIV Infection. Cannes, April 1999 [abstract P10].
14. Pettit R, Muurahainen N, Falutz J, et al. HIV-related truncal obesity: a comparison of physician and patient diagnoses in the SALSA (Self-Ascertained Lipodystrophy Syndrome Assessment) questionnaire. 6th Conference on Retroviruses and Opportunistic Infections, Chicago, January 1999 [abstract 651].
15. Carr A, Samaras K, Thorisdottir A, Kaufmann GR, Chisholm DJ, Cooper DA. Diagnosis and prediction of HIV protease inhibitor (PI)-induced lipodystrophy and impaired glucose tolerance.6th Conference on Retroviruses and Opportunistic Infections, Chicago, January 1999 [abstract 641].
16. Romeu J, Bonjoch A, Arno A, et al. Lipodystrophy index by anthropometric parameters (LIAP): an approach to the diagnosis of lipodystrophy by measurable parameters in patients taking protease inhibitors. 6th Conference on Retroviruses and Opportunistic Infection, Chicago, January 1999 [abstract 652].
17. Centers for Disease Control 1993 Revision of the CDC surveillance case definition for acquired immunodeficiency syndrome. MMWR 1992 41 (RR-17)1–19.
18. Boulier A Fricker J Thomasset AL et al Fat-free mass estimation by the two-electrode impedance method. Am J Clin Nutr 1990 52581–585.
19. Durnin JGVA Womersley J Body fat assessed from total body density and its estimation from skinfold thickness: measurements on 481 men and women aged from 16 to 72 years. Br J Nutr 1974 3777–97.
20. Enzi G Gasparo M Biondetti PR Fiore D Semisa M Zurlo F Subcutaneous and visceral fat distribution according to sex, age, and overweight, evaluated by computed tomography. Am J Clin Nutr 1986 44739–746
21. World Health Organization Diabetes mellitus: report of a WHO study group. World Health Organ Tech Rep Ser 1985 7279–17.
22. Intelligent Software Engineering. FoodPerfect. Nutritional analysis software. Victoria: Intelligent Software Engineering.
23. Viraben R Aquilina C Indinavir-associated lipodystrophy. AIDS 1998 12F37–F39.
24. Saint-Marc T Partisani M Poizot-Martin I et al A syndrome of peripheral fat wasting (lipodystrophy) in patients receiving long-term nucleoside-analogue (NRTI) therapy. AIDS 1999 131659–1667.
25. Saint-Marc T Touraine JL The effects of discontinuing stavudine therapy on clinical and metabolic abnormalities in patients suffering from lipodystrophy. AIDS 1999 132188–2189.
26. Kirk O Mocroft A Katzenstein TL et al Changes in use of antiretroviral therapy in regions of Europe over time. AIDS 1998 122031–2039.
27. Gharakhanian S, Salhi Y, Nguyen TH, et al. Frequency of lipodystrophy and factors associated with glucose/lipid abnormalities in a cohort of 650 patients treated by protease inhibitors. 6th Conference on Retroviruses and Opportunistic Infections, Chicago, January 1999 [abstract 642].
28. Hommes MJT, Romijn JA, Endert E, Karel J, Schattenkerk ME, Sauerwein HP. Insulin sensitivity and insulin clearance in human immunodeficiency virus-infected men.Metabolism 1991, 40, 6:651–656
29. Walli K, Michl GM, Segerer S, et al. Dyslipidemia and insulin resistance in HIV-infected patients treated with reverse transcriptase inhibitors alone and in combination with protease inhibitors. 6th Conference on Retroviruses and Opportunistic Infection, Chicago, January 1999 [abstract 645].
30. Madge S Kinloch-de-Loes S Mercey D Johnson MA Weller JVD Lipodystrophy in patients naive to HIV protease inhibitors. AIDS 1999 13735–737.
31. Carr A, Thorisdottir A, Samaras K, Kaufmann GR, Chisholm DJ, Cooper DA. Reversibility of protease inhibitor (PI) lipodystrophy syndrome on stopping PIs or switching to nelfinavir.6th Conference on Retroviruses and Opportunistic Infection, Chicago, January 1999 [abstract 668].


Multicentre Study Group on LIPOCO Bordeaux: N. Bernard, D. Malvy, Saint-André Hospital; Lyon: T. Saint-Marc, R. Avellaneda, Edouard Herriot Hospital; Marseille: I. Poizot-Martin, J. Fabre-Monges, Sainte-Marguerite Hospital; Strasbourg: M. Partisani, Strasbourg Hospital.


HIV; antiretroviral therapy; lipodystrophy

© 2000 Lippincott Williams & Wilkins, Inc.