Introduction
The development of combined antiretroviral (ARV) therapies represents a significant advance in the treatment of HIV infection [1-5]. Nevertheless, untoward effects of different degrees of severity have been reported [4-6], including various conditions characterized by fat tissue abnormalities [7-16]: the accumulation of adipose tissue in the dorsocervical region, commonly referred to as ‚buffalo hump‚ [10]; visceral abdominal-fat accumulation [11]; benign symmetric lipomatosis [7]; and fat wasting of the face and limbs with relative central adiposity, commonly referred to as ‚lipodystrophy‚ [12,13]. In the majority of cases, these abnormalities were observed in male patients, and only a few reports specifically regard women [9,14,16].
During 1997, a number of women patients being stably treated with combined ARV therapy at our Institute complained of a rapid and unexplained change in bodily appearance, characterized by greater breast size and abdominal girth, and accompanied by wasting of the glutei and lower limbs.
This study was designed to describe this syndrome, evaluate its prevalence in a cohort of 306 HIV-positive women receiving combined ARV therapy, and to assess the possible risk factors associated with its development.
Patients and methods
Study design
All the women consecutively seen at the outpatient clinic of the University of Milan‚s Institute of Infectious Disease and Tropical Medicine between December 1997 and February 1998 were evaluated, and those who had been observed for at least 1 year and were receiving a stable combined ARV treatment at the time of the clinic visit were included in the study. Stable treatment was defined as a regimen that had been administered for at least 3 months without any substitutions, dose changes or discontinuations. Patients who took less than 80% of the prescribed drugs or discontinued treatment for more than 36h were excluded from the study. A recording was made of the demographic and clinical data of all of the patients, including age, risk factors for HIV infection, stage of HIV infection (Centers for Disease Control and Prevention (CDC) classification), the type and duration of ARV therapy, and CD4 cell counts and HIV-RNA levels at the beginning of the stable ARV regimen and at the time of the last observation.
The occurrence of a fat redistribution (FR) syndrome was suspected on the basis of patient reports of progressive increase in abdominal girth and breast size accompanied by a wasting of the glutei, calves and thighs. The diagnosis of FR was confirmed by means of a physical examination (during which any changes in bodily appearance were compared with previous observations). The patients with FR also underwent dual-energy X-ray absorptiometry (DEXA) in order to measure the percentages of body fat and lean body mass (LBM). The examination was carried out using a Hologic QDR-2000 X-Ray Bone Densitomer (Waltham, MA, USA), with body composition being estimated using BCA 7.20 software. Additional estimates were made by evaluating the relative proportions of body fat and LBM in the trunk region and legs.
The possible association of FR with metabolic or endocrinological abnormalities was investigated by comparing the women with FR with an equally sized control group of HIV-infected women without FR, matched for age, body mass index (BMI), CD4 cell counts, HIV-RNA levels, CDC stage, and the type and duration of ARV therapy. The same subjects were used as a control group for the DEXA examinations. Blood and urine samples were taken from the study subjects and matched controls between 08.00 and 10.00h (under fasting conditions) in order to determine the plasma concentrations of cortisol, adrenocorticotropic hormone (ACTH), growth hormone (GH), C-peptide, testosterone and prolactin, the serum concentrations of cholesterol, triglycerides and glucose, glucose tolerance afterload, CD4 cell counts and HIV-RNA levels. In the case of the patients with FR, 24h urine collections were made in order to measure cortisol, creatinine and urea nitrogen excretion.
The role of ARV drugs in the occurrence of FR was evaluated in all of the enrolled women by analysing the type and duration of drug use during and before the stable ARV regimen they were taking at the time of the study. A comparison was also made of the plasma viraemia levels and CD4+ cell counts in the women with and without FR at the beginning of their stable ARV regimen and at the time of the last study visit.
Statistical analyses
The data were statistically analysed using SPSS software (release 7.5, SPSS, Chicago, USA). The significance of the differences in continuous variables was evaluated using the Kruskall-Wallis non-parametric test. When comparing the history of ARV treatment in the women with and without FR, the strength of the association between FR and each single drug administered was estimated by calculating the odds ratios. Statistical significance was determined using estimated χ2 as the likelihood ratio; the 95% confidence intervals (CI) of the odds ratios were calculated using the test-based method.
A final stepwise logistic regression was carried out in order to identify the variables that were significantly and independently associated with FR.
Results
The cohort of study patients consisted of 306 HIV-infected women with a median age of 34 years (range 23-61): 108 were or had been intravenous drug abusers; 195 were infected as a result of heterosexual intercourse; the remaining three had acquired the infection in other ways. The HIV infection was classified as stage III in 173 patients, and stage IV in the remaining 133 (44 IVC2, 87 IVC1 and two IVD).
FR ( Fig. 1 and Fig. 2) was observed in 32 of the 306 women (10.5%), none of whom reported variations of more than 2kg in body weight during the previous 3 months. None of these patients had a history of alcohol abuse or the taking of appetite stimulants, anabolic hormones or systemic glucocorticoids. The body changes were reported as having gradually emerged over a period of 3-18 months.
The percentage of total body fat revealed by DEXA was similar in the FR subjects and controls, with median values of 27% (range 21-39%) and 31% (range 11-43%), respectively; however, as percentages of total fat, the former had significantly more trunk fat (P<0.01) and significantly less leg fat (P<0.001) (Table 1). Moreover, LBM was similar in the two groups (data not shown). The levels of C-peptide, prolactin, GH, testosterone, ACTH, cortisol, glucose and cholesterol did not differ between the two groups, and were normal in the large majority of cases (Table 1); triglyceride levels were higher than normal in both groups, without any significant difference between them. Glucose tolerance afterload was analysed in 15 women (eight cases and seven controls), and was found to be normal in all of them. Twenty-four hour free cortisol urinary excretion was evaluated only in the women with FR, and was normal in all cases (median 44nmol/24h: range 21-58; normal values 10-90).
The ARV regimens were combinations of two nucleoside RT inhibitors in 162 patients, and three or more drugs including a PI in 144; two patients were being treated with a regimen containing two PIs (ritonavir plus saquinavir). All of the 32 women who developed FR were being treated with lamivudine (3TC)- containing regimens, whereas this was not the case for any of the other ARV drugs: 3TC was being taken with zidovudine (ZDV) (n=4) or stavudine (d4T) (n=8), or in triple combinations including a PI (15 d4T+3TC+indinavir, three d4T+3TC+ritonavir, and two ZDV+3TC+saquinavir). FR significantly correlated with 3TC-including regimens (P=0.017) with an odds ratio of infinity as the reference category in this calculation (the prevalence of FR among the patients not taking 3TC) was equal to zero. The absence of FR in patients treated with didanosine (ddI)- or zalcitabine (ddC)-including combinations was also statistically significant (P=0.019 and P=0.049, respectively). The use of d4T significantly correlated with FR (OR 3.9; 95% CI 1.5-9.7; P=0.0014), but six of the patients with FR did not take this drug. On the contrary, the risk was significantly lower in patients taking combinations including ZDV (OR 0.3; 95% CI 0.1-0.7; P=0.002). The prevalence of FR was 7.4% (12 out of 162) in the patients treated with two nucleoside RT inhibitors and 13.9% (20 out of 144) in those receiving combinations including at least one PI. The crude relative risk of FR was higher in the patients receiving a PI, but this difference was of borderline statistical significance (OR 2.0; 95% CI 0.9-4.3; P=0.064); the relative risk was 1.7 (95% CI 0.8-3.6; P=0.142) for the patients receiving indinavir, 1.6 (95% CI 0.4-5.7; P=0.514) for those receiving ritonavir, and 1.1 (95% CI 0.2-4.9; P=0.926) for those receiving saquinavir.
Given that our data showed that the use of 3TC was a necessary condition for the development of FR, we excluded the 41 subjects not taking this drug from the subsequent analyses; the prevalence of FR was 12.1% in the remaining 265 women. There was no difference in the age or modality of HIV infection in the FR patients (Table 2), but the use of ZDV significantly correlated with a risk reduction (OR 0.35; 95% CI 0.14-0.89; P=0.0171) and the use of d4T with a significantly greater risk (OR 2.98; 95% CI 1.18-7.52; P=0.0118). The prevalence of FR was 13.2% (12 out of 91) in the patients treated with two nucleoside analogues and 15.2% (20 out of 132) in those receiving PI-including combinations, but the increased risk associated with the use of PIs was not statistically significant (OR 1.80; 95% CI 0.84-3.85; P=0.124). The use of each single PI was always associated with a not significantly increased risk of developing FR. The CD4+ cell counts were dichotomized on the basis of a cut-off value of 200 cells/μl; although the women with FR tended to have higher CD4 cell counts both at the beginning of the stable ARV regimen and at the time of the last visit, these differences were never statistically significant. Viraemia values were dichotomized at a value of 10000 copies per ml; the risk of developing FR was twice as great in the women with a viral load of more than 10000 copies per ml, but this increase was only of borderline significance (OR 2.0; 95% CI 1.0-4.3; P=0.0672). At the time of the last observation, there was no difference in viral load between women with and without FR. Nevertheless, at the last visit more than 80% of the cohort had an undetectable viral load and more than 90% had fewer than 10000 copies per ml.
The median total duration of both ARV and 3TC treatment was longer in the case of the women with FR than in those without (1187 versus 976 days and 395 versus 337 days, respectively): only one of the 32 women with FR had received ARV treatment for less than 1000 days (the median value for the entire cohort). The risk of FR for the women who had received ARV treatment for more than 1000 days was 10 times greater than in those who had received shorter treatments (OR 10.8; 95% CI 1.4-80.5; P=0.0207). The estimated relative risk associated with the use of 3TC for more than 1 year was 2.1 (95% CI 1.0-4.5; P=0.0497).
Because of the close relationship of FR with the duration of therapy, we analysed the data controlling each variable for the length of ARV treatment. The analysis revealed some interesting differences from the results of uncontrolled estimates (Table 2). The increased risk associated with a viral load at enrolment of more than 10000 copies per ml was slightly but significantly higher (P=0.0300), whereas the difference in the risk associated with the use of ZDV and d4T was no longer significant (P=0.1121 and P=0.0807, respectively), nor was there any correlation between the duration of 3TC treatment and the occurrence of FR.
Finally, we carried out two stepwise logistic regression analyses of the patients taking 3TC (one forward and the other backward) using as variables the use of ZDV, d4T and PI, CD4 cell counts and viral load at enrolment and at the last visit, and risk factors for HIV infection. Both analyses indicated that a duration of ARV therapy of more than 1000 days and a viral load at enrolment of more than 10000 copies per ml were the only variables that significantly and independently correlated with the risk of FR.
Discussion
The natural history of HIV infection has been profoundly changed by the use of combination therapies, which have been found to be highly effective [1-5]. Reports of some unexpected side-effects have, however, begun to appear in the medical literature, including a group of conditions involving fat metabolism and distribution in patients treated with combined ARV regimens: benign symmetric lipomatosis [7], the accumulation of adipose tissue in the dorsocervical region [10], the accumulation of visceral abdominal fat [11], and an increase in the breast size and abdominal girth of female patients [9,14,16]. The majority of those reports indicate no changes in body weight, thus confirming that they are secondary to real FR; furthermore, no correlations were detected between FR and the examined endocrinological parameters.
On the contrary, Carr et al. [13] have recently described a syndrome characterized by fat loss of the face and limbs associated with central adiposity, hyperlipidemia and insulin resistance.
In the present study, we observed a particular FR picture characterized by breast and abdominal fat accumulation, and accompanied by a wasting of the glutei and lower limbs in approximately 10% of HIV-infected women undergoing combined ARV therapy. This frequency is similar to that described for ‚protease-paunch‚ (approximately 7%) [15], although less than that reported by Carr et al. [13] for lipodistrophy (approximately 50% ).
The glucose, cholesterol and triglyceride levels in the FR patients were not significantly different from those measured in the matched control group: the high triglyceride levels observed in both groups was generally associated with PI treatment (data not shown). Furthermore, none of the FR patients presented endocrinological alterations. Nevertheless, FR was associated with DEXA evidence of an increase in breast and girth fat and a decrease in leg fat.
All of the 32 women with FR were taking 3TC- containing regimens; only one of them had received ARV treatment for less than 1000 days. The duration of treatment with 3TC was associated with an increased risk of FR in the subgroup of women receiving this drug, but this association disappeared when the analysis was corrected in order to take the overall duration of ARV therapy into account. These results suggest that exposure to 3TC may cause FR in women who have received ARV treatment for a long time, but that the duration of the use of the drug does not affect the risk of developing the syndrome. In other words, FR occurs in women who have received long-term ARV treatment and, perhaps in predisposed patients, may be at least partly triggered by the use of 3TC.
FR also tends to occur more frequently in women who have a viral load of more than 10000 copies per ml when they begin receiving stable ARV combination therapy.
Although the treatment regimens including PI were associated with a slightly higher risk of developing FR, PIs themselves do not seem to be necessary for the development of the syndrome, because 12 of our patients were treated with 3TC plus another nucleoside analogue (d4T in eight out of cases) and had never received a PI. This finding is similar to that of Lo et al. [10], regarding four patients with buffalo hump who were not treated with PI, and also corresponds with that of Saint-Marc and Touraine [17], who described a case of buffalo hump in a man treated with d4T and 3TC. Moreover, in both of the previously published case reports regarding breast enlargement in women treated with indinavir [9,14], the ARV regimens also included d4Tand 3TC.
Conclusion
The mechanism responsible for the pathogenesis of fat tissue abnormalities remains unclear. It has been hypothesized that buffalo hump and visceral fat accumulation are secondary to the endocrinological alterations induced by PIs, but this has not yet been formally demonstrated. Our data are insufficient to allow us to state whether the FR abnormalities observed in our patients simply represent the manifestation in women of the FR disorders observed in male patients, or whether they are caused by a different metabolic pathway.
Carr et al. [18] have recently hypothesized that HIV PIs may induce peripheral lipodistrophy by binding and inhibiting a homologous human protein involved in lipid metabolism. Furthermore, as all of the reports concerning fat tissue abnormalities were published after the introduction of PIs in clinical practice, there was an almost automatic tendency to associate them with this class of drugs. Given the relatively late introduction of PIs in Italy, the large number of individuals stably treated with PI-free two-drug combination therapies in our Institute has made it possible to identify the fact that the overall duration of ARV and nucleoside RT inhibitor treatment plays a role in inducing FR, but not that the introduction of the latter is a determinant factor. Therefore, despite the strength of the evidence indicating that PI therapy and its duration are associated with the development of lipodistrophy [12,13], this is not necessarily true in the case of the conditions characterized by abnormal fat distribution without fat loss, as is documented by our findings and those of others [10,17],
It is also interesting to note that none of the women treated with a PI in our cohort presented the clinical and metabolic picture of peripheral lipodistrophy during the period of the study. Longitudinal studies are needed to determine whether PI-associated lipodistrophy is really less frequent in women or simply takes a longer time to appear.
Taken together, these considerations support the hypothesis that the mechanisms causing FR are largely independent of the action exerted by PIs, and suggest that it is worth extending the investigation to other drugs. Furthermore, as suggested by Miller et al. [11], it cannot be excluded that these phenomena may be an unusual side-product of effective virus control and treatment-induced immunological events, particularly in the case of patients who have received ARV therapy for a long time.
Regardless of these speculative aspects, however, the increasing incidence of the untoward effects involving lipid metabolism is alarming, and there is an urgent need for studies designed to evaluate the reversibility of such alterations by adopting different but equally effective therapeutic regimens.
Finally, regardless of the unknown long-term risks associated with these puzzling syndromes, their marked aesthetic alterations are beginning to arouse serious psychological problems in young women that may significantly reduce their compliance with ARV therapies in the near future.
References
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 or less.
N Engl J Med 1997, 337:725-733.
2. Gulick RM, Mellors JW, Havlir D,
et al.: Treatment with indinavir, zidovudine, and lamivudine in adults with human immunodeficiency virus infection and prior antiretroviral therapy.
N Engl J Med 1997, 337:734-739.
3. Mouton Y, Alfandari S, Valette M,
et al.: Impact of protease inhibitors on AIDS-defining events and hospitalizations in 10 French AIDS reference centres.
AIDS 1997, 11:F101-105.
4. Reinhard Brodt H, Kamps BS, Gute P, Knupp B, Staszewski S, Helm EB: Changing incidence of AIDS-defining illness in the era of antiretroviral combination therapy.
AIDS 1997, 11:1731-1738.
5. Hogg RS, Rhone SA, Yip B,
et al.: Antiviral effect of double and triple drug combinations amongst HIV-infected adults: lessons from the implementation of viral load-driven antiretroviral therapy.
AIDS 1998, 12:279-284.
6. Mehta S, Moore RD, Graham NMH: Potential factors affecting adherence with HIV therapy.
AIDS 1997, 11:1665-1670.
7. Hengel RL, Watts NB, Lennox JL: Benign symmetric lipomatosis associated with protease inhibitors.
Lancet 1997, 350:1596.
8. Ruane PJ: Atypical accumulation of fatty tissue.
37th Interscience Conference on Antimicrobial Agents and Chemotherapy. Toronto, September 1997 [abstract I-185].
9. Herry I, Bernard L, de Truchis P, Perronne C: Hypertrophy of the breasts in a patient treated with indinavir.
Clin Infect Dis 1997, 25:937-938.
10. Lo JC, Mulligan K, Tai VW, Algren H, Schambelan M: ‚Buffalo hump‚ in men with HIV-1 infection.
Lancet 1998, 351:867-870.
11. Miller KD, Jones E, Yanovski JA, Shankar R, Feuerstein I, Falloon J: Visceral abdominal-fat accumulation associated with use of indinavir.
Lancet 1998, 351:870-875.
12. Viraben R, Aquilina C: Indinavir-associated lipodystrophy.
AIDS 1998, 12:F37-39.
13. Carr A, Samaras K, Burton S,
et al.: A syndrome of peripheral lipodistrophy, hyperlipidaemia and insulin resistance in patients receiving HIV protease inhibitors.
AIDS 1998, 12:F51-58.
14. Lui A, Karter D, Turett G: Another case of breast hypertrophy in a patient treated with indinavir.
Clin Infect Dis 1998, 26:1482.
15. Rosenberg HE, Mulder J, Sepkowitz KA, Giordano MF: ‚Protease-paunch‚ in HIV+ persons receiving protease inhibitor therapy: incidence, risk and endocrinologic evaluation.
Fifth Conference on Retrovirus and Opportunistic Infections. Chicago, February 1998 [abstract 408].
16. Dong K, Flynn MM, Dickinson BP,
et al.: Changes in body habitus in HIV(+) women after initiation of protease inhibitor therapy.
12th World AIDS Conference. Geneva, June 1998 [Abstract 12373].
17. Saint-Marc T, Touraine JL: ‚Buffalo hump‚ in HIV-1 infection.
Lancet 1998, 352:319-320.
18. Carr A, Samaras K, Chisholm DJ, Cooper DA: Pathogenesis of HIV-1 protease inhibitor-associated peripheral lipodistrophy, hyperlipidaemia and insulin resistance.
Lancet 1998, 352:1881-1883.
