Metabolic Syndrome Among HIV-Infected Patients

Prevalence, characteristics, and related factors

  1. Carlos Jericó, MD12,
  2. Hernando Knobel, MD12,
  3. Milagro Montero, MD1,
  4. Jordi Ordoñez-Llanos, MD23,
  5. Ana Guelar, MD1,
  6. Juan L. Gimeno, MD12,
  7. Pere Saballs, MD12,
  8. Jose L. López-Colomés, MD12 and
  9. Juan Pedro-Botet, MD12
  1. 1Department of Medicine, Hospital del Mar, Barcelona, Spain
  2. 2Universidad Autónoma de Barcelona, Barcelona, Spain
  3. 3Department of Biochemistry, Institut de Recerca, Hospital de Sant Pau, Barcelona, Spain
  1. Address correspondence and reprint requests to Prof. J. Pedro-Botet, Department of Medicine, Paseo Marítimo, 25-29, 08003 Barcelona, Spain. E-mail: 86620{at}imas.imim.es

Abstract

OBJECTIVE—To assess the prevalence in HIV-infected patients of the metabolic syndrome as defined by the National Cholesterol Education Program, i.e., three or more of the following components: abdominal obesity, hypertriglyceridemia, low HDL cholesterol, high blood pressure, and high fasting glucose.

RESEARCH DESIGN AND METHODS—In this cross-sectional study, 710 HIV-infected patients managed at the outpatient clinic of a tertiary hospital during 2003 completed the study protocol consisting of a medical examination and laboratory analysis after a 12-h overnight fast.

RESULTS—Metabolic syndrome prevalence was 17% and increased from 5.1% among HIV-infected patients under age 30 years to 27.0% for those aged 50–59 years. Age (per 10-year increment) (odds ratio [OR] 1.41 [95% CI 1.12–1.77]), BMI (1.27 [1.19–1.36]), past and present protease inhibitor exposure (2.96 [1.03–3.55] and 4.18 [1.4–12.5], respectively) were independently associated with the metabolic syndrome on logistic regression analysis. Furthermore, only stavudine (d4T) (1.74 [1.01–2.98]) and lopinavir/ritonavir (2.46 [1.28–4.71]) were associated with the metabolic syndrome after adjustment for age and BMI.

CONCLUSIONS—The prevalence of metabolic syndrome among these HIV-infected patients is similar to that previously reported in uninfected individuals. Of specific concern is the association of protease inhibitor exposure with the metabolic syndrome and, more specifically, with exposure to stavudine and lopinavir/ritonavir when individual antiretroviral drugs were analyzed.

Combination antiretroviral therapy (ART) has positively modified the natural history of HIV infection, leading to a significant reduction in morbidity and mortality. However, long-term toxicity is becoming recognized, and a variety of metabolic abnormalities including dyslipidemia, fat redistribution, high blood pressure, and insulin resistance have frequently been associated with this therapy, particularly when it contains protease inhibitors (1). This fact has raised the concern that the HIV-infected population in the long term may be at increased risk for cardiovascular disease, as has recently been described in two large prospective studies (2,3).

The metabolic syndrome is affecting the general population in epidemic proportions and is frequently associated with increased risk of cardiovascular morbidity and mortality (46). This aspect has been emphasized by the National Cholesterol Education Program Adult Treatment Panel (ATP) III (7), and a recently reviewed working definition of this syndrome has been drawn up (8). Insulin resistance plays a key role in the pathogenesis of the metabolic syndrome (4) and is frequently detected among HIV patients on ART (9). Furthermore, several traits of the metabolic syndrome in the general population overlap with common features of metabolic side effects associated with combination ART in the HIV-infected population. Although cardiovascular risk factors have been extensively evaluated in HIV-infected patients (3,10), the identification and management of metabolic disorders in those receiving combination ART has become a critical issue. The present study focuses on the prevalence and characteristics of the metabolic syndrome in HIV-infected patients and possible related factors, including individual antiretroviral drug exposure.

RESEARCH DESIGN AND METHODS

A cross-sectional study was carried out on HIV-infected patients managed at the outpatient Infectious Disease Unit of the Hospital del Mar, Barcelona, over a period of 1 year, from January through December 2003. The protocol study approved by the local ethics committee consisted of physical examination and laboratory analysis after a 12-h overnight fast. All participants were 20 years of age or older and were evaluated by trained physicians after giving their informed consent. Exclusion criteria included withdrawal of combination ART and evidence of clinical signs of active AIDS in the 3 months before entry because of their possible impact on anthropometric and laboratory parameters.

Age, sex, HIV disease status according to the 1993 Centers for Disease Control and Prevention (CDC) classification of HIV disease (11), HIV exposure (mutually exclusive in the following order: intravenous drug use, male homosexual activity, heterosexual activity), and type and duration of ART were recorded. Lipodystrophy was defined and categorized by the blinded physician-assessed (H.K.) presence of peripheral lipoatrophy (face, arms, legs, buttocks, and prominent veins), central lipohypertrophy (abdomen, breasts, dorsocervical region), and mixed lipodystrophy. Weight, height, and waist circumference were measured by standard methods. After the patient had rested for 10 minutes seated in a quiet room, blood pressure was measured in the left arm with the elbow flexed at heart level by the same physician (C.J.) using a 1042 Riester sphygmomanometer (Jungingen, Germany) with diastolic pressures at Korotkoff phase V (disappearance of sounds). Three readings were obtained, and the average of the second and third systolic and diastolic blood pressure readings was used in the analyses.

Total cholesterol and triglycerides were determined using enzymatic methods in a Cobas Mira automatic analyzer (Baxter Diagnostics, Düdingen, Switzerland). HDL cholesterol was measured using separation by precipitation with phosphotungstic acid and magnesium chloride. Glucose was measured by the oxidase method. CD4 lymphocyte cell count and HIV RNA viral load (Nuclisens Easy Q HIV-1; Biomérieux, Boxtel, the Netherlands) were performed at the time of the study; the nadir of CD4 cell count and baseline viral load levels were recorded.

Definition of the metabolic syndrome

As detailed in the ATP III report (9), participants with three or more of the following criteria were defined as having the metabolic syndrome: waist circumference >102 cm in men and >88 cm in women; triglycerides ≥150 mg/dl (1.69 mmol/l); HDL cholesterol <40 mg/dl (1.04 mmol/l) in men and <50 mg/dl (1.29 mmol/l) in women; blood pressure ≥130/85 mmHg; and fasting glucose ≥110 mg/dl (6.1 mmol/l). Individuals met criteria for high blood pressure or high fasting glucose concentration if they were currently on antihypertensive or oral hypoglycemic therapies, respectively.

Statistical methods

Student’s t test was performed to assess differences between two means. When data were not normally distributed, the Mann-Whitney U test was used. Either χ2 test or Fisher’s exact test was used to test the degree of association of categorical variables. The 95% CIs for proportions are calculated according to the efficient score method (corrected for continuity): P ± 1.96 × sqrt[P(1 − P)/n] P ± 1.96 × square root[P(1 − P)/n]. Computed factors in the univariate analysis were age, sex, BMI, HIV transmission group (dichotomized as intravenous drug users versus sexual transmission), HIV clinical stage (dichotomized as asymptomatic, A stage of the CDC versus symptomatic, B and C stage of the CDC), current and nadir CD4 cell count, plasma HIV RNA categorized as detectable (>500 copies/ml) or undetectable, lipodystrophy, duration of ART and type of ART classified as antiretroviral naive, never protease inhibitor exposure, past protease inhibitor exposure, and current protease inhibitor exposure.

Variables demonstrating a univariate relationship (P < 0.05) with the outcome variable were included in the logistic regression analysis to assess the effect of independent variables on metabolic syndrome diagnosis. A P value <0.05 was considered statistically significant. Goodness-of-fit was verified with the Hosmer and Lemeshow statistic method. The variables included in the logistic regression model were age (per 10-year increment), BMI, HIV transmission group (intravenous drug users formed the reference group), CD4 nadir cell count, and type of ART classified as antiretroviral naive (the reference group), treated but never exposed to protease inhibitors, treated with past exposure to protease inhibitors, and treated with current exposure to protease inhibitors. The association between individual antiretroviral drug exposure and the metabolic syndrome was analyzed by the χ2 test. Drugs demonstrating a univariate relationship (P < 0.05) with the metabolic syndrome were included in the logistic regression analysis to assess their independent effect after adjustment for age and BMI. All statistical analyses of database results were performed with the Statistical Package for the Social Sciences (SPSS for Windows, v.11.5; Chicago, IL).

RESULTS

Among the 1,016 HIV-infected patients managed at the outpatient clinic of our hospital during 2003, 209 were excluded for age, ART withdrawal, or overt clinical disease that required hospital admission. Of the 807 eligible patients, only 710 (88%) completed the study protocol. Of these, 626 (88.2%) were on combination ART and 84 (11.8%) naive. Clinical characteristics of HIV infection and metabolic syndrome traits are listed in Tables 1 and 2, respectively. This shows that one or more features of metabolic syndrome were seen in 492 (69.3%) patients, two or more in 254 (35.8%), three or more in 121 (17%), four or more in 32 (4.5%), and five features were seen in 1 patient (0.1%). Thus, 121 patients (86 men, 35 women) met metabolic syndrome criteria, yielding a prevalence of 17% (95% CI 14–20%). This was significantly increased by age and rose from 5.1% in HIV-infected patients under age 30 to 27.0% in those aged 50–59. A total of 116 patients were on combination ART, and 5 were naive HIV-infected patients. Lipodystrophy was more common among participants with the metabolic syndrome compared with those without (50.4 vs. 33.8%; P = 0.0001). Hypertriglyceridemia (95%) was the most frequent trait of the metabolic syndrome, followed by low HDL cholesterol (71.1%), high blood pressure (67.8%), abdominal obesity (47.1%), and high blood glucose levels (46.3%).

Metabolic syndrome-related factors in the univariate analysis are shown in Table 3. Patients with the metabolic syndrome presented higher age and BMI, lower percentage of intravenous drug users, and lower CD4 nadir cell count compared with those without the metabolic syndrome. Moreover, lipodystrophy and ART were associated with the metabolic syndrome. Lipodystrophy could indirectly be considered a component trait of the metabolic syndrome because its definition involves fat redistribution and, frequently, insulin resistance; thus, it was not included in the logistic regression analysis. Besides age and BMI, past and current exposure to protease inhibitors emerged as significantly and independently associated with metabolic syndrome in the logistic regression model (Table 4) (OR 2.96 [95% CI 1.03–3.55] and 4.18 [1.4–12.5], respectively). The relationship between individual antiretroviral drug exposure and the metabolic syndrome in the univariate and logistic regression analysis adjusted for age and BMI is shown in Table 5. Only stavudine (d4T) (1.74 [1.01–2.98]) and lopinavir/ritonavir (2.46 [1.28–4.71]) were independently associated with the metabolic syndrome.

CONCLUSIONS

Using ATP III criteria (7), 17% of HIV-infected patients in this sample were estimated to have the metabolic syndrome. This estimate is somewhat lower than that reported for the American population (12) using the same clinical definition or for the Spanish population (13). This could be due at least in part to the low number of women (28%) and patients over age 60 years (5%) in the present study, population subgroups with known higher metabolic syndrome prevalence. However, when comparing the age-specific prevalence of the metabolic syndrome, the prevalence among participants aged 30 through 50 was nearly identical in the present study of HIV-infected patients to that of the uninfected persons (12,13).

Although the prevalence of metabolic syndrome has been assessed in several populations (1215), previous data on metabolic syndrome among the HIV-infected population are limited and show an impressively high prevalence (16,17). Potential explanations for these dissimilar results include differences in study design, methodological aspects, and differences in the patient populations studied. In this respect, the present study has a relatively high proportion of intravenous drug users with a low prevalence of the metabolic syndrome (13%). On the other hand, when the present HIV-infected group was reanalyzed by the Data Collection on Adverse Events of anti-HIV Drugs Study criteria for age and sex, obesity, hypertension, hypercholesterolemia, low HDL, hypertriglyceridemia, and diabetes (3), prevalences of these cardiovascular risk factors were quite similar. Nevertheless, we must emphasize that the metabolic syndrome is a heterogeneous disorder, with substantial variability in the prevalence of component traits within and across populations.

Some aspects of the present study need to be highlighted. First, to date, this is the largest sample size study conducted in HIV-infected patients focusing specifically on the metabolic syndrome. Second, from at least three clinical criteria of the metabolic syndrome recommended by different organizations (7,18,19), we used ATP III criteria (7) because all five components can easily be evaluated in the clinical setting. The main difference in metabolic syndrome components of the present study compared with other studies conducted with HIV-infected or uninfected subjects (12,13,16,17) was the low prevalence of abdominal obesity (12.5%). Third, we ensured that blood samples were obtained after a 12-h overnight fast to avoid hypertriglyceridemia and hyperglycemia overdiagnosis. Finally, concerning blood pressure measurement and according to the American Heart Association (20), blood pressure was measured with the patient’s elbow flexed at heart level. From a clinical viewpoint, this apparently insignificant fact has important implications. In this respect, Villegas et al. (21) reported that 73% of health care workers failed to use proper arm and cuff positions, and Hemingway et al. (22) recently found blood pressure readings to be higher when the arm was parallel to the torso and would decrease by 8.8 to 14.4 mmHg with the forearm raised to a perpendicular position.

As occurs in uninfected subjects (12,13), metabolic syndrome in HIV-infected patients is associated with age and BMI. Among the HIV-infected population, the new findings concern the additional independent association of metabolic syndrome in those with past and current exposure to protease inhibitors. The link between protease inhibitor exposure, lipodystrophy, and metabolic syndrome is not surprising because fat redistribution, hyperlipidemia, insulin resistance, and hyperglycemia have been extensively reported in subjects treated with protease inhibitors (23). In the present study, we went a step further because the association between individual antiretroviral drug exposure and the metabolic syndrome was evaluated. In this respect, only stavudine (d4T) and lopinavir/ritonavir were independently associated with the metabolic syndrome. Although a cross-sectional study is not the most appropriate method to assess this association, this finding is consistent with other studies in which lipid abnormalities tend to be more marked with ritonavir and lopinavir/ritonavir (24,25), even in uninfected subjects (26). In the present study, if the logistic regression analysis had been performed including lopinavir/ritonavir as ritonavir exposure, the adjusted OR for ritonavir would have been 2.23 (95% CI 1.38–3.61; P = 0.001). The possible role of specific nucleoside analogues to alter lipid profile is not well established; some studies suggest that stavudine may be associated with hypertriglyceridemia and hypercholesterolemia in comparison with other nucleosides or nucleotide analogues (27,28).

Limitations of the present study are mainly related to the observational design and cross-sectional nature of the current analyses as well as the clinical population studied. In this respect, the results reported herein are only associations from which no conclusions regarding causality can be drawn. Furthermore, it is not expected that many measurements will always be conducted in a uniform manner. This includes measurement of waist circumference and blood pressure and laboratory analyses of lipid and glucose levels. Finally, information on other environmental factors such as physical activity or diet was not collected.

Although the substantial benefits of combination ART clearly outweigh the increase in cardiovascular risk associated with this therapy, it must be borne in mind that with progressive aging of the HIV-infected population and the expected long-term use of combination ART, the need will arise to prevent an increased incidence of metabolic syndrome in this population. Because metabolic syndrome represents a cluster of modifiable cardiovascular risk factors, the present results may have significant implications for health care physicians.

Table 1—

Demographic, anthropometric, and HIV infection characteristics of the 710 patients

Table 2—

Component conditions of the metabolic syndrome of the 710 HIV-infected patients

Table 3—

Association between age, sex, HIV disease characteristics, lipodystrophy, ART, and the metabolic syndrome

Table 4—

OR (95% CI) for the metabolic syndrome from multivariate analysis for selected variables

Table 5—

Univariate and multivariate analyses of individual antiretroviral drugs and the metabolic syndrome

Acknowledgments

C.J. is the recipient of a grant from the Fundación Institut Municipal d’Investigació Mèdica.

We thank Miss Christine O’Hara for review of the English version of the manuscript.

Footnotes

  • A table elsewhere in this issue shows conventional and Système International (SI) units and conversion factors for many substances.

    • Accepted September 24, 2004.
    • Received July 20, 2004.

References

| Table of Contents