| |
Metabolic Syndrome in HIV-Infected Patients: No Different than the General Population? COMMENTARY
|
| |
| |
Clinical Infectious Diseases March 1, 2007;44:735-738
Clara Y. Jones
Department of Family Medicine and Community Health, Tufts University Medical School, Boston, Massachusetts
"....This study did not controvert the evidence that metabolic complications increase in persons after initiating HAART or with effective maintenance of weight and reduction of opportunistic infections during the post-HAART era, but showed that the estimated level of increased risk of metabolic syndrome in HIV-infected persons (diagnosed by NCEP criteria or Framingham 10-year risk score) is not significantly greater than the risk in HIV-uninfected persons of the same age, race, sex, and smoking status. The findings do not diminish the potential importance of addressing the risk factors in individual patients to decrease the risk of cardiovascular disease or type II diabetes...."
The study by Mondy et al. [1] in this issue of Clinical Infectious Diseases addresses the development of metabolic syndrome (diagnosed using National Cholesterol Education Program [NCEP] Adult Treatment Panel III criteria [2]) in HIV-infected persons. Metabolic syndrome has been associated with increased risk of cardiovascular disease and type II diabetes among a number of populations. The authors matched participants who completed a cardiovascular risk survey and laboratory examination at the Washington University HIV clinic (St. Louis, MO) with participants of the National Health and Nutrition Examination Survey (NHANES; 2001-2002) according to age, race, sex, and smoking status. They compared the prevalence of metabolic syndrome and the prevalence of individual components of metabolic syndrome between the case patients (HIV clinic patients) and matched control subjects (NHANES participants). HIV RNA levels and CD4 cell counts that were obtained within 3 months of survey completion were used for the HIV clinic patients. A Framingham 10-year risk score was calculated to give a quantitative estimate of the risk for cardiovascular disease [3]. The authors also evaluated the association of medication history (type of HAART), CD4 cell count, and HIV load with the prevalence of metabolic syndrome among the Washington University HIV clinic patients.
The Mondy et al. [1] study is a useful reference for several reasons. The case patients were an unselected, demographically mixed HIV clinic population (not specifically recruited because of lipodystrophy or known cardiovascular risk factors), and they were matched with control subjects from a recent national cohort (although not nationally representative, because the analysis ignores the complex multistage survey design of NHANES and the oversampling of the older population, children and adolescents, and Mexican Americans). The use of the NHANES cohort to obtain matched control subjects was more likely to reflect the pathology seen among the general population than the use of healthy blood donors or other healthy control subjects. In addition, the cohort of HIV-infected patients was a recent cohort (2005), and given the shifts in the patterns of antiretroviral use over time, it was more likely to be representative of the current patterns of antiretroviral use than some of the previous studies.
Mondy et al. [1] found the prevalence of metabolic syndrome to be similar among the HIV-infected patients (25.5%) that were enrolled in their study and among the matched NHANES participants (26.5%), although there was a difference in the patterns of the individual components leading to the diagnosis of metabolic syndrome between the 2 groups. A recently published study by Jacobson et al. [4] that compared metabolic syndrome in HIV-infected subjects in the Nutrition for Healthy Living (NFHL) cohort with NHANES (1999-2002) participants found a similar prevalence of metabolic syndrome among the HIV-infected patients (24%). During the Mondy et al. [1] study, HIV-infected subjects with metabolic syndrome had lower HDL cholesterol levels, higher triglyceride levels, smaller waist circumference, lower body mass index, and lower glucose levels, compared with the subjects in the NHANES cohort with metabolic syndrome. The finding of a high prevalence of lipid abnormalities has been reported in most studies of HIV-infected patients [4-10]. The occurrence of lower glucose levels among the HIV-infected patients than the subjects in the NHANES cohort was somewhat unexpected given the many studies that have shown decreased insulin sensitivity in HIV-infected persons receiving HAART (compared with either HIV-infected persons not receiving HAART or healthy control subjects). However, Jones et al. [9] showed a similar finding in nondiabetic subjects of the NFHL, compared with NHANES III participants; although fasting insulin levels were higher in NFHL, the fasting glucose levels were lower than in NHANES III.
Among the HIV-infected patients of the Mondy et al. [1] study, there was no significant difference in the percentage of persons receiving HAART between those with and without metabolic syndrome, and the type of HAART was not associated with the presence of metabolic syndrome. Jacobson et al. [4] found that HIV-infected patients (regardless of HAART use) were less likely to have metabolic syndrome than the NHANES participants on the basis of results from a multivariate analysis adjusting for age, sex, race, poverty, regular exercise, strength training, and daily caloric intake per kilogram of body weight. After additional adjustment for body mass index, the trend persisted but was not significant. As in the Mondy et al. [1] study, the HIV-infected patients of the NFHL cohort were less likely to have abdominal obesity or high glucose levels and were more likely to have low HDL cholesterol levels and high triglyceride levels than the NHANES participants. From an incidence analysis of metabolic syndrome, Jacobson et al. [4] found that HAART use (yes or no) was not significantly associated with the risk of developing metabolic syndrome. However, when they looked at specific individual medications, lopinavir-ritonavir and didanosine therapies were associated with an increased risk of developing metabolic syndrome.
Calculation of the Framingham 10-year cardiovascular risk score allows estimation of risk of incident cardiovascular disease in persons using quantitative values for lipid parameters, blood pressure and age, in addition to sex, smoking, and use of blood pressure medications and gives some indication of how differences in severity of different risk factors might be expected to influence cardiovascular risk. In the Mondy et al. [1] study, the comparison of the Framingham 10-year risk scores between the case patients and control subjects suggested that the different component patterns of metabolic syndrome yield the same level of overall risk. However, use of the Framingham risk score does not account for differences of the relative severity of increased triglyceride levels (total cholesterol level and HDL level are the lipid parameters that enter the regression equation). The increased triglyceride levels of the HIV-infected patients (despite the higher percent of patients undergoing lipid-lowering therapy) suggest an increased severity of hypertriglyceridemia. Hypertriglyceridemia was associated with increased risk of myocardial infarction during the Data Collection on Adverse Events of Anti-HIV Drugs Study [7]. The Framingham risk score also ignores cardiovascular disease-equivalent conditions, such as diabetes and kidney disease [11-14]. Mondy et al. [1] found that the prevalence of diagnosed diabetes was the same for both case patients and control subjects. Epidemiologic studies suggest that HIV-infected persons have increased rates of proteinuria and kidney disease, compared with similar HIV-uninfected populations [15-17], even for those receiving HAART [18]. Microalbuminuria and chronic kidney disease are associated with increased risk of cardiovascular death [14, 19, 20]. Other independent risk factors, such as an elevated C-reactive protein level, coagulation parameters, and adipokines, are not included in the Framingham equation. Although several published studies have assessed the Framingham risk score of HIV-infected patients [21-23], none has done so with longitudinal follow-up that allows evaluation of its actual predictive value. It is unclear how well the Framingham 10-year risk scores will predict cardiovascular risk for HIV-infected patients.
Do the results of this study make a difference in how we should implement recommendations for screening for cardiovascular risk factors, which has been suggested as a standard of HIV clinical care [24]? Should we be more concerned about looking for metabolic syndrome in patients receiving HAART? Mondy et al. [1] found no significant difference in the percentage of persons receiving HAART between those with and without metabolic syndrome, and the type of HAART was not a significant predictor of the presence of metabolic syndrome. Although HIV-infected subjects with metabolic syndrome were more likely to be white (47%) than HIV-infected subjects without metabolic syndrome (34%), the majority of HIV-infected subjects with metabolic syndrome in this study were African American. Mondy et al. [1] could not assess whether there was a difference of racial and sex demographics between the HIV-infected patients with metabolic syndrome and the control subjects with metabolic syndrome because they matched for race and sex. Their findings do not support the restriction of screening for metabolic syndrome for any demographic group of patients, nor for patients receiving any particular HIV medication regimens.
Are the NCEP criteria for metabolic syndrome the best to use for HIV infected persons? Criteria that include fasting insulin levels or microalbuminuria might give different estimates of the prevalence of metabolic syndrome than the NCEP criteria. Similarly, does the use of waist circumference criteria rather than waist-to-hip ratio tend to underestimate the proportion of subjects with central obesity among the HIV-infected population (given risk of peripheral lipoatrophy in patients infected with HIV)? Jones et al. [9] found that body mass index and waist circumference were lower for HIV-infected men than for men in the NHANES cohort, but they were not significantly different for HIV-infected women than for women in the NHANES cohort. However, waist-to-hip ratio was higher for HIV-infected women versus women in the NHANES cohort, although it was not significantly different for HIV-infected men versus men in the NHANES cohort [9]. Thus, the use of different criteria for central obesity might influence the number of men versus women who are identified with central obesity.
In a recent review, Kamin and Grinspoon [5] summarized studies that examined the relationships among HAART, HIV, and hard end points for cardiovascular disease. They concluded that a small but significant and increasing risk exists for cardiovascular disease related to HIV infection or antiretroviral therapy and/or PI use. HAART use in general-and PI use specifically-were not associated with increased risk of metabolic syndrome during this current study. Will diagnosis of metabolic syndrome (using whichever criteria) prove to be the best measure of cardiovascular risk? A recent commentary on metabolic syndrome by Kahn et al. [25] raised the issue of whether a diagnosis of metabolic syndrome adds anything in terms of greater cardiovascular risk than identification of the individual components of metabolic syndrome.
Metabolic syndrome has been assumed to be more common in HIV-infected persons because of the true increase of the components of metabolic syndrome, such as increased blood sugar level and diabetes, increased lipid levels, increased central obesity, and increased rates of hypertension from the pre-HAART to HAART era and in individual persons evaluated during the pre-HAART and HAART eras. The Mondy et al. [1] study compared an unselected cohort of HIV-infected persons with matched subjects from the NHANES (2001-2002) cohort and did not find a higher prevalence of metabolic syndrome among the patients from the HIV cohort. This study did not controvert the evidence that metabolic complications increase in persons after initiating HAART or with effective maintenance of weight and reduction of opportunistic infections during the post-HAART era, but showed that the estimated level of increased risk of metabolic syndrome in HIV-infected persons (diagnosed by NCEP criteria or Framingham 10-year risk score) is not significantly greater than the risk in HIV-uninfected persons of the same age, race, sex, and smoking status. The findings do not diminish the potential importance of addressing the risk factors in individual patients to decrease the risk of cardiovascular disease or type II diabetes.
The Mondy et al. [1] study extends our knowledge and understanding of the association of the components of metabolic syndrome with HIV infection parameters and suggests other specific studies that can be performed. It would be very interesting to compare different diagnostic criteria for metabolic syndrome among HIV-infected patients from the same HIV cohort (as was performed for HIV-uninfected patients during the San Antonio Heart Study [26]) to see whether the prevalence of metabolic syndrome depends substantially on the criteria used and which criteria have the greatest prognostic significance for HIV infection. Future research, such as a study looking at the predictive value of several criteria for metabolic syndrome (NCEP ATP III [2], World Health Organization [27], European Group for the Study of Insulin Resistance [28], and International Diabetes Federation Consensus [29]) and the Framingham 10-year cardiovascular risk score [3] for the same cohort, followed by the longitudinal follow-up of cardiovascular outcomes, would be very helpful to assess which method is best for identifying HIV-infected subjects at risk. Ongoing cohorts that have stored blood samples (for assessment of hyperinsulinemia) and that have assessed lipid levels, glucose levels, and proteinuria would be particularly useful.
|
|
| |
| |
|
|
|
