| |
Important differences in components of the metabolic syndrome between HIV-patients with and without HAART and healthy controls
|
| |
| |
Scandinavian Journal of Infectious Diseases, Vol. 38, No. 8, August 2006, pp. 682-689 (TABLES ARE IMBEDDED IN THE ARTICLE BELOW)
Bente Magny Bergersena,*, Anita Schumacherb, Leiv Sandvikc, Johan N. Bruund and Kare Birkelande
aFrom the Department of Infectious Diseases, Aker University Hospital, Oslo; bDepartment of Internal Medicine, Hospital of Vestfold, Tnsberg; cCentre for Clinical Research, Ulleval University Hospital, Oslo; dDepartment of Infectious Diseases, Ulleval University Hospital; eHormone Laboratory, Aker University Hospital and Faculty Division, Aker University Hospital, University of Oslo, Oslo, Norway.
"....Important differences in aspects of the metabolic syndrome and insulin resistance exist between HIV-positive HAART-naive, HIV-positive HAART-treated and HIV-negative controls. In this study we found that non-overweight HAART-treated patients had higher prevalence of metabolic syndrome (15% vs 2%) and insulin resistance (39% vs 19%) compared to HIV-negative controls. The prevalence in non-overweight HAART-treated patients with lipodystrophy was even higher; 21% had the metabolic syndrome and 49% were insulin resistant. In the total population of non-overweight and overweight subjects, only 2% of the HAART-naive patients fulfilled the criteria of the metabolic syndrome. Another important finding in this study was that a considerable number of HAART-treated patients without lipodystrophy had the metabolic syndrome and/or insulin resistance....The increased insulin resistance may be explained by the combined effects of HIV infection itself [19,20], HAART-induced effect on glucose uptake [21,22] and indirect effects through fat redistribution....
....As mitochondrial toxicity is 1 of the main explanatory factors for different HAART-induced adverse effects including skeletal myopathy, hyperlactataemia, peripheral neuropathy and lipoatrophy, mitochondrial toxicity may be another mutual link between HAART-induced lipodystrophy and insulin resistance...
.....25% of the patients with metabolic syndrome and/or insulin resistance did not have lipodystrophy. As metabolic syndrome and/or insulin resistance may be more closely associated with CHD risk than the heterogenic body composition changes of lipodystrophy, the lack of overlap between the syndromes may explain some of the inverse correlation between lipodystrophy and CHD risk in the DAD study [5]. The DAD study did not include fasting blood samples, thus prevalence of insulin resistance and metabolic syndrome were not included in the analysis...."
Lipodystrophy, metabolic syndrome and insulin resistance in HAART-treated patients
In the HAART group 41% had lipodystrophy, 16% metabolic syndrome and 45% insulin resistance (Figure 1).
Lipodystrophy was significantly more prevalent among those with the metabolic syndrome OR 3.39 (1.63-7.04) and among subjects with insulin resistance OR 2.56 (1.52-4.31). Eight percent had lipodystrophy, metabolic syndrome and insulin resistance. 25% had metabolic syndrome and/or insulin resistance without lipodystrophy. In the HAART-treated patients with lipodystrophy, 25% had the metabolic syndrome and 57% had insulin resistance. In non-overweight HAART-treated patients with lipodystrophy the metabolic syndrome was diagnosed in 21% and insulin resistance in 49% of cases.
".....In the HIV-positive group, HAART treatment, lipodystrophy, insulin resistance and high age were positively related to metabolic syndrome (Table IV). Overweight and HIV-related factors such as AIDS diagnosis, high HIV-RNA level, low CD4 count and long duration of HIV infection had no significant influence on metabolic syndrome....
....In non-overweight subjects, the prevalence of insulin resistance was significantly higher in HAART-treated (39%) than in controls (19%, p=0.012) (Table III). Also non-overweight, HAART-naive patients had a non-significant trend towards higher prevalence of insulin resistance than controls (33%, p = 0.22) (Table III). See table & additional information below)....
...In HIV-positive subjects with lipodystrophy (LD + ) insulin resistance was diagnosed in 53 of 91 subjects (58%), compared to 60 of 172 (35%) in HIV-positive subjects without lipodystrophy (LD - ) (p = 0.001). Both median insulin level (LD+ 120 pmol/l (IQR 92-173) vs LD- 97 pmol/l (IQR 69-125), p = 0.001) and mean fasting glucose (LD+ 5.4±0.8 mmol/l vs LD- 5.1±0.8 mmol/l, p = 0.01) was elevated in LD+ patients compared to LD- patients....
....Overweight was positively related to insulin resistance both in the HIV-infected group (OR 3.07: 1.66-5.67; p<0.001) and the entire study population (OR 2.51: 1.56-4.04; p<0.001).....
"....In HIV-positive subjects with lipodystrophy (LD + ), metabolic syndrome was diagnosed in 21 of 91 subjects (23%), compared to 14 of 172 (8%) in HIV-positive subjects without lipodystrophy (LD - ) (p = 0.001). All the individual factors were significantly elevated in the LD+ group compared to the LD- group: hyperglycaemia 17% vs 5% (p = 0.006), elevated blood pressure 28% vs 13% (p = 0.007), abdominal obesity 17% vs 5% (p = 0.002), elevated triglycerides 63% vs 39% (p<0.001) and low HDL cholesterol 46% vs 33% (p = 0.033).... After age adjustment, odds ratio for having the metabolic syndrome in non-overweight HAART-treated patients compared to non-overweight controls was OR 9.59: 1.26-73.31; p = 0.029.....The prevalence of metabolic syndrome in non-overweight subjects was significantly higher (15%) in HAART-treated patients compared to HAART-naive (2%, p=0.019) and controls (2%, p=0.020). The prevalence of hypertriglyceridaemia and low HDL-cholesterol was elevated in HAART-treated compared to controls (p<0.001 and p = 0.011), but the difference in elevated blood pressure was not statistically significant (p = 0.15). Prevalence of hyperglycaemia and abdominal obesity was similar in the 3 groups...."
Abstract
The aim of this study was to compare the prevalence of metabolic syndrome and insulin resistance in HIV-positive patients with and without HAART and healthy HIV-negative controls. In total 357 subjects were examined: 56 HIV-positive HAART-naive, 207 HIV-positive on HAART treatment and 94 HIV-negative controls. We measured blood pressure, abdominal circumference, weight and height, and fasting serum levels of glucose, insulin and lipids in all the subjects. The presence of lipodystrophy was assessed in the HAART-treated patients. In non-overweight subjects the prevalence of the metabolic syndrome was 15% (25 of 162) in HAART-treated patients, 2% (1 of 44) in HAART-naive (p=0.019) and 2% (1 of 45) in controls (p=0.020). The prevalence of insulin resistance in non-overweight subjects was also higher in HAART-treated than in controls, 39% vs 18% (p=0.012) but similar to HAART-naive, 32% (p = 0.48 vs HAART, p = 0.22 vs controls). In non-overweight patients with lipodystrophy the metabolic syndrome was diagnosed in 21% and insulin resistance in 49%. In the entire HAART group 25% had the metabolic syndrome and/or insulin resistance without having lipodystrophy. We conclude that fasting glucose, HDL-cholesterol, triglycerides and blood pressure should be closely monitored in all HAART-treated patients, not only in overweighed or lipodystrophic individuals.
Introduction
Metabolic cardiovascular syndrome includes hyperglycaemia, hypertension, abdominal obesity, elevated triglycerides and reduced high density lipoprotein (HDL) cholesterol level [1]. It is closely associated with insulin resistance, and has several common features with highly active antiretroviral therapy (HAART)-induced lipodystrophy [2,3]. The association between metabolic syndrome and coronary heart disease (CHD) is well documented [4]. Accumulating evidence supports also an association between HAART and CHD [5,6]. Surprisingly, the DAD study found an inverse correlation between lipodystrophy and CHD [5], and the mechanisms behind the effect of HAART on CHD are still unclear. HAART-induced lipodystrophy and the metabolic syndrome seem closely related, but few studies have described the metabolic syndrome in HAART-treated patients [7,8] and none have reported on the syndrome in HIV-positive, HAART-naive patients. The main aim of our study was to investigate the prevalence of the metabolic syndrome and insulin resistance in a Scandinavian cohort of HIV-positive patients on HAART treatment, HIV-positive HAART-naive patients, and in HIV-negative controls. The second aim was to study the associations between the metabolic syndrome, insulin resistance and HAART-induced lipodystrophy.
Materials and methods
Study population
All HIV-positive patients living in Oslo who attended our outpatient clinic from 1 April 2000 to 1 April 2001 were invited to participate in the study (Oslo HIV Cohort Study: OHIV). Of 407 patients approached, 308 patients (78%) were included. Details of the Oslo HIV Cohort Study have been described in previous papers [9-11]. Fasting blood samples were available in 286 of the 308 patients. In this study 23 of them were excluded because they had been on some kind of antiretroviral therapy previously, or had been on HAART for less than 6 months. We included 263 HIV-positive patients: 207 patients who had been on HAART for at least 6 months and were still on treatment (HAART-treated) and 56 patients who had never been on antiretroviral therapy (HAART-naive). Healthy, HIV-negative controls were selected from the Vestfold Heart Care Study (VHCS), a previously published study of CHD and Chlamydia infection [12] where 193 controls without symptoms or clinical evidence of atherosclerotic disease were recruited from 4 working sites in Vestfold. They were included after an interview and a clinical examination by a physician (AS). The mean age of the healthy controls of the VHCS (55 y) was considerably higher than in the Oslo HIV Cohort Study; thus only subjects ≦ 55 y of age (n = 94) from the original control group were selected.
The protocols for the Oslo HIV Cohort Study 2000 and the Vestfold Heart Care Study were both approved by the Regional Committee for Medical Research Ethics and the Norwegian Data Inspectorate. The studies have been conducted in full accordance with the World Medical Association Declaration of Helsinki.
Blood samples
Serum for glucose, insulin, triglycerides, total and HDL cholesterol analysis was drawn after at least 8 h fasting. Lipids and glucose in the OHIV patients were analysed at the Department of Clinical Chemistry, Ulleval University Hospital using standard assays (Cobas Integra 700, Roche Diagnostics, Germany). In the VHCS controls comparable standard assays were used (Vitros 950/Mira S Plus) for lipids and glucose. Insulin was measured in all samples at the hormone laboratory, Aker University Hospital using an in-house radioimmunoassay with specific antibody purchased from Linco Research Inc., St. Charles, MO, USA, as previously described [13].
Definitions
Metabolic syndrome
Participants having 3 or more of the following criteria were defined as having the metabolic syndrome [1]: 1) Abdominal obesity: waist circumference >102 cm in men and > 88 cm in women; 2) High level of fasting glucose ≥ 6.1 mmol/l or on anti-diabetic medication; 3) Low level of high density lipoprotein cholesterol: HDL < 1.0 mmol/l in men and <1.3 mmol/l in women; 4) Hypertriglyceridaemia: >1.7 mmol/l; 5) Elevated blood pressure ≥ 130/ ≥ 85 mm Hg (both) or on anti-hypertensive medication.
Lipodystrophy
For identification of lipodystrophy we applied the clinical method for evaluation of body composition changes described by The HIV Lipodystrophy Case Definition Study Group [10,14]. The degree of fat atrophy (face, arms, legs, buttocks) and fat accumulation (abdomen, breasts, neck, dorso-cervical spine), was rated independently both by the patient and the physician as absent (score 0), mild (score 1), moderate (score 2), or severe (score 3). To be classified as having lipodystrophy, at least 1 of the 8 body shape changes had to be rated by the patient as moderate or severe and independently confirmed by the physician as moderate or severe.
Insulin resistance
Insulin resistance was calculated as homeostasis model assessment of insulin resistance (HOMA-IR) from the fasting serum levels of glucose and insulin (insulin (pmol/l)/7.2)/(22.5/glucose (mmol/l)) according to Matthews et al. [15]. The use of HOMA-IR is validated in HIV-positive patients and has shown a fair and significant correlation with insulin-mediated glucose uptake estimated with invasive methods (r = 0.61) [16]. The lower limit of the top quintile of HOMA-IR distribution was 3.54 in non-overweight (BMI ≦ 25 kg/m2) subjects in the control group, and this value was chosen as the threshold for insulin resistance.
Statistical methods
Continuous variables in 2 groups were compared by Student's t-test. In case of multiple groups, the differences were analysed by ANOVA. If the distribution of a continuous variable was markedly skewed (duration of HIV infection, HIV-RNA, HDL cholesterol, triglycerides, insulin, HOMA-IR), the comparisons were performed using the Mann-Whitney test and median and inter-quartile ranges were presented. One patient with Type 1 diabetes using insulin was excluded in the analysis of insulin resistance. When comparing a dichotomous variable, an exact Fisher test was used. Logistic regression analysis was used to adjust for confounding factors.* Possible confounding factors were high age (>45 y), male gender, white race, overweight (BMI > 25 kg/m2), current use of lipid-lowering or anti-hypertensive drugs, current use of prednisolone, AIDS diagnosis, long duration of HIV infection (>7 y), CD4 count <0.20X106/l and HIV-RNA >400 copies/ml. Age was dichotomized by the median age of the total study population (n = 357), and duration of HIV infection was dichotomized by the median value of the HIV-positive cohort (n = 263). p<0.05 was considered statistically significant.
Results
Patient characteristics
The HIV-positive patients were somewhat younger and less overweight than the controls and included more non-white subjects (Table I). The non-whites of the HIV-positive group were African (36), Asian (14, mostly from Thailand), Latin-American (1) and 2 unknown. Mode of infection did not differ significantly between HAART-treated and HAART-naive (homosexual contact 59.9% vs 48.2%, heterosexual contact 27.1% vs 37.5%, intravenous drug use 10.7% vs 12.5%, p = 0.33). AIDS was more prevalent in HAART-treated than in HAART-naive and they had been HIV-infected for a longer period (Table I). Only 5 subjects (8.9%) in the HAART-naive group had a CD4 cell count below 0.20X109/l and thus below the CD4 cell count level where HAART is indicated.
The HAART-treated patients had a mean duration of nucleoside reverse transcriptase inhibitor (NRTI) treatment of 52±2 months. Stavudine (47%) and lamivudine (85%) were the most common current NRTIs. The mean duration of HAART was 33±2 months. Most of the patients (77%) used protease inhibitors (PIs) and 39% used non-nucleoside reverse transcriptase inhibitors (NNRTIs). Indinavir (25%) and nelfinavir (25%) were the most frequently used PIs; nevirapine (34%) and efavirenz (4%) were the only NNRTIs.
Metabolic syndrome
In the total population of both overweight and non-overweight subjects, the prevalence of metabolic syndrome in HAART-naive patients was low (1 of 56 subjects) compared to both HAART-treated (p = 0.003) and controls (p = 0.006), mainly because of few HAART-naive with hypertension or abdominal obesity (Table II).
High age (OR 2.68: 1.40-5.10; p = 0.003) and overweight (OR 2.28; 1.23-4.20; p = 0.009) was positively related to metabolic syndrome, while male gender (p=0.61) and white race (p = 0.15) was not. Odds ratio for having the metabolic syndrome in non-overweight HAART-treated patients compared to non-overweight controls was OR 8.02: 1.06-60.90; p = 0.044. After age adjustment, odds ratio for having the metabolic syndrome in non-overweight HAART-treated patients compared to non-overweight controls was OR 9.59: 1.26-73.31; p = 0.029.
As overweight and abdominal obesity usually are strongly related, we did not include BMI in the multiple regression analysis. Alternatively we compared prevalence of metabolic syndrome in the non-overweight subjects only (Table III). The prevalence of metabolic syndrome in non-overweight subjects was significantly higher (15%) in HAART-treated patients compared to HAART-naive (2%, p=0.019) and controls (2%, p=0.020). The prevalence of hypertriglyceridaemia and low HDL-cholesterol was elevated in HAART-treated compared to controls (p<0.001 and p = 0.011), but the difference in elevated blood pressure was not statistically significant (p = 0.15). Prevalence of hyperglycaemia and abdominal obesity was similar in the 3 groups.
In HIV-positive subjects with lipodystrophy (LD + ), metabolic syndrome was diagnosed in 21 of 91 subjects (23%), compared to 14 of 172 (8%) in HIV-positive subjects without lipodystrophy (LD -) (p = 0.001). All the individual factors were significantly elevated in the LD+ group compared to the LD- group: hyperglycaemia 17% vs 5% (p = 0.006), elevated blood pressure 28% vs 13% (p = 0.007), abdominal obesity 17% vs 5% (p = 0.002), elevated triglycerides 63% vs 39% (p<0.001) and low HDL cholesterol 46% vs 33% (p = 0.033).
Insulin resistance
In non-overweight subjects, the prevalence of insulin resistance was significantly higher in HAART-treated (39%) than in controls (19%, p=0.012) (Table III). Also non-overweight, HAART-naive patients had a non-significant trend towards higher prevalence of insulin resistance than controls (33%, p = 0.22) (Table III). When assessing both overweight and non-overweight subjects, no difference in prevalence of insulin resistance between the 3 groups was found (Table II). The insulin level (median (IQR) was somewhat higher in HAART-treated subjects compared to controls: 107 (83-146) pmol/l vs 99 (78-116) pmol/l, p = 0.017. In HAART-naive the level of insulin was 86 (56-115) pmol/l (p = 0.001 vs HAART-treated, p = 0.14 vs controls). Mean glucose (±SD) was 5.2±0.8 mmol/l in HAART-treated, 5.1±0.9 mmol/l in HAART-naive and 5.4±0.6 mmol/l. in controls (p = 0.003 vs HAART-treated and p = 0.001 vs HAART-naive).
In HIV-positive subjects with lipodystrophy (LD + ) insulin resistance was diagnosed in 53 of 91 subjects (58%), compared to 60 of 172 (35%) in HIV-positive subjects without lipodystrophy (LD - ) (p = 0.001). Both median insulin level (LD+ 120 pmol/l (IQR 92-173) vs LD- 97 pmol/l (IQR 69-125), p = 0.001) and mean fasting glucose (LD+ 5.4±0.8 mmol/l vs LD- 5.1±0.8 mmol/l, p = 0.01) was elevated in LD+ patients compared to LD- patients.
Overweight was positively related to insulin resistance both in the HIV-infected group (OR 3.07: 1.66-5.67; p<0.001) and the entire study population (OR 2.51: 1.56-4.04; p<0.001). No HIV-related factors were significantly related to insulin resistance: long (>7 y) vs short duration of HIV-infection OR 1.25 (0.76-2.04), AIDS-diagnosis OR 1.24 (0.69-2.23), HIV-RNA copies >400/ml OR 0.64 (0.39-1.05) or CD4 cell count <0.20X106/l OR 0.98 (0.52-1.83). HAART treatment was not significantly associated with insulin resistance: OR 1.76 (CI 0.94-3.29), p = 0.077.
In the HIV-positive group, HAART treatment, lipodystrophy, insulin resistance and high age were positively related to metabolic syndrome (Table IV). Overweight and HIV-related factors such as AIDS diagnosis, high HIV-RNA level, low CD4 count and long duration of HIV infection had no significant influence on metabolic syndrome (Table IV).
Discussion
Important differences in aspects of the metabolic syndrome and insulin resistance exist between HIV-positive HAART-naive, HIV-positive HAART-treated and HIV-negative controls. In this study we found that non-overweight HAART-treated patients had higher prevalence of metabolic syndrome (15% vs 2%) and insulin resistance (39% vs 19%) compared to HIV-negative controls. The prevalence in non-overweight HAART-treated patients with lipodystrophy was even higher; 21% had the metabolic syndrome and 49% were insulin resistant. In the total population of non-overweight and overweight subjects, only 2% of the HAART-naive patients fulfilled the criteria of the metabolic syndrome. Another important finding in this study was that a considerable number of HAART-treated patients without lipodystrophy had the metabolic syndrome and/or insulin resistance.
Elevated prevalence of metabolic syndrome and insulin resistance in patients on HAART, especially in patients with lipodystrophy, is shown in earlier studies [2,7,8]. The prevalence of metabolic syndrome in our HAART group was about half of that reported in the Italian cohort [7] and also lower than in HIV-negative Americans [18]. The main differences compared to the Italian and American populations were less abdominal obesity, less high fasting glucose and less hypertension.
The prevalence of metabolic syndrome among patients with lipodystrophy was somewhat higher than in the case patients of the Lipodystrophy Study [14]. The same definitions of metabolic syndrome [1] and lipodystrophy [14] were used, but as the individual metabolic factors were not presented, detailed comparisons cannot be made.
25% of the patients with metabolic syndrome and/or insulin resistance did not have lipodystrophy. As metabolic syndrome and/or insulin resistance may be more closely associated with CHD risk than the heterogenic body composition changes of lipodystrophy, the lack of overlap between the syndromes may explain some of the inverse correlation between lipodystrophy and CHD risk in the DAD study [5]. The DAD study did not include fasting blood samples, thus prevalence of insulin resistance and metabolic syndrome were not included in the analysis.
We have not found any other studies evaluating the prevalence of metabolic syndrome and insulin resistance in HIV-positive, HAART-naive patients. It is important to know whether the differences between HAART-naive and HAART-treated represent a change back to normal during HAART therapy, or an elevation above normal. In lack of pre- and post-HIV and pre- and post-HAART data, we compared unselected HAART-treated patients and HAART-naive with similar age, gender, race and BMI. The prevalence of metabolic syndrome was low in the HAART-naive because of lower blood pressure and few subjects with abdominal obesity. The low blood pressure in HAART-naive is presented and discussed in a previous paper [11]. As BMI in our study was similar in HAART-treated and HAART-naive patients, a potential explanation for more abdominal obesity in the HAART patients is HAART-induced visceral fat accumulation.
In HIV-negative controls BMI was higher than in HIV-positive patients. However, including BMI in the multiple regression analysis would be dubious as overweight and abdominal circumference (which is 1 of the 5 factors of metabolic syndrome) is intimately related in healthy subjects. Instead we divided the groups into overweight and non-overweight subjects, and found that metabolic syndrome and insulin resistance in patients on HAART were less related to overweight than in HIV-negative controls.
A limitation of this study is the lack of matched controls. Unfortunately, most of the age- and gender-matched controls from the Oslo Health Study (OHS) used in our previous studies [9-11] had blood samples taken in a non-fasting state. However, the fasting VHCS controls had HDL cholesterol level, systolic blood pressure and BMI similar to the age- and gender-matched HIV-negative controls from OHS; HDL cholesterol 1.35±0.37 mmol/l, systolic blood pressure 130±17 mmHg, BMI 26.0±3.9 kg/m2 [9]. In any event the difference in BMI between HIV-positive patients and HIV-negative controls is a limitation of our study and larger studies that include age-, gender- and BMI-matched HIV-negative controls are needed to confirm our findings. Another limitation is the subjective assessment of lipodystrophy. Application of CT scan, DEXA scan and implementation of the lipodystrophy case definition [14] had been beneficial, but these methods were too expensive for a cohort of 263 subjects.
The increased prevalence of metabolic syndrome in non-overweight patients on HAART is related to HIV- and HAART- induced hypertriglyceridaemia and lowering of HDL-cholesterol. The increased insulin resistance may be explained by the combined effects of HIV infection itself [19,20], HAART-induced effect on glucose uptake [21,22] and indirect effects through fat redistribution. Insulin resistance associated with intramuscular fat accumulation has been shown in patients with lipodystrophy [23,24]. Recently increased intramyocellular lipid content related to impaired mitochondrial activity was found in insulin resistant subjects [25]. As mitochondrial toxicity is 1 of the main explanatory factors for different HAART-induced adverse effects including skeletal myopathy, hyperlactataemia, peripheral neuropathy and lipoatrophy, mitochondrial toxicity may be another mutual link between HAART-induced lipodystrophy and insulin resistance. Today, new PIs such as lopinavir/ritonavir, fosamprenavir and atazanavir are introduced in the HAART-treated population. This change may have had an influence on the prevalence of metabolic syndrome and insulin resistance, as these PIs have different metabolic profiles compared to nelfinavir and indinavir. The higher level of HDL-cholesterol level in HAART-treated compared to HAART-naive in our study may be explained by the relatively frequent (39%) use of NNRTIs which are known to increase HDL-cholesterol [26].
We conclude that fasting glucose, HDL-cholesterol, triglycerides and blood pressure should be closely monitored in all HAART-treated patients, not only in overweight or lipodystrophic individuals.
|
|
| |
| |
|
|
|
