High Prevalence of Elevated Lipids Observed in Children on HAART
Serum lipids, glucose homeostasis and abdominal adipose tissue distribution in protease inhibitor-treated and naive HIV-infected children
AIDS 2003; 17(9):1319-1327
Ari Bitnun, From the Divisions of Infectious Diseases and Endocrinology and Diabetes, Department of Pediatrics. The Hospital for Sick Children, University of Toronto, Toronto, Canada
The morbidity and mortality associated with HIV infection has been dramatically reduced since the introduction of highly active antiretroviral therapy into routine clinical use. Unfortunately, in HIV-infected adults, such therapy has been linked to the development of a variety of metabolic abnormalities that include peripheral lipoatrophy, visceral adipose
tissue accumulation, hypercholesterolemia, hypertriglyceridemia, insulin resistance, impaired glucose tolerance and diabetes mellitus. Collectively, these metabolic abnormalities have been referred to as the HIV-associated
lipodystrophy syndrome. The protease inhibitors (PI), and to a lesser extent, the nucleoside reverse transcriptase inhibitors (NRTI) have been implicated as possible causes of the HIV-associated lipodystrophy syndrome.
There is currently a paucity of data pertaining to the HIV-associated lipodystrophy syndrome in children. Of the six paediatric studies published to date, three were cross-sectional analyses, two were longitudinal observational
studies and one was a physician survey. Only two of the studies compared PI-treated children with PI-naive children with respect to various aspects of the HIV-associated lipodystrophy syndrome. In the present report, we compare the glucose homeostasis and serum lipid profiles of PI-treated and PI-naive HIV-infected children and the abdominal adipose tissue distribution of PI-treated HIV-infected children, PI-naive HIV-infected children and HIV-uninfected children.
Objective: To determine the extent and degree of abnormalities of serum lipids, glucose homeostasis and abdominal adipose tissue distribution in protease inhibitor (PI)-treated and PI-naive HIV-infected children.
Design: A cross-sectional study involving HIV-infected children, 3-18 years of age, in a paediatric tertiary care centre.
Main outcome measures: Total, HDL and LDL-cholesterol, triglycerides, glucose, insulin, proinsulin and C-peptide were determined in the fasting state. Insulin resistance was assessed using the homeostatic model assessment-insulin resistance (HOMA-IR). Abdominal adipose tissue distribution was determined by single-slice computed tomography at the umbilical level.
Results: Thirty PI-treated and 20 PI-naive children were evaluated (76% prepubertal). PI-treated children had significantly higher total cholesterol (P = 0.0021), LDL-cholesterol (P = 0.019) and triglycerides (P = 0.0018). Serum glucose, insulin, proinsulin and C-peptide, the insulin : glucose ratio, HOMA-IR and abdominal adipose tissue distribution were similar in the two groups. Clinical and immunological HIV categories, viral load, CD4 cell count and stavudine therapy were not significantly associated with serum lipids, insulin resistance or abdominal adipose tissue distribution. The predictor variable most strongly associated with fasting serum insulin and HOMA-IR was the Tanner stage. Age was the most significant predictor variable of the visceral : subcutaneous adipose tissue ratio.
Serum lipids and glucose homeostasis
Fasting serum total cholesterol, LDL-cholesterol and triglyceride levels were significantly higher in PI-treated than PI-naive HIV-infected children in the unadjusted analysis. These differences were maintained after adjusting
for other variables. Total cholesterol levels above 5.44 mmol/l (210 mg/dl) were seen in 20% of PI-treated and 0% of PI-naive children. Thirty-three per cent of PI-treated and 15% of PI-naive children had LDL-cholesterol levels
above 3.37 mmol/l (130 mg/dl). Five of the PI-treated children had a total cholesterol level above 6.22 mmol/l (240 mg/dl); the LDL-cholesterol level of four of these children was greater than 4.40 mmol/l (170 mg/dl). Serum triglyceride levels of 1.64 mmol/l (145 mg/dl) or greater were seen in 40% of PI-treated and 5% of PI-naive subjects.
Factors that did correlate with serum insulin on univariate analysis included age (R = 0.61, P < 0.0001), Tanner stage (R = 0.64, P < 0.0001), BMI (R =
0.62, P < 0.0001), visceral adipose tissue area (R = 0.67, P < 0.0001), total adipose tissue area (R = 0.46, P = 0.0025) and serum triglyceride levels (R = 0.40, P = 0.0036).
The PI therapy category remained significant with respect to serum total cholesterol (P = 0.005) and triglyceride levels (P = 0.003), and remained non-significant with respect to serum insulin (P = 0.45), HOMA-IR (P = 0.94) and the visceral to subcutaneous adipose tissue ratio (P = 0.18). In each of these five analyses, the PI category-Tanner stage interaction term was not significant.
Explanatory variables significantly associated with total cholesterol included PI therapy, age and BMI. With respect to serum triglyceride levels, PI therapy was the only significant predictor variable. Tanner stage was the only significant predictor variable with respect to serum insulin and HOMA-IR, and age was the only significant predictor variable with respect to the visceral to subcutaneous adipose tissue ratio. Clinical and immunological HIV categories, viral load, CD4 cell count, sex and stavudine therapy were not significantly associated with serum total cholesterol, triglyceride or insulin levels, HOMA-IR or the visceral to subcutaneous adipose tissue ratio.
Conclusion: In this cohort of predominantly prepubertal HIV-infected children, PI therapy was associated with an atherogenic dyslipidemia but not with insulin resistance or abnormal abdominal adipose tissue distribution. The results suggest that children, particularly prepubertal children, are less susceptible than adults to PI-induced changes in glucose homeostasis and abdominal adipose tissue distribution.
In PI-treated children, the duration of PI therapy was not significantly associated with total, HDL or LDL-cholesterol, triglyceride levels, glucose, insulin, proinsulin, C-peptide, the insulin to glucose ratio, HOMA-IR, visceral,
subcutaneous or total adipose tissue areas or the visceral to subcutaneous adipose tissue ratio. Stavudine-exposed and unexposed children did not differ significantly with respect to serum lipids, glucose homeostasis or abdominal adipose tissue distribution.
Abdominal adipose tissue distribution
Abdominal adipose tissue distribution was determined for 30 PI-treated HIV-infected, 11 PI-naive HIV-infected and 52 HIV-uninfected children. The age and sex of PI-treated HIV-infected children (8.6 ± 4.4 years, 40% female), PI-naive HIV-infected children (9.2 ± 4.1 years, 55% female) and HIV-uninfected children (9.9 ± 4.2 years, 31% female) were similar (ANOVA; P = 0.29 and P = 0.54, respectively). The mean visceral to subcutaneous adipose tissue ratio in these
three groups was 0.312 ± 0.179, 0.247 ± 0.108 and 0.295 ± 0.178, respectively . The corresponding mean visceral adipose tissue area was 1562 ± 1989, 1500 ± 802 and 1250 ± 972 mm2 and the mean subcutaneous adipose tissue area was 5098 ± 3942, 8090 ± 6834 and 5857 ± 7101 mm2 for PI-treated HIV-infected, PI-naive HIV-infected and HIV-uninfected children, respectively. There were no significant differences between the three groups with respect to the visceral to subcutaneous adipose tissue ratio, visceral adipose tissue area or subcutaneous adipose tissue area.
The most striking finding of our study was the high prevalence of hypercholesterolemia and hypertriglyceridemia among PI-treated HIV-infected children. Similar findings have been noted in several previously published paediatric studies. In children, hypercholesterolemia has been associated with increased carotid artery intima-media thickness, increased coronary artery calcification scores, abnormal endothelial function, and more extensive fatty
streaks and fibrous plaques in the coronary arteries and aorta. In adults, elevated levels of total and LDL-cholesterol and triglycerides as well as low levels of HDL-cholesterol are well recognized as important risk factors for the
development of coronary artery disease, stroke and peripheral vascular disease. Preliminary data involving PI-treated HIV-infected adults suggest that those who develop dyslipidemia and other manifestations of the HIV-associated
lipodystrophy syndrome may be at increased risk of premature coronary artery disease. Clearly, the possibility that premature coronary artery disease, stroke and peripheral vascular disease will occur in dyslipidemic PI-treated
HIV-infected children and adolescents is of great concern.
The lack of effect of PI therapy on glucose homeostasis observed in our cohort is consistent with findings of previously published paediatric studies. In the two studies that compared PI-treated with PI-naive children, no significant differences in fasting serum glucose or insulin levels were observed. In the third study, glucose tolerance, determined by a standard oral glucose tolerance test, was normal in all 39 children studied, including 13 with clinically apparent lipodystrophy; 31 (79%) of these children were receiving PI therapy. Taken together, these results suggest that clinically significant increases in insulin resistance are relatively uncommon in HIV-infected children irrespective of antiretroviral therapy.
The lack of effect of PI therapy on abdominal adipose tissue distribution observed in our cohort is consistent with some, but not all, previously published studies. In the only study addressing this issue that specifically
compared PI-treated with PI-naive subjects, no significant differences in BMI, waist-hip ratio, percentage trunk adipose tissue and the trunk to limb adipose tissue ratio were detected between groups using anthropometric measurements and dual-energy X-ray absorptiometry. In another cross-sectional study, neither the proportion of subjects receiving PI therapy nor the duration of such therapy were significantly different in those with compared with those without clinically apparent lipodystrophy. On the other hand, in a longitudinal observational study of prepubertal children, PI therapy was found to be one of several risk factors for dual-energy X-ray absorptiometry-defined lipodystrophy. Unfortunately, because of the small numbers of subjects enrolled, it was not deemed possible to adjust for potential confounding variables such as age, Tanner stage and HIV-related virological and immunological factors.
The apparent lack of effect of PI therapy on glucose homeostasis and, to a lesser extent, on abdominal adipose tissue distribution observed in our cohort and other paediatric studies contrasted with published data pertaining to HIV-infected adults in which such abnormalities are substantially more common in PI-treated than PI-naive subjects. This apparent discrepancy may be related, at least partly, to normal physiological changes that occur during childhood.
Prepubertal children are inherently more insulin sensitive and tend to have proportionally more subcutaneous and less visceral adipose tissue than pubertal children and adults. The transition from the prepubertal state of relative
insulin sensitivity and the lower visceral to subcutaneous adipose tissue ratio to the more insulin resistant and higher visceral to subcutaneous adipose tissue ratio state of adulthood occurs principally during puberty and is mediated, at least partly, by hormonal changes that occur during this period. It is possible that PI-treated HIV-infected prepubertal children are able to maintain a state of normal glucose homeostasis, as determined by fasting insulin, proinsulin and C-peptide and HOMA-IR, by virtue of their relatively high insulin sensitivity. Hyperinsulinemia has been linked to visceral adipose tissue excess, and it may be that in the absence of hyperinsulinemia visceral adipose tissue accumulation does not occur or occurs only to a minimal degree.
In our PI-treated cohort, serum insulin, proinsulin, C-peptide, the insulin to glucose ratio and HOMA-IR correlated positively with age, Tanner stage, BMI, visceral and total abdominal adipose tissue area and serum triglyceride levels.
There was no significant association between these outcomes and the duration of PI therapy, stavudine therapy, clinical or immunological HIV categories, viral load or CD4 cell count. In the multiple regression analysis, Tanner stage was the only significant predictor variable with respect to serum insulin and HOMA-IR, and age was the only significant predictor variable with respect to the visceral to subcutaneous adipose tissue ratio. Taken together, these results suggest that in prepubertal HIV-infected children, glucose homeostasis and abdominal adipose tissue distribution are primarily affected by normal physiological changes that occur during childhood rather than by exposure to antiretroviral medications or HIV-related virological or immunological factors.
In evaluating adipose tissue distribution we elected to focus on the abdominal cavity because of the importance of the visceral adipose tissue depot in relation to insulin resistance, dyslipidemia and atherosclerotic cardiovascular disease. Visceral adipose tissue accumulation has been demonstrated in a substantial proportion of PI-treated HIV-infected adults, and has been associated, in these patients, with insulin resistance, impaired glucose tolerance and dyslipidemia. In contrast, the results of our study suggest that the excess accumulation of visceral adipose tissue does not
occur, or occurs at relatively low frequency and severity in prepubertal PI-treated HIV-infected children. Unfortunately, with the exception of one study, none of the previously published paediatric studies utilized imaging techniques
capable of discerning between visceral and subcutaneous adipose tissue. In this study of predominantly pubertal and post-pubertal children, the visceral adipose tissue area was significantly higher in those with than in those without clinically apparent lipodystrophy. The findings of that study as well as our own are consistent with those of another paediatric study that demonstrated less severe clinical lipodystrophy changes in prepubertal compared with pubertal children. These data suggest that prepubertal children are less susceptible to PI-induced adipose tissue distribution changes than pubertal children and adults.
The NRTI, particularly stavudine, and more advanced HIV disease have also been associated with the development of the HIV-associated lipodystrophy syndrome in adults as well as children. In the study by Arpadi et al, the presence of lipodystrophy in prepubertal HIV-infected children was associated with a lower baseline CD4 cell count, higher baseline viral load and treatment with stavudine as well as PI. In our cohort, there was no significant
correlation, on univariate analysis, between the severity of HIV disease, as determined by clinical and immunological HIV categories, viral load and CD4 cell count, or stavudine therapy with serum lipids, glucose, insulin, proinsulin
C-peptide or measures of abdominal adipose tissue distribution. In the multiple regression analyses, there were trends suggesting an inverse association between viral load and serum insulin, viral load and total cholesterol, and CD4 cell count and serum insulin. Although these results must be interpreted with caution, they do suggest that HIV-related virological and immunological factors may have some effect on glucose homeostasis, serum lipids and adipose tissue
distribution in HIV-infected children.
The principal limitations of our study were its cross-sectional design and small sample size. Although we attempted to adjust for the potential confounding effects of age, Tanner stage, race, BMI, HIV clinical and immunological disease
categories and stavudine therapy, it is possible that the two groups differed systematically in other important ways. With regard to sample size, it is conceivable that differences in glucose homeostasis and adipose tissue distribution between groups could have been detected with a larger sample size. However, our study was sufficiently powered to detect differences in glucose homeostasis parameters had the prevalence of and degree of insulin resistance of PI-treated children been similar to those observed in adults. With regard to abdominal adipose tissue distribution, given the small number of subjects in the PI-naive group and the wide variation in the visceral to subcutaneous adipose tissue ratio among subjects of all three groups, the power to detect a significant difference was limited.
Several other limitations deserve mention. First, our choices of glucose homeostasis and adipose tissue distribution outcomes although appropriate, were not without flaw. Fasting serum insulin, proinsulin and C-peptide and HOMA-IR may not be measures of sufficient sensitivity to allow the detection of small, but nevertheless physiologically important changes in insulin resistance. With regard to adipose tissue distribution, our study was designed solely to investigate the distribution of adipose tissue within the abdomen. As such, we are unable to comment on the potential effects of the PI on total body and peripheral adipose tissue distribution. Second, the incomplete application of CT scanning to the PI-naive group of children may have led to recruitment bias. However, we believe this is unlikely, as failure to perform CT scanning was not the result of a selection bias on the part of study investigators. Third, the findings of our study may not be generalizable to all PI, as the majority of children in our cohort were receiving either ritonavir or nelfinavir.
In conclusion, in this cross-sectional study of predominantly prepubertal HIV-infected children, PI therapy was associated with the development of an atherogenic dyslipidemia, but not insulin resistance or visceral adipose tissue
accumulation. Whether insulin resistance, adipose tissue redistribution and premature atherosclerosis become more prevalent in PI-treated HIV-infected children during puberty is unknown at the present time. Our results suggest that routine monitoring of serum lipids in all HIV-infected children, particularly in those receiving PI therapy, is warranted. Children found to be persistently dyslipidemic may benefit from dietary intervention and should be counselled regarding the benefits of regular exercise.