The United Nations Programme on HIV and AIDS global report 2011 estimated that 3 million children in Sub-Saharan Africa are HIV-infected.1 In South Africa alone, almost half a million children under 15 years of age are HIV-infected, and despite reductions in vertical transmission, new infections continue to occur.2 South Africa has the largest antiretroviral treatment (ART) program in the world, with an estimated 157,000 infants and children currently on ART (C. van der Walt and J. Stokes, South African National Department of Health, personal communication, October 15, 2013), with many more being added as the antiretroviral program rolls out. These children face a lifetime of ART exposure, extending into several decades. In addition, the rapid changes related to growth may make them more sensitive than adults to drug-induced changes in metabolism. Dyslipidemia is a common late adverse effect of ART in HIV-infected children in the developed world.3 Dyslipidemia and insulin resistance significantly increase the long-term risk of atherosclerotic vascular disease and may have major public health implications for HIV-infected children in Africa, who must already deal simultaneously with the complexities of social stigma, chronic illness and medication adherence. The public health importance of atherosclerotic vascular disease in this extremely vulnerable population will only increase as ART coverage expands. Despite this importance, very little is known about the prevalence and risk factors for dyslipidemia and insulin resistance in prepubertal African children on ART. Previous studies of ART-related dyslipidemia in children have focused mostly on children living outside of Africa, who make up less than 8% of the global burden of pediatric HIV infections.1
We investigated the prevalence and predictors of blood lipid abnormalities and insulin resistance in prepubertal South African children on ART. Specifically, we examined the correlation of lipid and insulin abnormalities with subcutaneous fat maldistribution [measured by dual-energy X-ray absorptiometry (DEXA) and formal grading of visually obvious lipoatrophy] and with ART exposures.
MATERIALS AND METHODS
This was a subanalysis of cross-sectional data from a longitudinal cohort study of lipoatrophy in children.4 The first 100 of 190 eligible (on ART, prepubertal) clinic attendees were recruited from a centralized pediatric ART service in Cape Town, South Africa. Participant selection and cohort characterization have been previously described, including a comparison with the clinic patients who were not available for recruitment, which showed no evidence of selection bias.4 Total cholesterol, low-density lipoprotein (LDL), high-density lipoprotein (HDL), triglycerides, glucose and insulin were measured after an overnight fast. Homeostatic model assessment (HOMA) insulin resistance index was calculated as fasting glucose (mmol/L) × insulin (mU/L)/22.5.5 The following thresholds were considered abnormal: total cholesterol > 200 mg/dL (>5.18 mmol/L)6; LDL cholesterol > 130 mg/dL (>3.37 mmol/L)6,7; HDL cholesterol < 35 mg/dL (<0.91 mmol/L)7; triglycerides > 150 mg/dL (>1.7 mmol/L)6,7 and fasting glucose > 100 mg/dL (>5.6 mmol/L).8 HOMA values were evaluated against norms presented by de Almeida et al,9 which are the most appropriate norms available in current literature for children in a developing country because norms for African children are not available. Lipoatrophy was formally graded by consensus between 2 HIV pediatricians, who were experienced in identifying lipoatrophy using a standardized visual grading scale similar to that described by Carr et al10–12: grade 0—no fat changes; grade 1—possible minor changes, noticeable only on close inspection; grade 2—moderate changes, unequivocably noticeable only to an experienced clinician or a close relative, who knows the child well and grade 3—major changes, readily noticeable to an uninformed observer. Visually obvious lipoatrophy was defined as a score of 2 or 3. Face, arms, legs and buttocks were assessed for loss of subcutaneous fat resulting in a lean, muscular appearance of limbs and face, abnormally prominent limb veins and loss of gluteal fat pad with reduction in buttock size and loss of gluteal contour. Where the assessment of the 2 investigators did not concur, the change was graded as the lower score to ensure that only those with unequivocal changes were labeled as having lipoatrophy. Durations of previous ART exposures, maximum previous World Health Organization (WHO) clinical stage, pre-ART nadir and most recent CD4 and HIV RNA viral load were recorded. Formal diet assessment was performed by an experienced dietician using a local standardized 48-hour recall questionnaire (see Appendix, Supplemental Digital Content 1, https://links.lww.com/INF/C288). Using data from this interview, the dietician categorized each recruit as receiving inadequate, appropriate or excessive daily refined carbohydrate intake, daily dietary fat intake and total daily energy intake, based on recommended daily allowances for each.13
DEXA was performed on a subset of 42 patients using a Hologic Discovery DXA scanner (Hologic, Bedford MA) as per manufacturer’s instructions. The machine was calibrated daily and weekly using appropriate manufacturer-supplied phantoms. DEXA was requested for all recruits, but was not obtained in many participants because of DEXA scanner breakdowns and prioritized use for patients with clinical indications. There was no difference in gender, CD4 or cumulative time on standard dose stavudine between the 42 subjects, who underwent DEXA scanning, and the 58 subjects, who did not (P > 0.50 for all). The children who underwent DEXA were marginally younger than those who did not (7.1 versus 8.0 years, P = 0.03).
A Student’s t test was used to compare continuous variables, and Pearson’s χ2 test for categorical comparisons. Multiple linear regression was used to correlate DEXA measures of body fat with total cholesterol, LDL, HDL, triglycerides and HOMA, adjusting for age, gender and ethnicity. Multiple regression models were constructed to identify which ART drug exposure variables remained predictive of lipid or HOMA abnormalities after adjustment for age, gender, body mass index (BMI), dietary intake of fats and refined carbohydrate, age at ART initiation and stavudine exposure. Because stavudine exposure was common in our setting, all models included adjustment for both recent and cumulative stavudine exposure. A separate model was constructed for each biochemical outcome variable (total cholesterol, HDL, LDL, triglycerides and HOMA). For each ART drug, recent drug exposure (defined as current exposure or exposure within the previous 6 months) and cumulative lifetime exposure were included as separate variables. All statistical analyses were performed using R (version 2.10.0, Bell Laboratories, NJ).
Demographic data on the study participants are presented in Table 1. Overall, 40% had at least 1 fasted lipid, insulin or glucose abnormality. Phlebotomy failed in 4 children. Of the remaining 96 children, 65 were on a lopinavir/r (LPV/r)-based ART regimen, and 31 on an efavirenz (EFV)-based regimen. Comparison of these 2 groups is shown in Table 1. Adjusted mean triglycerides were elevated in the LPV group with a trend towards higher total cholesterol and LDL cholesterol (Table 2). Local guidelines recommend LPV/r for children initiating ART below 3 years of age and EFV for children initiating ART over 3 years of age. Mixed ethnicity children tended to be older at diagnosis and, therefore, more commonly received EFV; the reasons for later diagnosis are unclear. Children with advanced WHO clinical disease stage had higher viral loads and tended to be diagnosed earlier and therefore were more likely to receive LPV/r. In line with local guidelines, over 90% had received stavudine. Around half had recently been switched to abacavir because of a change in national recommendations. All had received lamivudine as their third ART agent.
After adjusting for age, gender and ethnicity, analysis of DEXA measures revealed that greater trunk fat and lower peripheral subcutaneous fat were associated with elevated triglycerides but not with total cholesterol, LDL, HDL or HOMA (Table 3). Similarly, after adjusting for age, gender and ethnicity, the presence of visually obvious lipoatrophy was associated with elevated triglycerides but not with total cholesterol, LDL, HDL, HOMA or lactate (Table 4).
On dietary assessment, 4 of 96 children were determined by the dietician to have excessive daily dietary fat intake, and 11 of 96 had excessive daily refined carbohydrate intake. Analysis of dietary variables showed no significant correlation with lipid or HOMA values (P > 0.3 for all; Table 5).
Analysis of disease severity (pre-ART viral load, nadir CD4 count before starting ART, CD4 nearest to study visit and maximum previous WHO clinical stage) demonstrated no significant correlation to any of the lipid or HOMA measures (P > 0.15 for all; Table 6).
The multivariate regression model for LDL demonstrated a significant correlation with recent EFV exposure (P = 0.02) and cumulative LPV/r exposure (P = 0.03) after adjustment for age, gender, BMI, dietary intake of fat and refined carbohydrate, age at ART initiation and stavudine exposure (Table 7). The model for HDL demonstrated a significant correlation with cumulative LPV/r exposure (P = 0.03; Table 8). The model for total cholesterol demonstrated a near-significant correlation with cumulative LPV/r exposure (P = 0.08) and recent exposure to EFV (P = 0.09; Table 9). Similar models for triglycerides and HOMA demonstrated no significant drug exposure correlations (data not shown). Because recent exposure to LPV/r was strongly inversely correlated with recent exposure to EFV, these 2 input variables could not mathematically coexist in the models. Therefore, all multivariate models were repeated twice, retaining either recent exposure to LPV/r or recent exposure to EFV. These revealed (surprisingly) that recent exposure to EFV (defined as current exposure or exposure within the previous 6 months) was significantly associated with lipids, whereas recent exposure to LPV/r was not, and thus only the models incorporating recent exposure to EFV are shown.
Our study, which focuses exclusively on prepubertal African children on ART, presents 3 important findings: (1) the prevalence of dyslipidemia and insulin resistance is high; (2) the effect of ART drugs on blood lipids and insulin sensitivity appears to accumulate over time, that is, cumulative lifetime exposure exerts an effect that is independent of current or recent drug exposure and (3) ART-induced lipoatrophy, while associated with elevated triglycerides, was not associated with cholesterol abnormalities or insulin resistance.
Few previous studies of ART-induced lipoatrophy and dyslipidemia have included African children as >50% of their cohorts.14–20 Of these, 3 included mostly pubertal children and did not report their prepubertal data separately.15–17 This is an important distinction because puberty has profound effects on fat metabolism, and ART may have distinctly different metabolic effects on prepubertal compared with pubertal individuals.21 Musiime et al19 reported 109 prepubertal African children on ART with prevalence of total cholesterol, LDL, HDL and triglycerides of 10%, 13%, 17% and 6%, respectively, which are remarkably similar to our findings. However, they did not attempt to correlate these with their skinfold thickness measurements or specific ART drug exposures. Piloya et al18 reported 364 African children (57% prepubertal) with prevalence of abnormal fasting total cholesterol, triglycerides and any lipid abnormality of 6%, 28% and 34%, respectively. However, they did not report their prepubertal data separately, making comparison with our data difficult. In line with our findings, they found no association between visually assessed lipoatrophy and lipid abnormalities. Bwakura-Dangarembizi et al20 reported 256 African children on ART with prevalence of abnormal fasting total cholesterol and LDL of 25% and 15%, respectively (although it was not stated that what proportion of those 256 were prepubertal). Interestingly, they found, as we did, that current EFV exposure was associated with higher LDL cholesterol. In line with our findings, they found no association between body fat maldistribution (measured by skinfold thickness) and blood lipids. Arpadi et al14 reported 156 prepubertal South African children on ART with prevalence of abnormal fasting total cholesterol, LDL, HDL, triglycerides, glucose and HOMA of 14%, 12%, 6%, 8%, 1% and 2% respectively. These figures are almost identical to our figures, with the exception of HDL (13% in our study) and HOMA (10% in our study). Their study found a distribution of abnormalities between their LPV/r and nevirapine groups that was very similar to our comparison of LPV/r-exposed versus EFV-exposed. The notable exception was the lower prevalence of abnormal HDL (3%) and triglycerides (3%) in their nevirapine group compared with our EFV group (15% and 13%, respectively). Arpadi et al also attempted to correlate visible body fat maldistribution (lipoatrophy or lipohypertrophy) with blood lipids and HOMA and, as we did, found no relationship between body fat maldistribution and fasting total cholesterol, HDL, LDL, glucose or HOMA. They did find a difference in triglycerides (both mean value and proportion with abnormally raised triglycerides) between children with and without body fat maldistribution; however, their analysis did not differentiate between lipoatrophy and lipohypertrophy. Although their study did objectively measure abnormal body fat distribution using skinfold thickness, they did not attempt to correlate this with blood lipids or HOMA. Our study goes 1 step further than these studies by comparing both visually obvious lipoatrophy and objective DEXA measures of body fat amount and distribution (particularly limb fat loss, representing lipoatrophy) with blood lipids and HOMA. This analysis should be viewed in light of our previously published DEXA data on the same cohort of children with and without visually obvious lipoatrophy, whose total extremity fat, limb fat mass-to-limb lean mass ratio and limb fat mass-to-BMI ratio demonstrated clear and significant differences.4
Risk factors for ART-induced elevations in lipids or HOMA have been investigated previously,3 with stavudine and protease inhibitors being most commonly implicated. However, previous studies have not attempted to separate the effect of recent exposure from cumulative lifetime exposure, and our finding that these exert independent effects is novel. The mechanism of association between recent EFV exposure and elevated LDL is unclear, although this is consistent with data from the AIDS Clinical Trials Group A5142 trial, in which EFV was independently associated with elevations in total cholesterol.22
The correlation between cholesterol abnormalities and exposure to EFV (recent) and LPV/r (cumulative) presents a practical dilemma to clinicians on the ground because these 2 drugs are central to the first-line and second-line ART regimens recommended by both the WHO23 and the South African National Department of Health.24 This dilemma clearly reveals the pressing need to make alternative drug options available, particularly, formulations that are palatable and practical for children in Africa.
The prevalences of dyslipidemia and insulin resistance in our cohort were high (40% overall). This may have major public health implications related to long-term risk of atherosclerotic vascular disease, for which HIV-infected individuals are already at increased risk because of persistent immune activation.25 The true proportion with at-risk values may be even higher. Cook et al26 have suggested centile-related thresholds for lipids that appear to correlate with increased cardiovascular risk in adults (viz. the 88th, 89th and 90th centiles for total cholesterol, LDL and triglycerides in each age and gender group, respectively). If these thresholds were applied to our data, the proportion with at-risk values would almost double.
The apparent lack of association between lipoatrophy and either cholesterol abnormalities or insulin resistance is interesting. In existing literature, lipoatrophy and lipohypertrophy have not usually been analyzed separately. Accumulation of intra-abdominal visceral fat (the most metabolically active form of body fat) as occurs in lipohypertrophy has been clearly linked to dyslipidemias, especially cholesterol abnormalities.10 Previous data have shown that lipohypertrophy and lipoatrophy occur independently of one another;21,27,28 therefore, in our opinion, lipoatrophy and lipohypertrophy require separate analysis.
Limitations of this study include the single-center cross-sectional design and limited sample size, which may not have allowed the causes of the observed differences to be conclusively identified. Analysis of longitudinal follow-up data in this cohort is underway. Another limitation is the lack of local population reference ranges for HOMA and lipids, which may have led to incorrect interpretation of the lipid and HOMA profiles in our cohort. Although there were proportionally more children of mixed ethnicity in the EFV than in the LPV/r group, our multivariate models did not adjust for ethnicity. However, this is unlikely to have skewed our results because the proportions with a biochemical abnormality in the 2 ethnic groups were not significantly different (P = 0.94).
In our cohort of prepubertal African children on ART, the prevalences of dyslipidemia and insulin resistance are high. This may have major public health implications related to long-term risk of atherosclerotic vascular disease, for which HIV-infected individuals are already at increased risk because of persistent immune activation. The effect of ART drugs on blood lipids and insulin sensitivity appears to accumulate over time, that is, cumulative lifetime exposure exerts an effect that is independent of current or recent drug exposure. This presents clinicians on the ground with a dilemma as current alternative pediatric drug options are typically not formulated to withstand Sub-Saharan African conditions and are often unaffordably expensive. Alternative ART drug options are urgently needed for children in Africa.
1. UNAIDS, World Health Organization, UNICEF. Global HIV/AIDS Response Progress Report 2011: Epidemic Update and Health Sector Progress Towards Universal Access. 2011 Geneva WHO Press
2. UNAIDS. South Africa country report. 2012 Available at: http://www.unaids.org/en/regionscountries/countries/southafrica/
. Accessed May 13, 2014
3. Barlow-Mosha L, Eckard AR, McComsey GA, et al. Metabolic complications and treatment of perinatally HIV-infected children and adolescents. J Int AIDS Soc. 2013;16:18600
4. Innes S, Cotton MF, Haubrich R, et al. High prevalence of lipoatrophy in pre-pubertal South African children
on antiretroviral therapy: a cross-sectional study. BMC Pediatr. 2012;12:183
5. McAuley KA, Williams SM, Mann JI, et al. Diagnosing insulin resistance in the general population. Diabetes Care. 2001;24:460–464
6. Daniels SR, Greer FRCommittee on Nutrition. . Lipid screening and cardiovascular health in childhood. Pediatrics. 2008;122:198–208
7. Kavey RE, Daniels SR, Lauer RM, et al.American Heart Association. American Heart Association guidelines for primary prevention of atherosclerotic cardiovascular disease beginning in childhood. J Pediatr. 2003;142:368–372
8. Zimmet P, Alberti KG, Kaufman F, et al.IDF Consensus Group. The metabolic syndrome in children and adolescents - an IDF consensus report. Pediatr Diabetes. 2007;8:299–306
9. Almeida CA, Pinho AP, Ricco RG, et al. Determination of glycemia and insulinemia and the homeostasis model assessment (HOMA) in school children and adolescents with normal body mass index. J Pediatr (Rio J). 2008;84:136–140
10. Carr A, Emery S, Law M, et al.HIV Lipodystrophy Case Definition Study Group. An objective case definition of lipodystrophy in HIV-infected adults: a case-control study. Lancet. 2003;361:726–735
11. Aurpibul L, Puthanakit T, Taejaroenkul S, et al. Recovery from lipodystrophy in HIV-infected children after substitution of stavudine with zidovudine in a non-nucleoside reverse transcriptase inhibitor-based antiretroviral therapy. Pediatr Infect Dis J. 2012;31:384–388
12. Hartman K, Verweel G, de Groot R, et al. Detection of lipoatrophy in human immunodeficiency virus-1-infected children treated with highly active antiretroviral therapy. Pediatr Infect Dis J. 2006;25:427–431
13. Nicklas T, Hayes D.. Position of the American Dietetic Association: nutrition guidance for healthy children ages 2 to 11 years. J Am Diet Assoc. 2008;108(6):1038–1044
14. Arpadi S, Shiau S, Strehlau R, et al. Metabolic abnormalities and body composition of HIV-infected children on lopinavir or Nevirapine-based antiretroviral therapy. Arch Dis Child. 2013;98:258–264
15. Beregszaszi M, Dollfus C, Levine M, et al. Longitudinal evaluation and risk factors of lipodystrophy and associated metabolic changes in HIV-infected children. J Acquir Immune Defic Syndr. 2005;40:161–168
16. Dzwonek AB, Lawson MS, Cole TJ, et al. Body fat changes and lipodystrophy in HIV-infected children: impact of highly active antiretroviral therapy. J Acquir Immune Defic Syndr. 2006;43:121–123
17. Ene L, Goetghebuer T, Hainaut M, Peltier A, Toppet V, Levy J.. Prevalence of lipodystrophy in HIV-infected children: a cross-sectional study. Eur J Pediatr. 2007;166(6):13–21
18. Piloya T, Bakeera-Kitaka S, Kekitiinwa A, et al. Lipodystrophy among HIV-infected children and adolescents on highly active antiretroviral therapy in Uganda: a cross sectional study. J Int AIDS Soc. 2012;15:17427
19. Musiime V, Cook A, Kayiwa J, et al.CHAPAS-3 trial team. Anthropometric measurements and lipid profiles to detect early lipodystrophy in antiretroviral therapy experienced HIV-infected children in the CHAPAS-3 trial. Antivir Ther. 2014;19:269–276
20. Bwakura-Dangarembizi M, Musiime V, Szubert AJ, et al.ARROW Trial Team. Prevalence of lipodystrophy and metabolic abnormalities in HIV-infected African children
after 3 years on first-line antiretroviral therapy. Pediatr Infect Dis J. 2015;34:e23–e31
21. Alam N, Cortina-Borja M, Goetghebuer T, et al. Body fat abnormality in HIV-infected children and adolescents living in Europe: prevalence and risk factors. J Acquir Immune Defic Syndr. 2012;59(3):314–324
22. Haubrich RH, Riddler SA, DiRienzo AG, et al.AIDS Clinical Trials Group (ACTG) A5142 Study Team. Metabolic outcomes in a randomized trial of nucleoside, nonnucleoside and protease inhibitor-sparing regimens for initial HIV treatment. AIDS. 2009;23:1109–1118
23. World Health Organization. Consolidated Guidelines on the Use of Antiretroviral Drugs for Treating and Preventing HIV Infection. 2013 Geneva, Switzerland: World Health Organization
24. South African National Department of Health. The South African Antiretroviral Treatment Guidelines. 2013 Pretoria, South Africa: South African National Department of Health
25. Islam FM, Wu J, Jansson J, Wilson DP.. Relative risk of cardiovascular disease among people living with HIV: a systematic review and meta-analysis. HIV Med. 2012;13(8):453–468
26. Cook S, Auinger P, Huang TT.. Growth curves for cardio-metabolic risk factors in children and adolescents. J Pediatr. 2009;155:S6.e15–S6.e26
27. Aurpibul L, Puthanakit T, Lee B, Mangklabruks A, Sirisanthana T, Sirisanthana V.. Lipodystrophy and metabolic changes in HIV-infected children on non-nucleoside reverse transcriptase inhibitor-based antiretroviral therapy. Antivir Ther. 2007;12(8):1247–1254
28. Tien PC, Grunfeld C.. What is HIV-associated lipodystrophy? Defining fat distribution changes in HIV infection. Curr Opin Infect Dis. 2004;17:27–32