Body Fat Changes and Lipodystrophy in HIV-infected Children

Impact of Highly Active Antiretroviral Therapy

To the Editor:

HIV infection is a major cause of morbidity and mortality worldwide. In developed countries, life expectancy has increased considerably since the introduction of highly active antiretroviral therapy (HAART) in 1996.¹ However, a number of reports have documented some unexpected adverse effects from HAART, including a group of conditions characterized by altered lipid and glucose metabolism and alterations in body fat distribution in patients receiving combination antiretroviral therapy.^2,3 Clinical presentations have included the accumulation of adipose tissue in the dorsocervical region, commonly referred to as buffalo hump; fatty infiltration and bulging of supraclavicular fat pads, called Madelung disease; visceral abdominal fat accumulation ("crix belly"); benign symmetric lipomatosis; and subcutaneous fat wasting of the face and limbs.

The pathogenesis of body fat redistribution and metabolic changes occurring among HIV-infected patients treated with antiretroviral therapy remains obscure. The lipodystrophy syndrome was initially described in the adult HIV-infected population and was thought to be associated with the use of the protease inhibitors (PIs).² However, more recent studies (also in adults) have strongly suggested a role for nucleoside reverse transcriptase inhibitors (NRTIs) in the alteration of body fat distribution.⁴ Treatments combining a PI and NRTI increase the incidence and severity of the lipodystrophy syndrome.⁵

Although the lipodystrophy syndrome has been well characterized in the adult population, there are few pediatric data available.^6,7 We therefore studied children with perinatally acquired HIV infection to examine the effect of HAART on body fat distribution, as assessed by measurements of limb and waist circumference and skinfold thickness in a cross-sectional observational study.

We studied HIV-infected children older than 3 years attending the HIV Family Clinic at Great Ormond Street Hospital National Health Service Trust, London, UK, in November 2002 to 2003. Of the 107 HIV-infected children identified as eligible for the study, 98 agreed to participate. All children included were classified according to the Centers for Disease Control and Prevention (CDC, Atlanta, GA) revised classification system for HIV infection in children. Clinical details and therapy history were extracted from the patient's medical records. All children had a general physical examination, including specific observation for the presence of clinical features of altered body fat distribution (peripheral lipoatrophy: sunken cheeks, skinny arms, and legs with prominent veins; central lipohypertrophy: increased abdominal girth, "buffalo-hump," and breast enlargement).⁷ Blood samples and anthropometric measurements were also taken at the same visit. The latter included weight and height measurements and body mass index (BMI). Because these values vary with age, they were converted to age- and sex-adjusted z scores based on the British 1990 reference.^8,9 Skinfold measurements (triceps, biceps, subscapular, and suprailiac) were measured using the Harpenden skinfold caliper (Holtain Ltd, Crymych, UK). Upper arm, waist, hip, thigh, and calf circumferences were measured using a narrow 1-m tape. All skinfolds and circumference were measured by the same trained researcher (A.D.). Waist circumference measurements were converted to age- and sex-adjusted z scores based on UK data. Values for skinfold measurements were converted to age- and sex-adjusted z scores with the use of Dutch reference curves¹⁰ using an SPSS program developed at the Institute of Child Health, University College London.

Variables were tested for normality with the Kolmogorov-Smirnov test. For normal data, differences between groups were tested by t test or 1-way analysis of variance. Leptin levels were log₁₀ transformed before analysis. For nonnormal data, group differences were analyzed by the Mann-Whitney U test. Correlations between lipid abnormalities and HAART duration were assessed by Spearman coefficient. Regression analysis was used to look for associations between anthropometry and the type and duration of HAART treatment. Statistical analyses were performed using SPSS software (PC version 12.0; SPSS), and significance was defined as P < 0.05. Ethical approval was given by the Institute of Child Health and Great Ormond Street Hospital Research Ethics Committee.

Details of the 98 HIV-infected children enrolled in the study are shown in Table 1. All had vertically acquired HIV-1 infection, and 85% were of sub-Saharan African origin. At enrollment, 59 children (60%) were receiving combination HAART. Of these, 23 (39%) were receiving stavudine (d4T) as part of a backbone NRTI regimen. Twenty-four (41%) also received a PI, including 7 (12%) who were receiving both d4T and PI.

TABLE 1: Comparison of HAART Exposure in HIV-infected Children

Thirty-nine children (40%) were not on any treatment at enrollment, including 7 who had been previously treated, but had stopped therapy at least 6 months before entry. Thirty-two children had never been treated with HAART. Clinical features of lipodystrophy were seen in 5 (5%) of 98 children. The CDC clinical and CD4 immunologic categories of the patients are summarized in Table 1. A higher proportion of patients with CDC stage C (AIDS) received PIs as part of their HAART.

As far as anthropometric measurements are concerned, when the untreated group was compared with the reference sample, they were found to have a higher BMI, waist circumference, biceps, triceps, and subscapular skinfolds, but lower arm, thigh, and calf circumferences. Compared with the untreated group (Table 2), children receiving PIs had lower measurements for height, weight, arm, thigh, calf, and waist circumference and lower triceps, subscapular, and suprailiac skinfolds. Children receiving HAART without PIs had increased subscapular and biceps skinfolds and total fat (assessed by the sum of 4 skinfold thickness measurements) when compared with the untreated group, but these differences did not reach significance. This was because of the SDs in the treated group being much larger than the expected SD of unity. Comparing the 23 children receiving d4T in their regimen with the untreated children, the only difference was for suprailiac skinfold (P = 0.02), which was lower in those receiving d4T. Further regression analysis showed that fat redistribution was strongly and independently correlated with the duration of d4T treatment (P < 0.007), but not with the duration of HAART or PIs.

TABLE 2: PI- and Non-PI-treated Groups Compared With the Untreated Group

Nonfasting cholesterol measurements differed significantly between the groups: the untreated group showing lower values than the PI group (P < 0.01) and the non-PI group (P = 0.03). There was no significant difference in triglycerides and leptin values between groups (Table 2).

The results show that antiretroviral therapy appears to have an effect on body fat composition above any changes caused by the disease itself. Children on PI therapy showed a consistent decrease in body weight, BMI, arm circumference, and subcutaneous fat measurements compared with children who were infected with HIV/AIDS but not receiving treatment. Wasting of subcutaneous fat can therefore be seen in this pediatric population in a similar way as adults. The effect of d4T treatment in this group seemed to be less strongly associated with lipodystrophy in the short term compared with adults, although effects were seen with prolonged treatment. Lipohypertrophy in the form of visceral adipose tissue deposition (increased waist circumference) and buffalo hump (increased subscapular skinfold) was not seen to any degree in this age group. There was a trend toward greater subscapular skinfold thickness in children receiving HAART without PI, which may indicate early fat hypertrophy. Obvious clinical symptoms of lipodystrophy were less common in our population compared with adults (5% vs 15%-50%).¹¹ It may be that hormonal factors such as the growth hormone and differences between adults and children in terms of insulin sensitivity mitigate the development of the lipodystrophy syndrome in children.¹²

One limitation of the present study is that the reference population is composed of white European children. Because our population included both sexes and a wide age range, it is not possible to compare population means. The use of z scores, which measure the distance of the value from the age- and sex-appropriate mean, allows valid comparisons of such a diverse group to be made. No published data set of ethnically matched healthy children was found; the only reference data sets that enabled z scores to be calculated were composed of white British or Dutch children.^8-10 Body composition and fat distribution are known to differ between African and white populations¹³ and even between different ethnic groups within the African continent.¹³ Unsurprisingly, our untreated group differed from this reference population, and it is difficult to know what contribution was caused by ethnicity and what was caused by the disease process itself. However, there were significant differences between the untreated group and those receiving HAART, particularly if the regimen included PIs or long-term use of d4T, indicating that the regimen did affect body fat content and distribution. Longitudinal studies are required to differentiate the relative impact of different HAART regimens and to assess the potential for prevention. The use of simple anthropometric measurements to describe body fat distribution will be useful for diagnosis and monitoring of lipodystrophy in resource-poor environments where more sophisticated diagnostic tools, such as magnetic resonance imaging or computed tomography scans, are not available.

Agnieszka B. Dzwonek, MD*

Margaret S. Lawson, PhD†

Tim J. Cole, PhD‡

Vas Novelli, MD§

*Department of Infectious

Diseases and Microbiology

†Childhood Nutrition

Research Centre and

‡Centre for Paediatric

Epidemiology and Biostatistics

University College London Institute of

Child Health and

§Clinical Infectious Diseases Unit

Great Ormond Street Children's Hospital

National Health Service Trust

London, UK