What anthropometric indicators are associated with insulin resistance? Cross-sectional study on children and adolescents with diagnosed human immunodeficiency virus

ABSTRACT BACKGROUND: Studies that test associations between anthropometric indicators and insulin resistance (IR) need to provide better evidence in the context of the pediatric population (children and adolescents) with human immunodeficiency virus (HIV), as anthropometric indicators present a better explanation of the distribution of body fat. OBJECTIVE: To test the associations between anthropometric indicators and insulin resistance (IR) among children and adolescents diagnosed with HIV. DESIGN AND SETTING: Cross-sectional study on 65 children and adolescents (8-15 years) infected with HIV through vertical transmission conducted at the Joana de Gusmão Children's Hospital, Florianópolis, Brazil. METHODS: The anthropometric indicators measured were the abdominal (ASF), triceps (TSF), subscapular (SSF) and calf (CSF) skinfolds. The relaxed arm (RAC), waist (WC) and neck (NC) circumferences were also measured. Body mass index (BMI) was calculated from the relationship between body mass and height. IR was calculated through the Homeostasis Model Assessment for Insulin Resistance (HOMA-IR). Simple and multiple linear regression analyses were used. RESULTS: After adjustment for covariates (sex, bone age, CD4+ T lymphocytes, CD8+ T lymphocytes, viral load, and physical activity), associations between IR and models with SSF and CSF remained. Each of these explained 20% of IR variability. For females, in the adjusted analyses, direct associations between IR and models with ASF (R² = 0.26) and TSF (R² = 0.31) were observed. CONCLUSIONS: SSF and CSF in males and ASF and TSF in females were associated with IR in HIV-infected children and adolescents.


INTRODUCTION
Antiretroviral treatment (ART) and human immunodeficiency virus (HIV) infection itself can cause side effects in individuals with HIV. 1 Among the adverse effects, visible changes in the body such as lipodystrophy syndrome 2 and metabolic changes such as dyslipidemia and insulin resistance (IR) 3 are among the most common adverse effects. IR is defined as lower capacity of insulin to instigate use of glucose by adipose tissue and muscles, or which leads to expansion of pancreatic insulin formation. 4 HIV and continued use of ART are considered to be facilitators for development of IR in the pediatric population. 4 HIV infection, opportunistic infections and intestinal inflammation can culminate in changes to inflammatory cytokines, such as soluble tumor necrosis factor and hormones such as adiponectin and reduced leptin, which impairs glucose homeostasis. 5,6 In addition, changes to CD4+ and CD8+ T-cell functions may impair glycolysis, which may adversely influence glucose metabolism. 6 Specifically, ART protease inhibitors have been associated with hyperglycemia and glucose tolerance in adults diagnosed with HIV 7 and may inhibit the action of glucose transporter (GLUT4), thus resulting in decreased insulin-mediated glucose intake by muscle and adipose tissue. 1 In addition, changes to the body fat distribution pattern may result in changes to the hormonal secretory system of adipose tissue and generate a chronic inflammatory profile, which facilitates IR development. 8 Overweight is among the factors that contribute to the onset of IR in young people without HIV, 8 especially increased body fat.
To assess body fat, anthropometric indicators are commonly used. 9 Different measurements have been directly associated with IR, such as neck circumference, 10  Studies that test associations between skinfolds and IR need to provide better evidence in the context of the pediatric population (children and adolescents) with HIV, as anthropometric indicators present a better explanation of the distribution of body fat.

OBJECTIVE
The purpose of this study was to test associations between anthropometric indicators and IR in a pediatric population (eight to 15 years of age) with HIV.

Participants
Children and adolescents aged between eight and 15 years, with vertical transmission of HIV, were recruited for the study and were followed up clinically at the Hospital Infantil Joana de Gusmão, Florianópolis, Brazil. Eighty-three eligible patients were found. Three patients were excluded from the sample because they presented severe encephalopathy and because they were unable to walk. Three were excluded because we were unable to contact them, four because they had been transferred to another hospital and four because they refused to participate in the research; and another four were losses during the data collection.
The final sample consisted of 65 subjects.
The inclusion criteria were the following: a) presence of information in the medical record to prove that HIV infection had been

Dependent variable
To check IR, we used the Homeostasis Evaluation Model for Insulin Resistance Index (HOMA-IR), calculated through the mathematical model described by Matthews et al. 18 We applied the following equation: HOMA-IR = fasting blood glucose (mg/dl) x insulin (μIU/ml). In the mornings, fasting blood samples (

Statistical treatment
Firstly, descriptive analyses were performed on the data (median and interquartile range). Kurtosis and asymmetry analyses were then used to verify data normality (range from -2 to + 2), 23   There were differences between the sexes, such that the females had higher SSF (P < 0.001) and CSF (P = 0.050) than the males.

Sixty
Regarding physical activity, the male adolescents did more minutes/day than the females (P = 0.022) ( Table 1).
Among the males, direct associations were observed in simple lin- between IR and models with SSF and CSF remained, and each of these explained 20% of the IR variability (Table 2).
For the females, direct associations were observed in simple analyses, such that ASF and TSF explained 20% and 18% of IR variability, respectively. In addition, direct associations with RAC (R² = 0.10) and NC (R² = 0.25) were observed. In the adjusted analyses, direct associations between IR and models with ASF (R² = 0.26) and TSF (R² = 0.31) were observed.

DISCUSSION
The main results from the present study add to the current literature to show that higher values for peripheral and central skin folds are associated with IR.    [25][26][27][28] based on the assumption that accumulation of subcutaneous adiposity is associated with IR due to increased lipotoxicity. 29 In this context, insulin favors entry of glucose into adipose tissue, which activates lipoprotein lipase, thus promoting storage of triglycerides and preventing the action of protein kinase, an intracellular enzyme that is capable of blocking insulin signaling pathways. 29  Although there is no consensus on the association between different lipodystrophy phenotypes and IR in pediatric patients diagnosed with HIV, high insulin concentrations were found previously in children with lipohypertrophy, and less consistently in children with lipoatrophy. 4 The skinfolds associated with IR differed according to sex (male/ female). This may be explained by the existence of sexual dimorphism. In girls, as their age increases, estradiol hormone secretion also increases, which leads to fat accumulation in the arms and consequently increases the amount of adipocytes in the tricipital region. 24 In boys, increasing secretion of testosterone hormone inhibits abdominal fat accumulation. 24 Regarding the associations of anthropometric indicators with IR, the results from this study demonstrated the potential of skinfold analyses, such that associations were found with SSF and TSF among males, and with ASF and CSF among females. This is important from a practical point of view, for clinical use in monitoring the body composition and metabolic complications of HIV-infected children and adolescents, given that skinfold measurement is a low-cost alternative. This study had some limitations, such as the fact that HOMA-IR was used as an indicator of glycemic homeostasis impairment.
Nonetheless, this method is often used in clinical investigations.
Other limitations related to the absence of any clinical diagnosis for lipodystrophy.
Among the strengths of this study, the analyses were controlled for potential confounders (sex, bone age, CD4+ T lymphocytes, CD8+ T lymphocytes, viral load and physical activity) in the multiple linear regression analyses, a strategy that had not previously been addressed in studies making correlations between anthropometric indicators and body fat among children and adolescents diagnosed with IR.

CONCLUSIONS
In conclusion, SSF and CSF in males and ASF and TSF in females were directly associated with IR. It can be suggested that use of these anthropometric indicators should form part of the routine clinical follow-up for HIV-infected children and adolescents.
These low-cost anthropometric measurements can contribute to risk stratification among children and adolescents with IR, and consequently may prevent metabolic complications such as type 2 diabetes and other cardiovascular consequences.