Prevalence and Risk Factors of Malnutrition among HIV-Infected Children Aged 2-18 Years : A Cross-Sectional Study

Background: Malnutrition and underweightness are common problems among HIV-infected children in low and middle income countries. Methods and findings: A single-centered, retrospective, cross-sectional cohort survey was conducted between January 1, 2014 and June 30, 2014 to assess the prevalence and risk factors of malnutrition among HIV-infected children aged 2-18 years at the Bamrasnaradura Infectious Diseases Institute (BIDI), Thailand. Clinical and laboratory data were compared between a malnutrition group, (Group A): body mass index (BMI) for age <5th percentile and a nonmalnutrition group (Group B): BMI-for-age >5th percentile and <95th percentile. One hundred and ninety-five HIVinfected children were retained for analyses. The prevalence of malnutrition was 43% (83 of 195). The following proportions were higher in Group A than B, respectively: advanced clinical staging (83% vs. 12%), PI-based HAART (73% vs. 38 %) and anaemia (48% vs. 8%) (all p<0.05). The median of Log10 plasma HIV RNA was higher in Group A than B (1.7 vs. 1.6, respectively; p<0.05). The proportions of RNA ≤ 40 copies/mL and parent caregivers in Group A were lower than B (48% vs. 86% and 35% vs. 54%, respectively; all p<0.05). Other factors did not differ significantly. In multivariate analysis, malnutrition was significantly associated severity (WHO stages 3 and 4) (OR=184.50, 95% CI=[35.72-952.73]), PI (OR= 5.38, 95% CI=[1.97-14.67]) and parent as caregivers (OR= 0.56, 95% CI=[0.34-0.92]). No significant differences between Group A and B were found in Log10 plasma HIV RNA, RNA ≤ 40 copies/mL, hemoglobin and anemia. Conclusion: The prevalence of malnutrition was 43%. Malnutrition among HIV-infected children was associated with the severity of disease, PI-based HAART and nonparent caregiver.


Introduction
Malnutrition is estimated to contribute to more than 20 million childhood morbidity and mortality [1]. Weight loss and undernutrition are common in people living with HIV/AIDS, likely accelerating disease progression, increasing morbidity and reducing survival [2]. HIV and malnutrition are common comorbidities in low and middle income countries [3]. An African study demonstrated a prevalence of malnutrition among HIV-infected children of 42% [4]. The nutritional statuses of children living with HIV/AIDS are affected by frequent or persistent diarrhoea, pneumonia and tuberculosis (TB) [3]. In HIV/AIDS patients, malnutritional deficiencies and medical conditions are the major causes of morbidity and mortality [5][6][7][8]. Malnutritional worsens HIV disease by damaging lymphoid tissues and decreasing CD4 T-cell concentration, both of which weaken the immune system, especially in HIV-infected children [9]. Moreover, malnutrition leads to irreversible cognitive impairment, chronic diseases and growth failure [10].
In research and practice, a body mass index (BMI) cut-off of 18.5 kg/m 2 is used to identify malnutrition or underweightness in adults living with HIV [11,12]. Malnutrition or underweightness in children is defined as a BMI-for-age <5 th percentile at 2-18 years old [13]. Malnutrition and underweightness among HIV-infected children have been associated with greater risk of hospitalization, loss to follow-up (LTFU), TB, drug interaction, severe disease and anaemia [14][15][16][17][18]. In Malawi, acute malnutrition (26.6%) was the most common primary admission diagnosis, especially for HIVinfected children less than 3 years old and associated with longer admission [19]. The linkages between HIV/AIDS and malnutrition are many, strong and long-lasting through a vicious cycle of immune dysfunction, infectious disease and malnutrition [20]. TB is a most deadly infectious disease. A variety of causes increase susceptibility to Mycobacterium tuberculosis infection such as HIV/AIDS, malnutrition, very young age and drug [21]. Several studies have demonstrated the prevalence of anemia among HIV-1 infected children, which varies from 16 to 94%. This increases with advancing stage of HIV disease, but varies with sex, age, as well as the definition of anemia used [22][23][24]. In a meta-analysis of over 2000 HIV-infected children, the prevalences of mild (Hb 10-12.0 gm/dL) and moderate anemia (Hb 8.0-9.9 gm/dL) were 22%-94% and 3%-82%, respectively [25]. In Mozambique, there was 68% malnutrition and 51% WHO clinical stages 3 or 4 among all HIVinfected children under 15 years old [4]. Weight loss is associated with the progression of HIV infection, but it is unclear whether it causes progression or is a disease severity marker [26]. Distance from clinic causes LTFU, affecting disease severity. In Cameroon, retention in care was associated with good access to care and was defined as living ≤ 40 km of the clinic site [27].
The data of malnutrition or underweightness important for management are limited. Therefore, we evaluated the prevalence and risk factors of malnutrition among HIVinfected children 2-18 years old at the Bamrasnaradura Infectious Diseases Institute (BIDI), Thailand.

Study design
We conducted a retrospective, cross-sectional cohort survey of HIV-infected children seen at the BIDI, Thailand between January 1, 2014 and June 30, 2014. Participants were HIVinfected children aged 2-18 years with confirmed HIV-infection by a positive serological test. Children were excluded for missing data, no antiretroviral drug and a follow-up period ≤ 6 months. The primary objective was the prevalence of malnutrition .The secondary objective was identification of the factors associated with the malnutrition group (Group A) and non-malnutrition group (Group B). Group A was defined as BMI-for-age <5 th percentile and Group B as BMI for age >5 th percentile and <95 th percentile. We extracted all laboratory and clinical data, including established age, sex, weight, height, BMI, history of household TB contact, previous TB disease, adherence, disclosure or non-disclosure, WHO clinical staging, receipt of highly active antiretroviral therapy (HAART), CD4 cell counts, HIV RNA level, Hb, anaemia, cholesterol, triglyceride, death, study status, distance of residence and caregiver. All data were obtained from medical records at the first visit. Results of CD4, viral load and other laboratory tests were collected during survey. All children had vertical transmission.
BMI is the weight of a person in kilograms divided by the square of the height in meters. BMI for children and teenagers is age and sex-specific and is often referred to as BMI-for-age (2-18 years old). The BMI-for-age percentile growths charts are the most commonly used measures of size and growth patterns in children and teenagers.

HIV testing
Children over 18 months old were tested by serology only, which classified children as infected or uninfected. Children under 18 months old were also initially tested serologically. Afterwards positive and negative samples from HIV-exposed children were tested for DNA by PCR on dried blood spots. Those PCR detected viral DNA were classified as infected, and those PCR was negative were classified as uninfected.

Study population and sample size calculation
All continuous data were compared with the Student´s t-test and the Mann-Whitney U test. Categorical data were compared with the chi-square test. A p-value of <0.05 was considered statistically significant. A multinomial regression model studied the factors associated with malnutrition. All variables with pvalue <0.05 in bivariate analyses were selected for multivariate analyses. Factors considered were age, sex, history household TB contact, previous TB disease, a proportion of adherence ≥ 95%, disclosure, WHO clinical staging, the proportion of protease inhibitor (PI)-based HAART, CD4 cell counts, HIV RNA level, a proportion of RNA ≤ 40 copies/mL, Hb, anaemia, cholesterol, triglyceride, death, study status, distance of residence and caregiver during survey. In BIDI, the majority of HIV infected children are adolescents. Therefore, the sample size was calculated by a proportional formula assuming prevalence of malnutrition of 22.2% among HIV infected children [30], the calculated minimum sample size for the malnutrition group was 67.
The study protocol was reviewed and approved by Institutional Review Board of Bamrasnaradura Infections Diseases Institute, Ministry of Public Health, Thailand. Figure 1 shows a flow chart of the enrolment. Two hundred and eight children were enrolled. Thirteen (6%) were excluded: 7 follow-up for less than 6 months, 3 overweight and 3 no antiretroviral drug. One hundred and ninety-five were included in the final analysis. Of these, 83 (43%) were malnutrition or underweightness (Group A) and 112 (57%) were nonmalnutrition (Group B). Table 1 displays the demographics and characteristics of participants. The majority were male in Group A but female in B. The mean ages of Group A and B were 14.56 and 13.64 years, respectively. Sex and age of both groups were not significantly different (p<0.05).

Clinical and laboratory risk factors of study children
The risk factors of the participants are shown in Table 2. The following proportions were higher in Group A than in B, respectively: advanced clinical staging (83 vs. 12%), PI-based HAART (73% vs. 38 %) and anaemia (48% vs. 8%) (all p<0.05). The median of Log10 plasma HIV RNA was higher in Group A than B (1.7 vs. 1.6, respectively; p<0.05). The proportions of RNA ≤ 40 copies/mL and parent caregivers in Group A were lower than B (48% vs. 86%, respectively and 35% vs. 54%, respectively; all (p<0.05). No significant differences between Group A and B were found in history household TB contact, previous TB disease, the proportion of adherence ≥ 95%, mean CD4 levels, cholesterol, triglyceride, death, study status and distance of residence (all p>0.05).
Factors associated with malnutrition are shown in Table 3. In multivariate analysis, malnutrition was significantly associated with HIV disease severity (WHO stages 3

Discussion
We have demonstrated a prevalence of malnutrition as high as 43% among HIVinfected children. This finding is similar to a previous study. It demonstrated prevalences of malnutrition among HIV-infected children in central and West Africa aged 5-10 years and aged 10-19 years were 36 and 44%, respectively [4]. In contrast, a previous study in adolescents living with HIV/ AIDS showed a prevalence of malnutrition as low as 9.9% [31]. Nevertheless, studies are difficult to compare due to the differences in study populations, study designs and definitions of malnutrition. We have found that malnutrition among HIVinfected children was associated with the HIV/AIDS disease severity (advanced WHO clinical stagings), PI-based HAART and non-parent caregiver. The risk of malnutrition in HIV infection increases, because high pro-inflammatory cytokine activity can cause growth impairment among children [32]. Children or adolescents living with HIV/AIDS may acquire many infections that are conducive to malnutrition, underweightness, morbidity and mortality. Malnutrition also worsens HIV disease due to lymphoid-tissues breakdown and decreased CD4 T-cell concentration [4,9]. It is noteworthy that 83% of malnourished children in the present study were classified as WHO clinical stages 3 or 4 because BIDI is a HIV referral center. This contrasts to a previous study in which only 30% of malnourished children were WHO clinical stages 3 or 4 [3]. We found PI-based HAART was associated with malnutrition because PI is a second line drug regimen in Thailand, which is related to advanced stages of HIV disease.
We have demonstrated a low the proportion of parent as caregiver was associated with malnutrition, which is similar to other previous studies among HIV-infected children in sub-Saharan Africa [33][34][35]. In contrast, a previous study demonstrated that the proportions of parent caregivers or orphan status were not associated with malnutrition [4].
In our study, no significant differences between Group A and B were found in sex, age, history household TB contact, previous TB disease, the proportion of adherence ≥ 95%, disclosure, mean CD4 levels, cholesterol, triglyceride, death, study status and distance of residence (all p>0.05). In multivariate analysis, no significant differences between Group A and B were found in Log10 plasma HIV RNA, RNA ≤ 40 copies/mL, Hb and anemia (all p>0.05). However, a previous study demonstrated that boys were more likely to be malnourished than girls. The reasons for this association between gender and malnutrition remain unclear and need further study [4,36]. In Kenya, a further association between distance of residence and LTFU was report [37]. This may be related to severity of disease and malnutrition. However, we have found distance of residence did not affect malnutrition, which was similar to the finding of a Kansas City study in the USA [38]. Several studies have reported anaemia was associated with advancing stage of HIV disease [22][23][24]. The most important cause of anemia in HIV-infected children is insufficient production of erythrocytes caused by ineffective erythropoiesis, reduced erythropoietin production, associated infections, neoplasia, medications, and micronutrient deficiencies [39].
Our study has several limitations. Firstly, the study was a retrospective cross-sectional cohort study and short duration survey. Some data are missing such as timing from onset of diagnosis to initiating HAART, protein calorie malnutrition and data of LTFU along with the levels of macronutrients, trace minerals and microelements. Secondly, the sample size was small and possibly unrepresentative of the population of children living with HIV/AIDS in Thailand. Children included in our study were only those accessing the pediatric HIV service at BIDI, a referral centre. We had to exclude children less than 2 years old because we could not evaluate malnutrition or underweightness by BMI-for-age.

Conclusion
In conclusion, the prevalence of malnutrition among HIVinfected children aged 2-18 years was 43%. Malnutrition among HIV-infected children was associated with the severity of disease, PI-based HAART and non-parent caregiver.