Watch for malnutrition risk in children with HIV after starting ART

One in nine HIV-infected children with advanced illness was hospitalised with severe malnutrition within 12 weeks of starting antiretroviral and these children had a 15-fold increased risk of dying within the first six months compared to those children not hospitalised, Andrew Prendergast and colleagues reported in the ARROW study published in the advance online edition of AIDS.

The high frequency of new cases of severe malnutrition in children after starting antiretroviral therapy (ART) is not easy to explain say the researchers, and they warn that health care workers need to be alert for the condition, especially in children who are already malnourished.

Treatment programmes for children also need to integrate HIV and malnutrition care, and look at whether nutritional supplementation before starting treatment can protect against development of severe malnutrition once ART begins.

Antiretroviral Research for Watoto (ARROW) is an open-label, randomised trial of induction-maintenance and monitoring strategies for antiretroviral treatment in 1207 HIV-infected children with a median age of 6 (3-17) in three hospitals in Uganda and one in Zimbabwe recruited from March 2007 to November 2008.

HIV infection and malnutrition among children in resource-limited settings cannot be separated. HIV prevalence is high in children with severe malnutrition. The risks of dying are three times greater in HIV-infected children with severe malnutrition than in uninfected children with severe malnutrition.

In resource-poor settings HIV infected children are likely to be seriously ill and severely malnourished when they present for care. Studies show 5-10% early death in children starting ART in these settings

The two primary clinical manifestations of severe malnutrition are:

• non-oedematous (no swelling) malnutrition (marasmus) and
• oedematous (oedema or build up of swelling in the tissues or circulatory system) malnutrition (kwashiorkor and marasmic kwashiorkor).

The authors note that while marasmus is most often associated with HIV-infected children anecdotal reports suggest that oedematous malnutrition can occur in children soon after starting ART.

The authors wanted to better understand the effect of SM on early death in children soon after starting ART. They looked at the frequency of hospitalisations and change in CD4 percentages 12 weeks after the start of ART and death and change in Z score by six months.

Severe malnutrition (SM) in this study was defined as one or more of:

• Kwashiorkor: 60-80% of weight for age with oedema
• Marasmus: <60% of weight for age without oedema
• Marasmic kwashiorkor: <60% of weight for age with oedema.

Children were assessed for SM before enrolment. Children with SM before enrolment got 2-8 weeks of supplementary feeding.

At enrolment children began cotrimoxazole prophylaxis and started abacavir, lamivudine and either nevirapine or efavirenz. Children randomised to an induction-maintenance therapy also got zidovudine in the first 36 weeks. Children were followed up every four weeks for clinical evaluation, height and weight and CD4 count and percentage measurements.

After starting ART children were hospitalised for SM if they developed oedema, loss of appetite and/or concurrent infections needing treatment.

No child had oedema before starting ART.

3.2% (39) of children were hospitalised for SM, 20 with oedema (11 kwashiorkor and 9 marasmic kwashiorkor); 19 had marasmus.

The median time from starting ART to hospital admission for those with swelling (oedema) was 26 days (IQR: 14-56) and for those without  (marasmus) was 28 days (IQR: 14-36).

74% (29) of children with SM admitted to hospital had underlying infections. Similar proportions were seen in children with oedematous malnutrition or with non-oedematous malnutrition.

Children hospitalised for SM had significantly lower baseline weight-for-age, height-for-age, weight-for-height and mid-upper arm circumference than those not admitted. 

Among the 220 (18%) children with advanced illness 7.3% (95% CI: 3.8-10.7) developed kwashiorkor and 3.6% (95% CI: 1.2-6.1) developed marasmus within three months.

The authors highlighted that over half of the children hospitalised for SM developed oedema after starting ART. While marasmus is more common in HIV-infected children their findings support anecdotal evidence from sub-Saharan Africa suggesting starting ART may play a role in this contrary finding.

They suggest that since many children in resource-poor settings will start ART at an advanced stage it is wise for treatment programmes to anticipate oedematous malnutrition.

While the reasons for the onset of kwashiorkor are unclear, the authors offer several suggestions, which may not be mutually exclusive:

1)     The decline in immune activation and increase of CD4 cell count after starting ART may result in the development of oedema.

2)     The onset of kwashiorkor shortly after starting ART may be yet another presentation of immune reconstitution inflammatory syndrome (IRIS).

3)     Onset of oedema may be a manifestation of refeeding syndrome, a range of metabolic and physiological disturbances that can occur when food is introduced after a period of starvation. Refeeding syndrome might occur as a result of an increased appetite after starting antiretroviral therapy, itself a widespread phenomenon according to anecdotal reports.

4)     It may be a form of ART toxicity specific to malnourished children. Severe manifestations of well-known toxicities are more common in HIV-infected people with very low nadir CD4 cell counts. In those with oedematous malnutrition the median CD4 count was 2.5, an extremely low level.

Death rates among children hospitalised with SM were high. At six months the death rates were 32%, 20% and 1.7% among children hospitalised with marasmus, kwashiorkor and not hospitalised, respectively (p<0.001).

“There is an urgent need to understand better the complex interplay between malnutrition, infection and immunodeficiency” the authors note.

The authors conclude “integration of HIV/malnutrition services and further research to determine optimal ART timing, role of supplementary feeding and antimicrobial prophylaxis are urgently required.”

Prendergast A et al. Hospitalisation for severe malnutrition amongst HIV-infected children starting antiretroviral therapy in the ARROW trial. AIDS, Advance online publication, March 2011 DOI: 10.197/QAD.0b013e328345e56b