Water treatment does not reduce diarrhoea in infants born to Kenyan mothers with HIV

Further evidence that the provision of clean water is not enough to prevent weaning-associated diarrhoea among infants born to HIV-positive mothers who are weaned early in resource-poor settings was presented in a study by Julie Harris and colleagues in the September 16^th edition of The Journal of Infectious Diseases with an accompanying editorial by Louise Kuhn and Grace Aldrovandi.

Access to clean water is a major public health concern in resource-poor settings, and is of particular concern for the survival of children. Diarrhoea is one of the top three leading causes of infant mortality in these settings.

Interventions to improve water quality and hygiene practices have been introduced to help reduce the high rates of infant mortality. Kuhn and Aldrovandi stress that this takes on special significance for maternal HIV infection, where in the hopes of preventing HIV transmission the majority of women with HIV in resource-poor settings choose either not to breastfeed or else wean early.

While breastfeeding accounts for up to half of all cases of mother-to-child HIV transmission, shortening the duration of breastfeeding and substituting replacement feeding has resulted in increased or equal rates of infant mortality when compared to exclusive breastfeeding, as well as higher rates of diarrhoea and gastroenteritis.

Mixed feeding, in which infants are exposed to breast milk and to other liquids, as well as solid food and formula feed or cow’s milk, is associated with a higher rate of HIV transmission, probably due to exposure to allergens that irritate the gut and lead to inflammation, thus increasing the risk of HIV infection from breast milk.

The Zvitambo study, published in 2005, showed that infants who received mixed feeding were four times more likely to become infected by six months of age than infants who were exclusively breastfed.

Such increases are believed to be the result of reduced transmission of protective maternal antibodies, as well as increased exposure to contaminated water and complementary weaning foods.

In response to these outcomes the World Health Organization recommends exclusive breastfeeding for six months followed by rapid weaning if subsequent replacement feeding is affordable, feasible, available, safe and sustainable (AFASS) to reduce HIV transmission from mother to child. If not, then the recommendation is to continue breastfeeding beyond six months.

The water safety study

Begun in July 2003, the Kisumu (Kenya) Breastfeeding Study (KiBS) was a clinical trial designed to evaluate if antiretroviral treatment during the six-month postpartum period helps to prevent the transmission of HIV through breastfeeding.

Provision of antiretroviral treatment began during pregnancy and continued for the six months following delivery. Infants were given a single dose of nevirapine within 72 hours of birth. Women were encouraged to exclusively breastfeed for 5.5 months and then rapidly wean their infants from breast milk by six months of age (within two weeks). Preliminary results indicated that ART was highly effective.

However, 18 months into the study routine safety monitoring discovered higher than expected rates of diarrhoea among infants after weaning. Enrolment stopped and the data were reviewed.

To address the issue a household-based water quality intervention, Safe Water System (SWS), was introduced. Field workers provided mothers in the KiBS cohort with safe water storage vessels, instructions on food and water hygiene and on washing their hands with soap, and sodium hypochlorite (WaterGuard) for household water treatment. Instruction and replacements as needed were provided on an ongoing basis at no cost.

In field trials, the use of SWS has resulted in 25 to 85% reduction in the risk of diarrhoea. The World Health Organization recommends its use for people living with HIV and their households.

The authors compared incidence of diarrhoea in infants enrolled prior to the intervention from August 2003 to March 2005 (cohort A) and in a second cohort enrolled from August 2005 to January 2007 (cohort B).

Routine clinic and home visits were scheduled, beginning at one week of age at home and two weeks of age at the clinic, and then alternating week by week until 48 weeks of age. Free care was provided and all travel expenses reimbursed. Mothers were encouraged to bring their infants for medical attention as needed. Stored water was tested regularly from November 2006 to November 2007 for the presence of chlorine to indicate adherence to recommended SWS use.

A total of 234 cohort A infants and 257 cohort B infants were included in the analysis. Over 90% of the households in both cohorts had access to water supplies other than surface water.

230 infants from cohort A were included in the diarrhoea analysis and 252 from cohort B. Most infants did not have clinic visits for diarrhoea during the exclusive breastfeeding period, the weaning or post-weaning period.

However, frequency of clinic visits was highest during the weaning period in both cohorts. Sixteen per cent of cohort A infants and 16% of cohort B infants had ≥ 1 clinic visit for diarrhoea (P=0.89) during this month.

The introduction of SWS did not decrease the risk of diarrhoea among infants of HIV-infected mothers during this time of rapid and early weaning in spite of high adherence rates.

While there is limited understanding of the causes of infant diarrhoea, the authors note studies have indicated weaning as a risk factor. In addition other studies, the authors stress, have shown how both exclusive and non-exclusive breastfeeding are protective against severe diarrhoea.

Crucially breastfeeding, they argue, protects infants not only from exposure to bacteria found in water and food but also confers important protective immunologic benefits to fight illness as well as respiratory infection.

Louise Kuhn and Grace Aldrovandi note contaminated water is unlikely to be responsible for the latter and “further supports the idea that breastfeeding protects less by keeping the germs out and more by putting the good stuff in”. Weaning foods have been shown to contain bacteria, but many children with diarrhoea show no evidence of bacterial infection, suggesting that change of diet in itself is sufficient to cause serious diarrhoea.

In addition the authors note stopping breastfeeding has been associated with high rates of diarrhoea among infants of both HIV-positive and negative mothers. There are no known differences in the immunologic quality of breast milk in HIV-positive and negative mothers.

The authors note that while the risk of diarrhoea was similar in both cohorts at weaning, the risks before and after weaning were significantly lower in cohort B than in cohort A. They believe that greater exposure to exclusive breastfeeding in cohort B before weaning may have conferred additional protective benefits. SWS during non-weaning periods may also have reduced risk.

The authors suggest that a randomised controlled trial of the effectiveness of SWS on diarrhoea in rapidly weaned infants might provide further insight into its effectiveness in this age group.

Limitations cited by the authors include possible differences in the two cohorts due to one preceding the other by almost two years. Laboratory data on the causes of diarrhoea might have been helpful in better understanding the routes of transmission. The ability to appropriately assess the effect of SWS was limited as the study was not designed to determine the risk factors for transmission of diarrhoea in infants. And, data on SWS adherence was incomplete due to lateness of protocol approval.

An accompanying editorial by Louise Kuhn and Grace Aldrovandi supports the findings of the study that early weaning increases the risk of diarrhoeal disease among infants of HIV-positive women and underscores the importance of the protective benefits of breastfeeding.

Safe water interventions within this specific context can be beneficial, but as an addition to breastfeeding and not as a replacement, they argue. They stress the limitations of some public health interventions as revealed by the authors and note that few safe water interventions have focused on children under the age of one year and none on children not being breastfed. “Children are not little adults”, they caution.

Kuhn and Aldrovandi support the authors’ conclusion “that further investigation of extended maternal antiretroviral therapy is needed to address the longer duration of breastfeeding required in pathogen-rich environments”.

However, they add that in terms of proof of concept the capacity of extended maternal ART and extended infant prophylaxis to prevent HIV transmission after delivery has been demonstrated already. This also presents a means of decreasing the risks of HIV transmission through breast milk while retaining its benefits.

The problems associated with early weaning, known for many years among infants of uninfected women, will now they hope be rediscovered among infants exposed to HIV. This fact, they believe, provides an added incentive to make antiretroviral therapy available for all pregnant and lactating HIV-positive women.