Role of infant feeding and HIV in a severe diarrhoea outbreak in Botswana

A poster from Tracy Creek and coworkers (also presented in a special session prior to the conference), reported findings from an investigation conducted following a severe outbreak of diarrhoea in children, coinciding with heavy rains and widespread contaminated water in Botswana in 2006. [2]

Ministry of Health surveillance data reported 3.5 times the usual number of cases and 25 times the usual number of deaths between January and March (35,046 cases and 532 deaths compared to 9,166 and 21 in 2005). Cryptosporidium and enteropathogenic E. coli were the most common pathogens identified.

In 2005 surveillance data, 33.4% pregnant women were reported to be HIV-positive. The national programme for prevention of mother-to-child transmission (PMTCT) provides free infant formula and advises HIV-infected women not to breastfeed. Additionally, a proportion of HIV-negative women do not breastfeed. An estimated 35% of all infants under the age of 6 months are not breastfed.

In this study, the investigators focused on a paediatric ward in a hospital in Francistown (the countrys second largest referral hospital). All pediatric in-patients aged less than 5 years old with diarrhoea in hospital from February 20 to March 10, 2006 were followed by chart review until April 10. Their mothers were interviewed, and mothers and infants were tested for HIV.

The investigators found, of 153 diarrhoea in-patients, 96% were under 2 years old, 90% of those were not breastfed in the two weeks before illness, and 33 (22%) died. Among patients with HIV results (94% tested), 23 of 131 (18%) were HIV-positive and 93 of 144 (65%) had an HIV-positive mother. Malnutrition was common: of 99 children, 51 (52%) with growth records exhibited no growth in the previous 3 months; severe acute malnutrition (kwashiorkor or marasmus) developed in 38 of 147 (26%) during or after the diarrhoea.

Kwashiorkor was a significant risk factor for death (RR = 2.0, 95%CI 1.1 to 3.8) and tended to be more common in HIV-negative children (27% vs 9%, RR 3.0, 95% CI 0.75 to 12.5). No breastfeeding children died (RR 6.1, 95%CI 0.39 to 93.4). HIV in mother or child was not a risk factor for death.

Infants who died while receiving formula through the PMTCT programme had received an average of 51% of the formula they should have done from the clinics after birth, and 60% had not gained weight in the 3 months prior to illness.

Overall mortality was high and increased the national mortality rates in under-five-year-olds, for the year by at least 18%. Nearly all the severe morbidity and mortality was among HIV-negative children who were not breastfeeding and not growing well.