HIV-exposed infants with acute respiratory failure secondary to acute lower respiratory infections managed with and without mechanical ventilation.

OBJECTIVES The decision to provide mechanical ventilation (intermittent positive pressure ventilation (IPPV)) to HIV-exposed infants in resource-poor settings has remained difficult owing to problems in confirming HIV infection and the lack of data on outcome. We evaluated the predictive value of the HIV antibody test in confirming infection in infants requiring mechanical ventilation for acute lower respiratory infections (ALRIs), and compared the outcome for children denied access with the outcome for similar subjects who were ventilated. SETTING AND DESIGN This investigative study was conducted over a 12-month period at the paediatric intensive care unit (PICU) at King Edward VIII Hospital (KEH) in Durban, and at Ngwelezana Hospital in northern KwaZulu-Natal. SUBJECTS HIV-exposed patients with acute respiratory failure (ARF) secondary to ALRI entering the PICU at KEH were enrolled into the IPPV arm, while similar children who were refused such care at Ngwelezana Hospital were admitted into the non-IPPV arm. Standardised protocols for entry and management of enrolled subjects were utilised. OUTCOME MEASURES HIV DNA polymerase chain reaction (PCR) testing was performed to establish HIV status. Clinical and laboratory parameters were correlated with HIV status to determine predictors of infection and outcome (survival to discharge). RESULTS One hundred and sixteen HIV-exposed infants were enrolled, 49 into the IPPV arm and 67 into the non-IPPV arm. The median age of both groups was 3.0 months (0.5-11 months), and the male/female ratio and proportion of infants under 3 months of age were similar in both groups. The predictive values of the HIV antibody test in determining HIV infection in the IPPV and non-IPPV arms were 87.8% and 85.0% respectively. Splenomegaly and a serum globulin of > 35 g/l increased the likelihood of being HIV PCR-positive (p = 0.006 and p = 0.04 respectively). Survival to discharge rates for HIV-infected children in the IPPV and non-IPPV arms were 41.9% and 24.6% respectively (p = 0.08). Age less than 3 months (p = 0.04) and very severe pneumonia (p = 0.007) were the only indicators of poor outcome. CONCLUSION Mechanical ventilation provided little benefit in HIV-infected children with ARF from ALRI. An HIV antibody test in infants with ALRI and ARF is highly suggestive of HIV infection. Splenomegaly and a serum globulin of greater than 35 g/l were the only useful markers in identifying HIV infection.