BCG vaccination in children born to HIV-positive mothers [letter]

Dr. Athale and colleagues from Lusaka (Aug 15 p434) have suggested that in countries of high HIV prevalence BCG (bacillus Calmette-Guerin) vaccination should be given only to children over 12 months of age because of the risk of disseminated BCG infection. By that age it should be clear which children are infected. In a prospective study in Zaire of children born to HIV positive and negative mothers (seroprevalence 3.8% vertical transmission rate 23%) all received BCG at birth. During follow-up of 2 years 9 of 21 HIV-infected children developed tuberculosis which was disseminated in only 1 and 7 had a family contact. None had BCG adenitis. All responded well to short- course chemotherapy (avoiding thiacetazone). In the 2 matched control groups uninfected by HIV 12 developed tuberculosis--4 born to 21 HIV- positive mothers and 8 born to 21 HIV-negative mothers. None of these had disseminated infection and 1 had BCG adenitis (mother seronegative). The HIV-infected and noninfected groups did not differ significantly. In the past 10 years we have had only 1 case of probable disseminated BCG infection in a child of 5 months. He was vaccinated at 3 months and developed suppurating axillary BCG adenitis. His weight curve was static and he had persistent fever. Radiography showed typical miliary lesions and he responded well to short-course therapy. His mother was seropositive for HIV and he had no family tuberculosis contacts. We agreed that many cases of tuberculosis in children infected with HIV simulate the adult form with cavitation and extensive pneumonias but only in children over 2 years old. We suggest that this finding is probably due to progression of the primary focus rather than reactivation or progression of BCG vaccination. Such children respond well to short-course therapy but tend to relapse and have persisting problems because of lung damage (bronchiectasis). HIV infection in infants has a bimodal presentation. The early presenters will have symptoms by 12 months but only a few late presenters will have them. We would not expect nursing staff at a busy clinic or vaccination center to have time to pick out these with mildly symptomatic HIV infection. Although there is some evidence that BCG at 3 months produces a larger induration more frequent scar formation and less lymphadenopathy than if it is given at birth we suggest that in African countries this is not the time to change the policy of giving BCG near birth. Attendance at vaccination clinics falls off strikingly after the measles vaccine at 9 months and therefore many healthy children would not receive BCG. However a good case could be made for chemoprophylaxis for children for HIV-positive mothers who have had active tuberculosis in the previous 2 years. (full text)