Objective: Children <5 years old are at increased risk of miliary/meningeal tuberculosis, but the immunologic factors that place them at risk are unknown. BCG vaccine protects against miliary/meningeal tuberculosis, but the mechanism of protection is unknown. We assessed for abnormalities in immune response associated with miliary/meningeal or pulmonary tuberculosis in young children.

Patients and methods: We conducted a case-control study among HIV-seronegative Brazilian children who were <5 years old. Case subjects had previous culture-confirmed or clinical miliary/meningeal tuberculosis. There were 2 sets of control subjects: those with culture-confirmed pulmonary tuberculosis and purified protein derivative-positive household contacts. All of the children had completed treatment. Peripheral blood mononuclear cells were stimulated (phytohemagglutinin, phytohemagglutinin + interleukin 12, lipopolysaccharide, lipopolysaccharide + interferon-gamma, and purified protein derivative), and cytokine responses (interleukin 1beta, interleukin-4, interleukin-6, interleukin-8, interleukin 10, interleukin 12, interferon-gamma, tumor necrosis factor-alpha, and monocyte chemoattractant protein 1) were quantified by bead-based assay. Median cytokine responses were compared by the Kruskal-Wallis test. Multivariate analysis of variance accounted for multiple comparisons.

Results: There were 18 case subjects with miliary/meningeal tuberculosis, 28 pulmonary control subjects, and 29 purified protein derivative-positive control subjects. The median age was 4.2 years. There was no difference in case and control subjects by age, gender, race, BMI, or median CD4 count. Twelve (67%) of 18 case subjects, 26 (93%) of 28 pulmonary control subjects, and 28 (97%) of 29 purified protein derivative-positive subjects had received BCG vaccine. No cytokine defects were identified in case subjects with miliary/meningeal tuberculosis compared with either set of control subjects. Pulmonary control subjects had uniformly higher monocyte chemoattractant protein 1 levels than case subjects with miliary/meningeal tuberculosis and purified protein derivative-positive control subjects, both at rest and with lipopolysaccharide, lipopolysaccharide + interferon-gamma, and purified protein derivative stimulation. Pulmonary control subjects did not have a higher frequency of allele G in the -2518 monocyte chemoattractant protein 1 promoter polymorphism. Case subjects with miliary/meningeal tuberculosis who had received BCG vaccine (n = 12) had lower stimulated interleukin 8 production than children who did not receive BCG vaccine (n = 6).

Conclusions: Children with previous miliary/meningeal tuberculosis did not have a major defect in the cytokine pathways studied. Increased monocyte chemoattractant protein 1 levels were associated with pulmonary disease, occurred despite BCG vaccination, and were not associated with a polymorphism in the monocyte chemoattractant protein 1 promoter.