Results

Among the 180 mother–newborn pairs analysed were 149 babies born at term with adequate birthweight (above the 10th centile), six born at term of low birthweight, and 25 preterm (less than 37 weeks of gestation). All neonates were singletons. Table 1 shows that there was a higher incidence of both maternal HIV infection and placental malaria infection in mother–newborn pairs where the baby was either term and of low birthweight, or preterm. Maternal HIV infection and placental malaria coexisted in only 4% of pairs of term normal birthweight babies, but in 33.3% of term babies with low birthweight, and in 16% of preterm neonates. All HIV infected mothers were asymptomatic (category A, according to the Centers for Disease Control classification).

Mean total serum IgG and measles antibody titres were significantly higher in HIV infected mothers compared with those not infected. Mean serum S pneumoniae and tetanus toxoid IgG antibody titres were similar in HIV and non-HIV infected mothers (table 2). Mean maternal total serum IgG and antibody tires to S pneumoniae, measles, and tetanus toxoid were not significantly different between mother–cord pairs with or without placental malaria infection (table 3). These results refer to pairs with neonates born at term and with adequate birthweight, but similar figures were obtained when the whole sample was analysed (data not shown).

Placental transfer of specific IgG antibodies to S pneumoniae was reduced by placental malaria, maternal hypergammaglobulinaemia, and maternal HIV infection to 82%, 79%, and 86%, respectively. Placental transfer of specific IgG antibodies to measles was reduced by placental malaria to 81% and maternal hypergammaglobulinaemia to 87%. Maternal HIV seropositivity was not associated with reduced transfer of measles antibody. Placental transfer of specific IgG antibodies to tetanus toxoid was not reduced by any of the variables analysed.

Specific maternal IgG antibody titres were always positively associated with the corresponding cord antibody titres. Parity, prematurity, or previous malaria treatment did not influence the placental transfer of specific S pneumoniae, measles, or tetanus toxoid IgG antibodies, and these variables were excluded from the regression analyses. Table 4 shows the proportional reduction in neonatal specific IgG antibody titres due to placental malaria infection, maternal HIV infection, and high maternal total serum IgG titres. The estimated reduction is the ratio of the expected cord titre when maternal serum IgG is 22.9g/l to the expected cord titre when maternal serum IgG is 14.9 g/l; these total serum IgG titres are the third and first quartiles, respectively, in mothers for this sample. Thus a 1.5-fold increase in maternal total serum IgG titre is associated with a decrease in S pneumoniae IgG antibody titre in the neonate to 86%. Mean neonatal total serum IgG and serum IgG antibody titres toS pneumoniae, measles, and tetanus toxoid were not significantly different (p>0.05) between neonates born to HIV and non-HIV-infected mothers (table 2). Similarly, there were no significant differences between the mean neonatal total serum IgG and serum IgG antibody titres to S pneumoniae, measles, and tetanus toxoid when mother–cord pairs with and without placental malaria infection were compared (table 3). Here again, the data refer to pairs with neonates born at term and with adequate birthweight; results were comparable when the whole sample was analysed (data not shown).

Discussion

Our aim was to assess the influence of HIV infection and placental malaria on placental IgG antibody transfer. S pneumoniae, measles, and tetanus were the antibody specificities chosen because these infectious agents are important causes of morbidity and mortality in infants.

Although measles and tetanus immunisation has reduced the incidence of both diseases, there are still many children who die every year after acquiring natural infection in developing countries.13Similarly, factors affecting S pneumoniae antibody titres could affect vaccine efficacy to pneumococcal disease in infancy. In the absence of an effective pneumococcal vaccine for infants, maternal immunisation during pregnancy has also been evaluated as a strategy to improve passive antibody protection.16

Mean maternal total IgG titres were high in this Malawian population, and significantly increased in HIV infected mothers (tables 2 and 3). This could in part be due to repeated malaria infections and the coexistence of HIV infection. HIV seropositive mothers also had higher measles IgG antibody titres (tables 2 and 3). High measles and mumps antibody titres have been reported before in HIV infected patients.17 In a Brazilian population we have also shown that HIV infected mothers had significantly higher mean varicella zoster antibody titres than non-HIV infected mothers, but mean measles antibody titres were not significantly different between the two groups.11 Non-specific B cell activation could account for the high specific antibody titres in HIV infected individuals.17 Polyclonal activation is generally considered to be the explanation for the hypergammaglobulinaemia which occurs in subjects exposed to repeated malaria infections resulting from chronic mitogenic stimulation.18

Placental IgG antibody transfer was influenced independently by placental malaria infection, maternal HIV infection, and maternal total serum IgG titres. All three factors were associated with reduced placental transmission of pneumococcal IgG antibodies. Decreased transport of measles IgG antibodies was associated with placental malaria infection and maternal total IgG titres. Materno-fetal transmission of tetanus IgG antibodies was not affected by any of the variables tested, which was an unexpected result as a previous study has shown that high malaria parasite densities in the placenta are associated with decreased transplacental transfer of tetanus IgG antibodies using paired samples from Papua-New Guinea.12Coastal Papua New Guinea has one of the highest vectorial capacities for P falciparum malaria in the world and this might well explain the difference between the results from there and Malawi, as more frequent and higher density placental infections would be commonly expected to be more common in pregnant women in Papua New Guinea.

Maternal HIV infection and hypergammaglobulinaemia were associated with decreased placental IgG transfer for some antibody specificities, but not for others in our Brazilian study, with transfer of S pneumoniae antibodies being the most greatly affected.11 In the same study measles IgG antibody transfer was reduced by both maternal HIV infection and high maternal total serum IgG. High maternal total IgG was also associated with a reduction in placental IgG antibody transfer in the Malawian population. Tetanus IgG antibody titres were reduced by maternal HIV infection in the Brazilian study.11 The finding that parity was not associated with reduced placental transfer was similar to observations from Papua New Guinea.12 The lack of association between prematurity and reduced placental antibody transfer in this study has been described by others19 20Interestingly, all studies where prematurity was not examined in relation to placental antibody transfer have been undertaken in populations with high titres of maternal total serum IgG. Although this finding might be a characteristic of materno-fetal transmission of antibodies in the presence of maternal hypergammaglobulinaemia, it could also indicate a bias introduced when analysing preterm neonates born at a wide range of gestational ages—from 30 to 37 weeks. This is the reason why, in the present study, maternal and neonatal total IgG and specific IgG antibody titres were analysed only for samples from term babies of normal birthweight.

Although there was a reduction in placental transfer of IgG antibodies to S pneumoniae and measles, neonates from HIV infected mothers or from mothers with placental malaria infection did not have significantly lower measles and pneumococcal IgG antibody titres. This could relate to effects of maternal hypergammaglobulinaemia on antibody transfer. Michaux et al 21 were the first to notice that the pattern of placental IgG transfer may change when maternal total serum IgG titres reach 15 g/l: cord total IgG titres tend to be lower than the corresponding maternal values above that concentration. As all the Malawian sample had very high maternal total IgG titres, this factor might also be applicable in this population.

This study has shown that placental transfer to certain IgG antibody specificities is reduced and associated with maternal HIV infection, placental malaria, and maternal hypergammaglobulinaemia. It is not known why some antibody specificities are more affected and why interference with antibody transfer differs between populations. Also, it is not known how maternal HIV infection and placental malaria infection interfere with materno-fetal IgG antibody transmission, although malaria causes pathological damage in placental tissues.22 There is less evidence that HIV causes histological damage to the placenta.22

As factors influencing the efficiency of placental IgG antibody transfer are specific to particular populations, susceptibility to specific infections in infancy will vary correspondingly. Vaccine schedules might need to be modified to allow for these effects. However, overall IgG antibody transfer was good in this Malawian population, with adequate protective levels being achieved in most babies. It is difficult to identify that subgroup of infants who acquire the least maternal IgG antibody and who might benefit most from accelerated vaccination schedules. Screening tests would not be cost effective or possible in developing countries and, for this reason, an alternative emphasis might be placed on controlling maternal infection.