Results

Not surprisingly, the infants receiving dexamethasone were younger, smaller, required more ventilation and a longer period in oxygen than those who did not receive the drug (table 1). Immunisations were given at comparable times. The median timing of the second blood specimen in infants receiving dexamethasone was seven weeks later than that of infants who did not receive dexamethasone (table2).

Pre-immunisation GMTs were similar in both groups. The change in antibody titres was appreciably less in those infants who received dexamethasone. The differences in GMTs were significant (table 3). The rise in antibody titre for babies who received dexamethasone was approximately one fifth of that in those who were not given dexamethasone (p=0.047, ratio change 0.21, 95% CI 0.05–0.98). Numbers of infants with antibody titres greater than 1.0 mcg/ml were substantially reduced in those infants receiving dexamethasone (table4).

Birthweight, gestational age, days in oxygen, days ventilated, days between samples 1 and 2 and dose of dexamethasone were regarded as potential risk factors in the regression analysis but none was found to be a significant predictor of the change in antibody titres. The stepwise method suggested that the dose of dexamethasone was the most significant predictor of the change in antibody titre of all factors considered (p=0.014), but only accounted for 10.4% of the total variation.

Initial antibody titres of less than 0.15 mcg/ml were associated with greater rises in Hib titres—8.4 times greater (95% CI 2.3–30.1) This was seen in infants in the non-dexamethasone group (p=0.001) and in the dexamethasone group (p=0.013).

Discussion

Geometric mean titres in infants who did not receive dexamethasone were similar to those seen in term infants where Hib was given as a separate or combined vaccine.1 3 5 Dexamethasone substantially reduces GMT to H influenzae and this observation is in keeping with the effect of dexamethasone on antibodies to diphtheria, tetanus, and pertussis (unpublished observations). The higher antibody rise among infants with a lower initial titre may reflect the influence of a passive inhibitory maternal antibody effect, although it could be a spurious result. This finding agrees with the observations of some authors in relation to Hib,7 diphtheria,8 9pertussis,8 10 tetanus7 and polio,8 but not those of others in relation to Hib.2 This effect is reduced, although not abolished, by dexamethasone.

The reason for the delay in obtaining the second specimen in the dexamethasone group is not known. The dates of the immunisations in both groups of infants were concordant so that this delay is unlikely to have been due to demographic factors. Immunisation in term infants following the accelerated schedule for DTP offers significant protection during pre-school years.11 It is unlikely, although possible, that the lower Hib antibody concentration seen in the dexamethasone treated group after immunisation could simply reflect a rapid decrease from an earlier level initially comparable with that obtained in the untreated group.

The protective antibody concentration is currently unclear. Protective levels related to antibody titres to polysaccharide antigen are put at 0.15 mcg/ml in unvaccinated and 1.0 mcg/ml in vaccinated individuals.12-14 These levels may be inappropriate in individuals immunised with conjugated antigens in whom there is immunological memory and for whom a later antigenic exposure may stimulate the production of an adequate antibody titre, whatever the baseline level. There are no data on protective titres available for the current conjugated antigens. This study shows that 28% or 56% of infants in the dexamethasone treated group were unprotected againstH influenzae at 0.15 mcg/ml and 1.0 mcg/ml, respectively, as a protective antibody level. The numbers studied are small but if the findings were generally applicable there may be between 500 and 1100 unprotected infants each year in the United Kingdom. These infants would be protected by the high level of herd immunity. Natural immunity is assumed to build up following exposure to wild H influenzae and possibly through cross reactivity with specificE coli surface antigens.14 As H influenzae carriage rates would be expected to fall as protection in the childhood population rose, these infants may not develop natural immunity and may thus be particularly susceptible to subsequent H influenzae type b infection. Until such time as the protective Hib antibody titre is clarified, at risk infants will require further immunisation. A study is currently in progress to ascertain when such an additional immunisation should be given.