Peter Aaby, Henrik Jensen, Badara Samb, Badara Cisse, Morten Sodemann, Marianne Jakobsen, Anja Poulsen, Amabelia Rodrigues, Ida Maria Lisse, Francois Simondon, Hilton Whittle

Projecto de Saúde de Bandim, Danish Epidemiology Science Centre, Apartado 861, Bissau, Guinea-Bissau (Prof P Aaby MSc, H Jensen PhD, M Sodemann MD, M Jakobsen MD, A Poulsen MD, A Rodrigues PhD, I M Lisse MD); UR 24 Epidemiology and Prevention Research Unit, IRD, Dakar, Senegal (Prof P Aaby, B Samb MD, B Cisse MD, F Simondon MD); and MRC Laboratories, Banjul, Gambia (Prof H Whittle BSc)

Correspondence to: Prof Peter Aaby, Danish Epidemiology Science Centre, Statens Serum Institut, Artillerivej 5, 2300 Copenhagen S, Denmark (e-mail:psb@mail.gtelecom.gw)

Summary
Introduction
Methods
Results
Discussion
References
 

Background Females given high-titre measles vaccine (HTMV) have high mortality; diphtheria-tetanus-pertussis (DTP) vaccination might be associated with increased female mortality. We aimed to assess whether DTP or inactivated poliovirus (IPV) administered after HTMV was associated with increased female-male mortality ratio.

Methods In three trials from West Africa, 2000 children were randomised to HTMV or control vaccine at 4-5 months of age; a second vaccination was given at age 9-10 months (standard measles vaccine). Children in high-titre groups were given IPV or DTP-IPV. Another 944 children received HTMV as routine vaccination in Senegal.

Findings When we compared high-titre and control groups, no difference in mortality between the first and the second vaccination was noted. After the second vaccination, the female-male mortality ratio was 1·84 (95% CI 1·19-2·84) in children in the high-titre groups who received DTP-IPV or IPV, and 0·59 (0·34-1·04) in controls who received standard measles vaccine (p=0·007). Children who received HTMV but no additional DTP-IPV or IPV had a female-male mortality ratio of 0·83 (0·41-1·67). This ratio was 2·22 (1·04-4·71) for children who received DTP-IPV after routine HTMV and 1·00 (0·68-1·47) for those who did not. When we combined the results from all trials, the female-male mortality ratio was 1·93 (1·33-2·81) for those who received DTP or IPV after HTMV, and 0·96 (0·69-1·34) for those who did not (p=0·006).

Interpretation A change in sequence of vaccinations, rather than HTMV itself, may have been the cause of increased female mortality in these trials.

Lancet 2003; 361: 2183-88

See Commentary

 

In 1989, WHO recommended use of high-titre Edmonston-Zagreb (EZ-HT) measles vaccine at age 6 months for children living in countries in which the incidence of measles before age 9 months was high.¹ In 1992, after results of studies in Guinea-Bissau, Senegal, and Haiti had shown raised female mortality in recipients of high-titre measles vaccine (HTMV),^2-4 recommendation for this vaccine was rescinded.⁵ Since high-titre vaccines protect against measles infection, the results of the high-titre trials had shown that vaccines might reduce survival in geographical areas with high mortality from this infection.

In areas of high mortality, various vaccines might have non-specific effects on mortality; for example, measles^6-9 and BCG^9-11 vaccines reduce mortality from diseases other than measles and tuberculosis. On the other hand, inactivated vaccines such as diphtheria-tetanus-pertussis (DTP) and inactivated poliovirus (IPV) might amplify mortality from diseases other than diphtheria, tetanus, pertussis, and polio.^7,9,11,12 Non-specific effects are strongest in the first 3-6 months after immunisation¹³ and in girls,^2,3,6,12,14 and they are largely determined by the most recent vaccine received; for example, in Guinea-Bissau, the female-male mortality ratio was 3·08 (95% CI 1·11-8·56) for children who received DTP as their last vaccine, but only 0·63 (0·28-1·40) for those who received measles vaccine last.¹² Results of several studies from West Africa have shown that BCG could enhance the response to unrelated antigens.^10,15

The greatly reduced mortality after measles vaccination has been attributed to prevention of long-term effects of measles infection. However, although morbidity might be raised for a few months after measles infection,¹⁶ results of studies from Burundi,¹⁷ Ghana,¹⁸ Bangladesh,^13,16 and Senegal¹⁹ have shown no increase in mortality after the acute phase of infection. Mild measles infection might be associated with lower mortality,^13,19 thus both measles infection and measles vaccination might provide beneficial stimulation of the immune system, which enhances resistance to other infections.

HTMVs seem to result in a high female-male mortality ratio⁶ and an increased mortality when compared with female recipients of standard-titre measles vaccine.²⁰ In a meta-analysis of West African studies, girls who received HTMV had a mortality ratio of 1·86 (95% CI 1·28-2·70) compared with those who received standard measles vaccine, whereas boys had a mortality ratio of only 0·91 (0·61-1·35). The effect was not seen immediately, but several months later. Two different hypotheses have been proposed to account for these surprising observations.

Initially, HTMV was postulated to have come too close to the natural disease, thus inducing immune suppression, as happens in natural measles infection.^4,21 This hypothesis does not account for the delayed increase in mortality and why later, the effect was noted for girls only. Although measles mortality might be raised in older girls and women,^22,23 it is usually not higher in girls in the first 3 years of life, which is the period when high-titre vaccines are associated with increased female mortality. If anything, boys have higher measles mortality in this age range.²³ More importantly, measles infection is usually not associated with long-term excess mortality.^16-19 Hence, HTMV does not mimic natural measles disease. Furthermore, contradictory to the hypothesis, results of one large trial of HTMV in Zaire showed no increase in mortality of high-titre recipients when compared with recipients of medium-titre measles vaccine, and no increase in female-to-male mortality.²⁴

Second, we noted in geographical areas with high childhood mortality that standard measles vaccine was associated with a non-specific benefit on survival, which was especially strong for girls.^6,7 We therefore suggested that the detrimental effect of HTMV would only be seen in areas with high mortality,⁸ for high-titre vaccine did not provide the non-specific and sex-specific benefits of the standard measles vaccine.^3,8,20 The mortality difference would only be seen when girls in the control groups had received the standard measles vaccine, and it would not be noted in areas with low mortality since children in these areas had no non-specific survival benefit from standard measles vaccine. However, in the time before vaccination, girls did not have higher mortality than boys.⁶ Thus, our hypothesis did not fully explain why girls had a higher mortality than boys in the high-titre group^6,8 and why these effects did not arise in the HTMV trial in Zaire.²⁴

Since both of these interpretations are unable to account for all observations, we propose an alternative hypothesis. In the high-titre trials, many children received DTP or IPV after measles vaccination. Since DTP has been reported to be associated with an increase in female mortality,^12,14 we aimed to find out whether vaccination with DTP or IPV after high-titre measles vaccination might have contributed to the increased female-male mortality ratio. If excess female mortality was indeed caused by subsequent administration of DTP or IPV vaccines, these three deductions should hold. First, there should be no excess mortality for high-titre recipients compared with controls in the period between enrolment and subsequent reception of DTP or IPV vaccines. Second, recipients of high-titre vaccine, after being given DTP or IPV, should have a higher female-male mortality ratio than controls receiving standard measles vaccine at age 9-10 months, and a higher female-male mortality ratio than high-titre recipients who did not receive additional DTP or IPV. Finally, the hypothesis should account for contradictions encountered by the other two hypotheses.

 

We reanalysed data from three HTMV trials from Guinea-Bissau, Senegal, and the Gambia. These trials have been described in several publications,^2,3,25,26 and have been the subject of a combined meta-analysis.²⁰ The three randomised trials, although not identical in design, had similar features; the studies compared children receiving HTMV from age 4 or 5 months, with a randomised control group given IPV (Guinea-Bissau, the Gambia) or placebo (Senegal) at a similar age. At the same time as receiving trial vaccines, children received additional vaccines according to the national immunisation programme. In Guinea-Bissau, about two-thirds of children received DTP at the time of enrolment. In Senegal, all children received DTP-IPV at recruitment and, hence, all controls received IPV at the beginning of the study. At age 9-10 months, children were invited back; controls received a standard dose of measles vaccine whereas the children in the high-titre groups received IPV in Guinea-Bissau and the Gambia. In Senegal, children attending vaccination at 10 months received yellow-fever vaccine and the third dose of DTP-IPV at age 9-10 months, and the high-titre group therefore received DTP-IPV after measles vaccination. Survival information was obtained through demographic surveillance in the study areas, and children were followed up to age 3-5 years, as described previously.^2,3,20

To examine the effect of subsequent DTP vaccinations, we also used data from the post-trial period in Senegal, when EZ-HT was used as the routine measles vaccine.⁸ In this study, EZ-HT was administered to children born between March, 1989, and April, 1990. All children were called three times for vaccination at a health centre and measles vaccine was usually administered together with DTP-IPV or DTP. In the initial period, children born March-June, 1989, continued to receive EZ-HT at age 5 months (median age 4·8 months), usually together with the second dose of DTP-IPV, and most children received their third dose of DTP-IPV after measles vaccination. Subsequently, when WHO recommended EZ-HT from age 6 months,¹ this vaccination was postponed to the third vaccination session. Hence, children born July, 1989, to April, 1990, received EZ-HT usually together with their third dose of DTP-IPV; median age at vaccination was 6·4 months. This change in programme offered an opportunity to test whether DTP-IPV vaccination after high-titre measles vaccination affected the female-male mortality ratio.

At all sites, most vaccines, including additional DTP and IPV, were provided by the respective projects. Vaccines provided elsewhere have only been taken into consideration in the survival analysis from the date information on vaccinations was obtained. The combined effects for all trials were calculated with Mantel-Haenszel weights.

Role of the funding source

The sponsors of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report.

 

Features of the three randomised trials that are of relevance for our reanalysis are summarised in table 1 and figure 1. In these trials, we examined the mortality ratio between enrolment and subsequent vaccination with DTP or IPV vaccine--ie, the age interval when the Edmonston-Zagreb group had received HTMV and the control group had not yet received measles vaccine (comparison of groups A and B; figure 1). Over this period, mortality rates were as high as 5-8% (figure 2), but mortality of the high-titre groups was not increased when compared with that of the control groups (table 2). Although girls had slightly higher mortality rates than boys, the female-male mortality ratio for high-titre recipients was not higher than that for controls (table 2, figure 2).

Irrespective of whether they received the second vaccination in the trial at age 9-10 months, the high-titre groups had a mortality rate of 1·51 (1·10-2·07) from 9-10 months of age compared with the control groups (groups E, G, H vs groups F, I, J). High-titre recipients who also received DTP or IPV were compared with children in the control groups given standard measles vaccine and to other high-titre recipients who did not receive DTP or IPV vaccines after measles vaccine. In a combined analysis of the three trials (table 3), the high-titre recipients who received the DTP or IPV vaccine at age 9-10 months had a higher mortality rate than did controls (groups E vs F). The mortality ratio was 2·35 (95% CI 1·41-3·91) for girls and 0·74 (0·45-1·22) for boys (p=0·005). Female-male mortality ratios were significantly different in the high-titre groups receiving DTP or IPV vaccine compared with the standard-titre measles vaccine groups (p=0·007; table 3, figure 2).

In all three trials (table 3), we examined the female-male mortality ratio for high-titre recipients according to whether additional DTP or IPV vaccines had been received (groups E vs G). In the large trial in Senegal, the female-male mortality ratio was 0·83 (95% CI 0·41-1·67) for HTMV recipients who did not turn up to receive an additional vaccination at 10 months of age and 1·56 (0·98-2·49) for children who received an additional dose of DTP-IPV (p=0·142). In a combined analysis of the three trials, high-titre recipients who had not received additional DTP or IPV vaccines had a reduced female-male mortality ratio compared with that for those who received additional doses of DTP or IPV (p=0·059; table 3, figure 2).

An analysis of mortality after routine use of EZ-HT in Senegal offered a different way to test the effect of subsequent DTP-IPV vaccinations.⁸ As seen in table 4, increased female-male mortality was noted in the first part of the study for children who had received additional doses of DTP-IPV after HTMV. Overall, the difference in female-male mortality was very similar for children who received no subsequent DTP-IPV, whereas it was doubled among those who received further doses of DTP-IPV after measles vaccination (p=0·065; table 4).

In a combined analysis of all three randomised trials and the study of routine use of EZ-HT, the female-male mortality ratio was 1·93 (95% CI 1·33-2·81) for additional DTP or IPV vaccines after HTMV and 0·96 (0·69-1·34) for no additional vaccination (p=0·006).

 

On the basis of results of studies from West Africa and Haiti, high-titre measles vaccination is believed to cause increased female mortality, whereas no problem is recorded with medium-titre and standard-titre measles vaccines.^4,5,20,21 These trials were planned to study vaccine efficacy, not to examine an association with increased female mortality. Hence, present interpretations are at best post-hoc hypotheses. When hypothesis testing is not possible, as in this case, since increased female mortality could not ethically have been tested in a clinical trial, the preferred interpretation should account for all data and be capable of resolving contradictions in existing knowledge. The two present interpretations do not accord with available data.

Three strands of evidence suggest that DTP or IPV vaccines administered after high-titre measles vaccination might have contributed to the increased female-male mortality ratio noted in the high-titre trials.^2,3 First, although the maximum effect of a presumed deleterious high-titre vaccine would be expected soon after administration and during infancy when mortality is highest, no difference in female mortality was recorded between high-titre recipients and controls until the high-titre recipients received additional DTP or IPV vaccines. The three West African trials could be argued to have limited power to assess excess female mortality between 4 and 9 months of age before administration of other vaccines. However, this argument is not true of the study of routine use of EZ-HT and the trial in Zaire,²⁴ in which many children received EZ-HT at age 6 months without any subsequent DTP vaccination. In these situations, female mortality was not raised. Furthermore, in the Sudan trial, children vaccinated with EZ-HT had significantly lower mortality than controls between the first and second vaccination.²⁷

Second, in all the West African studies, increased mortality in the high-titre groups and raised female mortality compared with the control groups was only seen after high-titre recipients received additional vaccines of DTP, IPV, or both. No difference was seen in the proportion of boys and girls who came back to receive these vaccinations at age 9-10 months.³ Thus, bias seems to be unlikely to explain why girls had higher mortality than boys in the high-titre arm of the study and lower mortality in the control group. Likewise, socioeconomic background factors cannot account for why girls had higher mortality than boys in one arm of the study and lower mortality in the other. No drug, vaccine, or other health intervention was given in the same way as DTP, IPV, or both in all these studies.

Third, by contrast to the striking sex-specific effect after DTP or IPV vaccination, the female-male mortality ratio was not raised in high-titre recipients who did not receive DTP or IPV after measles vaccine. In Zaire,²⁴ where recipients of EZ-HT were not given subsequent DTP vaccine because they had already received three doses of DTP, the high-titre group had a mortality ratio of 0·5 (95% CI 0·2-1·5) compared with recipients of medium-titre Edmonston-Zagreb. Girls did not have higher mortality than did boys. Furthermore, trials in the Gambia,²⁰ Sudan,²⁷ and Zaire²⁴ included groups of children who received an additional measles vaccine at age 9 months after the initial high-titre vaccine; these groups had low mortality and a female-male mortality ratio of 1·0 (data available from authors).

Hence, the raised female mortality might not have been attributable to HTMV as such, but to subsequent DTP or IPV vaccinations. This interpretation, based on results of the West African studies, accords with the finding that an effect was reported in girls only, and that it was not recorded in studies of HTMV when DTP or IPV was not administered after measles vaccine.²⁴

The inversion of the female-male mortality ratios could be attributable to a combination of a benefit of standard measles vaccine for girls⁷ and an adverse effect of DTP or IPV that is more striking in girls.^9,11,12,14 In the present study, to separate the effects of DTP and IPV was not possible. In Senegal, a combined DTP-IPV vaccine was used. In the Gambia, only IPV was given after HTMV because children had received their DTP vaccinations before recruitment. In Guinea-Bissau, the two vaccines were given separately, but the high-titre group was too small to compare the effects because virtually every child received both DTP and IPV after high-titre measles vaccination. Since DTP has been associated with increased female mortality in several other studies,^12,14 DTP at least seems likely to be implicated.

The high-titre problem is likely to have been attributable to DTP or IPV, but to fully clarify the effect of HTMVs on child survival is not possible, because results from different studies are conflicting. In the West African trials, the mortality rate between 4 and 8 months of age after high-titre vaccination was high, and only slightly lower for the high-titre group than for controls, suggesting little benefit from high-titre vaccine. However, in the study from Zaire,²⁴ high-titre recipients had lower mortality than did medium-titre recipients. Furthermore, in the Sudan trial,²⁷ which had a design similar to the West African trials, EZ-HT vaccine was associated with significantly lower mortality than that for controls between the first and the second vaccination. If it is possible to use HTMV again, to examine whether high-titre vaccines have non-specific beneficial effects similar to standard measles vaccines would be necessary.

Our reinterpretation raises questions about the non-specific effects of vaccines and about the sequence in which they are given. The main purpose of the high-titre measles vaccination strategy was to enable early immunisation in the presence of maternal antibodies to prevent measles in infants. In the high-titre trials, the sequence of vaccinations was changed. Whereas the three doses of DTP and oral polio are usually administered before standard measles vaccine at age 9 months, the recipients of HTMV were much more likely to receive DTP or IPV after measles vaccination. Changes in the sequence of different vaccinations have not been analysed in other studies of childhood mortality. Increased mortality after HTMV has previously been pointed out to be only seen in areas with high mortality, whereas in areas with low mortality no effect was seen.⁸ Possibly some of this difference might be attributable to these areas having more efficient programmes providing all the DTP vaccinations before early measles vaccination.

The present study adds further weight to the observation that DTP in populations with a high burden of diseases may have little benefit, particularly for girls.^7,9,11,12,14 Very little basic research has been done to understand the biological mechanisms of non-specific immune stimulation caused by vaccines and why they should vary by sex. A few studies in animals have compared live and inactivated vaccines containing the same antigen^28,29 and reported that live vaccines promote a T-helper 1-type response, which provides better protection when challenged with live antigen. An aluminium-based inactivated vaccine has been reported to provoke stronger adverse reactions in women than in men.³⁰ As our results suggest that vaccines and the sequence in which they are given could have an effect on child survival, this topic deserves much greater attention when examining the effect of routine immunisations or new vaccines on child survival in developing countries. The challenge will be to design and implement appropriate trials to assess the non-specific effects and test the hypothesis.

The high-titre trials were planned by P Aaby and H Whittle. The trials in Guinea-Bissau were implemented by M Sodemann, M Jakobsen, A Poulsen, and I M Lisse. P Aaby, F Simondon, and B Samb completed the trial in Senegal. B Cisse and A Rodrigues assured long-term follow-up in Senegal and Guinea-Bissau. H Whittle was responsible for the trial in the Gambia. P Aaby and H Jensen did the reanalysis. All authors contributed to the final version of the report.

None declared.

The study received financial support from the Danish Council for Development Research, Danish Medical Research, DANIDA, and the EU Commission's INCO programme (IC18T95-0011). PA holds a research professorship funded by Novo Nordisk Foundation.

 

1 Expanded programme on immunization. Global advisory group: part I.  Wkly Epidemiol Rec 1990; 65: 5-11. [PubMed]

2 Aaby P, Knudsen K, Whittle H, et al. Long-term survival after Edmonston-Zagreb measles vaccination: increased female mortality.  J Pediatr 1993; 122: 904-08. [PubMed]

3 Aaby P, Samb B, Simondon F, et al. Sex specific mortality after high titre measles vaccines in rural Senegal.  Bull World Health Organ 1994; 72: 761-70. [PubMed]

4 Holt EA, Moulton LH, Siberry GK, Halsey NA. Differential mortality by measles vaccine titer and sex.  J Infect Dis 1993; 168: 1987-96. [PubMed]

5 Expanded programme on immunization. Safety and efficacy of high-titre measles vaccines.  Wkly Epidemiol Rec 1992; 67: 357-61. [PubMed]

6 Aaby P, Samb B, Simondon F, et al. Divergent mortality for male and female recipients of low-titre and high-titre measles vaccines in rural Senegal.  Am J Epidemiol 1993; 138: 746-55. [PubMed]

7 Aaby P, Samb B, Simondon F, et al. Non-specific beneficial effect of measles immunization: analysis of mortality studies from developing countries.  BMJ 1995; 311: 481-85. [PubMed]

8 Aaby P, Samb B, Simondon F, et al. A comparison of vaccine efficacy and mortality during routine use of high-titre Edmonston-Zagreb and Schwarz standard measles vaccines in rural Senegal.  Trans R Soc Trop Med Hyg 1996; 90: 326-30. [PubMed]

9 Kristensen I, Aaby P, Jensen H. Routine vaccinations and child survival: follow-up study in Guinea-Bissau.  BMJ 2000; 321: 1435-38. [PubMed]

10 Garly ML, Martins CL, Balé C, et al. BCG scar and positive tuberculin reaction associated with reduced child mortality in West Africa: a non-specific beneficial effect of BCG? Vaccine (in press).

11 Velerna JP, Alihonou EJ, Gandaho T, Hounye FH. Childhood mortality among users and non-users of primary health care in a rural West African community.  Int J EpidemioI 1991; 20: 474-79. [PubMed]

12 Aaby P, Garly ML, Balé C, et al. Routine vaccinations and child survival in war situation with high mortality: effect of gender.  Vaccine 2002; 21: 15-20. [PubMed]

13 Aaby P, Bhuyia A, Nahar L, et al. The survival benefit of measles immunisation may not be explained entirely by the prevention of measles disease.  Int J Epidemiol 2003; 32: 106-15. [PubMed]

14 Aaby P, Jensen H. Routine vaccinations and child survival: effect of gender. http://www.bmj.com/cgi/eletters/321/7274/ 1435#27803 (accessed Mar 27, 2003).

15 Ota M, Vekemans J, Schlegel S, et al. Neonatal immunisation with Mycobacterium bovis bacillus Calmette-Guérin promotes immune responses to unrelated vaccine antigens.  J Immunol 2002; 168: 919-25. [PubMed]

16 Akramuzzaman SA, Cutts FT, Wheeler JG, Hossain MJ. Increased morbidity after measles is short-term in urban Bangladesh.  Am J Epidemiol 2000; 151: 723-35. [PubMed]

17 Chen RT, Weierbach R, Bisoffi Z, et al. A 'post-honeymoon period' measles outbreak in Mayinga Sector, Burundi.  Int J Epidemiol 1994; 23: 185-93. [PubMed]

18 Dollimore N, Cutts F, Binka FN, et al. Measles incidence, case fatality, and delayed mortality in children with or without vitaimin A supplementation in rural Ghana.  Am J Epidemiol 1997; 146: 646-54. [PubMed]

19 Aaby P, Samb B, Andersen M, Simondon F. No long-term excess mortality after measles infection: a community study from Senegal.  Am J Epidemiol 1996; 143: 1035-41. [PubMed]

20 Knudsen KM, Aaby P, Whittle H, et al. Child mortality following standard, medium and high titre measles vaccination in West Africa.  Int J Epidemiol 1996; 25: 665-73. [PubMed]

21 Hussey GD, Goddard EA, Hughes J, et al. The effect of Edmonston-Zagreb and Schwarz measles vaccine on immune responses in infants.  J Inf Dis 1996; 173: 1320-26. [PubMed]

22 Garenne M. Sex differences in measles mortality: a world review.  Int J Epidemiol 1994; 23: 632-42. [PubMed]

23 Aaby P. Are men weaker or do their sisters talk too much? In: Basu AM, Aaby P, eds. The methods and uses of anthropological demography. Oxford: Oxford University Press, 1998: 223-45.

24 Cutts FT, Nyandu B, Markowitz LE, et al. Immunogenicity of high-titre AIK-C or Edmonston-Zagreb vaccines in 3.5-month-old infants, and of medium- or high-titre Edmonston-Zagreb vaccine in 6-month-old infants, in Kinshasa, Zaire.  Vaccine 1994; 12: 1311-16. [PubMed]

25 Garenne M, Leroy O, Beau JP, Sene I. Child mortality after high-titre measles vaccines: prospective study in Senegal.  Lancet 1991; 338: 903-07. [PubMed]

26 Aaby P, Samb B, Simondon F, et al. Child mortality after high-titre measles vaccines in Senegal: the complete data set.  Lancet 1991; 338: 1518-19. [PubMed]

27 Libmann MD, Ibrahim SA, Omer MIA, et al. No evidence for short or long term morbidity after increased titer measles vaccination in Sudan.  Pediatr Infect Dis J 2002; 21: 112-19. [PubMed]

28 Lindblad EB. Elhay MI, Silva R, Appelberg R, Andersen P. Adjuvant modulation of immune responses to tuberculosis subunit vaccines.  Infect Immun 1997; 65: 623-29. [PubMed]

29 Graham BS, Henderson GS, Tang YW, et al. Priming immunization determines T helper cytokine mRNA expression patterns in lungs of mice challenged with respiratory syncytial virus.  J Immunol 1993; 151: 2032-40. [PubMed]

30 Pittman PR. Aluminum-containing vaccine associated adverse events: role of route of administration and gender.  Vaccine 2002; 20: 48-50.  [PubMed]