Pertussis Immunisation and Food Allergy

BACKGROUND Atopic diseases have become the most important cause of chronic morbidity among children in wealthy countries. In particular, food allergy, including life-threatening food-related anaphylaxis, is growing at an alarming rate. For example, one Australian allergy clinic observed a 12-fold rise in consultations for food allergy in children between 1995 and 2006. In the 15 years to 2009/10, hospitalisations for food-related anaphylaxis in young Australian children increased more than 8-fold. No effective broadly implementable strategies to curb this 'second wave' of atopic disease have been discovered.

In Australia, an abrupt rise in food allergy coincided closely with the replacement of the antigenically complex whole-cell pertussis vaccine (wP) with the subunit acellular pertussis vaccine (aP) beginning in the late 1990s. Pertussis vaccine protects against Bordatella Pertussis, the causative agent of whooping cough.

Investigators previously shown the marked differences in the immune stimulating profile of wP compared with aP. wP has been shown to induce strong Th1 type immunity while aP induces a Th2 type response. Immune responses can be broadly grouped into cellular (Th1) and antibody mediated immunity (Th2). Natural protective immunity to pertussis relies on strong Th1 type responses to clear the pathogen. This is closely mimicked by the immunity induced by whole cell vaccine (wP). In contrast, the acellular vaccine relies on the development of antibody mediated protection. While antibody mediated responses are protective in the short term, the strong induction of Th2 responses may be deleterious to other aspects of the immune response.

The developing infant immune system has been shown to have a Th2 bias, which may be further skewed by the delivery of a vaccine with a strong Th2 immune profile. Dysregulation of the Th1/Th2 balance which should exist in the immune system can lead to disease states. Predominance of a Th2 environment leads to the development of atopy and allergy. A number of studies, exemplified by those from our centre, detail the Th2 immune bias induced in some infants who received DTaP-only schedules, resulting in excessive IgE production against vaccine antigens. This is likely due to the combined effect of 1) carry-over of the Th2-biased in utero phenotype, 2) the presence of alum and pertussis toxin which have Th2-adjuvantising properties, and 3) the absence of the balancing Th1-stimulating ligands present in DTwP. These effects manifest most strongly in infancy and among children with evidence of an underlying Th2-skewed phenotype.

Studies have demonstrated that the inclusion of even a single wP dose in a pertussis priming schedule is associated with a reduction in Th2-polarised T cell memory and the magnitude of the IgE response compared with a aP-only schedule. Based on these findings it is hypothesised that the inclusion of a single wP dose in the pertussis priming schedule might preserve the Th1/Th2 balance of the developing infant immune system.

Thus, this study will test the hypothesise that removal of wP from the infant vaccine schedule has contributed to the observed rise in IgE-mediated food allergy.

AIM To assess the possible food allergy-preventive benefit of using wP compared with aP for pertussis vaccination in childhood.

OBJECTIVES AND OUTCOME MEASURES

Primary objective To determine if Australian children born between 1997 to 1999 (inclusive) who received wP as their first pertussis vaccine dose in infancy were less likely to subsequently develop IgE-mediated food allergy compared with contemporaneous children who received aP as their first pertussis vaccine dose in infancy.

Secondary objectives To determine if Australian children born in the years 1997 to 1999 (inclusive) who received at least one dose of a wP pertussis vaccine at any age were less likely to subsequently develop IgE-mediated food allergy compared with contemporaneous children who received only aP pertussis vaccines.

To determine if Australian children born in the years 1997 to 1999 (inclusive) who received wP pertussis vaccines exclusively were less likely to subsequently develop IgE-mediated food allergy compared with contemporaneous children who received only aP pertussis vaccines.

STUDY DESIGN This is a retrospective individually-matched case-control study of Australian children born during the period of transition from use of wP containing pertussis vaccines to aP containing pertussis vaccines (year of birth 1997-1999 inclusive) and who are registered on ACIR and received their first dose of pertussis vaccine before age 16 weeks. Cases will be drawn from allergy clinics associated with tertiary teaching hospitals around Australia. 1000 cases, identified as having IgE mediated food allergy from reviewing case histories, and appearing on ACIR will be enrolled. Controls will be drawn from a de-identified database of the Australian Childhood Immunisation Register held by the National Centre for Immunisation Research and Surveillance of Vaccine Preventable Diseases. Each case will be individually matched to 10 controls.

Project coordination and epidemiological analysis will be conducted from the Wesfarmers Centre of Vaccines and Infectious Diseases at the Telethon Kids Institute Western Australia.

STATISTICS AND ANALYSIS Sample size and power considerations Sample size: A study with 1000 matched sets of cases and controls, with 10 matched controls for each case, will have at least 80% power to detect a 17% lower risk (odds ratio 0.83) of food allergy among children who received wP as their first dose of pertussis vaccine compared with those who received aP as their first dose. This assumes that: (i) 50% of controls receive a first pertussis vaccine dose of wP; (ii) the correlation coefficient for exposure (first dose of wP) between matched cases and controls is 0.5; and (iii) a significance level of 5% (α=0.05). Investigators believe that a small effect size will not influence policy or research direction. If only 500 cases meet the case criteria the study will still have at least 80% power to detect a 23% lower risk (odds ratio 0.77) of food allergy.

Description of Statistical Methods The cohort characteristics will be summarised by case or control status using means and standard deviations for symmetric distributions and medians and inter-quartile ranges for asymmetric distributions. Conditional logistic regression will be used to estimate odds ratios and the corresponding 95% confidence intervals for the associations between (exposure) pertussis vaccine type (wP or aP) and the risk of IgE-mediated food allergy. Because controls will be sampled from the register cohort irrespective of past or future case status, the odds ratio will be interpreted as an unbiased estimator of the relative risk of food allergy among children receiving wP compared with aP vaccines.