In response to an apparent resurgence of pertussis among older persons in several countries with well-established childhood pertussis immunization programs (North America, Australia, and some regions in Europe), adolescent/adult-formulated diphtheria-tetanus-acellular pertussis (dTap) vaccines have now been introduced.1,2 In Australia, 1 dTap vaccine (Boostrix, GlaxoSmithKline, Rixensart, Belgium) was licensed in 2000 but little used before 2004, when a funded program became available for adolescents.
From May to December 2004, dTap was given to 272,000 12–19 year olds during a mass vaccination program targeting all high school students in New South Wales (NSW),3 the largest Australian state with a population of 6.5 million. This was the first large-scale use of Boostrix vaccine in Australia and, to our knowledge, the largest cohort targeted over a short time period internationally.
A single dose of dTap Boostrix is highly immunogenic in adolescents and adults,4 with protective efficacy estimated at 92% (95% CI: 32–99%) for pertussis confirmed by culture, polymerase chain reaction (PCR) or serologic assay.5 In Australia, IgA serology based on the detection of antibodies to whole cell pertussis antigen has been increasingly available since the early 1990s. Laboratory reporting of positive results from this assay accounts for the majority of pertussis notifications to the National Notifiable Diseases Surveillance Scheme.2 Although sensitivity is low (24–64%), this method has been estimated to have a specificity of 93–98% in the presence of appropriate symptoms.6
In a randomized controlled trial, efficacy of dTap was 92% and varied according to whether pertussis diagnosed by serologic methods was included.5 However, there have been no field evaluations of vaccine effectiveness (VE) when dTap is administered to adolescents in routine practice. We applied the screening method described by Orenstein et al,7 using coverage data from the school program and routine pertussis surveillance data up to 1 year after completion of high school vaccination. Our preliminary assessment of VE in this large, rapidly immunized adolescent cohort provides a baseline for later measures and guidance to others considering such a program.
All notified cases of pertussis in persons aged 12–19 years with disease onset from January 1, 2005 to December 31, 2005 were selected from the NSW Notifiable Diseases Database (criteria for notification are shown in Table 1). Receipt of dTap vaccine was verified by school vaccination records or by telephone follow-up where records were unavailable.
Eligible cases attended a high school in 1 of 6 participating NSW health regions during 2004. Cases vaccinated outside of the school program (approximately 0.7%) and those whose vaccination status could not be confirmed were excluded. Cases were analyzed by sex, age at diagnosis, school location (regional or metropolitan), health region, and method of diagnosis.
The screening method estimates VE using the formula VE = 1 − [PCV/(1 − PCV)][(1 − PPV)/PPV], where PCV is the proportion of the cases vaccinated and PPV is the proportion of the study population vaccinated.7 The baseline VE estimate was adjusted for potential confounders by multivariate logistic regression using Proc Genmod in SAS version 9.1 (SAS Institute Inc., Cary, NC), as described by Farrington.8 Program coverage as assessed by school vaccination records was 56%3; this was used as the estimate of PPV in NSW adolescents. Overall and adjusted VE estimates were derived from average parameter estimates for variables in the final model (health region and method of diagnosis), as described by Torvaldsen et al.9 Parameter estimates for health region were weighted by the proportion of 12–19 year olds in the population.
During 2005, 341 cases of pertussis in persons aged 12–19 years were notified in NSW, of whom 207 attended high school in a participating health region in 2004. After exclusion of 40 cases, resulting from unknown vaccination status or vaccination outside of the school program, VE analysis was based on 167 cases. There were no significant differences between included and excluded cases with respect to age, sex, or location. Similar to notified cases, 82% of study cases were diagnosed by serology and 11% were diagnosed by PCR. However, a high proportion (10 of 11) of the cases notified without laboratory confirmation (clinically defined, as described in Table 1), were included in the study sample.
Overall, 43 study cases (26%) received dTap vaccine during the school program. Although estimated population coverage was higher among those aged 12–15 years than aged 16–19 (62% vs. 41%, respectively),3 the age distribution among vaccinated and unvaccinated study cases was similar. Vaccinated cases were predominantly female (65%) and diagnosed by serology (72%). A high proportion (50%) of clinically-defined study cases were vaccinated relative to those diagnosed by serology or PCR (23% and 33%, respectively).
Despite the overall predominance of serology based notifications, there was some variation in diagnostic method by health region. Both diagnostic method and health region were associated with VE in the univariate analysis and remained in the final model (P = 0.07 and P = 0.05, respectively). The overall adjusted VE estimated by this model was 78.0% (95% CI: 60.7–87.6%).
Table 1 presents health region-adjusted VE estimates by diagnostic method. The point estimate of VE in clinically diagnosed pertussis was considerably lower than estimates for notifications based on PCR or serologic tests and was imprecise, with confidence intervals substantially overlapping estimates for the other diagnostic methods.
When the analysis was limited to laboratory-confirmed cases (n = 155), only health region was retained in the model (P = 0.14), yielding an overall adjusted VE of 85.4% (95% CI: 83.0–87.5%). Estimates based on laboratory-confirmed cases only were higher for all health regions (data not shown).
Together with descriptive data from notifications demonstrating a substantial decrease in pertussis incidence among NSW adolescents after high school vaccination,2 our study supports the effectiveness of large-scale use of dTap to rapidly control pertussis in this age group.
Although our analysis was limited to 49% of pertussis notifications aged 12–19 years, there were no statistically significant demographic differences between notified cases residing in participating versus other health regions. Any selection bias would be expected to result in a more conservative estimate of vaccine effectiveness, as cases excluded due to unknown vaccination status are more likely to be unvaccinated, so their inclusion would have resulted in higher VE estimates. As all persons who meet laboratory criteria for pertussis (serology, culture, or PCR) are routinely notified by laboratories, only unrecognized or clinically diagnosed, but unreported cases would be missed. We had no evidence to suggest recent and/or substantial changes in physician diagnostic practices, but public health staff have encouraged physicians to obtain specimens for PCR or culture rather than serology in putative cases, especially in immunized cases.
We explored the impact of varying PPV values and found that overall program coverage, which also represented the best available estimate of NSW adolescents vaccinated with dTap in 2004, yielded the most statistically stable model. Despite older adolescents having had lower vaccination coverage3 and a high likelihood of prior exposure to pertussis from earlier NSW epidemics, we observed no significant change in VE estimates with increasing age.
Our VE estimates were substantially lower for cases notified on the basis of clinical symptoms only. This is expected as a clinical diagnosis of pertussis, in the absence of supporting laboratory data or epidemiologic contact, is likely to be nonspecific. Limiting the analysis to laboratory-confirmed cases increased the point estimate for VE by 7%.
The performance of the IgA antibody to whole cell antigen pertussis assays, currently predominantly used in Australia for diagnosis among persons older than 5 years,2 is unknown and questionable in persons recently vaccinated with dTap. Recent data indicate that IgA antibody to pertussis toxin may remain above the threshold of detection for more than 3 years after immunization.4 IgA serology available in the routine diagnostic setting may not differentiate between antibody response from immunization from that from infection within 1 year of vaccination.5 Our VE point estimate for cases diagnosed by PCR (84.3%) was slightly lower than that for notifications based on serology (88.8%), but had a higher upper confidence limit (94%). We conclude that the VE estimates from this study are most likely minimum estimates, as both clinically and serologically diagnosed cases are likely to have included false-positives.
Use of the screening method for serial monitoring of VE as a new program is facilitated by the availability of high-quality immunization coverage data, and is likely to be applicable for routine surveillance in other settings where such data are available. Despite the limitations of pertussis diagnosis in routine clinical or public health use, as long as these remain constant in any 1 setting, regular assessment of VE by the screening method is a useful means to monitor trends in the effectiveness of adolescent pertussis vaccination over time.
The authors thank NSW Health for provision of data from the NSW Notifiable Diseases Database; Sue Campbell Lloyd at NSW Health for supporting this project and coordinating the provision of vaccination data; and all NSW Health staff who assisted in the follow-up of school vaccination records.
1. Halperin SA. Pertussis
—a disease and vaccine for all ages. N Engl J Med
2. Quinn HE, McIntyre PB. Pertussis
epidemiology in Australia over the decade 1995–2005—trends by region and age group. Comm Dis Intell
3. Quinn HE, Boomer MJ. The NSW adolescent pertussis
vaccination program—program coverage and program acceptability among parents. Paper presented at: Public Health Association of Australia 10th National Immunisation/2nd Asia Pacific Vaccine Preventable Diseases Conference; July 31, 2006; Sydney, Australia.
4. Le T, Cherry JD, Chang S, et al. Immune response and antibody decay after immunization of adolescents and adults with an acellular pertussis
vaccine: the APERT study. J Infect Dis
5. Ward JI, Cherry JD, Swei-Ju C, et al. Efficacy of an acellular pertussis
vaccine among adolescents and adults. N Engl J Med
6. Poynten IM, Hanlon M, Irwig L, et al. Serological diagnosis of pertussis
: evaluation of IgA against whole cell and specific Bordetella pertussis
antigens as markers of recent infection. Epidemiol Infect
7. Orenstein WA, Bernier RH, Hinman AR. Assessing vaccine efficacy in the field: further observations. Epidemiol Rev
8. Farrington CP. Estimation of vaccine effectiveness
using the screening method
. Int J Epidemiol
9. Torvaldsen S, Simpson JM, McIntyre PB. Effectiveness of pertussis
vaccination in New South Wales, Australia, 1996–1998. Eur J Epidemiol