BACKGROUND: In the United States, the proportion of Bordetella pertussis isolates lacking pertactin, a component of acellular pertussis vaccines, increased from 14% in 2010 to 85% in 2012. The impact on vaccine effectiveness (VE) is unknown.
METHODS: We conducted 2 matched case-control evaluations in Vermont to assess VE of the 5-dose diphtheria, tetanus, and acellular pertussis vaccine (DTaP) series among 4- to 10-year-olds, and tetanus, diphtheria, and acellular pertussis vaccine (Tdap) among 11- to 19-year-olds. Cases reported during 2011 to 2013 were included. Three controls were matched to each case by medical home, and additionally by birth year for the Tdap evaluation. Vaccination history was obtained from medical records and parent interviews. Odds ratios (OR) were calculated by using conditional logistic regression; VE was estimated as (1-OR) × 100%. Pertactin status was determined for cases with available isolates.
RESULTS: Overall DTaP VE was 84% (95% confidence interval [CI] 58%–94%). VE within 12 months of dose 5 was 90% (95% CI 71%–97%), declining to 68% (95% CI 10%–88%) by 5–7 years post-vaccination. Overall Tdap VE was 70% (95% CI 54%–81%). Within 12 months of Tdap vaccination, VE was 76% (95% CI 60%–85%), declining to 56% (95% CI 16%–77%) by 2–4 years post-vaccination. Of cases with available isolates, >90% were pertactin-deficient.
CONCLUSIONS: Our DTaP and Tdap VE estimates remain similar to those found in other settings, despite high prevalence of pertactin deficiency in Vermont, suggesting these vaccines continue to be protective against reported pertussis disease.
- CDC —
- Centers for Disease Control and Prevention
- CI —
- confidence interval
- DTaP —
- diphtheria toxoid, tetanus toxoid, and acellular pertussis vaccine
- OR —
- odds ratio
- PCR —
- polymerase chain reaction
- Tdap —
- tetanus toxoid, reduced diphtheria toxoid, and acellular pertussis vaccine
- VE —
- vaccine effectiveness
What’s Known on This Subject:
The recent resurgence in pertussis disease may be associated with the previously demonstrated waning immunity from acellular vaccines. However, little is known about the impact of genetic changes in the bacteria, such as the loss of pertactin, on vaccine effectiveness.
What This Study Adds:
This is the first evaluation of pertussis vaccine effectiveness in the setting of pertactin deficiency. Our vaccine effectiveness estimates remain similar to those from other settings, despite the high prevalence of pertactin-deficient strains, suggesting pertussis vaccines continue to be protective.
The US pertussis vaccine program began in the 1940s, with widespread use of diphtheria, tetanus, and whole-cell pertussis vaccine for the 5-dose childhood series. Safety concerns prompted whole-cell vaccine to be replaced with diphtheria toxoid, tetanus toxoid, and acellular pertussis vaccine (DTaP) in 1992 for the fourth and fifth doses given at 15 to 18 months and 4 to 6 years of age, and ultimately for the complete childhood series in 1997, including the primary doses at 2, 4, and 6 months of age.1–3 Due to an increase of disease among adolescents and adults, a single adolescent booster dose of tetanus toxoid, reduced diphtheria toxoid, and acellular pertussis vaccine (Tdap) was recommended in 2006, with preferred administration at 11 to 12 years of age.4 Although coverage for both vaccines is high, pertussis incidence is increasing, with 48 000 cases reported in 2012.5–7 Of concern, the number of cases has increased among fully vaccinated children and adolescents, particularly among those who received only acellular vaccines as children.8–10 Possible explanations for this include waning immunity, less protective immune responses induced by acellular pertussis vaccines, increased provider awareness and reporting, and more sensitive diagnostic techniques.11–15 Additionally, genetic changes in the bacteria could have a role in the resurgence of disease. The proportion of pertussis strains in the United States lacking pertactin sharply increased from 14% in 2010 to 85% in 2012.16–18 Pertactin is an autotransporter potentially involved in bacterial adhesion to the respiratory tract and resistance to neutrophil-induced bacterial clearance.19,20 It is also a component of all acellular pertussis vaccines currently administered in the United States.21 Pertactin deficiency may have evolved in response to acellular vaccine–induced selection pressure, providing an advantage to the bacteria.18,22
In 2012, an epidemic year for pertussis in the United States, Vermont reported the second highest pertussis incidence rate (103/100 000 population) in the country.7 During this period, Vermont Department of Health Laboratory was one of a few laboratories in the nation that routinely cultured all specimens from suspected pertussis cases. This provided a unique opportunity to determine pertactin status for many pertussis cases in Vermont. An initial analysis demonstrated that of those cases with isolates available, >90% were pertactin-deficient.18 It was in this setting of high pertactin deficiency that we sought to assess vaccine effectiveness (VE) and duration of protection of the 5-dose DTaP childhood series and the adolescent Tdap dose.
Study Design and Population
To assess VE and duration of protection of DTaP and Tdap, we conducted 2 matched case-control studies. We included persons aged 4 to 10 years (born 2000–2009) and 11 to 19 years (born 1991–2002) for the DTaP and Tdap evaluations, respectively. All probable and confirmed pertussis cases reported to the Vermont Department of Health with cough onset from January 1, 2011, to December 31, 2013 were included, unless a medical home could not be identified for a case or it was based outside of Vermont. Cases were classified according to Vermont Department of Health definitions, a modified version of the Council of State and Territorial Epidemiologists definitions.23,24 A clinical case was considered any person with cough illness lasting ≥2 weeks with paroxysms of coughing, inspiratory “whoop,” or posttussive vomiting. A confirmed case was defined as any person with cough plus Bordetella pertussis isolation from a clinical specimen, or a clinical case with either a positive polymerase chain reaction (PCR) result or contact with a laboratory-confirmed case (epidemiologic link). A probable case was defined as a clinical case (ie, not laboratory-confirmed or epidemiologically linked to a laboratory-confirmed case).
Three controls were selected for each case, matched on medical home for the DTaP evaluation, and medical home and birth year for the Tdap evaluation. Controls were not matched on birth year for the DTaP evaluation due to the expected high correlation of age with vaccination status. Controls were randomly selected from patient rosters that included patients in the target age range with at least 1 clinic visit since January 1, 2006. Controls who had an out-of-state home zip code, only had an emergency or specialist visit at the medical home, or were suspected of having had pertussis in the previous 12 months were ineligible. The cough-onset date of the case was used as the enrollment date for each case and its matched controls.
Demographic information (age, gender, ethnicity, race, insurance, and Vaccines for Children program eligibility), and pertussis vaccination history were collected for all participants at their medical home by using a standardized abstraction form. Vaccine administration date, type, brand, manufacturer, and lot number were recorded for all pertussis-containing vaccines. If a participant’s pertussis vaccination history was incomplete, pertussis-containing vaccine receipt was verified by interviewing parents. Of history provided by parents, only vaccinations with exact dates (day, month, year) were considered verified. Access to the Vermont Immunization Registry was limited by state law, and could not be used to verify vaccination history.25
The number of pertussis-containing vaccines received at least 14 days before the enrollment date was determined for each participant, based on vaccination dates documented in the medical chart or provided by parents. For the 5-dose DTaP series, participants were considered vaccinated on-schedule if they had received doses 1 to 3 at age <1 year, dose 4 at age 1 to <2 years, and dose 5 at age 4 to <7 years.3 Participants who received 5 DTaP doses but did not meet this schedule were considered to be vaccinated off-schedule. Participants who received only 4 DTaP doses were considered vaccinated on a catch-up schedule if they had received doses 1 to 3 at age <4 years and dose 4 at age 4 to <7 years. Participants were considered unvaccinated with DTaP if no pertussis-containing vaccine was documented in their chart and nonreceipt was confirmed by a parent. If receipt of vaccination could not be confirmed for any of the DTaP doses by chart or parent interview, participants were classified with unknown vaccination status. Participants were considered vaccinated with Tdap if they had received 1 Tdap dose at age ≥11 years. They were considered unvaccinated if they had no Tdap documented in their chart and a parent confirmed nonreceipt. Participants without documentation of Tdap vaccination in their chart and for whom the parent interview was inconclusive were classified with unknown vaccination status.
B pertussis Pertactin Deficiency Determination
During the epidemic period, 66% of reported pertussis cases were confirmed by laboratory testing, and, of these, 80% were tested by the state laboratory, representing ∼50% of all reported pertussis cases. All specimens sent to the state laboratory underwent both culture and PCR testing; other laboratories conducted only PCR testing. Specimens were cultured per standard procedure, and pertussis-positive cultures (73% of cultured specimens) were tested for pertactin deficiency at the Centers for Disease Control and Prevention (CDC) as described previously.17,18
Participants were excluded from the analysis if their vaccination history was unknown, or if they had ever been reported to Vermont Department of Health as a previous pertussis case. Specific DTaP analysis exclusion criteria included the following: mis-administration of Tdap, Tdap receipt >14 days before enrollment to exclude individuals that had received both DTaP and Tdap vaccines, receipt of >5 DTaP doses, receipt of 5 DTaP doses off-schedule, or receipt of <5 DTaP doses. Tdap analysis exclusion criteria included the following: mis-administration of DTaP, receipt of >1 Tdap dose, receipt of Tdap before age 11 years, or receipt of Tdap before 2006. Individuals who had received a Tdap dose before 2006 were excluded as the record was likely to be erroneous because Vermont only introduced Tdap in 2006.
Demographics of cases and controls were compared by using matched odds ratios (ORs) from conditional logistic regression analyses; where cells had <5 observations, exact analyses were conducted. The 2-sided Wilcoxon rank-sum test assessed for differences in age-related characteristics.
For the primary analyses, conditional logistic regression was used to determine ORs for the association of pertussis with receipt of (1) 5-dose DTaP series on-schedule or (2) Tdap on-schedule. VE was calculated as (1 − OR) × 100%. Unvaccinated participants were used as the referent for all models. To evaluate duration of protection, ORs were calculated for time since receipt of fifth DTaP dose (classified as follows: <12, 12–23, 24–35, 36–47, 48–59, or >59 months) or Tdap (classified as follows: <12, 12–23, or >23 months).
To assess stability of the DTaP VE estimates, a number of sensitivity analyses were conducted by restricting the analytic population to the following: a combined group including 5 DTaP doses on-schedule and 4 doses on catch-up schedule; 5 DTaP doses regardless of schedule; 5 DTaP doses on-schedule, excluding public health districts with higher percentages of unvaccinated participants; and confirmed cases only and their matched controls.
Since 1993, the Vermont Immunization Program has purchased all recommended pertussis vaccines for children.26 We reviewed Immunization Program–provided vaccine distribution data, and determined the transition to acellular pertussis vaccines occurred during 1997, after which we assumed that whole-cell vaccines were no longer available. We verified this assumption by cross-checking study-obtained lot numbers against manufacturer data. The primary analysis for Tdap VE was restricted to participants who had received only acellular vaccines for all childhood series and adolescent doses (born after 1997), although we attempted to evaluate Tdap VE among participants who received a mix of whole-cell and acellular vaccines (born 1997 or earlier). The stability of the Tdap VE estimates was assessed by conducting sensitivity analyses restricting the analytic populations to the following: 5 DTaP doses on-schedule, 5 DTaP doses regardless of schedule, and confirmed cases only and their matched controls.
All analyses were conducted in SAS 9.3 (SAS Institute, Inc, Cary, NC).
This evaluation was determined to be a public health evaluation and designated as nonresearch by CDC Human Research Protection Office and the Vermont Agency of Human Services.
From January 1, 2011, to December 31, 2013, 1252 pertussis cases were reported to Vermont Department of Health. Sixty-eight percent of patients were aged 4 to 19 years (n = 848) and, of these, 73% (n = 624) were reported during 2012. Twenty-eight patients (3%) were ineligible because their medical home was located outside of Vermont (n = 20), could not be determined (n = 7), or declined to participate (n = 1). Data were collected on 820 patients aged 4 to 19 years and 2369 controls from 91 medical homes. For 2673 participants (83.8%), vaccine history was confirmed by medical chart review, for 160 participants (5.0%), nonreceipt or incomplete pertussis vaccine history was verified by parental recall.
Five-Dose DTaP Series VE and Duration of Protection
Data were collected on 382 cases and 1113 controls aged 4 to 10 years to evaluate DTaP VE. Overall, 119 cases (31%) and 387 controls (35%) were excluded from analyses (Table 1). The proportion of cases and controls excluded was similar for nearly all exclusion criteria, except excluded controls were more likely than excluded cases to have received <5 DTaP doses.
Of the 263 cases included in the analysis, 71% (n = 186) were classified as confirmed and 29% (n = 77) as probable. Of confirmed cases, 83% (n = 154) were laboratory-confirmed and 17% (n = 32) were epidemiologically linked. Eighty-five specimens (representing 32% of all cases included in the DTaP analysis) were tested for pertactin deficiency; 83 (98%) of these were pertactin-deficient. Cases and controls had similar demographics, but cases were more likely to be older (P < .01) and unvaccinated (P < .01; Table 2). Age at receipt of fifth DTaP dose was similar for cases and controls (P = .32); however, the median time since receipt of the fifth DTaP dose was longer for cases (P < .01; Table 2).
For the primary analysis, the estimated overall VE of the 5-dose DTaP series on-schedule was 84% (95% confidence interval [CI] 58%–94%; Table 3). When participants were stratified by time since fifth DTaP dose receipt, the estimated VE within 12 months was 90% (95% CI 71%–97%). By 60 to 83 months, VE declined to 68% (95% CI 10%–88%; Table 3). Sensitivity analyses, including VE among confirmed cases only, were not substantially different from the primary analysis, with overlapping confidence intervals noted (Supplemental Table 6).
Tdap VE and Duration of Protection
Data were collected on 438 cases and 1256 controls aged 11 to 19 years to evaluate Tdap VE. Overall, 66 cases (15%) and 166 controls (13%) were excluded from analyses (Table 4). The proportion of cases and controls excluded was similar for all reasons.
Of the remaining 372 cases, 80% (n = 297) were classified as confirmed and 20% (n = 70) as probable. Of confirmed cases, 90% (n = 266) were laboratory confirmed and 10% (n = 31) epidemiologically linked. A total of 110 specimens (representing 30% of all cases included in the Tdap analysis) were tested for pertactin deficiency and 104 (95%) of these were pertactin-deficient. Cases and controls had similar demographics, but cases were more likely to be unvaccinated (P < .01; Table 5). Age at receipt of Tdap was similar for cases and controls (P = .13); however, median time since Tdap receipt was slightly longer for cases (30 months vs 26 months, P < .01; Table 5).
For the primary analysis of Tdap VE, which comprised participants who received only acellular vaccines, 244 cases (66% of all eligible cases) and 714 controls (66% of all eligible controls) were included. The estimated overall VE of the Tdap was 70% (95% CI 54%–81%; Table 3). When participants were stratified by time since Tdap vaccination, VE within 12 months was 76% (95% CI 60%–85%). By 24 to 46 months, VE declined to 56% (95% CI 16%–77%; Table 3). The Tdap VE estimates remained stable for all sensitivity analyses, including VE among confirmed cases only, with CIs that overlapped those of the primary analysis (Supplemental Table 6). VE among participants who received a mix of whole-cell and acellular vaccines for the childhood series (born in 1997 or earlier) could not be estimated because of insufficient number in the referent group (n = 24).
Our results provide the first estimates of VE against reported pertussis predominately caused by pertactin-deficient B pertussis. Remarkably, our findings are consistent with earlier published VE studies, even though >90% of tested isolates included in our evaluation were pertactin deficient.11,12,27–31 For example, pertussis VE was assessed during the 2010 California outbreak (DTaP; VE 89%, 95% CI 79%–94%) and the 2012 Washington State outbreak (Tdap; VE 64%, 95% CI 50%–74%); the proportions of pertactin-deficient strains during these outbreaks were estimated to be 14% and 76%, respectively.11,12,17,18 Our findings suggest that both acellular pertussis vaccines remain protective against reported pertussis disease in the setting of high pertactin deficiency.
The role of pertactin in pathogenesis is unclear, with potential roles in adhesion of the bacterium to respiratory tract epithelium and resistance to neutrophil-mediated clearance.20,32 However, pertactin-deficient strains show no deficiency in growth or infection, suggesting a high level of redundancy in the genes responsible for these functions.33,34 Pertactin deficiency thus far has not been associated with any detectable changes in the clinical disease process or disease severity.18,35 In animal models, infection of a baboon with a pertactin-deficient strain showed no defect in colonization or leukocytosis compared with a pertactin-expressing strain, and infection of mice resulted in only minor deficiencies in lung colonization.36,37 In humans, a higher proportion of patients infected with pertactin-expressing strains of B pertussis reported apnea, but otherwise there was no difference.18,35
The rapid increase in pertactin-deficient B pertussis in multiple countries, combined with evidence that several mutations can result in lack of pertactin, suggest pertactin deficiency may confer a selective advantage to the bacteria.22,38–41 More specifically, several recent studies indicate pertactin deficiency may especially benefit the bacteria among a highly immunized population. For instance, pertactin-deficient strains were found to colonize mice primed with acellular pertussis vaccine more effectively than pertactin-expressing strains; a separate study found that acellular vaccinated individuals had two- to fourfold greater odds of being infected with a pertactin-deficient rather than a pertactin-expressing B pertussis strain.18,22 Pertactin deficiency may also facilitate transmission by supporting longer infections: acellular-vaccinated mice infected with pertactin-deficient pertussis sustained longer infections than those infected with pertactin-expressing strains.34 In addition, pertactin deficiency may result in improved transmission by avoidance of pertactin-specific vaccine-induced host immune responses. This would be especially relevant to acellular vaccine–primed individuals, in whom it has been demonstrated that antibodies to pertactin correlated with protection.42 Through these mechanisms, pertactin deficiency may be amplifying the pertussis disease resurgence, especially in the setting of exclusive use of acellular vaccines that are associated with waning immunity and that fail to prevent B pertussis colonization and transmission.43–45
Because our findings suggest pertussis vaccines remain protective in the setting of high prevalence of pertactin deficiency, other vaccine components (pertussis toxin, filamentous hemagglutinin, or fimbriae) are likely preventing symptomatic pertussis. Antibodies against pertussis toxin were shown to correlate modestly with protection in 2 trials, and Denmark has used a mono-component pertussis toxin vaccine for >15 years with an estimated VE of 78%, although this efficacy is now being questioned.42,46–48 However, data regarding protection elicited by filamentous hemagglutinin or fimbriae-specific immune responses are limited. Because correlates of protection are not well defined for pertussis, it is difficult to determine which vaccine components may be eliciting the necessary immune response.
Case-control study designs are susceptible to a number of limitations, including information, selection, and misclassification biases. To mitigate these, controls were selected from the same medical home as the cases to ensure exposure to similar circulating B pertussis strains and to limit provider-associated diagnostic and reporting biases. To control for information bias, we collected vaccination history in a standardized manner. Although the evaluation was limited to 1 state where pertactin deficiency was extremely high, recent analyses of specimens from around the country showed similar high levels of pertactin deficiency associated with the same insertion sequence, extending the applicability of these findings to across the United States.17,18 A further limitation was the low proportion of participants with confirmed pertactin status. All available culture isolates were tested for pertactin, but these represented only 40% of all cases included in the analyses: those that were Vermont state laboratory confirmed and B pertussis culture positive. Due to the low proportion of cases with confirmed pertactin status, the tested isolates may not be representative of all circulating pertussis strains. In addition, if there is a selective advantage to pertactin deficiency among vaccinated individuals, as suggested by Martin et al18 and Safarchi et al,22 we may expect more pertactin-expressing strains among unvaccinated cases. By including these cases, VE could be overestimated. Although a third of participants were excluded from the DTaP evaluation, the proportion of cases and controls excluded was similar for nearly all criteria, and is therefore unlikely to have biased our results. The one exception to this was the criteria for receipt of ≤4 doses DTaP: controls were more likely than cases to be excluded. This may be because controls were younger than cases, and therefore had less opportunity to complete their DTaP series before enrollment. Finally, our VE estimates are also unlikely to account for mild disease, which may be more prominent among vaccinated individuals and less likely to be reported.49,50
In summary, genetic changes in B pertussis may be one of the factors contributing to the recent resurgence in pertussis. We have shown that the current acellular vaccines continue to be effective in the setting of high pertactin-deficient B pertussis prevalence, and therefore remain the best way to protect against severe disease. However, further investigation is needed to better understand the implications of pertactin deficiency on pertussis pathogenesis and host immunologic response, which could provide insight into the development of novel pertussis vaccines.
This evaluation would not have been possible without the cooperation and support of the providers from 91 Vermont medical homes. We additionally thank the Vermont Department of Health (Judy Ashley, Sasha Bianchi, James Biernat, Christine Bongartz, Destiny Cadieux, Joanne Calvi, Brian Campion, Laura Cody McNaughton, Moira Cook, Heather Danis, Twyla Deneergaard, Jeff Heath, Patricia Hennard, Barb Kruml, Cindy Laraway, Joyce Larro, Megan Lausted, Prudence MacKinney, Sara Moran, Rebecca Olmstead, Sarah Orr, Jenny Rafuse, Monica Raymond, Tara Reil, Melissa Richards, Linda Seel, Shannon Tatro, Becky Thomas, Valerie Valcour, Wendy Walsh, Sydney White, Janet Wiatrowski), the following Epidemic Intelligence Service Officers (Dinorah Calles, Michelle Chevalier, Meredith Dixon, Rajni Gunnala, Candice Johnson, Hajime Kamiya, Samir Koirala, Gayathri Kumar, Michael Lowe, Almea Matanock, Lucy McNamara, Jonathan Meiman, Carla Mercado, Leigh Ann Miller, Tosin Ogunmoyero, Niyi Olayinka, Monica Patton, Tara Perti, Mary Puckett, Alison Ridpath, Candice Williams), and other CDC fellows and staff (Marieke Bierhoff, Amy Blain, Kayla Bronder, Hema Datwani, Amanda Faulkner, Jade Fettig, Adria Lee, Neil Murthy, Jarratt Pytell, Tess Wiskel) for their essential contributions in data collection. Additionally from the Vermont Department of Health we thank Bradley Tompkins for his help with the Vermont surveillance data and pertussis case identification, Karen Halverson for supplying Vermont’s historical vaccine information, and Margaret Wilson for her advice on evaluation logistics planning. From CDC we thank Tami Skoff for her critical review of the manuscript.
We also acknowledge the work of Keeley Weening from Vermont Department of Health Laboratory, and Pam Cassiday, Lucia Pawloski, Lucia Tondella, and Margaret Williams from the Meningitis and Vaccine Preventable Diseases Branch Pertussis Laboratory on pertussis isolate culture and pertactin testing.
- Accepted February 17, 2016.
- Address correspondence to Lucy Breakwell, PhD, Centers for Disease Control and Prevention, 1600 Clifton Rd, MS C25, Atlanta, GA 30333. E-mail:
The findings and conclusions are those of the authors and do not necessarily represent the views of the Centers for Disease Control and Prevention.
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: All phases of this evaluation were supported by the Centers for Disease Control and Prevention. The findings and conclusions are those of the authors and do not necessarily represent the views of the Centers for Disease Control and Prevention.
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.
- Cody CL,
- Baraff LJ,
- Cherry JD,
- Marcy SM,
- Manclark CR
- Long SS,
- Deforest A,
- Smith DG,
- Lazaro C,
- Wassilak GF
- Broder KR,
- Cortese MM,
- Iskander JK, et al; Advisory Committee on Immunization Practices (ACIP)
- Elam-Evans LD,
- Yankey D,
- Jeyarajah J, et al; Immunization Services Division, National Center for Immunization and Respiratory Diseases; Centers for Disease Control and Prevention (CDC)
- Acosta AM,
- DeBolt C,
- Tasslimi A, et al
- Pawloski LC,
- Queenan AM,
- Cassiday PK, et al
- Martin SW,
- Pawloski L,
- Williams M, et al
- Wells TJ,
- Tree JJ,
- Ulett GC,
- Schembri MA
- Inatsuka CS,
- Xu Q,
- Vujkovic-Cvijin I, et al
- American Academy of Pediatrics
- Vermont Department of Health
- ↵Vermont General Assembly. Immunization Registry Act of 1997. 18 VSA § 1129. Available at: http://legislature.vermont.gov/statutes/section/18/021/011292007. Accessed October 22, 2013
- Vermont Department of Health
- Baxter R,
- Bartlett J,
- Rowhani-Rahbar A,
- Fireman B,
- Klein NP
- Liko J,
- Robison SG,
- Cieslak PR
- Koepke R,
- Eickhoff JC,
- Ayele RA, et al
- Leininger E,
- Roberts M,
- Kenimer JG, et al
- Bowden KE,
- Williams MM,
- Cassiday PK, et al
- Warfel JM,
- Zimmerman LI,
- Merkel TJ
- Barlow RS,
- Reynolds LE,
- Cieslak PR,
- Sullivan AD
- Liese JG,
- Renner C,
- Stojanov S,
- Belohradsky BH; Munich Vaccine Study Group
- Copyright © 2016 by the American Academy of Pediatrics