Impact of the US Maternal Tetanus, Diphtheria, and Acellular Pertussis Vaccination Program on Preventing Pertussis in Infants <2 Months of Age: A Case-Control Evaluation

Abstract

Despite the dramatic impact of vaccines on the burden of Bordetella pertussis in the United States, pertussis remains endemic, and reported cases have increased steadily since the late 1980s. In 2012, >48000 pertussis cases were reported in the United States, the largest number since the mid-1950s. While causes for the increase are likely multifactorial, waning immunity from acellular pertussis vaccination has been documented to be a strong contributor [1].

Infants are at greatest risk for pertussis-related complications and mortality, especially during the first months of life [1]. Immunization against pertussis with the 5-dose childhood DTaP (diphtheria toxoid, tetanus toxoid, and acellular pertussis) vaccine series begins at 2 months of age, leaving young infants highly susceptible to pertussis. Among case patients <2 months old, approximately 75% are hospitalized and 1 in 100 die [2]. Studies suggest that parents and siblings play an important role in transmitting pertussis to vulnerable infants [3, 4].

In 2005, 2 tetanus toxoid, reduced diphtheria toxoid, and acellular pertussis (Tdap) vaccines were licensed for use as a single booster dose among US adolescents and adults [5, 6]. Although the primary goal of Tdap immunization was to provide direct protection to the vaccine recipient, recommendations also focused on indirectly protecting infants through vaccination of close contacts, a strategy known as “cocooning” [5, 6]. In 2011, the Advisory Committee on Immunization Practices (ACIP) recommended vaccinating women with Tdap during a single pregnancy as a strategy to protect young infants through the transplacental transfer of maternal antibodies [7]. Because maternal antibodies are short-lived and may not be sufficient to protect infants of subsequent pregnancies, the recommendation was expanded in 2012 to include a dose of Tdap during every pregnancy [8, 9]. Because the recommendation for vaccination during pregnancy was made with limited data, postimplementation evaluations are essential for monitoring effectiveness and longer-term success of the strategy.

We conducted a multistate, case-control evaluation to determine the effectiveness of maternal Tdap vaccination during pregnancy at preventing pertussis in US infants <2 months old.

METHODS

A case of pertussis was defined as the onset of cough illness and at least 1 of the following: laboratory confirmation (culture or polymerase chain reaction) of pertussis, epidemiological linkage to a laboratory-confirmed case, or clinically compatible illness (cough ≥2 weeks with paroxysms, inspiratory whoop, or posttussive vomiting) in an infant <2 months old between 1 January 2011 and 31 December 2014. Cases were identified through surveillance in 6 Emerging Infection Program Network sites [10]. The Emerging Infections Program Network is a collaborative network between the Centers for Disease Control and Prevention and state and local health departments, academic institutions, and laboratories that serves as a national resource for surveillance, prevention, and control of emerging infectious diseases. This case-control evaluation was conducted statewide in California, Connecticut, Minnesota, and New Mexico, and in select counties of New York (Albany, Allegany, Cattaraugus, Chautauqua, Chemung, Clinton, Columbia, Delaware, Erie, Essex, Franklin, Fulton, Genesee, Greene, Hamilton, Livingstone, Montgomery, Monroe, Niagara, Ontario, Orleans, Otsego, Rensselaer, Saratoga, Schenectady, Schoharie, Schuyler, Seneca, Steuben, Warren, Washington, Wayne, Wyoming, and Yates) and Oregon (Clackamas, Multnomah, and Washington). Infants were eligible for enrollment if they were at least 2 days old and resided in the catchment area on their cough onset date, were born in a hospital in their state of residence, were ≥37 weeks’ gestational age at birth, were not adopted or in foster care, and did not live in a residential care facility. For each enrolled case infant, we attempted to recruit 3 control infants from birth certificates of infants born at the same hospital as the case infant that were <2 months old on the case infant’s cough onset date; once all potential controls meeting these criteria were exhausted, control enrollment ended for that case infant. Eligibility criteria were the same for control infants as for case infants, and control infants were additionally considered ineligible if they had a pertussis diagnosis prior to the cough onset date of the corresponding case infant.

Mothers of case and control infants were interviewed by telephone to collect information on demographics, mother and infant healthcare providers, and infant household contacts. The reference period for a case infant and his or her matched controls was defined as the 30-day period prior to the case infant cough onset date. Data used in the comparison of enrolled and unenrolled infants with pertussis were obtained from surveillance case report forms, maternal interviews, and birth certificate records; surveillance case report form data, which included hospitalization status, were collected via patient and physician interview.

Pertussis vaccination status, including brand, manufacturer, and lot number, was collected through medical providers or state immunization registries for all enrolled case and control infants and their mothers. When complete vaccine history was unavailable in registries, all medical providers identified during the interview were contacted. Additionally, birth hospitals were contacted to obtain case and control maternal Tdap histories. Tdap doses were considered valid if received at least 2 weeks before the case infant’s cough onset date; for control infants, the date of cough onset for the matched case infant was used. Maternal vaccine history was considered complete when follow-up was exhausted with all providers; if there was incomplete follow-up with providers but at least 1 valid Tdap dose was identified, these individuals were included in the analysis. Mothers were considered unvaccinated if all medical providers were contacted and did not provide documentation of Tdap vaccination and no Tdap records were identified in the immunization registry. When >1 Tdap dose was verified, the most recent valid dose was included in the analysis. Mothers were classified as vaccinated before pregnancy if Tdap was received at any point prior to pregnancy with the case or control infant, vaccinated during pregnancy if they received Tdap ≥14 days before delivery, and vaccinated postpartum if Tdap was received <2 months following the case or control infant’s birth or in the 14 days before delivery; the trimester during which Tdap was administered was calculated from the vaccination date, infant’s date of birth, and infant’s gestational age at birth. When available, lot numbers were verified with vaccine manufacturers to confirm vaccine type and brand. An infant was considered enrolled when the maternal interview, and infant and maternal vaccine history were all completed.

Data were analyzed using SAS version 9.3 software. Odds ratios were calculated using multivariable conditional logistic regression; vaccine effectiveness (VE) was estimated as (1 – odds ratio) × 100%. Variables associated (P < .05) with maternal Tdap vaccination in bivariate analyses were included in the multivariable models; those that retained significance in multivariable analysis after backwards elimination were included in the final model. Although not significant, we retained infant age (weeks) in the final models, because cases and controls were not matched on age. Case and control mothers classified as unvaccinated were used as the reference group. Differences between proportions were tested using Pearson χ² test or Fisher’s exact test; differences in medians were assessed using the Wilcoxon rank-sum test.

RESULTS

A total of 788 infants <2 months old with pertussis were identified; of these, 29 (3.7%) were born prematurely, 6 (0.76%) were adopted or resided in foster care, 5 (0.63%) were born at a hospital outside their state of residence, and 3 (0.38%) were not born in a hospital. Of 745 infants eligible for enrollment, 251 (33.7%) were enrolled. The remaining 494 (66.3%) were not enrolled for the following reasons: 354 (71.7%) had mothers that could not be reached, 100 (20.2%) did not consent, 31 (6.3%) had incomplete maternal vaccination history follow-up, 7 (1.4%) did not speak English or Spanish, and 2 (0.4%) had another reason for nonenrollment. Compared with unenrolled infants with pertussis, mothers of enrolled infants were significantly more likely to have post–high school education; no significant differences were observed for sex, race, ethnicity, hospitalization, outcome, or insurance type (Table 1).

We identified 5507 infants eligible to be enrolled as controls, 682 (12.4%) of whom were enrolled in the evaluation. The remaining 4825 were not enrolled for the following reasons: 4124 (85.5%) had mothers that could not be reached, 623 (12.9%) did not consent, 53 (1.1%) had incomplete maternal vaccination history follow-up, 6 (0.1%) resided outside of the catchment area on corresponding case infant cough onset date, 4 (0.1%) did not speak English or Spanish, and 15 (0.3%) had another reason for nonenrollment. Demographics of enrolled case and control infants are shown in Table 2. Because there was a disproportionate number of enrolled control infants compared to enrolled case infants among those <2 weeks old, we restricted our analysis to case infants who were ≥2 weeks old on their cough onset date and control infants who were ≥2 weeks old on their corresponding case infant’s cough onset date (Table 2); 11 (4.4%) case infants and 147 (21.6%) control infants were excluded. The population for analysis included 240 case infants and 535 control infants.

Overall, 136 (56.7%) case mothers and 358 (66.9%) control mothers had at least 1 valid Tdap dose identified. Eighteen of 136 (13.2%) vaccinated case mothers and 50 of 358 (14.0%) vaccinated control mothers had >1 valid dose of Tdap reported (P = .83); 61 mothers had 2 documented doses of Tdap, and 7 had 3. Among the 61 women with 2 Tdap doses, the median time between doses was 1022 days (range, 7–2744 days) and did not differ significantly between case and control mothers (P = .32).

Figure 1A–C shows the distribution of Tdap doses included in the final model and their timing of administration in relation to pregnancy. Of Tdap doses received during pregnancy (associated with 22 cases and 117 controls), approximately 77% were received during the third trimester, most during the ACIP-recommended 27–36 weeks of gestation (Figure 1A). For the 24 case-associated and 67 control-associated doses received before pregnancy, 6 (25.0%) case mothers and 46 (68.7%) control mothers received Tdap ≤2 years before pregnancy (Figure 1B). For doses classified as postpartum, 75.9% of control-associated doses and 74.4% of case-associated doses were given within the first 2 days following birth; 5 (5.6%) doses among case mothers and 7 (4.0%) among control mothers were received during the last 2 weeks of pregnancy and therefore classified as postpartum (Figure 1C).

Figure 1.

A, Timing of tetanus toxoid, reduced diphtheria toxoid, and acellular pertussis (Tdap) vaccine doses classified as during pregnancy. The white bars represent Tdap doses received by control-associated mothers during pregnancy, and the gray bars represent Tdap doses received by case-associated mothers during pregnancy. Two cases were missing gestational age so the exact week of Tdap administration could not be calculated; based on date of birth and Tdap date, both cases were included in the second trimester of pregnancy in the analysis models. B, Timing of Tdap doses classified as before pregnancy. The white bars represent Tdap doses received by control-associated mothers before pregnancy, and the gray bars represent Tdap doses received by case-associated mothers before pregnancy. One case was missing gestational age so the exact month of Tdap administration before pregnancy could not be calculated; based on date of birth and Tdap date, this case was included in the ≤2 years group in the analysis models. C, Timing of Tdap doses classified as postpartum. The white bars represent Tdap doses received by control-associated mothers during the postpartum period, and the gray bars represent Tdap doses received by case-associated mothers during the postpartum period.

Open in new tabDownload slide

The overall effectiveness of vaccination at any time during the third trimester of pregnancy was 77.7% (95% confidence interval [CI], 48.3%–90.4%; Table 3); the effectiveness of Tdap given during the first or second trimester was 64.3% (95% CI, –13.8% to 88.8%), but confidence bounds overlapped with those for the third trimester. When restricting third-trimester doses to the recommended window of 27–36 weeks, VE was 78.4% (95% CI, 49.8%–90.7%). There was no effectiveness when a dose of Tdap was given to the mother during the postpartum period (4.9%; 95% CI, –49.3% to 39.5%; Table 3).

Tdap given at any point before pregnancy was 50.8% (95% CI, 2.1%–75.2%) effective at preventing infant pertussis (Table 3). We further stratified this analysis by examining Tdap doses given ≤2 years before pregnancy and Tdap doses given >2 years before pregnancy; the effectiveness of Tdap given >2 years before pregnancy was –25.6% (95% CI, –207.4% to 48.7%) and the effectiveness of Tdap given ≤2 years before pregnancy was 83.0% (95% CI, 49.6%–94.3%). The point estimate for Tdap given ≤2 years before pregnancy was not significantly different than the point estimate for Tdap given during the third trimester (P = .6034).

Approximately 65.4% (157/240) of case-infants were hospitalized during the course of their pertussis infection. When we examined the effectiveness of Tdap at preventing infant pertussis hospitalizations, the VE point estimate for vaccination during the third trimester increased to 90.5% (95% CI, 65.2%–97.4%). An increase in VE point estimates was also observed for doses given at any point before pregnancy, during the first or second trimester, and postpartum; however, the postpartum estimate remained nonsignificant (Table 4).

A total of 43 of 136 (31.6%) vaccinated case mothers and 102 of 358 (28.5%) vaccinated control mothers received Boostrix, and 76 of 136 (55.9%) vaccinated case mothers and 207 of 358 (57.8%) vaccinated control mothers received Adacel; brand information was not available for the remaining 17 of 136 (12.5%) case mothers and 49 of 358 (13.7%) control mothers who had record of Tdap receipt. When we calculated the effectiveness of either vaccine product at preventing infant disease when administered during the third trimester, point estimates were not statistically different from one another (P = .85).

DISCUSSION

Our findings, using data from multiple US states, add to the growing body of evidence that vaccination during pregnancy is effective, underscoring its importance as a key strategy for preventing pertussis in young infants. The United Kingdom was first to evaluate Tdap vaccination during pregnancy, demonstrating high effectiveness at preventing pertussis during the first 2 months of life [11, 12]. Vaccination during pregnancy was recommended in the United Kingdom in late 2012 owing to a sudden increase in pertussis morbidity and mortality among infants, and high Tdap coverage (60%) was rapidly achieved among pregnant women [11]. Three years postintroduction, high effectiveness has been sustained [13]. Contrary to the UK experience, US uptake of the recommendation has progressed at a much slower pace. A recent survey reports that 48.8% of US pregnant women received Tdap during the 2015–2016 influenza season, an increase of 21.8% from the 2013–2014 season [14]. Although coverage continues to increase, there is likely considerable variation in vaccine uptake across the United States and among medical providers caring for pregnant women. Emphasis should be placed on maximizing Tdap uptake during pregnancy in order to optimize the benefits of the strategy.

Although our results are consistent with those from a US cohort analysis, we found slightly lower effectiveness of vaccination during pregnancy than in the United Kingdom, where estimates were obtained using similar case-control methodology; a second, more recent US cohort study also found higher point estimates of VE, but the confidence intervals were wide and included our estimated effectiveness [11, 15, 16]. Ninety percent of infants <1 year old with pertussis reported in the United Kingdom between 2002 and 2009 were hospitalized [17]. In contrast, US data indicate that 34%–69% of infants with pertussis are hospitalized [3, 6, 18, 19], with the proportion of hospitalizations decreasing in recent years [2]. While there could be real differences in the epidemiology of severe disease or the threshold for hospitalization of suspected infant cases, this disparity more likely reflects differences in case ascertainment practices and the ability of US surveillance to capture more outpatient illness [10]. Interestingly, when our models were restricted to hospitalized cases, higher vaccine effectiveness was observed, closely mirroring the UK estimates. In the UK study, all cases were laboratory-confirmed, which could also lead to a higher VE estimate; however, the majority of our cases were also laboratory-confirmed (94%), suggesting that this does not account for our lower estimate.

Sixty-eight percent of our cases were from California (Table 1), which has experienced large pertussis epidemics in recent years, with a proportion of cases in Hispanic infants that is greater than the proportion of Hispanics in the California birth cohort [20]. As a result, 62% of our enrolled cases were Hispanic (Table 1), which is higher than US national estimates (49% of pertussis cases aged <2 months with known ethnicity, US National Notifiable Disease Surveillance System, 2011–2014, unpublished data). While an increased risk of pertussis in Hispanic infants has been reported, additional study is needed to fully understand why Hispanic infants are at increased risk [21]. It is reassuring that vaccination during pregnancy is highly effective in a population with a high proportion of Hispanic cases, further underscoring the importance of this strategy.

The recommendation for cocooning has been in place since the US introduction of Tdap; however, programmatic implementation of the strategy has encountered substantial logistical challenges resulting in poor Tdap uptake and incomplete coverage among infant close contacts [22, 23]. Consistent with published evaluations of the cocooning strategy, point estimates from our analysis, although not reaching significance, show no benefit of a dose of Tdap administered to the mother postpartum [16, 24, 25]. Recent data from non-human primates have suggested that acellular vaccination, unlike vaccination with whole-cell pertussis vaccine, may not preclude B. pertussis colonization and further transmission of the organism [26]. In an era of exclusively acellular vaccine use in the United States, vaccinated individuals could remain a significant source of pertussis transmission to young infants, highlighting additional shortcomings of cocooning and further underscoring the critical role of effective strategies such as vaccination during pregnancy.

Recent studies have brought into question the ideal timing of Tdap administration during pregnancy, providing evidence that vaccination during the second or early third trimester may maximize transplacental transfer of maternal pertussis antibodies to the infant [27–29]. While active transport of maternal antibodies is thought to be minimal before 30 weeks’ gestation, vaccination too late in pregnancy could diminish protection provided to the infant [30]. Because most Tdap-vaccinated mothers in our evaluation received vaccine during the recommended third trimester, our sample size was not adequate to detect a significant difference in VE between third-trimester doses and doses administered earlier in pregnancy, and small numbers left us unable to evaluate narrower time periods. Our analysis did find benefit of a dose of Tdap given in the 2 years prior to pregnancy, with a VE point estimate not significantly different than that of a dose administered during the third trimester. While we were not powered to detect a significant difference between point estimates, this finding is consistent with other VE studies that have shown benefit of Tdap prior to pregnancy [16, 31]. Future studies should not only continue to evaluate the infant immune response to Tdap doses administered prior to the third trimester of pregnancy, but also assess how transferred antibody correlates with protection against clinical disease.

Although Tdap vaccination during pregnancy is effective at preventing infant pertussis, important questions remain. Reduced Tdap effectiveness has been documented among adolescents primed with acellular pertussis vaccines compared to cohorts vaccinated with whole-cell vaccine [32–34]. Whether the effectiveness of vaccination during or in the 2 years prior to pregnancy will diminish as the number of aP-primed pregnant women increases is yet to be determined; the exact timing of Tdap administration could become more critical among aP-vaccinated mothers if antibodies diminish more rapidly in these women. Maternal immunization during pregnancy also has the potential to blunt an infant’s immune response to the DTaP childhood series. While circulating maternal antibodies decline rapidly and potential interference is expected to be short-lived, studies have generated conflicting evidence on blunting, and its clinical relevance is still largely unknown [35–40]. Continued monitoring of the epidemiology of pertussis will be important to detect potential shifts in the age distribution of childhood disease.

Maximizing the protection of young infants remains a high priority, especially in the setting of increased pertussis activity. Efforts should focus on the promotion of maternal immunization through provider and patient education to increase Tdap uptake among pregnant women (www.cdc.gov/pertussis/pregnant). Pertussis vaccines with improved duration of immunity are needed, but new vaccines remain on the distant horizon. While maternal immunization during pregnancy will help bridge the gap until next-generation pertussis vaccines are licensed and available for use, this highly effective strategy will likely remain an integral component of pertussis prevention and control, even in the setting of new vaccines.

Notes

Acknowledgments. We thank the following individuals for their work on the evaluation: Matt Griffith, Nancy Messonnier, Thomas Clark, Adria Lee (Centers for Disease Control and Prevention [CDC]); Pam Daily, Mohammed Khan, Sarah New, Tara Scheuer, Kathy Harriman, Kathleen Winter (California); Roxanne Ryan (Connecticut); Emily Banerjee, Rachel Ostadkar, Cynthia Kenyon, Pam Gahr (Minnesota); Kevin Aicher, Julianna Ferreira, Bernadette Gutierrez, Dana Moore-Smith, Emily Hancock (New Mexico); Prescela Perez, Kathrine Woodworth (New York); Juventila Liko, Beletshachew Shiferaw, Allison Ryan, Laura Reynolds, Rita McConathy (Oregon).

Disclaimer. The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the CDC.

Financial support. This work was supported by a CDC cooperative agreement with the Emerging Infections Program Network.

Potential conflicts of interest. All authors: No potential conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.

References

Author notes