A systematic review of the burden of pertussis disease in infants and the effectiveness of maternal immunization against pertussis

ABSTRACT

Introduction

Infants too young to be fully immunized are the most vulnerable to severe pertussis disease. To close this susceptibility gap, passive infant immunization through vaccination of pregnant women against pertussis was first introduced in 2011 in the United States and has been extended since then to more than 40 countries.

Areas covered

We conducted two systematic literature searches to describe the worldwide burden of pertussis disease in infants <6 months of age since 2005, and the effectiveness and impact of maternal pertussis vaccination in preventing infant pertussis since 2011.

Expert opinion

Pertussis disease incidence rates in infants aged <2-3 months were substantial in all countries with available data, exceeding 1000 cases per 100,000 population during outbreaks. Virtually all pertussis deaths occurred in this age group. Data from Africa, Eastern Mediterranean, and Asia were limited, but suggest a similar or higher disease burden than in Europe or the Americas. Estimates of effectiveness of second/third trimester pertussis vaccination in preventing pertussis disease in <2-3 months old infants were consistently high (69%–93%) across the observational studies reviewed, conducted in various settings with different designs. Maternal vaccination programs appear to be achieving their goal of reducing the burden of disease in very young infants.

Plain language summary

What is the context?

• Pertussis, also known as whooping cough, is a highly contagious disease of the respiratory tract.

• Infants too young to be fully vaccinated are at the highest risk of severe pertussis disease, hospitalization, and death.

• Vaccinating pregnant women against pertussis with a Tdap vaccine is recommended in more than 40 countries as a safe and effective strategy to protect infants for the first months of life.

What is new?

• This review summarizes recent literature describing the burden of pertussis disease in infants worldwide prior to the introduction of maternal vaccination programs; pertussis disease incidence rates in infants aged <2-3 months were substantial in all countries with available data, exceeding 1000 cases per 100,000 population during outbreaks.

• Immunization of pregnant women with a Tdap vaccine can prevent about 70–90% of pertussis disease and up to 90.5% of pertussis hospitalizations in infants under 3 months of age.

What is the impact?

• Limited available data suggest that incidence rates of pertussis disease after the introduction of Tdap maternal immunization have declined in infants.

• Current knowledge supports the implementation of Tdap maternal immunization programs.

KEYWORDS:

• Acellular pertussis vaccine
• burden of disease
• epidemiology
• incidence
• maternal immunization
• Bordetella pertussis
• vaccine effectiveness
• vaccine impact
• whooping cough
• infant

1. Introduction

Pertussis (whooping cough) is a highly contagious disease of the respiratory tract caused by Bordetella pertussis. Pertussis outbreaks have a cyclical pattern with natural peaks that occur every 3–5 years. Pertussis is endemic worldwide and is still difficult to control, despite decades of universal childhood vaccination [1,2]. Despite high vaccine coverage with either whole-cell (wP) or acellular (aP) pediatric pertussis vaccines in most parts of the world, pertussis continues to cause a substantial disease burden. Pertussis incidence has also been steadily rising in the last two decades in several countries with high vaccination coverage. This resurgence may be explained by multiple factors, such as increased disease awareness and improved diagnosis, faster waning of immunity following vaccination with aP compared to wP vaccines, and/or genetic changes in the bacterium. While approximately 60% of pertussis cases occur in adults and adolescents, the highest incidence of disease and the highest mortality rates occur in infants aged <2 months who are too young to be vaccinated [3,4]. In 2013, the Global Burden of Disease Study estimated mortality due to pertussis in the first year of life to be approximately 400 per million live births, or approximately 56,000 deaths annually [5].

‘Cocooning’ is a strategy that has been implemented to reduce the risk of pertussis transmission from mothers or family members to infants, by vaccinating women immediately postpartum, and all other close contacts with a combined diphtheria-tetanus-aP booster vaccine (Tdap). However, cocooning alone has proven to be insufficient to prevent pertussis morbidity and mortality in newborn infants [6]. In 2011, the United States (US) was the first country to recommend Tdap vaccination during pregnancy to protect infants in the first months of life through the passive transfer of maternal antibodies [7]. Several countries followed suite in 2012, including the United Kingdom (UK) in response to rising infant deaths during a pertussis epidemic in 2012 [8]. Vaccine effectiveness in preventing laboratory-confirmed pertussis in infants <3 months of age was more than 90% in the first 3 years of the UK program [9,10]. This was the first published evidence of the clinical benefit of a Tdap maternal immunization program [11]. Recommendations for maternal Tdap immunization now exist, with some differences in the recommended vaccination window, in over 40 countries (Table 1) [12–14]. Though evidence of the safety and benefits of maternal Tdap immunization is steadily increasing, many countries still do not have any recommendation concerning Tdap immunization in pregnancy. Where recommendations are in place, information on vaccine coverage is not always available or varies widely, ranging, for example, from 48.8% in the US in 2017 to 93.7% in the Bizkaia province of Spain [15,16].

Table 1. Recommendations for maternal immunization with tetanus toxoid, reduced diphtheria toxoid, and acellular pertussis vaccine (Tdap) for each pregnancy in countries with a Tdap maternal immunization program (accurate to our knowledge at the time of writing).

Country	Recommendation for Tdap maternal immunization	Year of introduction	Reference
Argentina*	From 20 weeks gestation	2012	[17]
Australia*	Between 28 and 32 weeks gestation	2014	[18]
Bahamas	Tdap during each pregnancy	No information	[19]
Belgium	Tdap during each pregnancy	2012	[20]
Brazil*	A 20 weeks gestation	2017	[17]
Canada	During each pregnancy	2012	[21]
Chile	Between 27 and 36 weeks gestation	2016	[22]
Colombia*	From 26 weeks gestation	2012	[17]
Costa Rica	From 20 weeks gestation	2012	[17]
Czech Republic	Between 28 and 36 weeks gestation	2015	[23]
Denmark	From 32 weeks gestation	2019	[24]
El Salvador	From 28 weeks gestation	Not known	[25]
England*	From 16 weeks gestational	2016	[26]
French Territories (Mayotte)	From 18 weeks and ideally before 39 weeks gestation	2018	[27]
Honduras	Between 26 and 37 weeks gestation	2012	[17]
Hong Kong	Anytime in the second or third trimester, preferably before 35 weeks gestation	2019	[28]
India	Between 27 and 36 weeks gestation	2013	[29]
Ireland*	Between 27 and 36 weeks gestation	2013	[30]
Israel*	Pregnant women	2015	[31]
Italy*	In the third trimester (ideally 28 weeks gestation)	2017	[20]
Korea*	Between 27 and 36 weeks gestation	2014	[32]
Mexico*	After 20 weeks gestation	2012	[17]
Netherlands*	At 22 weeks gestation	2018	[33]
New Zealand*	Between 28 and 38 weeks gestation	2013	[34]
Panama*	Pregnant women	2012	[17]
Paraguay	Pregnant women		[17]
Peru*	Between 26 and 37 weeks gestation		[35]
Philippines*	Pregnant women and post-partum	2015	[36]
Portugal	Between 20 and 36 weeks gestation		[37]
Qatar	Between 27 and 36 weeks gestation	2017	[38]
Singapore*	Between 16 and 32 weeks gestation)	2017	[39]
Slovenia	As soon as possible after 24 weeks gestation	2017	[40]
Spain*	Pregnant women	2014	[37]
Switzerland*	Pregnant women	2013	[41]
Taiwan	Between 28 and 36 weeks gestation	2012	[42]
UK*	Between 28 and 38 weeks gestation	2012	[43]
Uruguay	Pregnant women		[17]
USA*	Between 27 and 36 weeks gestation	2012	[44]

UK, United Kingdom; USA, United States of America.

* Countries for which this review presents burden of disease results.

Overall, studies of Tdap immunization of pregnant women showed that vaccination is well tolerated and immunogenic, resulting in transfer of high antibody levels to the fetus ([45–47] and reviewed in [48–50]). Fewer studies have evaluated maternal Tdap vaccine effectiveness in preventing pertussis disease in infants, or the potential impact of the timing of vaccination, or underlying medical conditions such as human immunodeficiency virus (HIV) infection, on effectiveness. In addition, numerous studies describe the burden of disease due to pertussis in infants. However, systematic reviews of the available data are lacking. We conducted two independent systematic literature searches to identify, analyze, and critically appraise the most recent evidence describing the burden of pertussis disease in infants <6 months of age worldwide; Tdap vaccine effectiveness in preventing pertussis in young infants (<2-3 months of age) before commencement of the primary DTaP immunization series; and the impact of Tdap maternal immunization programs on the epidemiology of pertussis in infants up until 1 year of age.

2. Methods

2.1. Objectives

The first study objective was to evaluate the global burden of pertussis disease in infants <6 months of age, stratified by country and by age <3 months and 3–6 months when data allowed. Information on the study design, pertussis surveillance system, the number and incidence of reported/estimated pertussis cases, pertussis-related hospitalizations, deaths, and duration of hospitalization was retrieved.

The second objectives were to summarize worldwide data describing the effectiveness of Tdap immunization during pregnancy at preventing pertussis in young infants (less than 2–3 months of age) before commencement of the primary immunization series; and to summarize the impact of Tdap vaccination during pregnancy on the epidemiology of pertussis disease in infants up to 1 year of age.

The selected literature was restricted to prospective or retrospective observational studies or data from notifiable disease surveillance systems. Information on the other aspects of maternal immunization including safety and immunogenicity was not considered in this review and have been described elsewhere [48–54].

2.2. Article identification, selection and data extraction

The literature search for each objective was carried out according to a pre-established protocol and followed The Cochrane Collaboration guidelines for performing systematic reviews [55]. We searched the PubMed, Embase, and Cochrane Library databases for peer-reviewed articles published between 1 January 2005 and 4 July 2018 for objective 1, and 1 January 2011 and 4 July 2018 for objective 2, using queries specific to each objective. Search queries are provided in Supplementary Table S1 and S2. Grey literature including health authority websites was searched in addition to the peer-reviewed data as described in the on-line supplement. Searches were not limited by language or geography.

Inclusion and exclusion criteria for articles relevant to the study objectives are provided in the online supplement. In brief, articles were selected if they reported outcomes relevant to the study objective. Systematic reviews, meta-analyses, narrative reviews, case reports, and modeling studies were not included. However, systematic reviews and meta-analyses were checked for any potentially overlooked articles.

For each search, titles and abstracts were screened for relevance in duplicate by two researchers, and articles of possible interest underwent full-text screening. The first 10% of articles were appraised in duplicate, with any disagreement between the appraisers being adjudicated by a third researcher. Further scrutiny of the articles was performed during the data-extraction phase. For each literature search, evidence tables were compiled by a senior epidemiologist and reviewed by a second epidemiologist.

The burden of disease was assessed using the following outcomes: (1) number and incidence of reported cases and estimated cases by year, (2) number and incidence of reported and estimated hospitalizations by year, (3) duration of hospitalization, and (4) number and incidence of pertussis-related deaths by year. Denominators of incidence, hospitalization, and mortality rates extracted during the burden of disease assessment referred to the general population (age-specific) unless stated otherwise.

For the second objective, vaccine effectiveness (VE) of Tdap administered during pregnancy against pertussis in infants was determined. If VE was not reported in an article, it was derived from the association between Tdap vaccination during pregnancy and the endpoint of interest (pertussis disease, or pertussis hospitalization, or pertussis death) in infants, measured either as relative risks (RRs), odds ratios (ORs), or hazard ratios (HRs), where VE = (1-RR)x100, (1-OR)x100, or (1-HR)x100. The impact of Tdap vaccination during pregnancy on the epidemiology of pertussis in infants was assessed from studies that reported the following outcomes: (1) (change in) number and incidence of pertussis cases and (2) number and incidence of pertussis-related deaths by year, according to the time point of maternal immunization for each country for which data were available.

2.3. Study quality assessment

Existing quality assessment tools were not suited to many of the study designs reported here. A GRADE analysis (Grading of Recommendations Assessment, Development, and Evaluation) [56] was planned for studies of vaccine effectiveness. However, this proved not possible because the outcomes differed per study and could not be synthesized. Furthermore, data as presented in the included articles on effectiveness needed some recalculation in order to fit a GRADE profile, and we preferred to present the data as in the original articles.

The quality of studies reporting outcomes in terms of population-based rates was therefore evaluated using an in-house quality assessment tool provided in Section 3.0 in the online supplement. According to this tool, any article that contained a ‘fatal error’ (an element that could throw doubt on the main study conclusions) was excluded from the search, while all other articles were included irrespective of the overall rating score (strong, moderate, or weak). Studies describing follow-up of a group of patients were not suited for appraisal with existing checklists or the quality assessment tool mentioned above, and thus were not critically appraised.

3. Results

3.1. Burden of pertussis disease in infants

The search for articles on the burden of pertussis disease yielded 3,359 unique records in the electronic databases, of which 67 articles were included in the final analysis (Figure 1). Additional information was found in the gray literature for four countries. Many of the articles were conducted in one or several hospitals and reported absolute case numbers, hospitalizations, or deaths. Key characteristics and results of all included studies are provided in Table S3. There were 22 articles that reported population-based incidence rates of pertussis disease, hospitalization, or deaths, and these results are summarized below by WHO region in Figures 2–4.  Data are expressed as either per 1000 person-months or per 100,000 population.

Figure 1. PRISMA flow diagrams for the two systematic literature searches performed for this review.

Figure 2. Pertussis disease incidence and hospitalization rates in infants <6 months of age in studies reporting results by person-time [57–59,63,69,78,81,83].

Figure 4. Pertussis-related hospitalizations per 100,000 population in infants <6 months of age [71,72,75,76,86].

Direct comparison of results needs to be made cautiously because of heterogeneity between study designs, data sources, pertussis case definitions, periods analyzed and populations and study settings considered. Several studies were conducted over short time periods or a single season and were unable to take into account the cyclic pattern of pertussis disease outbreaks. Estimates of the disease burden are markedly affected by case detection rates, and the studies we reviewed used clinical symptoms, laboratory-confirmation, epidemiological links, and/or International Classification of Disease (ICD) codes in combination or alone, which may differ in their sensitivity and specificity, as methods of case detection.

3.1.1. Region of Africa

Eight studies describe the burden of pertussis disease data in children <6 months of age from 4 countries in Africa: Algeria, Niger, South Africa (five studies), and Zambia. None of the countries had recommendations for maternal Tdap vaccination in place during the study periods.

Six of the eight studies were hospital-based observational studies and 2 were prospective longitudinal cohort studies; one conducted in mothers previously enrolled in an influenza vaccine trial in South Africa, and one in a cohort of mother-infant pairs from a peri-urban slum in Zambia [57,58]. The studies were conducted between 2008 and 2015 and in 5 studies, the observation period was <1 year.

Population-based estimates of the disease burden were reported in three studies in South Africa and Zambia (Figure 2) [57–59]. The incidence of PCR-positive pertussis per 1,000 person-months ranged from 2.4 to 6.5 in infants aged <3 or <3.5 months, and was 5.5 in infants aged <6 months [57,58]. The incidence of pertussis and pertussis hospitalizations was higher in children who were born to HIV-positive mothers versus HIV-unexposed infants [58,59]. The pertussis death rate in infants aged <6 months was reported in one study (conducted in a single clinic) as 10 deaths (95% confidence of interval (CI) 2–30) per 100,000 infants <6 months of age in the general population [59]. Compared to some other world regions (Figure 2), the observed pertussis incidence rates and hospitalizations were higher in African countries.

3.1.2. Region of the Americas

Sixteen studies were included from Argentina, Brazil, Colombia, Mexico, Panama, Peru, and the US. Data were also found from the gray literature for the US. Six studies were retrospective cohort studies using national surveillance data or large healthcare claims databases, and 10 studies described clinical features of pertussis cases in hospital or pediatric intensive care unit (PICU) settings. The studies were conducted between 1997 and 2018.

Three studies and one report from the gray literature reported population-based rates in Argentina, Panama, and the US (Figures 2 and 3) [60–63]. The incidence of laboratory-confirmed pertussis in Argentina was assessed using cases notified to the National Health Surveillance System between 2002 and 2011 [60]. The incidence of reported confirmed pertussis was approximately 550 per 100,000 population in infants <2 months of age, 800 per 100,000 population in infants 2-<4 months, and 400 per 100,000 population in infants 4-<6 months [60].

Figure 3. Annual pertussis disease incidence per 100,000 population in infants <6 months of age [60,64–70,72–77,81,84,86].

A retrospective hospital-based study in Panama reported the incidence of hospitalizations due to laboratory-confirmed pertussis between 2001 and 2008. The incidence of pertussis hospitalizations per 10,000 all-cause hospitalizations was 29.9 in infants aged <1 month, 88.1 among infants aged 2–3 months, and 28.5 in infants aged 4–6 months [61].

Pertussis is a notifiable disease in the US and the Centers for Disease Control and Prevention report the annual incidence of confirmed and probable cases. Maternal Tdap vaccination was introduced in 2012. The incidence of pertussis in infants <6 months decreased from 169.0 per 100,000 population in 2014 to 57.2 per 100,000 in 2018 (provisional data) (Figure 3) [62]. Between 42.9% and 44.3% of infants were hospitalized each year, and the number of deaths in infants until the age of 1 year decreased from 16 in 2012 (of which 15 were infants <3 months of age) to 4 deaths in 2018 (provisional, age not reported).

A large, retrospective study in the US using claims databases used ICD-Revision 9, Clinical Modification (ICD-9-CM) codes to identify pertussis cases in infants aged <1 year. The incidence of pertussis per 1000 person-months was highest in infants aged <3 months (0.21) and 1-<2 months (0.20) [63].

3.1.3. Eastern Mediterranean region

Seven studies conducted in five countries of the Eastern Mediterranean region were included Iran, Morocco, Oman, Pakistan, and Tunisia. Three studies used surveillance notification data, three were conducted in hospital settings and one was a prospective, active community-based surveillance study conducted in four primary healthcare centers in Pakistan. No reports from the gray literature were found for this region. The studies were conducted between 2008 and 2016.

Population-based rates were reported in three studies from Iran, Oman, and Pakistan (Figures 2 and 3) [67–69]. Two studies in Iran used pertussis surveillance data from the Iranian Center for Diseases Control. The incidence of clinically diagnosed and laboratory-confirmed pertussis in <2 month-old infants between 2008 and 2011 ranged between 132 per 100,000 population (2010) and 463 per 100,000 population (2008) [67]. A study using surveillance data from several government sources in Oman (2011–2015) showed that the incidence of suspected or laboratory-confirmed pertussis ranged between 295 and 1390 per 100,000 infants, peaking during an outbreak in 2013 [68]. In Pakistan, an active community-based surveillance study spanning 14 months reported that the incidence of laboratory-confirmed pertussis in infants aged <2 months was 1.14 per 1,000 person-months (95% CI 0.57–2.28), or 3.96 per 1,000 infants [69].

3.1.4. Region of Europe

Twenty-five studies from 16 European countries were included Bulgaria, Denmark, Finland, France, Germany, Greece, Ireland, Israel, Italy, Norway, Palestine, Spain, Sweden, Switzerland, the Netherlands, and the UK. Data from gray literature was found for France, Spain, and the UK.

Seventeen articles reported data from national pertussis surveillance systems, including one case–control study based on surveillance data. The remainder of the articles described cases in hospital-based studies. The length of periods analyzed ranged from 2 to 21 years, but 15 studies included data covering 10 years or longer. The studies were conducted between 1990 and 2017.

Population-based rates were reported in 10 studies from France, Germany, Israel, Norway, Spain, Sweden, the Netherlands, and the UK, including two reports from France and Spain in the gray literature (Figures 2 and 3) [70–81]. The pertussis incidence rate in the <3 month age-group ranged from 155.2 per 100,000 population (Netherlands 2005–2007) to 1114.3 per 100,000 population in Spain during a pertussis epidemic [75,76]. The incidence of pertussis was higher in preterm than in term infants after the age of 1 month (range 2.99–4.31 versus 1.3–2.3 per 1000 person-months) and was highest in 1- and 2-month-old infants in two of 3 studies with results stratified by month of age (Figures 2 and 3). Compared to very young infants, the incidence of pertussis was lower in the 3–6 months age group, ranging from 130 to 193 per 100,000 population [77].

Hospitalization rates for pertussis in children in the <3 month age group ranged from 103 per 100,000 population (Germany 2013–2015) to 997 per 100,000 population during the 2015 epidemic in Spain. Hospitalization rates were lower in 3–5 months old children, ranging from 45.7 (Netherlands 2011–2012) to 110 per 100,000 population (Spain 2012) (Figure 4).

The incidence of pertussis-related death was reported using national surveillance data in the UK. The mean number of deaths per 100,000 maternities was 0.708 (95% CI, 0.690–0.751) in infants <6 months old, and 0.601 (95% CI, 0.597–0.613) in infants aged <66 days [79]. The authors of the UK study noted that death certificates may not always mention pertussis as a cause of death in infants, potentially leading to underestimation of the fatality rate [79]. In all other studies reporting pertussis-associated deaths, the majority occurred in children <3 months of age [72,74,76,79].

3.1.5. Region of South-East Asia

Two studies were identified from this region. One was a retrospective observational study at one hospital in Indonesia between 2008 and 2014 [82]. The other was a prospective, population-based active home surveillance study nested within a randomized controlled trial of influenza vaccination during pregnancy in one region of Nepal [83]. A total of 3.483 infants were followed-up over 40 months between 2011 and 2014, and the estimated pertussis incidence rate was 13.3 per 1,000 infant-years [83].

3.1.6. Region of the Western Pacific

Ten studies were included from six countries of the Western Pacific region: Australia, Japan, New Zealand, Philippines, Singapore, and South Korea. For New Zealand, data from the gray literature were also found. Five articles reported data from country-wide notifiable disease surveillance systems, one was an active hospital-based surveillance study, one was a hospital-based study investigating severe pneumonia cases caused by B. pertussis; two analyzed infant pertussis hospitalization either in a nationwide inpatient database or in hospital records; two described pertussis cases reported to an intensive care unit (ICU) admission registry. The study periods encompassed year 1995 to 2018.

Population-based incidence rates in infants <6 months of age were only available for Australia and New Zealand, and varied over time from 137 per 100,000 population (2006–2012) in a region of Australia to 801 per 100,000 population in 2011–2013 in New Zealand during a pertussis epidemic [84–86]. The pertussis hospitalization rate was estimated in one study (Australia 2006–2012) as 257.9 per 100,000 population [85]. The majority of pertussis-related deaths reported in Australia and New Zealand occurred in infants <2 months of age; 10/10 total pertussis deaths between 2006 and 2010 in Australia, and 17/19 total pertussis deaths between 2002 and 2014 in New Zealand [86,87].

3.2. Effectiveness of maternal immunization and impact on infant pertussis disease

The search of articles on the effectiveness and wider impact of maternal immunization on the epidemiology of pertussis disease in infants <12 months of age yielded 688 unique records, of which the full text of 32 articles was assessed and 11 articles included after applying the inclusion and exclusion criteria (Figure 1). No additional references were included from the gray literature.

Ten out of 11 articles reported on the vaccine effectiveness of maternal Tdap vaccination in preventing pertussis disease in infants [9,10,88–96] and 3 out of 11 provided information on the impact of the maternal immunization program on pertussis disease in infants [9,10,88] (Table 2 and Table 3). All of the included studies were observational; six were cohort studies, four were case–control studies, and one was a time-series analysis that compared pertussis incidence between states with a high (>50%) and low (≤50%) Tdap vaccine coverage in Argentina [9,10,88–96]. Case–control studies used surveillance and notification data to identify cases and three of the four studies matched controls for age and place of birth [89–92]. Three cohort studies used large healthcare claims databases or statewide surveillance data and two used the screening method to compare maternal Tdap vaccination coverage in pertussis cases identified through national surveillance, with average national coverage rates [9,10,93–96]. One cohort study recruited mothers from clinics and followed them for 6 months after delivery [96]. Studies of VE were conducted in Argentina, Australia, Brazil, Colombia, Spain, the UK, and the US between 2010 and 2016, and studies of impact were conducted in Argentina and the UK between 2010 and 2015.

Table 2. Key design features and results of studies describing the effectiveness of maternal pertussis immunization.

						Vaccine effectiveness (95% CI)
Country, reference	Population	Maternal vaccination recommendation (vaccine)	Study design; period	Sample size (no. of mothers vaccinated)	Outcome	Unadjusted	Adjusted
Australia Saul, 2018 [90]	Cases notified through public health units in local health districts Controls were born in a public hospital in the same district	28–32 weeks (Boostrix)	1:1 matched case–control study 16 August 2015–17 August 2016	Cases: n = 117 (52) Controls: n = 117 (72)	Laboratory-confirmed pertussis: culture, PCR or seroconversion in paired sera in the absence of recent vaccination.	<6 months: 51% (16–72) <3 months: 64% (18–84)	<6 months: 39% (−12 – 66) <3 months: 69% (13–89)
					Pertussis requiring hospitalization	6 months: 84% (47–95)	<6 months: 94% (59–99)
Spain Bellido-Blasco, 2017 [92]	Notified cases Controls: 3 for each case with anage difference <15 days	28–36 weeks and 15–89 days before delivery (NR)	Prospective matched case–control study 1 March 2015–29 February 2016	Cases n = 22 (5) Controls n = 66 (41)	PCR confirmed pertussis	<3 months: 92% (63–98)	<3 months: 90.9% (56.6–98.1)
United Kingdom Amirthalingam, 2014 [9]	Infants <3 months born 1 October 2012–3 September 2013 with laboratory confirmed pertussis	28–38 weeks or at least 7 days before delivery (Repevax)	Observational study (screening method; CPRD, PHE enhanced pertussis surveillance program, HES)	Cases n = 82 (15% vs 62% nationally)	Laboratory-confirmed pertussis: culture, PCR or serology	<2 months: 90% (82–95) <3 months: 91% (84–95)	-
United Kingdom Dabrera, 2015 [89]	Cases: infants <2 months with pertussis born during the study period Controls: 2 for each case, born consecutively after the case, from the same practice	Median 31.5 wks (cases), 33 wks (controls) (NR)	Case-(unmatched) control study (country-level, multiple primary care centers); 22 October 2012–11 July 2013	Cases: n = 58 (10) Controls: n = 55 (39)	Laboratory-confirmed pertussis: PCR or culture	<2 months: 91% (77–97)	<2 months: 93% (81–97)
United Kingdom Amirthalingam, 2016 [10]	Infants <3 months born 1 October2012–3 September 2013 with laboratory confirmed pertussis	28–38 weeks or at least 7 days before delivery (Repevax until 1 July 2014 then Boostrix-IPV)	Cohort study (screening method Immform, CPRD, PHE enhanced pertussis surveillance program); 2008–2013	n = 243 infants with pertussis (35% vs Immform: 50–62% Jan 2013-Dec 2015; CPRD: 78% Jan 2013; 70% Aug 2015)	Laboratory-confirmed pertussis using culture or PCR	<2 months: 90% (86–93) <3 months: 91% (88–94)	-
					Pertussis-related deaths	<3 months: 95% (79%–100%)	-
					Laboratory-confirmed pertussis (mothers vaccinated ≥ 28 days before delivery)	91% (88–94)	-
					Laboratory-confirmed pertussis (mothers vaccinated 7–27 days before delivery)	91% (80–96)	-
United States Baxter, 2017 [93]	Full term infants born at KPNC hospitals 2010–2015 to wP-vaccinated mothers	≥20 weeks (NR)	Retrospective cohort study (Kaiser Permanente Northern California claims database)	2 months follow-up n = 17 cases (1) 12 month follow-up n = 103 cases (22)	PCR-positive pertussis	<2 months 92.0% (57–100) <12 months 65% (45–79)	<2 months: 91.4% (19.5–99.1) <12 months: 69.0% (43.6–82.9)
United States Skoff, 2017 [91]	Cases: Full term Infants at least 2 days old 1 January 2011–31 December 2014 Controls: infants born at the same hospital aged <2 months on the case infant’s cough onset date	77% in the 3^rd trimester (Adacel, Boostrix)	Case–control study (multiple hospitals at 6 Emerging Infection Program Network sites)	n = 240 cases (17 in the 3^rd trimester) n = 535 controls (90 in the 3rd trimester)	Pertussis confirmed by culture, PCR or epidemiological link to a laboratory-confirmed case, or clinically compatible illness	-	<2 months: 77.7% (48.3–90.4)
					Pertussishospitalizations	-	<2 months: 90.5% (65.2–97.4)
United States Winter, 2017 [94]	Infants born 1 January 2011–Dec 2015 with known maternal vaccination status	76% during the third trimester (NR)	Retrospective cohort study (California public health surveillance)	n = 420 infants (354 during the 3^rd trimester)	Pertussis hospitalizations using clinical definition or laboratory confirmed. Mothers vaccinated in the 3rd trimester	<2 months: 75.4 (49.8–88.0)	<2 months: 52.1% (−0.16–80.3)
					Mothers vaccinated at any time during pregnancy	<2 months: 72.3 (49.0–85.0)	<2 months: 58.3 (14.9–79.6)
United States Becker-Dreps, 2018 [95]	Singleton deliveries >26 weeks gestation Jun 2010-Dec 2014	pregnancy onset to 2 weeks prior to delivery (NR)	Cohort study (Truven Market-Scan claims databases) Follow-up until 18 months	n = 332 (22)	Either inpatient or outpatient diagnosis + appropriate antibiotic treatment within 7 days	<6 months: 55% (24–73)	<6 months: 46% (6–69)
				n = 151 (6)	Inpatient-only cases (95% CI)	<6 months: 81% (40%–94%)	<6 months: 75% (20%–92%)
				n = 449 (30)	Possible pertussis: also includes outpatients with a cough diagnosis + a claim for a pertussis laboratory test, + appropriate antibiotic treatment within 7 days	<6 months: 49% (21–66)	<6 months: 35% (−2–59)
Columbia Hincapie-Palacio, 2018 [96]	Mothers recruited prior to vaccination at 8 urban hospitals Dec 2015–Apr 2016	Mean 28.59 ± 3.51 weeks (NR)	Cohort study 6 months follow-up	n = 745 vaccinated mothers (2 cases) n = 260 non-vaccinated mothers (0 cases)	Confirmed pertussis: met the case definition and PCR positive, or epidemiology linked to a laboratory-confirmed case, or confirmed by clinical criteria	Unable to be calculated due to low attack rate

ACIP, Advisory Committee on Immunization Practices; CI, confidence interval; CPRD, Clinical Practice Research Datalink; PHE, Public Health England; HES, Hospital Episode Statistics; n, number of participants/cases; MI, maternal immunization; Tdap, reduced tetanus toxoid, diphtheria toxoid, and acellular pertussis vaccine; PCR, polymerase chain reaction; VE, vaccine effectiveness.

Adacel, reduced antigen content diphtheria-tetanus-acellular pertussis vaccine; Boostrix, reduced antigen content diphtheria-tetanus-acellular pertussis vaccine; Repevax, reduced antigen diphtheria-tetanus-acellular pertussis-inactivated poliovirus vaccine.

Table 3. Key design features and results of studies describing the impact of maternal pertussis immunization.

Country, reference	Maternal vaccination coverage	Study design; period	Sample size	Outcome	% Reduction (95% CI)
Argentina Vizzotti, 2016 [88]	High ≤50% Low = <50%	Surveillance study (Argentinean Health Surveillance System) Jan 2010–Dec 2013	5,657 cases	Confirmed pertussis: suspected illness positive by culture, PCR, or serology, or, an epidemiological link to a laboratory-confirmed case Suspected pertussis in infants <6: ARI and at least one of: apnea, cyanosis, stridor, post-tussive vomiting or paroxysmal cough	Reduction between high and low vaccination regions: <2 months: 51% (35%–67%; p = 0.001) 2–6 months: 44% (24%–66%)
				Deaths per 100,000 population	<12 months: 2.1 (year 2010); 9.8 (2011); 3.9 (2012) and 1.3 (2013).
				Case fatality rate	<12 months: 1.4 (year 2010); 2.7 (2011); 1.7 (2012) and 0.9 (2013).
United Kingdom Amirthalingam, 2014 [9]	64%	Observational study (screening method; CPRD, PHE enhanced pertussis surveillance program, HES) 1 January 2008–30 September 2013	90 cases	Percentage change in number of confirmed pertussis cases, 2013 vs 2012 (95% CI)	<1 month: −77% (−90% – −53%) 1 month: −77% (−84% – −67%) 2 months: −80% (−87% – −69%) 3–5 months: −65% (−79% – −41%) 6–1 months: −68% (−89% – −23%)
				Percentage in number of confirmed pertussis cases, 2013 vs 2011 (95% CI)	<1 month: −38% (−75 – 46) 1 month: −35% (−58 – 0) 2 months: −44% (−67 – 7) 3–5 months: 5% (−45 – 100) 6–11 months: 0% (−70 – 234%)
United Kingdom Amirthalingam, 2016 [10]	78% (Jan 2013) 70% (Aug 2015)	Cohort study (screening method; PHE); Oct 2012-Sept 2015	243 cases	Annual incidence (cases per 100,000 population) before maternal immunization (2009–2011) vs after (2013–2015)	<3 months: 62.5 (pre-vaccine) vs 60.4 (vaccine era) 3–5 months: 14.0 (pre-vaccine) vs 12.7 (vaccine era)

CPRD, Clinical Practice Research Datalink; PHE, Public Health England; HES, Hospital Episode Statistics.

In most studies, VE was adjusted for one or more confounding factors, such as breastfeeding status, household size, and gestational age at birth. As these factors differed between the studies, the adjusted outcomes could not be compared. Unadjusted and adjusted estimates of VE are presented in Table 2.

3.2.1. Vaccine effectiveness

3.2.1.1. Infants up to 2 months of age

Estimates of VE (adjusted) in preventing confirmed pertussis disease in infants up to 2 months of age ranged from 77.7% to 93% across studies [9,10,89,91,93]. VE estimates from the three studies conducted in the UK were similar (90%–93%). VE estimates from two US studies were 91.4% and 77.7%, with wide CIs (Figure 5). In one study, VE against pertussis disease was lower when vaccination occurred in the first or second versus the third trimester of pregnancy (VE 64.3% versus 77.7%) [91].

Figure 5. Effectiveness of maternal Tdap vaccination against pertussis disease and hospitalizations in infants up to 6 months of age [9,10,89–95].

3.2.1.2. Infants up to 3 months of age

Estimates of VE in preventing confirmed pertussis in infants up to 3 months of age were similar to estimates for <2 month-olds in 3 out of 4 studies (range 90.9%–91%) including the two UK studies that used the screening method and a case–control study in Spain [9,10,92]. Saul et al. [90] estimated VE against laboratory-confirmed pertussis at 69% (95% CI, 13–89) in a case–control study in Australia.

VE in preventing pertussis-related deaths was estimated in one UK study [10]. There were 11 deaths among 243 infants with pertussis aged <3 months. The only death attributed to pertussis among the group of maternally immunized infants was in an infant whose mother was vaccinated <10 days before delivery (VE 95%, 95% CI, 79–100).

3.2.1.3. Infants up to 6 months of age

Two studies assessed the VE of maternal Tdap in preventing pertussis infection and pertussis-related hospitalization in infants up to 6 months of age [90,95] (Figure 2). Estimates of VE against pertussis disease were lower than those in younger infants (39% not significant, and 46% in the respective studies).

A cohort study in Colombia by Hincapie-Palacio et al., followed infants born to vaccinated and non-vaccinated mothers from birth until 6 months of age [96]. The primary study objectives related to immunogenicity. Only two pertussis cases occurred among 1005 enrolled mothers and VE could not be estimated [96].

3.2.1.4. Hospitalization

Winter et al. [94] assessed whether pertussis-infected infants whose mothers received Tdap vaccine during the third trimester of pregnancy had less severe pertussis compared to pertussis-infected infants whose mother did not receive Tdap vaccine (Figure 5). The crude VE estimate for preventing hospitalization as a measure of severity was 75.4% (95% CI, 49.8–88.0), but was not significant (VE 52.1%, 95% CI, −0.16–80.3) after adjusting for chronological age, gestational age, and pediatric aP vaccines.

Three studies observed that estimates of VE against pertussis disease were lower than VE against hospitalization, suggesting a greater impact of maternal vaccination on severe disease than on milder disease [90,91,95]. Skoff et al. [91] found that VE in preventing pertussis hospitalization was 91.4% (95% CI, 24.8–99.0) when Tdap was administered during the first or second trimesters, and 90.5% (95% CI, 65.2–97.4) when administered during the third trimester. Saul et al. [90] found that the point-estimate for VE was substantially higher in preventing pertussis hospitalization (94%, 95% CI, 59–91 in infants up to 6 months of age) than against laboratory-confirmed pertussis 69% (95% CI, 13–89). Becker Dreps et al. [95] demonstrated significant VE against hospitalization due to pertussis in infants up to 6 months of age (75%) versus 46% against pertussis disease in this age group.

3.2.2. Impact of maternal immunization on pertussis incidence

Three articles reported on the impact of maternal Tdap vaccination on the epidemiology of pertussis disease in infants <12 months of age [9,10,88].

Vizzotti et al. compared the incidence of pertussis disease in Argentina between states with high (>50%) and low (≤50%) maternal Tdap vaccine coverage using data from the Argentinean Health Surveillance System in a time-series model [88]. Low coverage regions were used as a control group, and the time series were compared before and after the implementation of the Tdap program in 2012. Pertussis cases were included if they were confirmed or suspected. There was a relative reduction in pertussis cases of 51% (95% CI 35%–67%) between the high and low vaccine coverage states. In children, 2–6 months of age, the relative reduction in cases was 44% (95% CI 24%–66%). The observed reductions could have been influenced by the cyclic nature of pertussis disease, but the peak in disease amongst <12-month-old in low coverage states in 2012 was not observed in high coverage states.

Country-wide surveillance data from the UK span the period before, during and after the implementation of a nationwide pertussis vaccination campaign in pregnant women [9]. Between 2008 and 2011 the annual number of laboratory-confirmed pertussis cases in infants aged 1 month ranged from 22 to 67. This number increased to 161 in 2012 during a large pertussis epidemic, but reduced to 10 cases in 2013 after implementation and rapid uptake of the maternal vaccination program [9]. The number of hospital admissions also peaked at 209 in 1-month-olds in 2012 and reduced to 68 in 2013.

The greatest reduction in confirmed pertussis cases was observed in infants <3 months (328 cases in 2012 versus 72 cases in 2013, −78%; 95% CI, 72%–83%), although this age group still showed the highest incidence across all age categories. Maternal Tdap coverage in January 2013 was 78%. Further surveillance data showed increasing pertussis disease in all age-groups by 2015, consistent with a cyclical upsurge in disease activity and possibly by changes to pertussis detection methods in some age groups in the UK (PCR testing previously only available for hospitalized infants was extended to all age-groups; oral fluid testing was extended from 8–16 years old to 5–16 years old) [10]. Despite increases in all other age groups, the incidence of pertussis in infants <6 months of age remained lower than incidence rates observed prior to commencement of the maternal vaccination program. Of 16 pertussis-associated deaths in infants between 2013 and 2015, 14 were in infants of mothers who had not received vaccination, and two had mothers who had been vaccinated <10 days prior to delivery [10]. Surveillance data in the UK available since the cutoff date of our review show that the decline in incidence in infants under 3 months of age has continued. The incidence of confirmed pertussis in infants aged <3 months was 30 per 100,000 population in 2018, versus 234 per 100,000 in 2012 [97].

4. Conclusion

Of all age groups, infants bear the greatest pertussis burden in terms of disease incidence, severity, and mortality [4]. The studies we reviewed reported a wide range of disease incidence rates in infants <2-3 months of age that exceeded 1.000 per 100,000 population in some countries during outbreaks, and virtually all pertussis deaths occurred in this age group. We found data from a wide range of low and upper income countries but observed that, in countries where pertussis is not notifiable, hospital-based studies were the main source of information about infant pertussis. These studies often lacked population denominators to estimate the incidence of disease and can only provide a partial view of the disease burden. For countries with existing national surveillance systems, some did not publish data of enough granularity (for example, data for infants all combined as <1 year) to be included in this review. There was a paucity of data for some regions including Africa, the Eastern Mediterranean, and Asia, although the available information point to a disease burden in these countries that is similar or higher than the European region or the Americas.

Maternal immunization aims to prevent pertussis disease in the most vulnerable infants until they are old enough to receive and respond to routine vaccination. We conducted an exhaustive review of the available effectiveness data and found 10 studies, all of which reported significant vaccine effectiveness in preventing pertussis disease and hospitalization in infants <3 months of age. The majority of studies showed that maternal Tdap vaccination was approximately 90% effective in preventing pertussis until the age of 3 months, and reduced thereafter. Maternal Tdap vaccination was effective in preventing between 58.3% and 90.5% of pertussis-associated hospitalizations in 2 months olds, and between 75% and 94% in <6 months old, which suggests that despite waning protection after 3 months of age, maternal vaccination continues to have an impact on severe disease requiring hospitalization until at least 6 months of age. To date, fewer studies have evaluated the impact of maternal Tdap on pertussis epidemiology. Analyses following the introduction of maternal immunization in Argentina and England suggest that maternal vaccination programs are achieving their goal of reducing the burden of disease in young infants before the initiation of primary vaccination (Figure 6).

Figure 6. Plain language summary.

Studies published after the date of our review support these conclusions. Burden of disease studies conducted in Brazil and Canada continues to see the highest rates of pertussis disease in infants, with nearly all pertussis-related deaths, hospitalizations, and ICU admissions in children aged <1 year [98,99]. A large health claims database study in the US observed a 48% decrease in infant pertussis hospitalizations; incidence 8.4 per 100,000 (95% CI 7.2–9.7) in 2009–2012 versus 3.3 per 100,000 (95% CI 2.6–4.3) in 2013–2017 [100], coinciding with the 2012 ACIP recommendation for maternal Tdap vaccination. No information on maternal vaccination rates was available in this study.

5. Expert opinion

In the absence of licensed pertussis vaccines for newborns that can be administered in the first days after birth, maternal Tdap vaccination is proving an effective means by which to reduce pertussis disease, hospitalization, and death in infants who are too young to be fully vaccinated. Remarkably, estimates of VE in preventing pertussis disease in <2-3 months old infants were consistently high across the studies we reviewed, despite differences in case detection and confirmation methods, study design, setting, and maternal vaccination program. The results suggest a real benefit of maternal Tdap vaccination for newborns and young infants. Overall, a growing body of publications has found no association between maternal Tdap vaccination and the evaluated adverse pregnancy and/or fetal outcomes [48–54]. An increased risk of chorioamnionitis in women exposed to Tdap during pregnancy was observed in three studies [101–103]. However, these results should be interpreted with caution, especially since no increased risk of preterm births or other adverse neonatal outcomes were observed in the studies. Additionally, other studies by Morgan et al., [104], Berenson et al., [105], and Griffin et al., [106], found no increased risk of chorioamnionitis in women exposed to Tdap during pregnancy. Continued careful evaluation and communication of safety are needed to monitor the benefit/risk of Tdap maternal immunization and maintain confidence amongst mothers and vaccine providers. More data are needed to further characterize the safety of repeated Tdap immunization in successive pregnancies.

Nevertheless, a number of questions remain: What is the optimal timing of Tdap vaccination during pregnancy to maximize protection of the newborn? Vaccine effectiveness estimates by timing of administration included in this review come from only one study which showed a higher point estimate of VE against pertussis disease when Tdap was administered in the third trimester than earlier in pregnancy, but equally high VE point estimates against hospitalization [91]. However, the 95% CI was wide, and no conclusions on the potential difference of effectiveness by the timing of administration can be drawn on the basis of these results. A more recent study conducted in Argentina observed similar VE against confirmed pertussis disease in infants when Tdap was administered during the second or third trimester [107]. The optimal timing for Tdap vaccination during pregnancy needs to balance the desire to achieve the highest vaccine-induced antibody levels during the period of highest disease susceptibility, the need to vaccinate when safest in pregnancy, a desire to minimize any impact on subsequent responses to routine infant vaccination, and the practical consideration of electing a time when vaccine uptake will be maximal [108].

Does maternal Tdap immunization warrant a change in infant vaccination schedules? In a randomized clinical trial, Barug et al. [109] observed the effect of blunting is likely to be higher when starting primary vaccination at age 2 months compared with starting primary vaccinations at 3 months, and concluded that maternal vaccination supports a delay of the first pertussis vaccination in infants until at least age 3 months. In 2018, the Health Council of the Netherlands has recommended that infants of mothers vaccinated in pregnancy receive a 2-dose primary schedule at 3 and 5 months of age, while infants whose mothers were not vaccinated continue to receive the current 3-dose schedule at 2, 3 and 4 months of age [110]. This strategy was also considered as less burdensome for the baby and less expensive [111]. The Netherlands is the first country to adjust the immunization schedule for infants born to women immunized with Tdap during pregnancy. Parallel national schedules for infants may present programmatic challenges in terms of implementation and potentially risk confusion for parents and providers. Furthermore, careful consideration is needed when using a delayed primary immunization schedule given that non-Tdap antigens (hepatitis B, Haemophilus influenzae type band poliomyelitis) are also delayed, and protection by maternal antibodies might not be high enough to prevent disease caused by these pathogens. Ongoing surveillance of pertussis epidemiology might help understand the impact of maternal immunization and modifications to infant schedule on levels of childhood disease.

The burden of pertussis disease in infants is substantial in all parts of the globe but remains poorly documented in some regions where surveillance infrastructure is lacking. Additional work is needed to understand the disease burden in infants in countries where access to healthcare services may be limited. In the future, we can expect that more countries will introduce recommendations for maternal pertussis vaccination, according to national capacity. The first studies in aP-primed mothers may be possible; the optimal timing of maternal vaccination to maximize VE may be defined; and other countries may implement changes to infant pertussis schedules in response to successful maternal vaccination programs. Vaccination of newborns against pertussis has proven problematic in terms of the long-term immune response, and large studies of immunogenicity, safety, and efficacy are currently lacking [112–114]. Maternal vaccination currently stands as the most effective means of protecting newborns and young infants against pertussis disease before starting primary vaccination.

Article highlights

• Worldwide, the highest incidence rates of pertussis disease are in infants aged <6 months, with the majority of pertussis-related deaths, hospitalizations, and ICU admissions in this age group.

• Infants aged <2-3 months are most at risk. Annual pertussis disease incidence rates in this age group ranged from 132 per 100,000 population to as high as 1390 per 100,000 during outbreaks.

• The point estimates of maternal Tdap effectiveness in preventing pertussis in infants below 3 months of age ranged between 69% and 93%.

• Maternal Tdap vaccination has been shown to prevent between 58.3% and 94% of pertussis-associated hospitalizations up until the age of 6 months.

Maternal Tdap vaccination is proving an effective means by which to reduce pertussis disease, hospitalization, and death in young infants who are too young to be fully vaccinated.

Author contributions

All authors participated in the design or implementation or analysis, and interpretation of the study. CvdE and EB performed the search strategy, article selection, and data extraction. All authors had full access to the data and participated in the development of the manuscript, revised it critically for important intellectual content, approved the final version before submission, and agreed to be accountable for all aspects of the work.

VJ: contributed to the design of the study, interpretation of the data, and writing the manuscript.

Declaration of interest

V. A. Jenkins, M. A. Ceregido, and A. Guignard are employed by the GlaxoSmithKline group of companies and V. A. Jenkins and A. Guignard hold shares in the GSK group of companies. W. Kandeil was an employee of GlaxoSmithKline at the time of the study. C. van den Ende and E. Bunge are employed by the Pallas company grants from GlaxoSmithKline Vaccines during the conduct of the study; as well as grants from GlaxoSmithKline Vaccines and from Sanofi Pasteur outside the submitted work. The Pallas company received consultancy fees from GlaxoSmithKline for the performed source reports. Writing assistance was utilized in the production of this manuscript: J Fraczek-Goffin and A McGillivray both of GlaxoSmithKline, and J van den Bosch (Pallas, Health Research and Consulting, Rotterdam, The Netherlands), contributed to the source documents (Pallas reports); J Wolter (independent medical writer c/o GlaxoSmithKline) and S Rubio (Modis c/o GlaxoSmithKline) provided writing support services; F Danhier (Modis c/o GlaxoSmithKline) provided publication coordinating services; all services were funded by GlaxoSmithKline. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Trademark statement

Boostrix and Boostrix-IPV are registered trademarks of the GlaxoSmithKline group of companies. Adacel and Repevax are registered trademarks of Sanofi Pasteur.

Acknowledgments

The authors would like to thank Joanna Fraczek-Goffin, Amanda McGillivray, and Judith van den Bosch for their contribution to the source documents (Pallas reports).

Writing support was provided by Sara Rubio and Joanne Wolter (Modis on behalf of GlaxoSmithKline); editorial support and publication management was provided by Fabienne Danhier (Modis on behalf of GlaxoSmithKline).

Supplementary material

Supplemental data for this article can be accessed here.

Additional information

Funding

GlaxoSmithKline Biologicals SA funded all costs associated with the development and the publishing of the present manuscript.