- Email: [email protected]
Effectiveness of parental cocooning as a vaccination strategy to prevent pertussis infection in infants: A case-control study
Vaccine xxx (2018) xxx–xxx
Contents lists available at ScienceDirect
Vaccine journal homepage: www.elsevier.com/locate/vaccine
Effectiveness of parental cocooning as a vaccination strategy to prevent pertussis infection in infants: A case-control study Stacey L. Rowe a,⇑, Ee Laine Tay a, Lucinda J. Franklin a, Nicola Stephens a, Robert S. Ware b,c, Marlena C. Kaczmarek c,d, Rosemary A. Lester a, Stephen B. Lambert d a
State Government Department of Health and Human Services, 50 Lonsdale Street, Melbourne, Victoria, Australia Menzies Health Institute Queensland, Griffith University, Brisbane, Queensland, Australia School of Public Health, The University of Queensland, Brisbane, Queensland, Australia d UQ Child Health Research Centre, School of Medicine, The University of Queensland, Brisbane, Queensland, Australia b c
a r t i c l e
i n f o
Article history: Received 15 October 2017 Received in revised form 18 January 2018 Accepted 22 February 2018 Available online xxxx Keywords: Pertussis Cocoon Immunization Vaccine Vaccination Infant
a b s t r a c t Background: During a pertussis epidemic in 2009, the Department of Health, Victoria, Australia, implemented a cocoon program offering parents of new babies a funded-dose of pertussis-containing vaccine. We assessed vaccine effectiveness (VE) of the program in reducing pertussis infection in infants. Methods: Using a matched case-control design, infants aged <12 months that were notified with pertussis between 1 January 2010 and 31 December 2011, and born during the time that the cocoon program was in place, were identified. Controls were matched by area of residence and date of birth. Telephone interviews we conducted to ascertain parents’ vaccination status, and if vaccinated, timing of vaccination receipt relative to the birth of their baby. Odds ratios (ORs) were calculated for the association between vaccination and pertussis infection, with VE calculated as (1 – OR) 100%. Results: The study recruited 215 cases and 240 controls (response rates 67% and 25% of eligible participants, respectively). Vaccination of both parents after delivery of the infant and 28 days prior to illness onset reduced pertussis infection by 77% (Vaccine Effectiveness [VE] = 77% (confidence interval [95% CI], 18–93%). After adjusting for maternal education, presence of a sibling within the household, and the infants’ primary course vaccination status, the adjusted VE was 64% (95% CI, 58–92%). Conclusions: Although not reaching statistical significance, our results demonstrated that cocoon immunisation – where both parents are vaccinated in the post-partum period – may offer some protection again infant pertussis infection. Cocoon immunisation could be considered in circumstances where antenatal vaccination of the mother has not occurred. Crown Copyright Ó 2018 Published by Elsevier Ltd. All rights reserved.
1. Background Bordetella pertussis (pertussis) has re-emerged as a global threat to public health, with recent epidemics reported in several countries [1–4] including Australia [5]. In Victoria, Australia, medical practitioners and laboratories are required to notify the State Government Department of Health and Human Services cases of pertussis in accordance with the Public Health and Wellbeing Act 2008 [6]. In the latter half of 2008, an increase in the number of pertussis cases was observed [7]. In subsequent years, annual case numbers more than doubled from 3698 in 2009 to 8831 in 2011 [8]. Whilst epidemics of pertussis are known to occur in 3–5 yearly cycles [9], the greatest concern with the most recent resurgence ⇑ Corresponding author at: Department of Health and Human Services, GPO Box 4057, Melbourne, Victoria, Australia. E-mail address: [email protected] (S.L. Rowe).
was an increase in cases among infants too young to be vaccinated [8] who are vulnerable to severe disease and death [10]. A range of vaccination strategies to prevent pertussis in infants, children, and adolescents have been applied by governments worldwide. Cocooning is one strategy whereby infants who are too young to be vaccinated are indirectly protected against disease by vaccinating their parents and other close contacts. In 2005, the Global Pertussis Initiative recommended that cocooning be implemented in countries where it was considered economically feasible [11]. Cocooning strategies were adopted by several countries, and in all Australian state and territories as an epidemic response at the time [12]. Between June 2009 and June 2012, the Victorian Government Department of Health introduced a cocoon program, whereby parents of new babies were offered a dose of free pertussis-containing vaccine (dTpa). Under the program, fathers could receive the vaccine immediately before or during pregnancy or after delivery of
https://doi.org/10.1016/j.vaccine.2018.02.094 0264-410X/Crown Copyright Ó 2018 Published by Elsevier Ltd. All rights reserved.
Please cite this article in press as: Rowe SL et al. Effectiveness of parental cocooning as a vaccination strategy to prevent pertussis infection in infants: A case-control study. Vaccine (2018), https://doi.org/10.1016/j.vaccine.2018.02.094
2
S.L. Rowe et al. / Vaccine xxx (2018) xxx–xxx
the infant. Mothers were recommended to receive the vaccine as soon as possible after delivery. Estimated uptake was good with 80% of mothers and 70% of fathers reporting receipt of vaccine in relation to the birth of their baby [13]. The aim of this study was to assess vaccine effectiveness (VE) of cocoon immunisation in reducing the risk of pertussis in infants. 2. Materials & methods We conducted a matched case-control study to examine the effectiveness of the cocoon program in preventing pertussis in infants born between 15 June 2009 and 31 December 2011. Approval to conduct the study was obtained from the Medical Research Ethics Committee of The University of Queensland, Australia. Cases were defined as infants aged younger than 12 months with confirmed pertussis (case definition, Fig. 1 [14]) notified to the then State Government Department of Health between 01 January 2010 and 31 December 2011. Controls were babies born between 15 June 2009 and 7 January 2011 and registered on the Victorian Perinatal Data Collection (VPDC) – a population-based registry containing all births in Victoria. Data sources were linked with the Victorian Death Index – a registry of all deaths occurring in Victoria – to identify and exclude infants who had died. Our per protocol approach was to over-sample the number of controls (n = 6) per case to allow for loss to follow up or declinations to participate. All potential controls identified from the VPDC were matched to cases by statistical sub-division (SSD, socially and economically homogeneous areas of Victoria characterised by identifiable links between the inhabitants) and date of birth (+/ 7 days). Potential controls for each case were then randomly ordered using the binomial distribution, and the first six controls were selected for inclusion in the study. Completed interviews were sought from all cases and at least one control per case. The call protocol involved up to six call attempts to establish contact with each participant over varying times of the day, including weekends. For controls, if contact was not established, or if the person refused to participate or was ineligible, they were replaced by the next control participant from the ordered pool of six. 2.1. Data collection, management, and validation Participant information letters were sent to potential participants to advise them of the objectives of the study, to provide them
with an opportunity to opt out, and to request them to have their children’s immunisation records available during the interview. Computer-assisted telephone interviews were carried out between September 2013 and June 2014. Case participants were interviewed, and if completed, interviews were attempted with their corresponding controls. Data collected during interview included socio-demographic details of the parent(s); household composition (including number of resident siblings); the household’s vaccination status (parents, infants and siblings); and other potential risk factors such as breastfeeding history and childcare attendance in the first 12 months of life. Parity and smoking status during pregnancy were obtained from VPDC. Date of illness onset for cases were obtained from the notification record. Parents were asked about receipt of a pertussis-containing vaccine in relation to the birth of the infant included in the study. Exact date of vaccination was sought. If the exact date was not known, month and year of vaccination, or an estimated timing of vaccination relative to the birth of their infant was requested. For these participants, a derived date of vaccination was calculated by adding the estimated timing of vaccination in days to the infants’ date of birth using the following methodology: Where only month and year of parental vaccination was available and this was the same as the month and year of birth of the infant, we assumed that the parent was vaccinated (1) within seven days of the infant date of birth if the vaccination took place in hospital; or (2) within two weeks of the infant date of birth if the vaccination took place elsewhere. In Australia, infants are recommended to receive a pertussiscontaining vaccine at two, four, and six months of age, with the first dose able to be administered from six weeks of age. During telephone interviews, participants were asked to recall their infant’s (and associated siblings’) vaccination status and the dates at which each of the three primary course vaccinations were received. Responses were validated using the then Australian Childhood Immunisation Register (ACIR) – a national register that recorded vaccinations given to children younger than 7 years of age [15]. If the self-reported vaccination date was discordant with the ACIR, we used the date of vaccination sourced from ACIR in the analysis. Infant vaccination status was included in the analysis as a categorical covariate (0, 1, or 2 doses). To be classified as vaccinated with 1 or 2 doses, we applied a refractory period whereby infants were considered vaccinated only if receipt of the vaccine was 14 days prior to illness onset. For controls, the onset date of their matched case was used.
Fig. 1. Case definition for a confirmed pertussis case.
Please cite this article in press as: Rowe SL et al. Effectiveness of parental cocooning as a vaccination strategy to prevent pertussis infection in infants: A case-control study. Vaccine (2018), https://doi.org/10.1016/j.vaccine.2018.02.094
S.L. Rowe et al. / Vaccine xxx (2018) xxx–xxx
Among parents who reported being vaccinated during interview but were unable to provide an exact date of vaccination, consent was provided for the investigators to ascertain date of vaccination with their immunisation provider. For mothers, the primary exposure of interest was defined as vaccination after delivery and 28 days prior to illness onset in the infant. For fathers, the primary exposure of interest was defined as ever being vaccinated: either immediately before or during pregnancy, or after delivery and 28 days prior to illness onset in the infant. This was reflective of the program in place in Victoria. The reference category against which all exposures of interest were compared was ‘‘Not vaccinated (or vaccinated after illness onset)”. For households containing two parents, the primary exposure of interest was defined as both parents vaccinated after delivery of the infant and 28 days prior to illness onset in the infant, and the reference category was: ‘‘Neither parent vaccinated (or vaccinated after illness onset)”. 2.2. Statistical analyses To calculate the detectable difference in pertussis by vaccination status we assumed 250 case infants would be recruited, with 3 matched controls per case, that the probability of maternal cocoon vaccination among control infants was 80%, and that the correlation coefficient for vaccination between matched cases and controls was 0.2. With 80% power and alpha = 0.05 we would be able to detect a true vaccine effectiveness of 40% or more in vaccinated compared to unvaccinated mothers. Statistics were summarised as frequencies (%). As planned a priori conditional logistic regression was used to identify factors associated with disease in three separate models assessing (1) effectiveness of maternal vaccination alone; (2) paternal vaccination alone; and (3) combined parental vaccination status among two-parent households. Models were conditioned on their individually-matched case-control groups. Unadjusted odds ratios (OR) and 95% confidence intervals (CI) were calculated where the exposure was vaccination status and the outcome was pertussis infection. Adjusted odds ratios (aOR) were calculated using multivariable conditional logistic regression. Variables were explored as potential confounding factors by first examining their independent associations with the exposure (vaccination status) and outcome (pertussis infection) during univariate analysis. Factors that were significantly associated with both the exposure and outcome at p < 0.05 were further explored as potential confounding factors in the multivariable model. Crude and adjusted VE was calculated as VE = (1 - OR) 100%, where OR is the odds ratio for vaccination between cases and controls. Three separate sensitivity analyses were conducted. First, analyses were re-run using frequency matching. This approach was not planned a priori, but was conducted after it was identified that a number of the interviewed case participants were unable to be matched with one or more of their corresponding controls (due to a low response among control participants). Participants were grouped in two-month blocks according to the infants’ date of birth and the conditional logistic regression models were re-run. Second, analyses were re-run after excluding infants aged 4 months. This was conducted to explore the effectiveness of parental vaccination in reducing pertussis infection among infants who were either too young to receive any of their recommended pertussis-containing vaccine doses, or had only received their first dose. A third sensitivity analysis was conducted to account for possible misclassification bias introduced due to the application of a 14-day refractory period around the infants’ vaccination status. In this latter analysis, the multivariable model included all covariates included in the primary analysis except for the infants’ vaccination status. All data analyses were completed using Stata Version 13.1 (StataCorp, College Station, TX).
3
3. Results 3.1. Study participants A total of 393 cases and 2254 controls met the criteria for inclusion in the study and were invited to participate. After excluding participants who had opted out or were subsequently found to be ineligible, 382 (97%) cases and 1981 controls (88%) were deemed eligible for interview. Interviews were completed with 215 cases (response rate 67% of eligible cases) and 240 controls (response rate, 25% of eligible controls). At least one matched control (n = 223) was available for 146 interviewed cases and were included in the analyses. A total of 369, 351 and 430 participant mothers, fathers and siblings, respectively, were also included in the analysis (Fig. 2). Among the 146 cases included in the analysis, the median age at illness onset was 2.8 months (Inter Quartile Range [IQR] 1.4–5.6 months). Seventy-three percent sought hospital-based care: 36 cases presented to an emergency department and an additional 71 cases were admitted. Among cases that were admitted to hospital, the median length of stay was six days (IQR: 3–11), and 24% were admitted to an Intensive Care Unit (ICU). Cases requiring ICU care were younger (median age: 1 month [IQR: 0.6–1.5 mont hs]) than those who did not (median age: 2.2 months [IQR: 1.3 – 3.2 months]). None of the cases died. Differences in participant characteristics are shown in Table 1. Maternal education, maternal work status and household income were all were positively associated with parental vaccination status and negatively associated with infant pertussis diagnosis. Maternal education was selected as a potential confounding factor for inclusion in the multivariable models. Having a resident sibling was negatively associated with vaccination status and positively associated with pertussis diagnosis. The infants’ own primary vaccination status was also positively associated with parental vaccination status and negatively associated with pertussis diagnosis. These factors were all explored and ultimately included in the adjusted model. 3.2. Vaccination validation Of the 369 study infants who were included in the primary analysis, vaccination was validated for 344 (93%) infants. There was 96% to 98% concordance found between self-reported vaccination status (i.e. whether or not the infant was vaccinated at 2, 4 or 6 months of age) and ACIR. Of those, self-reported vaccination date differed from ACIR in between 8% and 11% of infants. For parents included in the primary analysis, 165 (50%) had provided an estimated timing of their vaccination during interview and also provided consent for the investigators to ascertain the exact date of vaccination from their immunisation provider. Of these, exact date of vaccination was ascertained from immunisation providers in 40 of 62 (65%) case parents and 73 of 103 (71%) control parents (p = 0.395). 3.3. Vaccination status of parents Ninety-one (62%) case mothers and 149 (67%) control mothers reported being vaccinated prior illness onset in the infant (Table 2). Infants of mothers vaccinated after delivery and 28 days prior to illness onset had a lower odds of pertussis infection compared with infants of mothers who were not vaccinated (unadjusted OR = 0.40 [95% CI, 0.20–0.82], VE = 60% [95% CI, 18%–80%]). After adjustment for confounding variables, the protective effect of vaccinating the mother alone was not retained (aOR = 1.17 [95% CI, 0.41–3.36]). Maternal pre-pregnancy vaccination was associated with a lower
Please cite this article in press as: Rowe SL et al. Effectiveness of parental cocooning as a vaccination strategy to prevent pertussis infection in infants: A case-control study. Vaccine (2018), https://doi.org/10.1016/j.vaccine.2018.02.094
4
S.L. Rowe et al. / Vaccine xxx (2018) xxx–xxx
Cases
Controls
393 (100%)
2,254 (100%)
Eligible to participate* Opt out: n=8 (2%) Ineligible: n=3 (0.8%)
Opt out: n=183 (8%) Case opt out: n=29 (1%) Ineligible: n=61 (3%)
Eligible for interview Phone no. not functioning / found: n=61 (16%)
382 (97%)
Matched case not interviewed: n=682 (34%) Phone no. not functioning / found: n=345 (10%)
1,981 (88%)
Call attempts made Call attempt / interview unsuccessful: n=106 (26%)
321 (82%)
954 (42%)
Call attempt / interview unsuccessful: n=1027 (46%)
Completed interview Matched control not available: n=69 (18%)
215 (67%^)
240 (25%^)
Matched case not available: n=17 (<1%)
Infants included in analysis Mothers: 146 Fathers: 133 Siblings: 220
146 (45%^) 223 (23%^)
Mothers: 223 Fathers: 218 Siblings: 212
* Ineligible case participants included two participants with incorrect date of birth and one participant with an incorrect address (non-Victorian resident); ineligible control participants included participants whose corresponding matched case was not eligible, or who were stillborn. ^ Denominator = Call attempts made (i.e. examined for eligibly) Fig. 2. Study attrition.
Table 1 Study participant characteristics, cases and controls. Total participants
Male infant Mother born in Australia Father born in Australia Mother of Aboriginal or Torres Strait Islander ethnicity Father of Aboriginal or Torres Strait Islander ethnicity Mother, tertiary education Father, tertiary education Mother in paid employment (including part time employment) Father in paid employment (including part time employment) Household income of at least AUD$60,001 Language other than English used in household Attended childcare in first 12 months of life Breastfed (exclusively) for at least 2 weeks after birth Breastfed (exclusively) for 4–6 months after birth Ever smoked during pregnancy Parity > 1 Resident sibling at home Sibling < 5 years Sibling 5–17 years Resident father at homea a
Cases
Controls
p-value
n 146
%
n 223
%
77 120 110 2 3 43 28 73 120 90 28 22 127 74 21 112 114 70 69 133
52.7 82.2 75.3 1.4 2.1 29.5 19.2 50.0 82.2 61.6 19.2 15.1 87.0 58.7 14.4 76.7 78.0 47.9 47.2 91.1
117 196 193 0 4 100 76 154 208 167 27 59 196 125 27 134 141 118 46 218
52.4 87.9 86.6 0.0 1.8 44.8 34.1 69.1 93.3 74.9 12.1 26.5 87.9 63.8 12.1 60.1 63.2 52.9 20.1 97.8
0.959 0.187 0.008 0.113 0.030 0.009 0.001 0.001 0.003 0.010 0.131 0.027 0.797 0.363 0.004 0.001 0.003 <0.001 <0.001 0.004
Denotes a two-parent household where the second parent was male. There were no same sex households identified in the study.
odds of infant pertussis: unadjusted OR = 0.20 (95% CI, 0.04–0.88) and VE of 80% (95% CI, 12%–96%). After adjustment, the point estimate was attenuated (aOR = 0.47 [95% CI, 0.04–6.02] and aVE = 53% [95% CI, 502%–96%]) (Table 2). Among the mothers reporting pre-pregnancy vaccination, 75% of the case mothers and 85% of the control mothers reported receiving the vaccine within the year prior to becoming pregnant (data not shown). Sixty-one (41%) case fathers and 121 (60%) control fathers reported being vaccinated prior to onset of illness in the infant (Table 2). Fathers vaccination before, during, or after delivery and 28 days prior to onset was associated with a lower odds of infant
pertussis (OR = 0.36 [95% CI, 0.19–0.69], VE = 64% [95% CI, 31–81%]. The adjusted VE was 37% [95% CI, 0.61–75%]) (Table 2). Vaccination of both parents after delivery and 28 days prior to illness onset in the infant was associated with a lower odds of infant pertussis (OR = 0.23 [95% CI, 0.07–0.82], VE = 77% [95% CI, 18%–93%]). This association was reduced after adjustment for possible confounders (aOR = 0.36 [95% CI, 0.08–1.58] and VE = 64% [95%CI, 58–92%]). The sensitivity analysis using frequency matching included data from 215 cases and 240 controls. Results were comparable to those found in the primary analysis. However, in contrast to the primary
Please cite this article in press as: Rowe SL et al. Effectiveness of parental cocooning as a vaccination strategy to prevent pertussis infection in infants: A case-control study. Vaccine (2018), https://doi.org/10.1016/j.vaccine.2018.02.094
Model
Cases n (%)
Controls n (%)
Multivariate modelb
Univariate model OR (95% CI)
VE (95% CI)
aOR (95% CI)
aVE (95% CI)
62 (27.8) 27 (12.1) 7 (3.1) 3 (1.4) 91 (40.8) 10 (4.5) 7 (3.1) 4 (1.8) 12 (5.4)
1.00 0.20 (0.04, 0.88) NC NC 0.40 (0.2, 0.82) 2.14 (0.62, 7.41) 6.30 (0.74, 53.26) NC NC
80 (12, NC NC 60 (18, 114 ( 530 ( NC NC
1.00 0.47 (0.04, NC NC 1.17 (0.41, 0.20 (0.01, 6.50 (0.28, NC NC
53 ( 502, 96) NC NC 17 ( 236, 59) 80 ( 357, 99) 550 ( 14992, 72) NC NC
67 (50.4) 32 (24.1) 17 (12.8) 12 (9.0) 5 (3.8)
82 (37.6) 94 (43.1) 6 (2.8) 21 (9.6) 15 (6.9)
1.00 0.36 9.66 0.57 0.66
64 (31, 81) 866 ( 7707, 43 ( 54, 79) 34 ( 137, 82)
(3) Combined parental vaccination status Both parents not vaccinated (or vaccinated after illness onset) (REF) Both parents vaccinated after delivery & 28 days before illness onset
36 (27.1) 20 (15.0)
52 (23.9) 50 (22.9)
1.00 0.23 (0.07, 0.82)
Other risk factors Education of mother Resident sibling at home Infant, not vaccinated (REF) Infant vaccinated, 1 dose Infant vaccinated, 2 doses
43 (29.5) 113 (77.4) 81 (55.5) 32 (21.9) 33 (22.6)
100 (44.8) 140 (62.8) 91 (40.8) 60 (26.9) 72 (32.3)
0.48 2.24 1.00 0.23 0.12
(1) Maternal vaccination status Not vaccinated (or vaccinated after illness onset) (REF) Vaccinated before pregnancy Vaccinated during pregnancy Vaccinated before birth, timing unknown Vaccinated after delivery & 28 days before illness onset Vaccinated after delivery & <28 days before illness onset Vaccinated after delivery, timing unknown Vaccinated, timing unknown Vaccination status unknown
54 (37.0) 8 (5.5) 3 (2.1) 1 (0.7) 49 (33.6) 20 (13.7) 10 (6.9) 0 (0.0) 1 (0.7)
(2) Paternal vaccination status Not vaccinated (or vaccinated after illness onset) (REF) Vaccinated (any time & 28 days before illness onset)a Vaccinated after delivery & <28 days before illness onset Vaccinated, timing unknown Vaccination status unknown
(0.19, (1.19, (0.21, (0.18,
0.69) 78.07) 1.54) 2.37)
96)
80) 641, 38) 5226, 26)
77 (18, 93)
19)
6.02)
3.36) 4.57) 150.92)
1.00 0.63 (0.25, 1.61) NC 1.11 (0.26, 4.85) 0.52 (0.07, 3.68)
37 ( 61, 75) NC 11 ( 385, 74) 48 ( 268, 93)
1.00 0.36 (0.08, 1.58)
64 ( 58, 92)
S.L. Rowe et al. / Vaccine xxx (2018) xxx–xxx
(0.30, 0.77) (1.32, 3.80) (0.09, 0.60) (0.03, 0.41)
Abbreviations: CI – confidence interval; NC – not calculable due to zero cells; OR – odds ratio; aOR – adjusted odds ratio; REF – reference; VE – vaccine effectiveness. a Vaccinated before or during pregnancy, or after delivery and <28 days prior to illness onset. b Model (1) adjusted for maternal education, paternal and infant vaccination status, and resident sibling. Model (2) adjusted for paternal education, maternal and infant vaccination status, and resident sibling. Model (3) adjusted for maternal education, infant vaccination status, and resident sibling.
5
Please cite this article in press as: Rowe SL et al. Effectiveness of parental cocooning as a vaccination strategy to prevent pertussis infection in infants: A case-control study. Vaccine (2018), https://doi.org/10.1016/j.vaccine.2018.02.094
Table 2 Vaccination status and other risk factors associated with infant pertussis, odds ratios and associated vaccine effectiveness estimates.
6
S.L. Rowe et al. / Vaccine xxx (2018) xxx–xxx
analysis, infants of fathers who were vaccinated before, during pregnancy, or after delivery were significantly less likely to have pertussis even after adjustment for confounding variables (aOR = 0.42, 95% CI, 0.20–0.78) (Supplementary Table 1). The analysis excluding infants aged 4 months included 70 cases and 102 controls. Results were similar to those obtained in the primary analysis but were imprecise due to smaller numbers (Supplementary Table 2). The modified multivariable model built to account for possible misclassification bias relating to the 14-day refractory period applied to infants’ vaccination status demonstrated that vaccination of both parents after delivery and 28 days prior to illness onset was associated with a significantly lower odds of infant pertussis (aOR = 0.24 [95% CI, 0.06–0.95], VE = 76% [95%CI, 5–94%]) (Supplementary Table 3).
4. Discussion This study demonstrates that cocoon immunisation – defined as both parents being vaccinated after delivery and 28 days prior to illness onset – reduced pertussis infection in infants by 64% (95% CI, 58–92%). Whilst the point estimate in the adjusted model was indicative of a moderately protective effect, the confidence intervals were wide and included the null value, and should therefore be interpreted with caution. This study was unique in the use of a case-control study design to examine the effectiveness of cocoon immunisation where the timeframe to define the cocoon was distinctly postpartum and included paternal vaccination status. During the most recent global pertussis epidemic, at least eight countries adopted a cocoon strategy [16]. However data from field studies relating to its effectiveness were lacking. Many cocoon programs were implemented on the basis of modelling studies [17]. Modelling showed cocooning may be effective in reducing pertussis in infants [18], preventing up to 55% of infant cases [19]. In 2009, when Victoria’s program was implemented, cocooning was considered an appropriate epidemic response designed to protect infants most vulnerable to pertussis infection. Since that time, a range of studies exploring the effectiveness of cocooning to prevent pertussis in infants have been published with conflicting results. A case-control study showed that when both parents are immunised before or after birth and 4 week before disease onset, risk of pertussis among infants aged <4 months was reduced by 51% [20]. In contrast, other studies have suggested that cocooning may be ineffectual, inefficient, and resource intensive [21,22]. Using a hospital-based cross-sectional study, Castagnini et al. (2012) found that rates of pertussis infection did not differ after the introduction of postpartum pertussis vaccination for mothers [23]. Another hospital-based study found no difference in the severity of pertussis illness in infants aged 6 months after implementation of postpartum cocooning [24]. An Australian study using a populationbased linkage study demonstrated no difference in the incidence of pertussis among infants born to parents who were vaccinated postpartum compared to unvaccinated parents [25]. Only a small number of mothers in our study reported being vaccinated before pregnancy. We found that maternal prepregnancy vaccination was associated with an unadjusted VE of 80% (95% CI, 12–96%) however, after adjustment the VE reduced and statistical significance was not retained. Of these mothers, a majority reported receiving the vaccine within a year prior to becoming pregnant. It is plausible that the protective effect seen in this pre-pregnancy cohort may relate to transplacental transfer of pertussis-specific antibodies in infants; however, the literature provides conflicting evidence on this subject [20,26]. Maternal vaccination with a pertussis-containing vaccine is increasingly being adopted worldwide in light of the compelling
evidence demonstrating the effectiveness of prenatal vaccination [27–29]. Amirthalingam et al. (2014) demonstrated a VE of 91% (95% CI, 84%–95%) using a screening method, and Dabrera et al. (2015) reported a VE of 93% (95% CI, 81–97%) using a casecontrol design [27,28]. Winter et al. (2017) [29] in a linkage study demonstrated that prenatal vaccination was 85% more effective than postpartum vaccination in prevention pertussis in infants <8 weeks of age. At the time that our cocoon program was in place, vaccination during pregnancy was not specifically recommended in Victoria nor Australia-wide [30]. Not surprisingly, only three (2%) case mothers and seven (3%) control mothers indicated that they had received the vaccine during pregnancy. Vaccination during pregnancy was recommended in the United States in 2012 [31]. Australian guidelines now recommend vaccination of women in the third trimester of pregnancy [32]. Additionally, in Victoria, the current program includes vaccination of partners of pregnant women (who are at least 28 weeks pregnant) [33]. Our study found that the presence of a sibling within the household – particularly among those aged 5–17 years – was positively associated with pertussis infection among infants. This finding is consistent with recent evidence indicating that siblings are important sources of pertussis infection among infants [29,34,35]. Additional research is needed to further elucidate the role of siblings in the transmission of pertussis among infants and how this may change relative to local pertussis epidemiology. Updated guidelines published in the United States recommends that – in addition to pertussis vaccination of pregnant women – other people who expect to be in contact with an infant should be ‘‘up to date with pertussis vaccination” [36]. Depending on the age of siblings at the time of an infant’s birth, siblings could be considered ‘‘up to date” having received a pertussis-containing vaccine up to 10 years prior to an infant’s birth. The role of targeted sibling vaccination programs delivered alongside programs supporting vaccination during pregnancy warrants further exploration. 4.1. Strengths and limitations The strength of our study was in its robust study design. We used two population-based, reliable registers [37,38] ensuring good ascertainment of laboratory-confirmed pertussis cases and eligible controls. Our earlier work identified that uptake in the parental pertussis vaccination program varied by time and area of residence [39,13], and our matching methodology accounted for these factors. Another strength related to our ability to explore the effectiveness the parental vaccination individually or combined among two-parent households. This enabled us to evaluate the effect of a complete ‘‘cocoon”, which was central to the program’s design and is lacking in similar studies examining cocoon immunisation. One of the major limiting factors in this study was our low participation rate which limited generalizability of our study. Low participation may have been influenced by the considerable delay between program implementation and evaluation. Whilst an estimation of vaccine uptake was assessed at 6-months [39] and 18months [13] post program implementation, this vaccine effectiveness study was not conducted until after the program ceased (June 2012). The delay between potentially receiving the vaccine under the program (as early as 15 June 2009) and being interviewed (as late as June 2014) may have contributed to the poor response rate among both cases and controls. This delay may have introduced issues relating to the general recall of vaccination receipt. However, our matching methodology would have minimised any systematic differences in recall between case and control participants. Some residual systematic recall or misclassification bias between case and control participants may exist in our study due to the recognised limitation of case control studies generally: Case participants may be more likely to recall pertinent informa-
Please cite this article in press as: Rowe SL et al. Effectiveness of parental cocooning as a vaccination strategy to prevent pertussis infection in infants: A case-control study. Vaccine (2018), https://doi.org/10.1016/j.vaccine.2018.02.094
S.L. Rowe et al. / Vaccine xxx (2018) xxx–xxx
tion relating to receipt or timing of vaccination by virtue of being affected by the condition of interest. Conversely, control participants may have been less likely to recall details around vaccine receipt and timing. This may have led to more control participants being classified as ‘‘Vaccinated, timing unknown”, and thus being excluded from our main outcome measure. Another limiting factor in our study was that the low response rate in controls meant that 69 case participants who had completed an interview were excluded from the matched analysis. A sensitivity analysis incorporating these cases using frequency matching found comparable results to those obtained in our primary analysis. Low participation is a common limitation in many case-control studies, however systematic biases are not an inevitable consequence of such a problem [40,41]. A range of methodological approaches were adopt to minimise and assess the impact of low participation. For example, and in addition to our careful study design involving appropriately matched controls, we adopted an exhaustive call protocol to contact participants. We also assessed the differences between participating and non-participating cases and controls. Whilst we were unable to assess whether there were any differences between the two groups with regard to parental vaccination status, we found no systematic differences between participating and nonparticipating cases or controls with respect to the following factors: baby’s year of birth, area of residence, parity, birth plurality and smoking status during pregnancy. 5. Conclusions Results demonstrated that cocoon immunisation reduced pertussis in infants by 64%. Whilst not reaching statistical significance, this estimate suggests that cocoon immunisation – when defined as the vaccination of both parents after delivery and 28 days prior to illness onset – may offer some protection against infant pertussis infection. The compelling research evidence derived from other studies [27,28] indicates that vaccination during pregnancy is at present the most effective approach for preventing cases in early infancy. However, barriers relating to acceptance of vaccination during pregnancy and challenges with program implementation at a population level mean that uptake remains suboptimal in many countries [42,43]. Where vaccination has not been given during pregnancy, protecting the infant through cocoon vaccination of both parents may afford clinically meaningful protection and should be considered in this context. Funding This work was supported by the State Government Department of Health, Victoria, Australia. Conflicts of interest None declared. Acknowledgements The authors wish to acknowledge Drs Helen Quinn, Tom Snelling and Peter McIntyre of the National Centre for Immunisation Research and Surveillance. The Hunter Valley Research Institute for the conduct of computer assisted telephone interviews, particularly Shanthi Ramanathan, Russ Redford, and David Shellard. We also acknowledge staff at the Clinical Councils Unit, particularly Katharine Gibson and Dr Mary-Ann Davey who oversaw the release of Victorian Perinatal Data Collection data, as well as members of the Consultative Council on Obstetric Morbidity and Mortality for granting access to VPDC data. We acknowledge staff of
7
the Data Linkages Unit, particularly Ying Chen, Mee Lee Easton and Simon Waters who were involved in linkage of data against the Victorian Death Index and selection of controls. We also gratefully acknowledge staff in the Health Protection Branch, Victorian Government Department of Health, in particular we thank Michael Batchelor, Helen Pitcher, Rosemary Morey, and Kana Volin, Joshua Stapleton, Alannah Moon, Michael Brown, and Michael Panagiotis. Finally, we sincerely thank the external members of the Pertussis Expert Advisory Group: Professor Terry Nolan, Dr Nigel Crawford, and Dr Jim Buttery for their advice relating to the conduct, analysis, and reporting of the study. A/Prof Lambert was supported by an Early Career Fellowship from the Australian Government National Health and Medical Research Council and a people support grant from the Queensland Children’s Hospital Foundation during the conduct of this work. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at https://doi.org/10.1016/j.vaccine.2018.02. 094. References [1] Fisman DN, Tang P, Hauck T, Richardson S, Drews SJ, Low DE, et al. Pertussis resurgence in Toronto, Canada: a population-based study including testincidence feedback modeling. BMC Public Health 2011;11:694. [2] Takeuchi M, Yasunaga H, Horiguchi H, Matsuda S. The incidence of pertussis hospitalizations among Japanese infants: excess hospitalizations and complications? Epidemiol Infect 2012;140:1497–502. [3] Celentano LP, Massari M, Paramatti D, Salmaso S, Tozzi AE, Group E-N. Resurgence of pertussis in Europe. Pediatr Infect Dis J 2005;24:761-5 [4] Centers for Disease Control and Prevention. Pertussis epidemic–Washington, 2012. MMWR Morb Mortal Wkly Rep 2012;61:517-22 [5] Australian Government Department of Health and Ageing. National Notifiable Diseases Surveillance System, Number of notifications of Pertussis, Australia, in the period of 1991 to 2012 and year-to-date notifications for 2013. 2013. [6] Victorian Government Health Information. Public Health and Wellbeing Act 2008 and Public Health and Wellbeing Regulations 2009. 2010. [7] Donnan EJ, Fielding JE, Simpson K, Moloney M, Vally H. A continuing pertussis epidemic in Victoria, 2009. Victorian Infect Dis Bullet 2009;13:46–51. [8] Franklin LJ, Cowie B. The Victorian 2009–2011 pertussis epidemic in review. Victorian Infect Dis Bullet 2012;15:7–10. [9] Heymann DL. Control of communicable diseases manual. 18 ed. Washington USA: American Public Health Association; 2004. [10] Wood N, Quinn HE, McIntyre P, Elliott E. Pertussis in infants: preventing deaths and hospitalisations in the very young. J Paediatr Child Health 2008;44:161–5. [11] Forsyth KD, Wirsing von Konig CH, Tan T, Caro J, Plotkin S. Prevention of pertussis: recommendations derived from the second Global Pertussis Initiative roundtable meeting. Vaccine 2007;25:2634–42. [12] National Centre for Immunisation Research & Surveillance. Significant events in diptheria, tetanus and pertussis vaccination practice in Australia. 2016. [13] Rowe SL, Cunningham HM, Franklin LJ, Lester RA. Uptake of a governmentfunded pertussis-containing booster vaccination program for parents of new babies in Victoria. Australia Vaccine 2015;33:1791–6. [14] Australian National Notifiable Disease Surveillance System. Surveillance Case Definitions for the Australian National Notifiable Diseases Surveillance System, 1 January 2004 to 1 January 2015. 2015. [15] Hull BP, Deeks SL, McIntyre PB. The australian childhood immunisation register-a model for universal immunisation registers? Vaccine 2009;27:5054–60. [16] Chiappini E, Stival A, Galli L, de Martino M. Pertussis re-emergence in the postvaccination era. BMC Infect Dis 2013;13:151. [17] Centers for Disease Control and Prevention. Updated recommendations for use of tetanus toxoid, reduced diphtheria toxoid and acellular pertussis vaccine (Tdap) in pregnant women and persons who have or anticipate having close contact with an infant aged <12 months —Advisory Committee on Immunization Practices (ACIP), 2011. MMWR Morb Mortal Wkly Rep 2011;60:1424–6. [18] Van Rie A, Hethcote HW. Adolescent and adult pertussis vaccination: computer simulations of five new strategies. Vaccine 2004;22:3154–65. [19] de Greeff SC, Mooi FR, Westerhof A, Verbakel JM, Peeters MF, Heuvelman CJ, et al. Pertussis disease burden in the household: how to protect young infants. Clin Infect Dis 2010;50:1339–45. [20] Quinn HE, Snelling TL, Habig A, Chiu C, Spokes PJ, McIntyre PB. Parental Tdap boosters and infant pertussis: a case-control study. Pediatrics 2014;134:713–20.
Please cite this article in press as: Rowe SL et al. Effectiveness of parental cocooning as a vaccination strategy to prevent pertussis infection in infants: A case-control study. Vaccine (2018), https://doi.org/10.1016/j.vaccine.2018.02.094
8
S.L. Rowe et al. / Vaccine xxx (2018) xxx–xxx
[21] Blain AE, Lewis M, Banerjee E, Kudish K, Liko J, McGuire S, et al. An assessment of the cocooning strategy for preventing infant pertussis-united states, 2011. Clin Infect Dis 2016;63:S221–6. [22] Skowronski DM, Janjua NZ, Tsafack EP, Ouakki M, Hoang L, De Serres G. The number needed to vaccinate to prevent infant pertussis hospitalization and death through parent cocoon immunization. Clin Infect Dis 2012;54:318–27. [23] Castagnini LA, Healy CM, Rench MA, Wootton SH, Munoz FM, Baker CJ. Impact of maternal postpartum tetanus and diphtheria toxoids and acellular pertussis immunization on infant pertussis infection. Clin Infect Dis 2012;54:78–84. [24] Healy CM, Rench MA, Wootton SH, Castagnini LA. Evaluation of the impact of a pertussis cocooning program on infant pertussis infection. Pediatr Infect Dis J 2015;34:22–6. [25] Carcione D, Regan AK, Tracey L, Mak DB, Gibbs R, Dowse GK, et al. The impact of parental postpartum pertussis vaccination on infection in infants: A population-based study of cocooning in Western Australia. Vaccine 2015;33:5654–61. [26] Healy CM, Rench MA, Baker CJ. Importance of timing of maternal combined tetanus, diphtheria, and acellular pertussis (Tdap) immunization and protection of young infants. Clin Infect Dis 2013;56:539–44. [27] Amirthalingam G, Andrews N, Campbell H, Ribeiro S, Kara E, Donegan K, et al. Effectiveness of maternal pertussis vaccination in England: an observational study. Lancet 2014;384:1521–8. [28] Dabrera G, Amirthalingam G, Andrews N, Campbell H, Ribeiro S, Kara E, et al. A case-control study to estimate the effectiveness of maternal pertussis vaccination in protecting newborn infants in England and wales, 2012–2013. Clin Infect Dis 2015;60:333–7. [29] Winter K, Nickell S, Powell M, Harriman K. Effectiveness of prenatal versus postpartum tetanus, diphtheria, and acellular pertussis vaccination in preventing infant pertussis. Clin Infect Dis 2017;64:3–8. [30] Australian Government Department of Health and Ageing. The Australian Immunisation Handbook. 9 ed. Canberra, Australia: Australian Government; 2008. [31] Centers for Disease Control and Prevention. Updated recommendations for use of tetanus toxoid, reduced diphtheria toxoid, and acellular pertussis vaccine
[32]
[33] [34] [35]
[36] [37]
[38]
[39]
[40]
[41]
[42] [43]
(Tdap) in pregnant women–Advisory Committee on Immunization Practices (ACIP), 2012. MMWR Morb Mortal Wkly Rep 2013;62:131–5. Australian Government Department of Health and Ageing. The Australian Immunisation Handbook. 10 ed. Canberra, Australia: Australian Government; 2014. Victoria State Government. Immunisation and pregnancy. 2016. Skoff TH, Kenyon C, Cocoros N, Liko J, Miller L, Kudish K, et al. Sources of Infant Pertussis Infection in the United States. Pediatrics 2015;136:635–41. Levri KM, Reynolds L, Liko J, Dott M, Robinson BF, Cieslak PR. Risk factors for pertussis among hispanic infants: metropolitan portland, oregon, 2010–2012. Pediatr Infect Dis J 2016;35:488–93. Centers for Disease Control and Prevention. Vaccinating Pregnant Patients. Georgia: Atlanta; 2017. Gibney KB, Cheng AC, Hall R, Leder K. Australia’s national notifiable diseases surveillance system 1991–2011: expanding, adapting and improving. Epidemiol Infect 2017;145:1006–17. Davey MA, Sloan ML, Palma S, Riley M, King J. Methodological processes in validating and analysing the quality of population-based data: a case study using the Victorian perinatal data collection. The HIM J 2013;42:12–9. Donnan EJ, Fielding JE, Rowe SL, Franklin LJ, Vally H. A cross sectional survey of attitudes, awareness and uptake of the parental pertussis booster vaccine as part of a cocooning strategy, Victoria. Australia BMC Public Health 2013;13:676. Pandeya N, Williams GM, Green AC, Webb PM, Whiteman DC. Do low control response rates always affect the findings? Assessments of smoking and obesity in two Australian case-control studies of cancer. Aust N Z J Public Health 2009;33:312–9. Stang A, Jockel KH. Studies with low response proportions may be less biased than studies with high response proportions. Am J Epidemiol 2004;159:204–10. MacDougall DM, Halperin SA. Improving rates of maternal immunization: Challenges and opportunities. Hum Vaccin Immunother 2016;12:857–65. Healy CM. Pertussis vaccination in pregnancy. Hum Vaccin Immunother 2016;12:1972–81.
Please cite this article in press as: Rowe SL et al. Effectiveness of parental cocooning as a vaccination strategy to prevent pertussis infection in infants: A case-control study. Vaccine (2018), https://doi.org/10.1016/j.vaccine.2018.02.094
Contents lists available at ScienceDirect
Vaccine journal homepage: www.elsevier.com/locate/vaccine
Effectiveness of parental cocooning as a vaccination strategy to prevent pertussis infection in infants: A case-control study Stacey L. Rowe a,⇑, Ee Laine Tay a, Lucinda J. Franklin a, Nicola Stephens a, Robert S. Ware b,c, Marlena C. Kaczmarek c,d, Rosemary A. Lester a, Stephen B. Lambert d a
State Government Department of Health and Human Services, 50 Lonsdale Street, Melbourne, Victoria, Australia Menzies Health Institute Queensland, Griffith University, Brisbane, Queensland, Australia School of Public Health, The University of Queensland, Brisbane, Queensland, Australia d UQ Child Health Research Centre, School of Medicine, The University of Queensland, Brisbane, Queensland, Australia b c
a r t i c l e
i n f o
Article history: Received 15 October 2017 Received in revised form 18 January 2018 Accepted 22 February 2018 Available online xxxx Keywords: Pertussis Cocoon Immunization Vaccine Vaccination Infant
a b s t r a c t Background: During a pertussis epidemic in 2009, the Department of Health, Victoria, Australia, implemented a cocoon program offering parents of new babies a funded-dose of pertussis-containing vaccine. We assessed vaccine effectiveness (VE) of the program in reducing pertussis infection in infants. Methods: Using a matched case-control design, infants aged <12 months that were notified with pertussis between 1 January 2010 and 31 December 2011, and born during the time that the cocoon program was in place, were identified. Controls were matched by area of residence and date of birth. Telephone interviews we conducted to ascertain parents’ vaccination status, and if vaccinated, timing of vaccination receipt relative to the birth of their baby. Odds ratios (ORs) were calculated for the association between vaccination and pertussis infection, with VE calculated as (1 – OR) 100%. Results: The study recruited 215 cases and 240 controls (response rates 67% and 25% of eligible participants, respectively). Vaccination of both parents after delivery of the infant and 28 days prior to illness onset reduced pertussis infection by 77% (Vaccine Effectiveness [VE] = 77% (confidence interval [95% CI], 18–93%). After adjusting for maternal education, presence of a sibling within the household, and the infants’ primary course vaccination status, the adjusted VE was 64% (95% CI, 58–92%). Conclusions: Although not reaching statistical significance, our results demonstrated that cocoon immunisation – where both parents are vaccinated in the post-partum period – may offer some protection again infant pertussis infection. Cocoon immunisation could be considered in circumstances where antenatal vaccination of the mother has not occurred. Crown Copyright Ó 2018 Published by Elsevier Ltd. All rights reserved.
1. Background Bordetella pertussis (pertussis) has re-emerged as a global threat to public health, with recent epidemics reported in several countries [1–4] including Australia [5]. In Victoria, Australia, medical practitioners and laboratories are required to notify the State Government Department of Health and Human Services cases of pertussis in accordance with the Public Health and Wellbeing Act 2008 [6]. In the latter half of 2008, an increase in the number of pertussis cases was observed [7]. In subsequent years, annual case numbers more than doubled from 3698 in 2009 to 8831 in 2011 [8]. Whilst epidemics of pertussis are known to occur in 3–5 yearly cycles [9], the greatest concern with the most recent resurgence ⇑ Corresponding author at: Department of Health and Human Services, GPO Box 4057, Melbourne, Victoria, Australia. E-mail address: [email protected] (S.L. Rowe).
was an increase in cases among infants too young to be vaccinated [8] who are vulnerable to severe disease and death [10]. A range of vaccination strategies to prevent pertussis in infants, children, and adolescents have been applied by governments worldwide. Cocooning is one strategy whereby infants who are too young to be vaccinated are indirectly protected against disease by vaccinating their parents and other close contacts. In 2005, the Global Pertussis Initiative recommended that cocooning be implemented in countries where it was considered economically feasible [11]. Cocooning strategies were adopted by several countries, and in all Australian state and territories as an epidemic response at the time [12]. Between June 2009 and June 2012, the Victorian Government Department of Health introduced a cocoon program, whereby parents of new babies were offered a dose of free pertussis-containing vaccine (dTpa). Under the program, fathers could receive the vaccine immediately before or during pregnancy or after delivery of
https://doi.org/10.1016/j.vaccine.2018.02.094 0264-410X/Crown Copyright Ó 2018 Published by Elsevier Ltd. All rights reserved.
Please cite this article in press as: Rowe SL et al. Effectiveness of parental cocooning as a vaccination strategy to prevent pertussis infection in infants: A case-control study. Vaccine (2018), https://doi.org/10.1016/j.vaccine.2018.02.094
2
S.L. Rowe et al. / Vaccine xxx (2018) xxx–xxx
the infant. Mothers were recommended to receive the vaccine as soon as possible after delivery. Estimated uptake was good with 80% of mothers and 70% of fathers reporting receipt of vaccine in relation to the birth of their baby [13]. The aim of this study was to assess vaccine effectiveness (VE) of cocoon immunisation in reducing the risk of pertussis in infants. 2. Materials & methods We conducted a matched case-control study to examine the effectiveness of the cocoon program in preventing pertussis in infants born between 15 June 2009 and 31 December 2011. Approval to conduct the study was obtained from the Medical Research Ethics Committee of The University of Queensland, Australia. Cases were defined as infants aged younger than 12 months with confirmed pertussis (case definition, Fig. 1 [14]) notified to the then State Government Department of Health between 01 January 2010 and 31 December 2011. Controls were babies born between 15 June 2009 and 7 January 2011 and registered on the Victorian Perinatal Data Collection (VPDC) – a population-based registry containing all births in Victoria. Data sources were linked with the Victorian Death Index – a registry of all deaths occurring in Victoria – to identify and exclude infants who had died. Our per protocol approach was to over-sample the number of controls (n = 6) per case to allow for loss to follow up or declinations to participate. All potential controls identified from the VPDC were matched to cases by statistical sub-division (SSD, socially and economically homogeneous areas of Victoria characterised by identifiable links between the inhabitants) and date of birth (+/ 7 days). Potential controls for each case were then randomly ordered using the binomial distribution, and the first six controls were selected for inclusion in the study. Completed interviews were sought from all cases and at least one control per case. The call protocol involved up to six call attempts to establish contact with each participant over varying times of the day, including weekends. For controls, if contact was not established, or if the person refused to participate or was ineligible, they were replaced by the next control participant from the ordered pool of six. 2.1. Data collection, management, and validation Participant information letters were sent to potential participants to advise them of the objectives of the study, to provide them
with an opportunity to opt out, and to request them to have their children’s immunisation records available during the interview. Computer-assisted telephone interviews were carried out between September 2013 and June 2014. Case participants were interviewed, and if completed, interviews were attempted with their corresponding controls. Data collected during interview included socio-demographic details of the parent(s); household composition (including number of resident siblings); the household’s vaccination status (parents, infants and siblings); and other potential risk factors such as breastfeeding history and childcare attendance in the first 12 months of life. Parity and smoking status during pregnancy were obtained from VPDC. Date of illness onset for cases were obtained from the notification record. Parents were asked about receipt of a pertussis-containing vaccine in relation to the birth of the infant included in the study. Exact date of vaccination was sought. If the exact date was not known, month and year of vaccination, or an estimated timing of vaccination relative to the birth of their infant was requested. For these participants, a derived date of vaccination was calculated by adding the estimated timing of vaccination in days to the infants’ date of birth using the following methodology: Where only month and year of parental vaccination was available and this was the same as the month and year of birth of the infant, we assumed that the parent was vaccinated (1) within seven days of the infant date of birth if the vaccination took place in hospital; or (2) within two weeks of the infant date of birth if the vaccination took place elsewhere. In Australia, infants are recommended to receive a pertussiscontaining vaccine at two, four, and six months of age, with the first dose able to be administered from six weeks of age. During telephone interviews, participants were asked to recall their infant’s (and associated siblings’) vaccination status and the dates at which each of the three primary course vaccinations were received. Responses were validated using the then Australian Childhood Immunisation Register (ACIR) – a national register that recorded vaccinations given to children younger than 7 years of age [15]. If the self-reported vaccination date was discordant with the ACIR, we used the date of vaccination sourced from ACIR in the analysis. Infant vaccination status was included in the analysis as a categorical covariate (0, 1, or 2 doses). To be classified as vaccinated with 1 or 2 doses, we applied a refractory period whereby infants were considered vaccinated only if receipt of the vaccine was 14 days prior to illness onset. For controls, the onset date of their matched case was used.
Fig. 1. Case definition for a confirmed pertussis case.
Please cite this article in press as: Rowe SL et al. Effectiveness of parental cocooning as a vaccination strategy to prevent pertussis infection in infants: A case-control study. Vaccine (2018), https://doi.org/10.1016/j.vaccine.2018.02.094
S.L. Rowe et al. / Vaccine xxx (2018) xxx–xxx
Among parents who reported being vaccinated during interview but were unable to provide an exact date of vaccination, consent was provided for the investigators to ascertain date of vaccination with their immunisation provider. For mothers, the primary exposure of interest was defined as vaccination after delivery and 28 days prior to illness onset in the infant. For fathers, the primary exposure of interest was defined as ever being vaccinated: either immediately before or during pregnancy, or after delivery and 28 days prior to illness onset in the infant. This was reflective of the program in place in Victoria. The reference category against which all exposures of interest were compared was ‘‘Not vaccinated (or vaccinated after illness onset)”. For households containing two parents, the primary exposure of interest was defined as both parents vaccinated after delivery of the infant and 28 days prior to illness onset in the infant, and the reference category was: ‘‘Neither parent vaccinated (or vaccinated after illness onset)”. 2.2. Statistical analyses To calculate the detectable difference in pertussis by vaccination status we assumed 250 case infants would be recruited, with 3 matched controls per case, that the probability of maternal cocoon vaccination among control infants was 80%, and that the correlation coefficient for vaccination between matched cases and controls was 0.2. With 80% power and alpha = 0.05 we would be able to detect a true vaccine effectiveness of 40% or more in vaccinated compared to unvaccinated mothers. Statistics were summarised as frequencies (%). As planned a priori conditional logistic regression was used to identify factors associated with disease in three separate models assessing (1) effectiveness of maternal vaccination alone; (2) paternal vaccination alone; and (3) combined parental vaccination status among two-parent households. Models were conditioned on their individually-matched case-control groups. Unadjusted odds ratios (OR) and 95% confidence intervals (CI) were calculated where the exposure was vaccination status and the outcome was pertussis infection. Adjusted odds ratios (aOR) were calculated using multivariable conditional logistic regression. Variables were explored as potential confounding factors by first examining their independent associations with the exposure (vaccination status) and outcome (pertussis infection) during univariate analysis. Factors that were significantly associated with both the exposure and outcome at p < 0.05 were further explored as potential confounding factors in the multivariable model. Crude and adjusted VE was calculated as VE = (1 - OR) 100%, where OR is the odds ratio for vaccination between cases and controls. Three separate sensitivity analyses were conducted. First, analyses were re-run using frequency matching. This approach was not planned a priori, but was conducted after it was identified that a number of the interviewed case participants were unable to be matched with one or more of their corresponding controls (due to a low response among control participants). Participants were grouped in two-month blocks according to the infants’ date of birth and the conditional logistic regression models were re-run. Second, analyses were re-run after excluding infants aged 4 months. This was conducted to explore the effectiveness of parental vaccination in reducing pertussis infection among infants who were either too young to receive any of their recommended pertussis-containing vaccine doses, or had only received their first dose. A third sensitivity analysis was conducted to account for possible misclassification bias introduced due to the application of a 14-day refractory period around the infants’ vaccination status. In this latter analysis, the multivariable model included all covariates included in the primary analysis except for the infants’ vaccination status. All data analyses were completed using Stata Version 13.1 (StataCorp, College Station, TX).
3
3. Results 3.1. Study participants A total of 393 cases and 2254 controls met the criteria for inclusion in the study and were invited to participate. After excluding participants who had opted out or were subsequently found to be ineligible, 382 (97%) cases and 1981 controls (88%) were deemed eligible for interview. Interviews were completed with 215 cases (response rate 67% of eligible cases) and 240 controls (response rate, 25% of eligible controls). At least one matched control (n = 223) was available for 146 interviewed cases and were included in the analyses. A total of 369, 351 and 430 participant mothers, fathers and siblings, respectively, were also included in the analysis (Fig. 2). Among the 146 cases included in the analysis, the median age at illness onset was 2.8 months (Inter Quartile Range [IQR] 1.4–5.6 months). Seventy-three percent sought hospital-based care: 36 cases presented to an emergency department and an additional 71 cases were admitted. Among cases that were admitted to hospital, the median length of stay was six days (IQR: 3–11), and 24% were admitted to an Intensive Care Unit (ICU). Cases requiring ICU care were younger (median age: 1 month [IQR: 0.6–1.5 mont hs]) than those who did not (median age: 2.2 months [IQR: 1.3 – 3.2 months]). None of the cases died. Differences in participant characteristics are shown in Table 1. Maternal education, maternal work status and household income were all were positively associated with parental vaccination status and negatively associated with infant pertussis diagnosis. Maternal education was selected as a potential confounding factor for inclusion in the multivariable models. Having a resident sibling was negatively associated with vaccination status and positively associated with pertussis diagnosis. The infants’ own primary vaccination status was also positively associated with parental vaccination status and negatively associated with pertussis diagnosis. These factors were all explored and ultimately included in the adjusted model. 3.2. Vaccination validation Of the 369 study infants who were included in the primary analysis, vaccination was validated for 344 (93%) infants. There was 96% to 98% concordance found between self-reported vaccination status (i.e. whether or not the infant was vaccinated at 2, 4 or 6 months of age) and ACIR. Of those, self-reported vaccination date differed from ACIR in between 8% and 11% of infants. For parents included in the primary analysis, 165 (50%) had provided an estimated timing of their vaccination during interview and also provided consent for the investigators to ascertain the exact date of vaccination from their immunisation provider. Of these, exact date of vaccination was ascertained from immunisation providers in 40 of 62 (65%) case parents and 73 of 103 (71%) control parents (p = 0.395). 3.3. Vaccination status of parents Ninety-one (62%) case mothers and 149 (67%) control mothers reported being vaccinated prior illness onset in the infant (Table 2). Infants of mothers vaccinated after delivery and 28 days prior to illness onset had a lower odds of pertussis infection compared with infants of mothers who were not vaccinated (unadjusted OR = 0.40 [95% CI, 0.20–0.82], VE = 60% [95% CI, 18%–80%]). After adjustment for confounding variables, the protective effect of vaccinating the mother alone was not retained (aOR = 1.17 [95% CI, 0.41–3.36]). Maternal pre-pregnancy vaccination was associated with a lower
Please cite this article in press as: Rowe SL et al. Effectiveness of parental cocooning as a vaccination strategy to prevent pertussis infection in infants: A case-control study. Vaccine (2018), https://doi.org/10.1016/j.vaccine.2018.02.094
4
S.L. Rowe et al. / Vaccine xxx (2018) xxx–xxx
Cases
Controls
393 (100%)
2,254 (100%)
Eligible to participate* Opt out: n=8 (2%) Ineligible: n=3 (0.8%)
Opt out: n=183 (8%) Case opt out: n=29 (1%) Ineligible: n=61 (3%)
Eligible for interview Phone no. not functioning / found: n=61 (16%)
382 (97%)
Matched case not interviewed: n=682 (34%) Phone no. not functioning / found: n=345 (10%)
1,981 (88%)
Call attempts made Call attempt / interview unsuccessful: n=106 (26%)
321 (82%)
954 (42%)
Call attempt / interview unsuccessful: n=1027 (46%)
Completed interview Matched control not available: n=69 (18%)
215 (67%^)
240 (25%^)
Matched case not available: n=17 (<1%)
Infants included in analysis Mothers: 146 Fathers: 133 Siblings: 220
146 (45%^) 223 (23%^)
Mothers: 223 Fathers: 218 Siblings: 212
* Ineligible case participants included two participants with incorrect date of birth and one participant with an incorrect address (non-Victorian resident); ineligible control participants included participants whose corresponding matched case was not eligible, or who were stillborn. ^ Denominator = Call attempts made (i.e. examined for eligibly) Fig. 2. Study attrition.
Table 1 Study participant characteristics, cases and controls. Total participants
Male infant Mother born in Australia Father born in Australia Mother of Aboriginal or Torres Strait Islander ethnicity Father of Aboriginal or Torres Strait Islander ethnicity Mother, tertiary education Father, tertiary education Mother in paid employment (including part time employment) Father in paid employment (including part time employment) Household income of at least AUD$60,001 Language other than English used in household Attended childcare in first 12 months of life Breastfed (exclusively) for at least 2 weeks after birth Breastfed (exclusively) for 4–6 months after birth Ever smoked during pregnancy Parity > 1 Resident sibling at home Sibling < 5 years Sibling 5–17 years Resident father at homea a
Cases
Controls
p-value
n 146
%
n 223
%
77 120 110 2 3 43 28 73 120 90 28 22 127 74 21 112 114 70 69 133
52.7 82.2 75.3 1.4 2.1 29.5 19.2 50.0 82.2 61.6 19.2 15.1 87.0 58.7 14.4 76.7 78.0 47.9 47.2 91.1
117 196 193 0 4 100 76 154 208 167 27 59 196 125 27 134 141 118 46 218
52.4 87.9 86.6 0.0 1.8 44.8 34.1 69.1 93.3 74.9 12.1 26.5 87.9 63.8 12.1 60.1 63.2 52.9 20.1 97.8
0.959 0.187 0.008 0.113 0.030 0.009 0.001 0.001 0.003 0.010 0.131 0.027 0.797 0.363 0.004 0.001 0.003 <0.001 <0.001 0.004
Denotes a two-parent household where the second parent was male. There were no same sex households identified in the study.
odds of infant pertussis: unadjusted OR = 0.20 (95% CI, 0.04–0.88) and VE of 80% (95% CI, 12%–96%). After adjustment, the point estimate was attenuated (aOR = 0.47 [95% CI, 0.04–6.02] and aVE = 53% [95% CI, 502%–96%]) (Table 2). Among the mothers reporting pre-pregnancy vaccination, 75% of the case mothers and 85% of the control mothers reported receiving the vaccine within the year prior to becoming pregnant (data not shown). Sixty-one (41%) case fathers and 121 (60%) control fathers reported being vaccinated prior to onset of illness in the infant (Table 2). Fathers vaccination before, during, or after delivery and 28 days prior to onset was associated with a lower odds of infant
pertussis (OR = 0.36 [95% CI, 0.19–0.69], VE = 64% [95% CI, 31–81%]. The adjusted VE was 37% [95% CI, 0.61–75%]) (Table 2). Vaccination of both parents after delivery and 28 days prior to illness onset in the infant was associated with a lower odds of infant pertussis (OR = 0.23 [95% CI, 0.07–0.82], VE = 77% [95% CI, 18%–93%]). This association was reduced after adjustment for possible confounders (aOR = 0.36 [95% CI, 0.08–1.58] and VE = 64% [95%CI, 58–92%]). The sensitivity analysis using frequency matching included data from 215 cases and 240 controls. Results were comparable to those found in the primary analysis. However, in contrast to the primary
Please cite this article in press as: Rowe SL et al. Effectiveness of parental cocooning as a vaccination strategy to prevent pertussis infection in infants: A case-control study. Vaccine (2018), https://doi.org/10.1016/j.vaccine.2018.02.094
Model
Cases n (%)
Controls n (%)
Multivariate modelb
Univariate model OR (95% CI)
VE (95% CI)
aOR (95% CI)
aVE (95% CI)
62 (27.8) 27 (12.1) 7 (3.1) 3 (1.4) 91 (40.8) 10 (4.5) 7 (3.1) 4 (1.8) 12 (5.4)
1.00 0.20 (0.04, 0.88) NC NC 0.40 (0.2, 0.82) 2.14 (0.62, 7.41) 6.30 (0.74, 53.26) NC NC
80 (12, NC NC 60 (18, 114 ( 530 ( NC NC
1.00 0.47 (0.04, NC NC 1.17 (0.41, 0.20 (0.01, 6.50 (0.28, NC NC
53 ( 502, 96) NC NC 17 ( 236, 59) 80 ( 357, 99) 550 ( 14992, 72) NC NC
67 (50.4) 32 (24.1) 17 (12.8) 12 (9.0) 5 (3.8)
82 (37.6) 94 (43.1) 6 (2.8) 21 (9.6) 15 (6.9)
1.00 0.36 9.66 0.57 0.66
64 (31, 81) 866 ( 7707, 43 ( 54, 79) 34 ( 137, 82)
(3) Combined parental vaccination status Both parents not vaccinated (or vaccinated after illness onset) (REF) Both parents vaccinated after delivery & 28 days before illness onset
36 (27.1) 20 (15.0)
52 (23.9) 50 (22.9)
1.00 0.23 (0.07, 0.82)
Other risk factors Education of mother Resident sibling at home Infant, not vaccinated (REF) Infant vaccinated, 1 dose Infant vaccinated, 2 doses
43 (29.5) 113 (77.4) 81 (55.5) 32 (21.9) 33 (22.6)
100 (44.8) 140 (62.8) 91 (40.8) 60 (26.9) 72 (32.3)
0.48 2.24 1.00 0.23 0.12
(1) Maternal vaccination status Not vaccinated (or vaccinated after illness onset) (REF) Vaccinated before pregnancy Vaccinated during pregnancy Vaccinated before birth, timing unknown Vaccinated after delivery & 28 days before illness onset Vaccinated after delivery & <28 days before illness onset Vaccinated after delivery, timing unknown Vaccinated, timing unknown Vaccination status unknown
54 (37.0) 8 (5.5) 3 (2.1) 1 (0.7) 49 (33.6) 20 (13.7) 10 (6.9) 0 (0.0) 1 (0.7)
(2) Paternal vaccination status Not vaccinated (or vaccinated after illness onset) (REF) Vaccinated (any time & 28 days before illness onset)a Vaccinated after delivery & <28 days before illness onset Vaccinated, timing unknown Vaccination status unknown
(0.19, (1.19, (0.21, (0.18,
0.69) 78.07) 1.54) 2.37)
96)
80) 641, 38) 5226, 26)
77 (18, 93)
19)
6.02)
3.36) 4.57) 150.92)
1.00 0.63 (0.25, 1.61) NC 1.11 (0.26, 4.85) 0.52 (0.07, 3.68)
37 ( 61, 75) NC 11 ( 385, 74) 48 ( 268, 93)
1.00 0.36 (0.08, 1.58)
64 ( 58, 92)
S.L. Rowe et al. / Vaccine xxx (2018) xxx–xxx
(0.30, 0.77) (1.32, 3.80) (0.09, 0.60) (0.03, 0.41)
Abbreviations: CI – confidence interval; NC – not calculable due to zero cells; OR – odds ratio; aOR – adjusted odds ratio; REF – reference; VE – vaccine effectiveness. a Vaccinated before or during pregnancy, or after delivery and <28 days prior to illness onset. b Model (1) adjusted for maternal education, paternal and infant vaccination status, and resident sibling. Model (2) adjusted for paternal education, maternal and infant vaccination status, and resident sibling. Model (3) adjusted for maternal education, infant vaccination status, and resident sibling.
5
Please cite this article in press as: Rowe SL et al. Effectiveness of parental cocooning as a vaccination strategy to prevent pertussis infection in infants: A case-control study. Vaccine (2018), https://doi.org/10.1016/j.vaccine.2018.02.094
Table 2 Vaccination status and other risk factors associated with infant pertussis, odds ratios and associated vaccine effectiveness estimates.
6
S.L. Rowe et al. / Vaccine xxx (2018) xxx–xxx
analysis, infants of fathers who were vaccinated before, during pregnancy, or after delivery were significantly less likely to have pertussis even after adjustment for confounding variables (aOR = 0.42, 95% CI, 0.20–0.78) (Supplementary Table 1). The analysis excluding infants aged 4 months included 70 cases and 102 controls. Results were similar to those obtained in the primary analysis but were imprecise due to smaller numbers (Supplementary Table 2). The modified multivariable model built to account for possible misclassification bias relating to the 14-day refractory period applied to infants’ vaccination status demonstrated that vaccination of both parents after delivery and 28 days prior to illness onset was associated with a significantly lower odds of infant pertussis (aOR = 0.24 [95% CI, 0.06–0.95], VE = 76% [95%CI, 5–94%]) (Supplementary Table 3).
4. Discussion This study demonstrates that cocoon immunisation – defined as both parents being vaccinated after delivery and 28 days prior to illness onset – reduced pertussis infection in infants by 64% (95% CI, 58–92%). Whilst the point estimate in the adjusted model was indicative of a moderately protective effect, the confidence intervals were wide and included the null value, and should therefore be interpreted with caution. This study was unique in the use of a case-control study design to examine the effectiveness of cocoon immunisation where the timeframe to define the cocoon was distinctly postpartum and included paternal vaccination status. During the most recent global pertussis epidemic, at least eight countries adopted a cocoon strategy [16]. However data from field studies relating to its effectiveness were lacking. Many cocoon programs were implemented on the basis of modelling studies [17]. Modelling showed cocooning may be effective in reducing pertussis in infants [18], preventing up to 55% of infant cases [19]. In 2009, when Victoria’s program was implemented, cocooning was considered an appropriate epidemic response designed to protect infants most vulnerable to pertussis infection. Since that time, a range of studies exploring the effectiveness of cocooning to prevent pertussis in infants have been published with conflicting results. A case-control study showed that when both parents are immunised before or after birth and 4 week before disease onset, risk of pertussis among infants aged <4 months was reduced by 51% [20]. In contrast, other studies have suggested that cocooning may be ineffectual, inefficient, and resource intensive [21,22]. Using a hospital-based cross-sectional study, Castagnini et al. (2012) found that rates of pertussis infection did not differ after the introduction of postpartum pertussis vaccination for mothers [23]. Another hospital-based study found no difference in the severity of pertussis illness in infants aged 6 months after implementation of postpartum cocooning [24]. An Australian study using a populationbased linkage study demonstrated no difference in the incidence of pertussis among infants born to parents who were vaccinated postpartum compared to unvaccinated parents [25]. Only a small number of mothers in our study reported being vaccinated before pregnancy. We found that maternal prepregnancy vaccination was associated with an unadjusted VE of 80% (95% CI, 12–96%) however, after adjustment the VE reduced and statistical significance was not retained. Of these mothers, a majority reported receiving the vaccine within a year prior to becoming pregnant. It is plausible that the protective effect seen in this pre-pregnancy cohort may relate to transplacental transfer of pertussis-specific antibodies in infants; however, the literature provides conflicting evidence on this subject [20,26]. Maternal vaccination with a pertussis-containing vaccine is increasingly being adopted worldwide in light of the compelling
evidence demonstrating the effectiveness of prenatal vaccination [27–29]. Amirthalingam et al. (2014) demonstrated a VE of 91% (95% CI, 84%–95%) using a screening method, and Dabrera et al. (2015) reported a VE of 93% (95% CI, 81–97%) using a casecontrol design [27,28]. Winter et al. (2017) [29] in a linkage study demonstrated that prenatal vaccination was 85% more effective than postpartum vaccination in prevention pertussis in infants <8 weeks of age. At the time that our cocoon program was in place, vaccination during pregnancy was not specifically recommended in Victoria nor Australia-wide [30]. Not surprisingly, only three (2%) case mothers and seven (3%) control mothers indicated that they had received the vaccine during pregnancy. Vaccination during pregnancy was recommended in the United States in 2012 [31]. Australian guidelines now recommend vaccination of women in the third trimester of pregnancy [32]. Additionally, in Victoria, the current program includes vaccination of partners of pregnant women (who are at least 28 weeks pregnant) [33]. Our study found that the presence of a sibling within the household – particularly among those aged 5–17 years – was positively associated with pertussis infection among infants. This finding is consistent with recent evidence indicating that siblings are important sources of pertussis infection among infants [29,34,35]. Additional research is needed to further elucidate the role of siblings in the transmission of pertussis among infants and how this may change relative to local pertussis epidemiology. Updated guidelines published in the United States recommends that – in addition to pertussis vaccination of pregnant women – other people who expect to be in contact with an infant should be ‘‘up to date with pertussis vaccination” [36]. Depending on the age of siblings at the time of an infant’s birth, siblings could be considered ‘‘up to date” having received a pertussis-containing vaccine up to 10 years prior to an infant’s birth. The role of targeted sibling vaccination programs delivered alongside programs supporting vaccination during pregnancy warrants further exploration. 4.1. Strengths and limitations The strength of our study was in its robust study design. We used two population-based, reliable registers [37,38] ensuring good ascertainment of laboratory-confirmed pertussis cases and eligible controls. Our earlier work identified that uptake in the parental pertussis vaccination program varied by time and area of residence [39,13], and our matching methodology accounted for these factors. Another strength related to our ability to explore the effectiveness the parental vaccination individually or combined among two-parent households. This enabled us to evaluate the effect of a complete ‘‘cocoon”, which was central to the program’s design and is lacking in similar studies examining cocoon immunisation. One of the major limiting factors in this study was our low participation rate which limited generalizability of our study. Low participation may have been influenced by the considerable delay between program implementation and evaluation. Whilst an estimation of vaccine uptake was assessed at 6-months [39] and 18months [13] post program implementation, this vaccine effectiveness study was not conducted until after the program ceased (June 2012). The delay between potentially receiving the vaccine under the program (as early as 15 June 2009) and being interviewed (as late as June 2014) may have contributed to the poor response rate among both cases and controls. This delay may have introduced issues relating to the general recall of vaccination receipt. However, our matching methodology would have minimised any systematic differences in recall between case and control participants. Some residual systematic recall or misclassification bias between case and control participants may exist in our study due to the recognised limitation of case control studies generally: Case participants may be more likely to recall pertinent informa-
Please cite this article in press as: Rowe SL et al. Effectiveness of parental cocooning as a vaccination strategy to prevent pertussis infection in infants: A case-control study. Vaccine (2018), https://doi.org/10.1016/j.vaccine.2018.02.094
S.L. Rowe et al. / Vaccine xxx (2018) xxx–xxx
tion relating to receipt or timing of vaccination by virtue of being affected by the condition of interest. Conversely, control participants may have been less likely to recall details around vaccine receipt and timing. This may have led to more control participants being classified as ‘‘Vaccinated, timing unknown”, and thus being excluded from our main outcome measure. Another limiting factor in our study was that the low response rate in controls meant that 69 case participants who had completed an interview were excluded from the matched analysis. A sensitivity analysis incorporating these cases using frequency matching found comparable results to those obtained in our primary analysis. Low participation is a common limitation in many case-control studies, however systematic biases are not an inevitable consequence of such a problem [40,41]. A range of methodological approaches were adopt to minimise and assess the impact of low participation. For example, and in addition to our careful study design involving appropriately matched controls, we adopted an exhaustive call protocol to contact participants. We also assessed the differences between participating and non-participating cases and controls. Whilst we were unable to assess whether there were any differences between the two groups with regard to parental vaccination status, we found no systematic differences between participating and nonparticipating cases or controls with respect to the following factors: baby’s year of birth, area of residence, parity, birth plurality and smoking status during pregnancy. 5. Conclusions Results demonstrated that cocoon immunisation reduced pertussis in infants by 64%. Whilst not reaching statistical significance, this estimate suggests that cocoon immunisation – when defined as the vaccination of both parents after delivery and 28 days prior to illness onset – may offer some protection against infant pertussis infection. The compelling research evidence derived from other studies [27,28] indicates that vaccination during pregnancy is at present the most effective approach for preventing cases in early infancy. However, barriers relating to acceptance of vaccination during pregnancy and challenges with program implementation at a population level mean that uptake remains suboptimal in many countries [42,43]. Where vaccination has not been given during pregnancy, protecting the infant through cocoon vaccination of both parents may afford clinically meaningful protection and should be considered in this context. Funding This work was supported by the State Government Department of Health, Victoria, Australia. Conflicts of interest None declared. Acknowledgements The authors wish to acknowledge Drs Helen Quinn, Tom Snelling and Peter McIntyre of the National Centre for Immunisation Research and Surveillance. The Hunter Valley Research Institute for the conduct of computer assisted telephone interviews, particularly Shanthi Ramanathan, Russ Redford, and David Shellard. We also acknowledge staff at the Clinical Councils Unit, particularly Katharine Gibson and Dr Mary-Ann Davey who oversaw the release of Victorian Perinatal Data Collection data, as well as members of the Consultative Council on Obstetric Morbidity and Mortality for granting access to VPDC data. We acknowledge staff of
7
the Data Linkages Unit, particularly Ying Chen, Mee Lee Easton and Simon Waters who were involved in linkage of data against the Victorian Death Index and selection of controls. We also gratefully acknowledge staff in the Health Protection Branch, Victorian Government Department of Health, in particular we thank Michael Batchelor, Helen Pitcher, Rosemary Morey, and Kana Volin, Joshua Stapleton, Alannah Moon, Michael Brown, and Michael Panagiotis. Finally, we sincerely thank the external members of the Pertussis Expert Advisory Group: Professor Terry Nolan, Dr Nigel Crawford, and Dr Jim Buttery for their advice relating to the conduct, analysis, and reporting of the study. A/Prof Lambert was supported by an Early Career Fellowship from the Australian Government National Health and Medical Research Council and a people support grant from the Queensland Children’s Hospital Foundation during the conduct of this work. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at https://doi.org/10.1016/j.vaccine.2018.02. 094. References [1] Fisman DN, Tang P, Hauck T, Richardson S, Drews SJ, Low DE, et al. Pertussis resurgence in Toronto, Canada: a population-based study including testincidence feedback modeling. BMC Public Health 2011;11:694. [2] Takeuchi M, Yasunaga H, Horiguchi H, Matsuda S. The incidence of pertussis hospitalizations among Japanese infants: excess hospitalizations and complications? Epidemiol Infect 2012;140:1497–502. [3] Celentano LP, Massari M, Paramatti D, Salmaso S, Tozzi AE, Group E-N. Resurgence of pertussis in Europe. Pediatr Infect Dis J 2005;24:761-5 [4] Centers for Disease Control and Prevention. Pertussis epidemic–Washington, 2012. MMWR Morb Mortal Wkly Rep 2012;61:517-22 [5] Australian Government Department of Health and Ageing. National Notifiable Diseases Surveillance System, Number of notifications of Pertussis, Australia, in the period of 1991 to 2012 and year-to-date notifications for 2013. 2013. [6] Victorian Government Health Information. Public Health and Wellbeing Act 2008 and Public Health and Wellbeing Regulations 2009. 2010. [7] Donnan EJ, Fielding JE, Simpson K, Moloney M, Vally H. A continuing pertussis epidemic in Victoria, 2009. Victorian Infect Dis Bullet 2009;13:46–51. [8] Franklin LJ, Cowie B. The Victorian 2009–2011 pertussis epidemic in review. Victorian Infect Dis Bullet 2012;15:7–10. [9] Heymann DL. Control of communicable diseases manual. 18 ed. Washington USA: American Public Health Association; 2004. [10] Wood N, Quinn HE, McIntyre P, Elliott E. Pertussis in infants: preventing deaths and hospitalisations in the very young. J Paediatr Child Health 2008;44:161–5. [11] Forsyth KD, Wirsing von Konig CH, Tan T, Caro J, Plotkin S. Prevention of pertussis: recommendations derived from the second Global Pertussis Initiative roundtable meeting. Vaccine 2007;25:2634–42. [12] National Centre for Immunisation Research & Surveillance. Significant events in diptheria, tetanus and pertussis vaccination practice in Australia. 2016. [13] Rowe SL, Cunningham HM, Franklin LJ, Lester RA. Uptake of a governmentfunded pertussis-containing booster vaccination program for parents of new babies in Victoria. Australia Vaccine 2015;33:1791–6. [14] Australian National Notifiable Disease Surveillance System. Surveillance Case Definitions for the Australian National Notifiable Diseases Surveillance System, 1 January 2004 to 1 January 2015. 2015. [15] Hull BP, Deeks SL, McIntyre PB. The australian childhood immunisation register-a model for universal immunisation registers? Vaccine 2009;27:5054–60. [16] Chiappini E, Stival A, Galli L, de Martino M. Pertussis re-emergence in the postvaccination era. BMC Infect Dis 2013;13:151. [17] Centers for Disease Control and Prevention. Updated recommendations for use of tetanus toxoid, reduced diphtheria toxoid and acellular pertussis vaccine (Tdap) in pregnant women and persons who have or anticipate having close contact with an infant aged <12 months —Advisory Committee on Immunization Practices (ACIP), 2011. MMWR Morb Mortal Wkly Rep 2011;60:1424–6. [18] Van Rie A, Hethcote HW. Adolescent and adult pertussis vaccination: computer simulations of five new strategies. Vaccine 2004;22:3154–65. [19] de Greeff SC, Mooi FR, Westerhof A, Verbakel JM, Peeters MF, Heuvelman CJ, et al. Pertussis disease burden in the household: how to protect young infants. Clin Infect Dis 2010;50:1339–45. [20] Quinn HE, Snelling TL, Habig A, Chiu C, Spokes PJ, McIntyre PB. Parental Tdap boosters and infant pertussis: a case-control study. Pediatrics 2014;134:713–20.
Please cite this article in press as: Rowe SL et al. Effectiveness of parental cocooning as a vaccination strategy to prevent pertussis infection in infants: A case-control study. Vaccine (2018), https://doi.org/10.1016/j.vaccine.2018.02.094
8
S.L. Rowe et al. / Vaccine xxx (2018) xxx–xxx
[21] Blain AE, Lewis M, Banerjee E, Kudish K, Liko J, McGuire S, et al. An assessment of the cocooning strategy for preventing infant pertussis-united states, 2011. Clin Infect Dis 2016;63:S221–6. [22] Skowronski DM, Janjua NZ, Tsafack EP, Ouakki M, Hoang L, De Serres G. The number needed to vaccinate to prevent infant pertussis hospitalization and death through parent cocoon immunization. Clin Infect Dis 2012;54:318–27. [23] Castagnini LA, Healy CM, Rench MA, Wootton SH, Munoz FM, Baker CJ. Impact of maternal postpartum tetanus and diphtheria toxoids and acellular pertussis immunization on infant pertussis infection. Clin Infect Dis 2012;54:78–84. [24] Healy CM, Rench MA, Wootton SH, Castagnini LA. Evaluation of the impact of a pertussis cocooning program on infant pertussis infection. Pediatr Infect Dis J 2015;34:22–6. [25] Carcione D, Regan AK, Tracey L, Mak DB, Gibbs R, Dowse GK, et al. The impact of parental postpartum pertussis vaccination on infection in infants: A population-based study of cocooning in Western Australia. Vaccine 2015;33:5654–61. [26] Healy CM, Rench MA, Baker CJ. Importance of timing of maternal combined tetanus, diphtheria, and acellular pertussis (Tdap) immunization and protection of young infants. Clin Infect Dis 2013;56:539–44. [27] Amirthalingam G, Andrews N, Campbell H, Ribeiro S, Kara E, Donegan K, et al. Effectiveness of maternal pertussis vaccination in England: an observational study. Lancet 2014;384:1521–8. [28] Dabrera G, Amirthalingam G, Andrews N, Campbell H, Ribeiro S, Kara E, et al. A case-control study to estimate the effectiveness of maternal pertussis vaccination in protecting newborn infants in England and wales, 2012–2013. Clin Infect Dis 2015;60:333–7. [29] Winter K, Nickell S, Powell M, Harriman K. Effectiveness of prenatal versus postpartum tetanus, diphtheria, and acellular pertussis vaccination in preventing infant pertussis. Clin Infect Dis 2017;64:3–8. [30] Australian Government Department of Health and Ageing. The Australian Immunisation Handbook. 9 ed. Canberra, Australia: Australian Government; 2008. [31] Centers for Disease Control and Prevention. Updated recommendations for use of tetanus toxoid, reduced diphtheria toxoid, and acellular pertussis vaccine
[32]
[33] [34] [35]
[36] [37]
[38]
[39]
[40]
[41]
[42] [43]
(Tdap) in pregnant women–Advisory Committee on Immunization Practices (ACIP), 2012. MMWR Morb Mortal Wkly Rep 2013;62:131–5. Australian Government Department of Health and Ageing. The Australian Immunisation Handbook. 10 ed. Canberra, Australia: Australian Government; 2014. Victoria State Government. Immunisation and pregnancy. 2016. Skoff TH, Kenyon C, Cocoros N, Liko J, Miller L, Kudish K, et al. Sources of Infant Pertussis Infection in the United States. Pediatrics 2015;136:635–41. Levri KM, Reynolds L, Liko J, Dott M, Robinson BF, Cieslak PR. Risk factors for pertussis among hispanic infants: metropolitan portland, oregon, 2010–2012. Pediatr Infect Dis J 2016;35:488–93. Centers for Disease Control and Prevention. Vaccinating Pregnant Patients. Georgia: Atlanta; 2017. Gibney KB, Cheng AC, Hall R, Leder K. Australia’s national notifiable diseases surveillance system 1991–2011: expanding, adapting and improving. Epidemiol Infect 2017;145:1006–17. Davey MA, Sloan ML, Palma S, Riley M, King J. Methodological processes in validating and analysing the quality of population-based data: a case study using the Victorian perinatal data collection. The HIM J 2013;42:12–9. Donnan EJ, Fielding JE, Rowe SL, Franklin LJ, Vally H. A cross sectional survey of attitudes, awareness and uptake of the parental pertussis booster vaccine as part of a cocooning strategy, Victoria. Australia BMC Public Health 2013;13:676. Pandeya N, Williams GM, Green AC, Webb PM, Whiteman DC. Do low control response rates always affect the findings? Assessments of smoking and obesity in two Australian case-control studies of cancer. Aust N Z J Public Health 2009;33:312–9. Stang A, Jockel KH. Studies with low response proportions may be less biased than studies with high response proportions. Am J Epidemiol 2004;159:204–10. MacDougall DM, Halperin SA. Improving rates of maternal immunization: Challenges and opportunities. Hum Vaccin Immunother 2016;12:857–65. Healy CM. Pertussis vaccination in pregnancy. Hum Vaccin Immunother 2016;12:1972–81.
Please cite this article in press as: Rowe SL et al. Effectiveness of parental cocooning as a vaccination strategy to prevent pertussis infection in infants: A case-control study. Vaccine (2018), https://doi.org/10.1016/j.vaccine.2018.02.094