- Email: [email protected]
19, England
Vaccine xxx (xxxx) xxx
Contents lists available at ScienceDirect
Vaccine journal homepage: www.elsevier.com/locate/vaccine
Effectiveness of influenza vaccine in children in preventing influenza associated hospitalisation, 2018/19, England R.G. Pebody a,⇑, H. Zhao a, H.J. Whitaker a, J. Ellis a, M. Donati b, M. Zambon a, N. Andrews a a b
Public Health England National Infection Service, Colindale, London, UK Public Health England National Infection Service, Bristol, UK
a r t i c l e
i n f o
Article history: Received 9 August 2019 Received in revised form 12 October 2019 Accepted 14 October 2019 Available online xxxx Keywords: Influenza vaccine effectiveness Children LAIV Hospitalisation
a b s t r a c t 2013/14 saw the start of the introduction of a new live attenuated influenza vaccine (LAIV) programme for children in England. 2018/19 saw co-circulation of both A(H1N1)pdm09 and A(H3N2), when LAIV was offered to all healthy children 2–9 years of age. LAIV effectiveness against influenza hospitalisation is not well described. This paper presents the 2018/19 end-of-season adjusted vaccine effectiveness (aVE) against laboratory confirmed influenza related hospitalisation in children aged 2–17. The test negative case control approach was used to estimate aVE by influenza A subtype and vaccine type. Cases and controls were selected from a sentinel laboratory surveillance system which collates details of individuals tested for influenza with reverse-transcription polymerase chain reaction (RT-PCR) on respiratory samples. Vaccine and clinical history was obtained from general practitioners of study participants. There were 307 hospitalised cases and 679 hospitalised controls. End-of-season influenza aVE was 53.0% (95% CI: 33.3, 66.8) against influenza confirmed hospitalisation; 63.5% (95% CI: 34.4, 79.7) against influenza A(H1N1)pdm09 hospitalisation and 31.1% (95% CI: 53.9, 69.2) against influenza A(H3N2). LAIV aVE was 49.1% (95% CI: 25.9, 65.0) for any influenza and 70.7% (95% CI: 41.8, 85.3) for A(H1N1)pdm09, whereas for those receiving quadrivalent inactivated influenza vaccine (QIV), aVE was 64.4% (95% CI: 29.4, 82.0) and 44.4% (95% CI: 51.9, 79.6) respectively. We provide evidence of overall significant VE for both LAIV and QIV against influenza associated hospitalisation in children 2–17 years of age, most notably against influenza A(H1N1)pdm09, with non-significant protection against A(H3N2). Crown Copyright Ó 2019 Published by Elsevier Ltd. All rights reserved.
1. Introduction The United Kingdom (UK) started the incremental introduction of a universal paediatric influenza vaccination programme in the 2013/14 influenza flu season with a newly licensed live attenuated influenza vaccine (LAIV). The aim is to ultimately offer annual influenza vaccination to all children 2–16 years of age to both directly protect them, and by reducing their risk of infection, indirectly protect others in the community who may be at higher risk of severe disease [1]. The 2018/2019 influenza season in the UK was the sixth consecutive season of roll-out of LAIV to healthy children, with the offer of LAIV vaccination extended in England to all healthy children aged two to 9 years of age [2]. Children contraindicated LAIV are offered quadrivalent inactivated vaccine (QIV). In 2015/16, following circulation of mainly influenza A(H1N1) pdm09 in the Northern hemisphere, the Centres for Disease ⇑ Corresponding author. E-mail address: [email protected] (R.G. Pebody).
Control (CDC) of the United States reported evidence of poor effectiveness of LAIV in protecting children against influenza [3]. Although other countries, including the UK, published evidence that LAIV vaccine had been effective in protecting children against laboratory confirmed infection in primary care [4], the American Committee on Immunisation Practice (ACIP) temporarily withdrew their recommendation to use LAIV [5]. The reasons for the discrepancy in VE between the US and elsewhere remains unclear, with several hypotheses proposed, though lower effectiveness of LAIV against the (H1N1)pdm09 strain compared to the inactivated vaccine was seen in all settings. Reduced replicative fitness of the A/Bolivia/559/2013 (H1N1)pdm09 vaccine strain contained in LAIV at that time was the leading theory, together with prior LAIV vaccination cited as a potential contributory factor and/or early life exposure to inactivated influenza vaccine which is recommended for children 6 months to 2 years of age [6]. The A/Bolivia strain was updated for the 2017/18 season to the A/Slovenia/2903/2015 (H1N1)pdm09 vaccine strain [4]. Finally, in recent seasons, reduced A(H3N2) specific VE has been observed in both adults and children vaccinated with egg-grown influenza vaccines, both
https://doi.org/10.1016/j.vaccine.2019.10.035 0264-410X/Crown Copyright Ó 2019 Published by Elsevier Ltd. All rights reserved.
Please cite this article as: R. G. Pebody, H. Zhao, H. J. Whitaker et al., Effectiveness of influenza vaccine in children in preventing influenza associated hospitalisation, 2018/19, England, Vaccine, https://doi.org/10.1016/j.vaccine.2019.10.035
2
R.G. Pebody et al. / Vaccine xxx (xxxx) xxx
inactivated and live attenuated. It has been suggested this poor VE may be related to egg adaption of A(H3N2) vaccine viruses [7]. Only limited studies have been published on the recent effectiveness of influenza vaccine against hospitalisation in children [8,9]. The 2018/19 season in England was characterised by cocirculation of both A(H1N1)pdm09 and A(H3N2)[10]. This paper presents the results of an evaluation of influenza vaccine effectiveness in children of 2–17 years in England in protecting against laboratory confirmed infection resulting in hospitalisation during the 2018/19 season.
2. Methods A frequency matched test-negative case-control study was undertaken using the Respiratory DataMart Surveillance system (RDS) according to a pre-defined protocol. Cases and controls were identified from RDS. This is a national sentinel laboratory surveillance system which records details of individuals tested for influenza infection with real-time reverse transcriptase polymerase chain reaction (RT-PCR) on respiratory samples (including both positive and negative results) from 13 NHS and PHE laboratories located across England [11]. These laboratories routinely test for influenza A and B, though many do not undertake influenza A sub-typing. The study population was defined as residents in England 2– 17 years of age (at August 31st 2018) who were admitted to hospital and who had a respiratory swab taken between week 40 2018 and week 20 2019 which was tested for influenza with RT-PCR by one of the RDS laboratories. Those who were swabbed more than seven days after onset were excluded (when onset was known). A case was defined as a child with RT-PCR confirmed influenza infection. A control was defined as a child with RT-PCR influenza negative infection. Controls were group-matched to cases by age group (2–4, 5–8, 9–11, 12–17) and week of sample (+/ 1 week) with up to 3 controls randomly selected per case within these groups. Details of cases and controls from RDS were used to identify the general practitioners of these children, using the patient demographic service (PDS) system. PDS provides the postcode of the child and their GP address. The residential postcode was linked to the Index of Multiple Deprivation (IMD) to identify which IMD quintile the child is resident in and their PHE region. A child who was not identifiable by the national PDS system as being registered with a GP or not resident in England was excluded. The project team sent a short questionnaire to the identified GP to collect key demographic and clinical information on the child including date of onset of illness; fact and date of hospitalisation; underlying clinical risk factor; influenza vaccine history including date and type of vaccination. Statistical analysis was undertaken in Stata 15 (StataCorp, College Station, TX, USA). A descriptive and analytical analysis was undertaken both overall and by age-group (2–8 years and 9– 17 years). For analysis of the case-control study, logistic regression was used to calculate the unadjusted odds ratios for influenza vaccination in cases compared to controls, with a 95% confidence interval, as unadjusted VE = 1-OR. Logistic regression was also used to calculate the odds ratio for vaccination, adjusted for relevant characteristics which changed the adjusted vaccine effectiveness (aVE) against influenza by 3% or more (these could include sex, region, risk group and deprivation). Analysis was undertaken stratified, where numbers were sufficient, by key factors including age group (2–9 years and 10– 17 years); underlying clinical risk factor; type of vaccine (intranasal and inactivated); influenza type and subtype and prior season’s vaccination. Sensitivity analyses were undertaken - specifically
including only swabs where onset date was present; including those with onsets >7 days; removing those with unknown vaccination date and using multiple imputation for missing risk-group information. A multiple imputation approach was used for missing vaccination dates by sampling known vaccination dates from individuals whose event occurred in the same week. VE estimates are reported when there is sufficient precision, based on requiring an expected number of vaccinated cases of at least 8 under the null hypothesis of zero VE. This is calculated as either Tc * proportion of controls vaccinated or, when using an adjusted VE as Tc * Oc/[(1-aVE) + Oc], where Tc is total cases for the VE estimate, Oc is the observed odds of vaccination in the cases and aVE the adjusted VE estimate. Public Health England (PHE) holds permissions under section 251 of the 2006 NHS Act and the 2002 Health Service (Control of Patient Information) Regulations, to process patient information relating to this investigation. 3. Results In all, 2685 questionnaires were sent to GPs of which 713 were for cases and 1972 were for controls. Replies were received for 2089 patients (78%) aged 2–17 years of age. Of these 379 were excluded: nine were not resident in England; one had unknown postcode; 111 had unknown 2018/19 vaccination status; 219 had more than 7 days from disease onset to sampling and 39 were vaccinated within 14 days of onset. After exclusions, 482 cases and 1228 controls were left of which 307 cases and 679 controls were hospitalised (Fig. 1). The description of hospitalised cases and controls by age-group, sex, risk-group and vaccine status are provided in Table 1. Of the 307 hospitalised cases, 104 were due to A(H1N1)pdm09, 40 were due to A(H3N2), 156 were influenza A (subtype not tested) and six were due to influenza B. There was one A and B co-infection. 65% (443) of the hospitalised controls had no other respiratory virus detected, with rhinovirus followed by RSV the main viruses detected (in 64 and 49 controls respectively). 3.1. Model fitting for vaccine effectiveness estimates To estimate VE, age-group, month, PHE region, deprivation (IMD score) and risk group were adjusted for in the multivariable model as they changed the overall estimates by more than 3%. Gender was not included as it changed the vaccine effect by less than 1%. The crude VE and aVE against hospitalisation are shown in Table 2, with an aVE against all influenza, influenza A, influenza A(H1N1)pdm09 and influenza A(H3N2) of 53.0% (95% CI 33.3, 66.8); 55.2% (95% CI: 36.2, 68.5); 63.5% (95% CI: 34.4, 79.7) and 31.1% (95% CI: 53.9, 69.2) respectively. In addition, influenza A (not subtyped) was 51.4% (95% CI: 32.1, 65.2). There were inadequate samples to estimate influenza B VE. The aVE estimates for influenza by type/sub-type are shown by age-group in Table 3. There were no significant differences in aVE between 2–9 and 10–17 year olds with similar aVE point estimates and overlapping 95% confidence intervals in estimates for all influenza and influenza A (Table 3). By risk-group, the aVE point estimates were higher for healthy children compared to those with an underlying clinical risk factor, albeit with overlapping 95% confidence intervals for all influenza and restricted to influenza A (Table 3). For (H1N1)pdm09, the aVE was 81.6% (95% CI 57.3, 92.1) for healthy children compared to 4.8% (95%CI: 176.8, 60.3) for children with an underlying clinical risk factor. By vaccine type, the aVE point estimates were broadly similar with overlapping 95% confidence intervals when comparing LAIV and QIV for
Please cite this article as: R. G. Pebody, H. Zhao, H. J. Whitaker et al., Effectiveness of influenza vaccine in children in preventing influenza associated hospitalisation, 2018/19, England, Vaccine, https://doi.org/10.1016/j.vaccine.2019.10.035
R.G. Pebody et al. / Vaccine xxx (xxxx) xxx
3
Fig. 1. Swabbing results of patients in secondary care in England, October 2018 to April 2019.
all influenza and restricting to influenza A. For A(H1N1)pdm09, LAIV aVE was non-significantly higher compared to QIV. The aVE estimates by prior vaccine history are shown in Table 4. The lowest aVE was seen in those who were vaccinated only in 2017/18 for all influenza and influenza A(H1N1)pdm09. Significant aVE was seen for those vaccinated in 2018/19 regardless of prior vaccine status in 2017/18 for all influenza and influenza A(H1N1) pdm09. Non-significant results were seen for A(H3N2), with 95% confidence intervals spanning zero. Sensitivity analyses found similar estimates if restricted to only patients where onset date was present; removing swabs with unknown vaccination date; including onsets >7 days and imputing missing risk-group information. 4. Discussion In summary, we report that the overall effectiveness of influenza vaccine in children 2–17 years of age in preventing influenza
confirmed hospitalisation was moderate. Effectiveness was high against A(H1N1)pdm09 hospitalisation, but no significant protection was found against A(H3N2). There was no significant difference in effectiveness by age-group or risk-group, except a higher aVE against A(H1N1)pdm09 in healthy children compared to those with an underlying clinical risk factor. Overall effectiveness of LAIV and QIV were similar, though there were some potential differences by influenza A subtype. There was no evidence of residual protection against hospitalisation from vaccination only in the prior season. This study has a number of strengths and weaknesses. Although children without a registered GP have been excluded, as it will not be possible to collect relevant information, only a small number fall into this group. Laboratory testing for influenza infection in the age group studied tends to occur mainly among those presenting to secondary care. As the study is restricted to those children who have been hospitalised, this will have a limited effect on the estimate of VE as cases and (test-negative) controls are likely to
Please cite this article as: R. G. Pebody, H. Zhao, H. J. Whitaker et al., Effectiveness of influenza vaccine in children in preventing influenza associated hospitalisation, 2018/19, England, Vaccine, https://doi.org/10.1016/j.vaccine.2019.10.035
4
R.G. Pebody et al. / Vaccine xxx (xxxx) xxx
Table 1 Details for influenza A and B hospitalised cases and controls, United Kingdom, week 40 2018–week 20 2019 Numbers and row percentages (to indicate positivity ratesa) are shown. Χ2 p-value
Negative
A/H1
A/H3
A (unknown)
B
Age (years) 2–4 5–9 10–17
344 (68%) 151 (70%) 184 (70%)
58 (11%) 19 (9%) 27 (10%)
17 (3%) 6 (3%) 17 (6%)
86 (17%) 37 (17%) 33 (13%)
3 (1%) 2 (1%) 1 (0%)
Sex Female Male
324 (71%) 355 (67%)
46 (10%) 58 (11%)
18 (4%) 22 (4%)
69 (15%) 87 (17%)
2 (0%) 4 (1%)
Risk group No Yes Missing
299 (60%) 333 (81%) 47 (64%)
71 (14%) 23 (6%) 10 (14%)
24 (5%) 14 (3%) 2 (3%)
105 (21%) 39 (9%) 12 (16%)
2 (0%) 2 (0%) 2 (3%)
Onset to swab (days) 0–1 2–4 5–7 missing
155 (61%) 139 (63%) 82 (62%) 303 (40%)
25 31 23 25
(10%) (14%) (17%) (3%)
17 (7%) 12 (5%) 4 (3%) 7 (1%)
56 38 24 38
2 1 0 3
Vaccination Status Unvaccinated at onset Vaccinated >14 days before Vaccination date missing
390 (64%) 270 (78%) 19 (73%)
78 (13%) 26 (7%) 0 (0%)
28 (5%) 11 (3%) 1 (4%)
112 (18%) 40 (11%) 4 (15%)
4 (1%) 1 (0%) 1 (4%)
Previous 2017/18 vaccination No Yes Unknown
418 (64%) 239 (80%) 22 (73%)
84 (13%) 16 (5%) 4 (13%)
26 (4%) 12 (4%) 2 (7%)
121 (18%) 33 (11%) 2 (7%)
6 (1%) 0 (0%) 0 (0%)
Month of event October November December January February March
10 (67%) 57 (80%) 107 (82%) 237 (62%) 260 (68%) 8 (100%)
0 (0%) 5 (7%) 13 (10%) 53 (14%) 33 (9%) 0 (0%)
1 (7%) 0 (0%) 5 (4%) 12 (3%) 22 (6%) 0 (0%)
1 (7%) 7 (10%) 3 (2%) 79 (21%) 66 (17%) 0 (0%)
3 1 2 0 0 0
(20%) (1%) (2%) (0%) (0%) (0%)
IMD (deprivation index) 1 2 3 4 5
192 132 107 126 122
24 19 18 22 21
(9%) (10%) (11%) (13%) (11%)
13 (5%) 5 (3%) 7 (4%) 3 (2%) 12 (7%)
51 30 23 25 27
2 2 1 0 1
(1%) (1%) (1%) (0%) (1%)
Vaccine type Unvaccinated LAIV IIV Vaccinated (unknown)
391 (64%) 194 (74%) 87 (87%) 7 (78%)
78 (13%) 19 (7%) 6 (6%) 1 (11%)
28 (5%) 10 (4%) 2 (2%) 0 (0%)
112 (18%) 38 (15%) 5 (5%) 1 (11%)
5 1 0 0
(1%) (0%) (0%) (0%)
0.723
0.275
<0.001
0.871 (22%) (17%) (18%) (5%)
(1%) (0%) (0%) (50%) P < 0.001
P < 0.001
P < 0.001
P = 0.862 (68%) (70%) (68%) (72%) (67%)
(18%) (16%) (15%) (14%) (15%)
P < 0.001
p-values relate to a Chi-square test on all influenza positives (A and B) and influenza negatives.
Table 2 Samples positive (cases N = 307) and negative (controls N = 679) for influenza A and B according to vaccination status and VE estimates, England, 1 October 2018–8 April 2019. Influenza
A and B A A/H1N1pdm09 A/H3N2
Cases
Controls
Vac
Unvac
Vac
Unvac
83 82 26 12
224 219 78 28
288 288 288 288
391 391 391 391
Crude VE (95%CI)
Adjusteda VE (95% CI)
48.7 49.0 54.7 41.8
53.0 55.2 63.5 31.1
(31.2, 61.8) (31.4, 62.1) (27.6, 71.7) ( 16.4, 70.9)
(33.3, 66.8) (36.2, 68.5) (34.4, 79.7) ( 53.9, 69.2)
CI: confidence interval; VE: vaccine effectiveness; Vac: vaccinated; Unvac: unvaccinated. a Adjusted for age-group, month, region, IMD and risk group.
have similar severity of illness in order to be tested and hospitalised. A relatively large proportion of influenza A detections remain unsubtyped, which reflects local NHS laboratory practice in many instances. Clinicians will not know whether a person admitted to hospital with acute respiratory illness has H1N1pdm09 or H3N2, and so likelihood of sampling will be similar. Flu A sub-typing is undertaken locally regardless of factors such as vaccine history and dependent upon test platforms in use in the laboratory. We have also looked at aVE for flu A unsubtyped,
with the estimate lying, as anticipated between that seen for H3N2 and H1N1pdm09. Finally, although primary care is recommended to keep vaccine history up to date for primary school agechildren, it is possible that vaccine history may not have been captured on some children vaccinated in schools. The missing vaccine history will result in non-differential exposure misclassification resulting in bias to the null. We found that influenza vaccine in children 2–17 years of age provided moderate levels of protection in preventing influenza
Please cite this article as: R. G. Pebody, H. Zhao, H. J. Whitaker et al., Effectiveness of influenza vaccine in children in preventing influenza associated hospitalisation, 2018/19, England, Vaccine, https://doi.org/10.1016/j.vaccine.2019.10.035
5
R.G. Pebody et al. / Vaccine xxx (xxxx) xxx Table 3 Adjusted vaccine effectiveness estimates for influenza associated hospitalisation by sub-type, age-group and vaccine type, England, 1 Oct 2018–8 April 2019. Cases
Adjusteda VE
Controls
Vac
Unvac
Vac
Unvac
(95% CI)
A&B By age-group 2–9 10–17
61 11
147 62
212 61
245 114
52.3 (29.4, 67.8) 62.6 (12.9, 84.0)
By risk group None In risk group
41 31
162 47
100 173
199 160
65.4 (43.8, 78.7) 35.9 ( 9.0, 62.3)
By vaccine type LAIV IIV
58 12
209 209
183 83
359 359
49.1 (25.9, 65.0) 64.4 (29.4, 82.0)
A By age-group 2–9 10–17
60 11
145 61
212 61
245 114
55.6 (33.8, 70.2) 61.2 (8.9, 83.5)
By risk group None In risk group
40 31
161 45
100 173
199 160
67.3 (46.5, 80.0) 38.2 ( 5.8, 64.0)
By vaccine type LAIV IIV
57 12
206 206
183 83
359 359
52.2 (30.0, 67.3) 64.2 (28.7, 82.0)
A/(H1N1)pdm09 By age-group 2–9 10–17
19 2
52 21
212 61
245 114
61.4 (26.2, 79.8) NR
By risk group None In risk group
10 11
61 12
100 173
199 160
81.6 (57.3, 92.1) 4.8 ( 176.8, 60.3)
By vaccine type LAIV IIV
14 6
73 73
183 83
359 359
70.7 (41.8, 85.3) 44.4 ( 51.9, 79.6)
A/H3N2 By age-group 2–9 10–17
8 3
13 14
212 61
245 114
14.5 ( 137.1, 69.2) NR
By risk group None In risk group
6 5
18 9
100 173
199 160
NR NR
By vaccine type LAIV IIV
9 2
27 27
183 83
359 359
20.1 ( 87.3, 65.9) NR
CI: confidence interval; VE: vaccine effectiveness; Vac: vaccinated; Unvac: unvaccinated; LAIV – live attenuated vaccine, aTIV – adjuvanted trivalent inactivated vaccine, QIV – quadrivalent inactivated vaccine; NR – not reported. a Adjusted for age-group, month, region, IMD and risk group.
Table 4 Adjusted vaccine effectiveness estimates for influenza hospitalisation by subtype and prior vaccine history, England in 2–17 year olds, 1 Oct 2018–8 April 2019. Casesa
Controlsa
Adjustedb VE
Vac
Unvac
Vac
Unvac
(95% CI)
A and B 2017/18 only 2018/19 only Both seasons
22 35 33
185 185 185
53 95 171
292 292 292
11.3 ( 58.1, 50.3) 44.5 (10.5, 65.5) 64.8 (44.5, 77.7)
A/(H1N1)pdm09 2017/18 only 2018/19 only Both seasons
6 10 9
65 65 65
53 95 171
292 292 292
10.8 ( 138.0, 66.6) 64.9 (20.0, 84.6) 71.0 (35.2, 87.0)
A/H3N2 2017/18 only 2018/19 only Both seasons
5 3 7
22 22 22
53 95 171
292 292 292
NR NR 21.7 ( 111.1, 70.9)
CI: confidence interval; VE: vaccine effectiveness; Vac: vaccinated; Unvac: unvaccinated; NR not reported. a Frequencies were calculated substituting the median vaccination date where unknown and by excluding those with unknown risk status. b Adjusted for age-group, month, region, IMD and risk group.
Please cite this article as: R. G. Pebody, H. Zhao, H. J. Whitaker et al., Effectiveness of influenza vaccine in children in preventing influenza associated hospitalisation, 2018/19, England, Vaccine, https://doi.org/10.1016/j.vaccine.2019.10.035
6
R.G. Pebody et al. / Vaccine xxx (xxxx) xxx
confirmed hospitalisation during a season with co-circulation of A (H1N1)pdm09 and A(H3N2). The aVE figures we describe are similar to those reported elsewhere in earlier seasons [12,13], and provide on-going encouragement that influenza vaccination of children provides important direct protection against severe disease. The level of aVE was high against A(H1N1)pdm09, though we found no significant protection against A(H3N2). This apparent lower VE against A(H3N2) was also seen in 2017/18 and 2018/19 for both QIV and LAIV for laboratory confirmed infection in primary care [14,15]. It has been hypothesised that this may be related to egg adaption of A(H3N2) vaccine viruses during the vaccine production process [19]. We report that the overall effectiveness of LAIV and QIV in preventing influenza confirmed hospitalisation were similar, though there were some potential differences by influenza A subtype, with a higher, non-significant, effectiveness of LAIV against A(H1N1) pdm09 compared to QIV. It is important to note though that those children receiving QIV will be those contraindicated LAIV for medical reasons – including severe asthma or immunosuppression. These findings are still encouraging and support the update in the vaccine composition of A(H1N1)pdm09 from A/Bolivia/559/2013 in 2016/17 to A/Slovenia/2903/2015 in 2017/18. This latter strain was intended to replicate more efficiently compared to the antecedent strain – and our findings of likely effectiveness are supported by early season aVE reports from elsewhere in the Northern hemisphere, such as Hong Kong [16]. In 2019/20, the A (H1N1)dm09 vaccine strain will be updated again to a A/Brisbane/02/2018 (H1N1)pdm09-like virus. The reasons for the apparent divergent A(H1N1)pdm09 findings in 2015/16 between the United States and elsewhere, including the UK, remain unclear, though early life priming with inactivated influenza vaccine in the USA, could be one hypothesis [20]. It will be important to ensure vaccine effectiveness surveillance continues to provide assurance of on-going clinical protection. We found no evidence of a significant difference in effectiveness by age-group. However, there was a suggestion of higher aVE for healthy children compared to those with an underlying clinical risk factor. Lower VE seemed particularly apparent for A(H1N1)pdm09 infection in at-risk children. Our results of significant VE against influenza in secondary care are supported by similar observations in primary care [17] and emphasises the important protection that is provided by influenza vaccine for healthy children, but that vaccine protection may be sub-optimal for groups who are at elevated risk of poor outcome following infection. It is not clear what is driving this, for example LAIV induces immunity mainly in the upper-respiratory tract, which may not be as protective against infection in the lower respiratory tract. Further work is required to understand the specific risk factors for vaccine failure in these groups. Finally, our findings continue to support the need for annual revaccination, with no evidence of residual protection against hospitalisation from prior season vaccination or that repeat vaccination had impaired vaccine induced protection the following season in 2018/19 [18]. Some investigators have suggested that repeat vaccination may impair vaccine related immunity, particularly when A(H3N2) is circulating. Further epidemiological and immunological studies over several seasons are required to better understand these phenomena and their policy implications. In summary, our study provides evidence of important protection that annual influenza vaccination provides against hospitalisation in eligible children. There is strong evidence that protection remains good against A(H1N1)pdm09, with reduced effectiveness against A(H3N2) for both live attenuated and inactivated vaccines.
Declaration of Competing Interest The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: MD received lecturing fee from Sanofi Pasteur MSD; SpeeDx provided partial financial support for an educational meeting and UK Clinical Virology Network (UK CVN) which he chairs is a registered charity which includes a number of commercial partners. No other co-authors had conflicts to declare.
Appendix A. Supplementary material Supplementary data to this article can be found online at https://doi.org/10.1016/j.vaccine.2019.10.035.
References [1] Joint Committee on Vaccination and Immunisation. Meeting minute.5 Oct 2011. London. Available from: http://webarchive.nationalarchives.gov.uk/ 20120907090205/http://www.dh.gov.uk/prod_consum_dh/groups/dh_ digitalassets/@dh/@ab/documents/digitalasset/dh_133598.pdf. [2] The national flu immunisation programme 2018/19 https://assets.publishing. service.gov.uk/government/uploads/system/uploads/attachment_data/file/ 694779/Annual_national_flu_programme_2018-2019.pdf. [3] Chung JR, Flannery B, Thompson MG, Gaglani M, Jackson ML, Monto AS, et al. Seasonal effectiveness of live attenuated and inactivated influenza vaccine. Pediatrics 2016;137(2):e20153279. [4] Pebody R, McMenamin J, Nohynek H. Live attenuated influenza vaccine (LAIV): recent effectiveness results from the USA and implications for LAIV programmes elsewhere. Arch Dis Child 2018;103(1):101–5. https://doi.org/ 10.1136/archdischild-2016-312165. [5] ACIP votes down use of LAIV for 2016-2017 flu season http://www. cdc.gov/media/releases/2016/s0622-laiv-flu.html. [6] Penttinen PM, Friede MH. Decreased effectiveness of the influenza A(H1N1) pdm09 strain in live attenuated influenza vaccines: an observational bias or a technical challenge?. Euro Surveill 2016;22(21):38. https://doi.org/10.2807/ 1560-7917.ES.2016.21.38.30350. [7] Belongia EA, McLean HQ. Influenza vaccine effectiveness: defining the H3N2 problem. Clin Infect Dis 2019. https://doi.org/10.1093/cid/ciz411. [8] Pebody R, Sile B, Warburton F, Sinnathamby M, Tsang C, Zhao H, Ellis J, Andrews N. Live attenuated influenza vaccine effectiveness against hospitalisation due to laboratory-confirmed influenza in children two to six years of age in England in the 2015/16 season. Euro Surveill 2017;22(4):30450. https://doi.org/10.2807/1560-7917.ES.2017.22.4.30450. [9] Jefferson T, Rivetti A, Di Pietrantonj C, Demicheli V. Vaccines for preventing influenza in healthy children. Cochrane Database Syst Rev 2018(2). CD004879. [10] Surveillance of influenza and other respiratory viruses in the UK Winter 2018 to 2019 https://assets.publishing.service.gov.uk/government/uploads/system/ uploads/attachment_data/file/807472/Surveillance_of_influenza_and_other_ respiratory_viruses_in_the_UK_2018_to_2019-FINAL.pdf. [11] Zhao H, Green HK, Lackenby A, Donati M, Ellis J, Thompson C, et al. A new laboratory-based surveillance system (Respiratory DataMart System) for influenza and other respiratory viruses in England: results and experience from 2009-2012. Euro Surveill 2014;19(3):20680. [12] El Omeiri N, Azziz-Baumgartner E, Thompson MG. REVELAC-i network participants, Clará W, Cerpa M, Palekar R, Mirza S, Ropero-Álvarez AM. Seasonal influenza vaccine effectiveness against laboratory-confirmed influenza hospitalizations - Latin America, 2013. Vaccine 2018;36 (24):3555–66. [13] Buchan SA, Chung H, Campitelli MA, Crowcroft NS, Gubbay JB, Karnauchow T, et al. Vaccine effectiveness against laboratory-confirmed influenza hospitalizations among young children during the 2010–11 to 2013–14 influenza seasons in Ontario, Canada. PLoS One 2017;12(11):e0187834. [14] PHE Surveillance of influenza and other respiratory viruses in the UK: winter 2017 to 2018. https://assets.publishing.service.gov.uk/government/uploads/ system/uploads/attachment_data/file/740606/Surveillance_of_influenza_and_ other_respiratory_viruses_in_the_UK_2017_to_2018.pdf. [15] PHE surveillance of influenza and other respiratory viruses in the UK: winter 2018 to 2019. https://assets.publishing.service.gov.uk/government/uploads/ system/uploads/attachment_data/file/807472/Surveillance_of_influenza_and_ other_respiratory_viruses_in_the_UK_2018_to_2019-FINAL.pdf. [16] Chiu SS, Kwan MY, Feng S, Chan EL, Chua H, Wong JS, et al. Early season estimate of influenza vaccination effectiveness against influenza hospitalisation in children, Hong Kong, winter influenza season 2018/19. Euro Surveill 2019;24(5). https://doi.org/10.2807/1560-7917. [17] Shang M, Chung JR, Jackson ML, Jackson LA, Monto AS, Martin ET, et al. Influenza vaccine effectiveness among patients with high-risk medical
Please cite this article as: R. G. Pebody, H. Zhao, H. J. Whitaker et al., Effectiveness of influenza vaccine in children in preventing influenza associated hospitalisation, 2018/19, England, Vaccine, https://doi.org/10.1016/j.vaccine.2019.10.035
R.G. Pebody et al. / Vaccine xxx (xxxx) xxx conditions in the United States, 2012–2016. Vaccine 2018;36(52):8047–53. https://doi.org/10.1016/j.vaccine. [18] Belongia EA, Skowronski DM, McLean HQ, Chambers C, Sundaram ME, De Serres G. Repeated annual influenza vaccination and vaccine effectiveness: review of evidence. Expert Rev Vaccines 2017;16(7):1–14. Epub 2017 Jun 9. Review.
7
[19] https://www.pnas.org/content/114/47/12578.long. [20] Ilyushina NA, Haynes BC, Hoen AG, Khalenkov AM, Housman ML, Brown EP, et al. Live attenuated and inactivated influenza vaccines in children. J Infect Dis 2015;211(3):352–60.
Please cite this article as: R. G. Pebody, H. Zhao, H. J. Whitaker et al., Effectiveness of influenza vaccine in children in preventing influenza associated hospitalisation, 2018/19, England, Vaccine, https://doi.org/10.1016/j.vaccine.2019.10.035
Contents lists available at ScienceDirect
Vaccine journal homepage: www.elsevier.com/locate/vaccine
Effectiveness of influenza vaccine in children in preventing influenza associated hospitalisation, 2018/19, England R.G. Pebody a,⇑, H. Zhao a, H.J. Whitaker a, J. Ellis a, M. Donati b, M. Zambon a, N. Andrews a a b
Public Health England National Infection Service, Colindale, London, UK Public Health England National Infection Service, Bristol, UK
a r t i c l e
i n f o
Article history: Received 9 August 2019 Received in revised form 12 October 2019 Accepted 14 October 2019 Available online xxxx Keywords: Influenza vaccine effectiveness Children LAIV Hospitalisation
a b s t r a c t 2013/14 saw the start of the introduction of a new live attenuated influenza vaccine (LAIV) programme for children in England. 2018/19 saw co-circulation of both A(H1N1)pdm09 and A(H3N2), when LAIV was offered to all healthy children 2–9 years of age. LAIV effectiveness against influenza hospitalisation is not well described. This paper presents the 2018/19 end-of-season adjusted vaccine effectiveness (aVE) against laboratory confirmed influenza related hospitalisation in children aged 2–17. The test negative case control approach was used to estimate aVE by influenza A subtype and vaccine type. Cases and controls were selected from a sentinel laboratory surveillance system which collates details of individuals tested for influenza with reverse-transcription polymerase chain reaction (RT-PCR) on respiratory samples. Vaccine and clinical history was obtained from general practitioners of study participants. There were 307 hospitalised cases and 679 hospitalised controls. End-of-season influenza aVE was 53.0% (95% CI: 33.3, 66.8) against influenza confirmed hospitalisation; 63.5% (95% CI: 34.4, 79.7) against influenza A(H1N1)pdm09 hospitalisation and 31.1% (95% CI: 53.9, 69.2) against influenza A(H3N2). LAIV aVE was 49.1% (95% CI: 25.9, 65.0) for any influenza and 70.7% (95% CI: 41.8, 85.3) for A(H1N1)pdm09, whereas for those receiving quadrivalent inactivated influenza vaccine (QIV), aVE was 64.4% (95% CI: 29.4, 82.0) and 44.4% (95% CI: 51.9, 79.6) respectively. We provide evidence of overall significant VE for both LAIV and QIV against influenza associated hospitalisation in children 2–17 years of age, most notably against influenza A(H1N1)pdm09, with non-significant protection against A(H3N2). Crown Copyright Ó 2019 Published by Elsevier Ltd. All rights reserved.
1. Introduction The United Kingdom (UK) started the incremental introduction of a universal paediatric influenza vaccination programme in the 2013/14 influenza flu season with a newly licensed live attenuated influenza vaccine (LAIV). The aim is to ultimately offer annual influenza vaccination to all children 2–16 years of age to both directly protect them, and by reducing their risk of infection, indirectly protect others in the community who may be at higher risk of severe disease [1]. The 2018/2019 influenza season in the UK was the sixth consecutive season of roll-out of LAIV to healthy children, with the offer of LAIV vaccination extended in England to all healthy children aged two to 9 years of age [2]. Children contraindicated LAIV are offered quadrivalent inactivated vaccine (QIV). In 2015/16, following circulation of mainly influenza A(H1N1) pdm09 in the Northern hemisphere, the Centres for Disease ⇑ Corresponding author. E-mail address: [email protected] (R.G. Pebody).
Control (CDC) of the United States reported evidence of poor effectiveness of LAIV in protecting children against influenza [3]. Although other countries, including the UK, published evidence that LAIV vaccine had been effective in protecting children against laboratory confirmed infection in primary care [4], the American Committee on Immunisation Practice (ACIP) temporarily withdrew their recommendation to use LAIV [5]. The reasons for the discrepancy in VE between the US and elsewhere remains unclear, with several hypotheses proposed, though lower effectiveness of LAIV against the (H1N1)pdm09 strain compared to the inactivated vaccine was seen in all settings. Reduced replicative fitness of the A/Bolivia/559/2013 (H1N1)pdm09 vaccine strain contained in LAIV at that time was the leading theory, together with prior LAIV vaccination cited as a potential contributory factor and/or early life exposure to inactivated influenza vaccine which is recommended for children 6 months to 2 years of age [6]. The A/Bolivia strain was updated for the 2017/18 season to the A/Slovenia/2903/2015 (H1N1)pdm09 vaccine strain [4]. Finally, in recent seasons, reduced A(H3N2) specific VE has been observed in both adults and children vaccinated with egg-grown influenza vaccines, both
https://doi.org/10.1016/j.vaccine.2019.10.035 0264-410X/Crown Copyright Ó 2019 Published by Elsevier Ltd. All rights reserved.
Please cite this article as: R. G. Pebody, H. Zhao, H. J. Whitaker et al., Effectiveness of influenza vaccine in children in preventing influenza associated hospitalisation, 2018/19, England, Vaccine, https://doi.org/10.1016/j.vaccine.2019.10.035
2
R.G. Pebody et al. / Vaccine xxx (xxxx) xxx
inactivated and live attenuated. It has been suggested this poor VE may be related to egg adaption of A(H3N2) vaccine viruses [7]. Only limited studies have been published on the recent effectiveness of influenza vaccine against hospitalisation in children [8,9]. The 2018/19 season in England was characterised by cocirculation of both A(H1N1)pdm09 and A(H3N2)[10]. This paper presents the results of an evaluation of influenza vaccine effectiveness in children of 2–17 years in England in protecting against laboratory confirmed infection resulting in hospitalisation during the 2018/19 season.
2. Methods A frequency matched test-negative case-control study was undertaken using the Respiratory DataMart Surveillance system (RDS) according to a pre-defined protocol. Cases and controls were identified from RDS. This is a national sentinel laboratory surveillance system which records details of individuals tested for influenza infection with real-time reverse transcriptase polymerase chain reaction (RT-PCR) on respiratory samples (including both positive and negative results) from 13 NHS and PHE laboratories located across England [11]. These laboratories routinely test for influenza A and B, though many do not undertake influenza A sub-typing. The study population was defined as residents in England 2– 17 years of age (at August 31st 2018) who were admitted to hospital and who had a respiratory swab taken between week 40 2018 and week 20 2019 which was tested for influenza with RT-PCR by one of the RDS laboratories. Those who were swabbed more than seven days after onset were excluded (when onset was known). A case was defined as a child with RT-PCR confirmed influenza infection. A control was defined as a child with RT-PCR influenza negative infection. Controls were group-matched to cases by age group (2–4, 5–8, 9–11, 12–17) and week of sample (+/ 1 week) with up to 3 controls randomly selected per case within these groups. Details of cases and controls from RDS were used to identify the general practitioners of these children, using the patient demographic service (PDS) system. PDS provides the postcode of the child and their GP address. The residential postcode was linked to the Index of Multiple Deprivation (IMD) to identify which IMD quintile the child is resident in and their PHE region. A child who was not identifiable by the national PDS system as being registered with a GP or not resident in England was excluded. The project team sent a short questionnaire to the identified GP to collect key demographic and clinical information on the child including date of onset of illness; fact and date of hospitalisation; underlying clinical risk factor; influenza vaccine history including date and type of vaccination. Statistical analysis was undertaken in Stata 15 (StataCorp, College Station, TX, USA). A descriptive and analytical analysis was undertaken both overall and by age-group (2–8 years and 9– 17 years). For analysis of the case-control study, logistic regression was used to calculate the unadjusted odds ratios for influenza vaccination in cases compared to controls, with a 95% confidence interval, as unadjusted VE = 1-OR. Logistic regression was also used to calculate the odds ratio for vaccination, adjusted for relevant characteristics which changed the adjusted vaccine effectiveness (aVE) against influenza by 3% or more (these could include sex, region, risk group and deprivation). Analysis was undertaken stratified, where numbers were sufficient, by key factors including age group (2–9 years and 10– 17 years); underlying clinical risk factor; type of vaccine (intranasal and inactivated); influenza type and subtype and prior season’s vaccination. Sensitivity analyses were undertaken - specifically
including only swabs where onset date was present; including those with onsets >7 days; removing those with unknown vaccination date and using multiple imputation for missing risk-group information. A multiple imputation approach was used for missing vaccination dates by sampling known vaccination dates from individuals whose event occurred in the same week. VE estimates are reported when there is sufficient precision, based on requiring an expected number of vaccinated cases of at least 8 under the null hypothesis of zero VE. This is calculated as either Tc * proportion of controls vaccinated or, when using an adjusted VE as Tc * Oc/[(1-aVE) + Oc], where Tc is total cases for the VE estimate, Oc is the observed odds of vaccination in the cases and aVE the adjusted VE estimate. Public Health England (PHE) holds permissions under section 251 of the 2006 NHS Act and the 2002 Health Service (Control of Patient Information) Regulations, to process patient information relating to this investigation. 3. Results In all, 2685 questionnaires were sent to GPs of which 713 were for cases and 1972 were for controls. Replies were received for 2089 patients (78%) aged 2–17 years of age. Of these 379 were excluded: nine were not resident in England; one had unknown postcode; 111 had unknown 2018/19 vaccination status; 219 had more than 7 days from disease onset to sampling and 39 were vaccinated within 14 days of onset. After exclusions, 482 cases and 1228 controls were left of which 307 cases and 679 controls were hospitalised (Fig. 1). The description of hospitalised cases and controls by age-group, sex, risk-group and vaccine status are provided in Table 1. Of the 307 hospitalised cases, 104 were due to A(H1N1)pdm09, 40 were due to A(H3N2), 156 were influenza A (subtype not tested) and six were due to influenza B. There was one A and B co-infection. 65% (443) of the hospitalised controls had no other respiratory virus detected, with rhinovirus followed by RSV the main viruses detected (in 64 and 49 controls respectively). 3.1. Model fitting for vaccine effectiveness estimates To estimate VE, age-group, month, PHE region, deprivation (IMD score) and risk group were adjusted for in the multivariable model as they changed the overall estimates by more than 3%. Gender was not included as it changed the vaccine effect by less than 1%. The crude VE and aVE against hospitalisation are shown in Table 2, with an aVE against all influenza, influenza A, influenza A(H1N1)pdm09 and influenza A(H3N2) of 53.0% (95% CI 33.3, 66.8); 55.2% (95% CI: 36.2, 68.5); 63.5% (95% CI: 34.4, 79.7) and 31.1% (95% CI: 53.9, 69.2) respectively. In addition, influenza A (not subtyped) was 51.4% (95% CI: 32.1, 65.2). There were inadequate samples to estimate influenza B VE. The aVE estimates for influenza by type/sub-type are shown by age-group in Table 3. There were no significant differences in aVE between 2–9 and 10–17 year olds with similar aVE point estimates and overlapping 95% confidence intervals in estimates for all influenza and influenza A (Table 3). By risk-group, the aVE point estimates were higher for healthy children compared to those with an underlying clinical risk factor, albeit with overlapping 95% confidence intervals for all influenza and restricted to influenza A (Table 3). For (H1N1)pdm09, the aVE was 81.6% (95% CI 57.3, 92.1) for healthy children compared to 4.8% (95%CI: 176.8, 60.3) for children with an underlying clinical risk factor. By vaccine type, the aVE point estimates were broadly similar with overlapping 95% confidence intervals when comparing LAIV and QIV for
Please cite this article as: R. G. Pebody, H. Zhao, H. J. Whitaker et al., Effectiveness of influenza vaccine in children in preventing influenza associated hospitalisation, 2018/19, England, Vaccine, https://doi.org/10.1016/j.vaccine.2019.10.035
R.G. Pebody et al. / Vaccine xxx (xxxx) xxx
3
Fig. 1. Swabbing results of patients in secondary care in England, October 2018 to April 2019.
all influenza and restricting to influenza A. For A(H1N1)pdm09, LAIV aVE was non-significantly higher compared to QIV. The aVE estimates by prior vaccine history are shown in Table 4. The lowest aVE was seen in those who were vaccinated only in 2017/18 for all influenza and influenza A(H1N1)pdm09. Significant aVE was seen for those vaccinated in 2018/19 regardless of prior vaccine status in 2017/18 for all influenza and influenza A(H1N1) pdm09. Non-significant results were seen for A(H3N2), with 95% confidence intervals spanning zero. Sensitivity analyses found similar estimates if restricted to only patients where onset date was present; removing swabs with unknown vaccination date; including onsets >7 days and imputing missing risk-group information. 4. Discussion In summary, we report that the overall effectiveness of influenza vaccine in children 2–17 years of age in preventing influenza
confirmed hospitalisation was moderate. Effectiveness was high against A(H1N1)pdm09 hospitalisation, but no significant protection was found against A(H3N2). There was no significant difference in effectiveness by age-group or risk-group, except a higher aVE against A(H1N1)pdm09 in healthy children compared to those with an underlying clinical risk factor. Overall effectiveness of LAIV and QIV were similar, though there were some potential differences by influenza A subtype. There was no evidence of residual protection against hospitalisation from vaccination only in the prior season. This study has a number of strengths and weaknesses. Although children without a registered GP have been excluded, as it will not be possible to collect relevant information, only a small number fall into this group. Laboratory testing for influenza infection in the age group studied tends to occur mainly among those presenting to secondary care. As the study is restricted to those children who have been hospitalised, this will have a limited effect on the estimate of VE as cases and (test-negative) controls are likely to
Please cite this article as: R. G. Pebody, H. Zhao, H. J. Whitaker et al., Effectiveness of influenza vaccine in children in preventing influenza associated hospitalisation, 2018/19, England, Vaccine, https://doi.org/10.1016/j.vaccine.2019.10.035
4
R.G. Pebody et al. / Vaccine xxx (xxxx) xxx
Table 1 Details for influenza A and B hospitalised cases and controls, United Kingdom, week 40 2018–week 20 2019 Numbers and row percentages (to indicate positivity ratesa) are shown. Χ2 p-value
Negative
A/H1
A/H3
A (unknown)
B
Age (years) 2–4 5–9 10–17
344 (68%) 151 (70%) 184 (70%)
58 (11%) 19 (9%) 27 (10%)
17 (3%) 6 (3%) 17 (6%)
86 (17%) 37 (17%) 33 (13%)
3 (1%) 2 (1%) 1 (0%)
Sex Female Male
324 (71%) 355 (67%)
46 (10%) 58 (11%)
18 (4%) 22 (4%)
69 (15%) 87 (17%)
2 (0%) 4 (1%)
Risk group No Yes Missing
299 (60%) 333 (81%) 47 (64%)
71 (14%) 23 (6%) 10 (14%)
24 (5%) 14 (3%) 2 (3%)
105 (21%) 39 (9%) 12 (16%)
2 (0%) 2 (0%) 2 (3%)
Onset to swab (days) 0–1 2–4 5–7 missing
155 (61%) 139 (63%) 82 (62%) 303 (40%)
25 31 23 25
(10%) (14%) (17%) (3%)
17 (7%) 12 (5%) 4 (3%) 7 (1%)
56 38 24 38
2 1 0 3
Vaccination Status Unvaccinated at onset Vaccinated >14 days before Vaccination date missing
390 (64%) 270 (78%) 19 (73%)
78 (13%) 26 (7%) 0 (0%)
28 (5%) 11 (3%) 1 (4%)
112 (18%) 40 (11%) 4 (15%)
4 (1%) 1 (0%) 1 (4%)
Previous 2017/18 vaccination No Yes Unknown
418 (64%) 239 (80%) 22 (73%)
84 (13%) 16 (5%) 4 (13%)
26 (4%) 12 (4%) 2 (7%)
121 (18%) 33 (11%) 2 (7%)
6 (1%) 0 (0%) 0 (0%)
Month of event October November December January February March
10 (67%) 57 (80%) 107 (82%) 237 (62%) 260 (68%) 8 (100%)
0 (0%) 5 (7%) 13 (10%) 53 (14%) 33 (9%) 0 (0%)
1 (7%) 0 (0%) 5 (4%) 12 (3%) 22 (6%) 0 (0%)
1 (7%) 7 (10%) 3 (2%) 79 (21%) 66 (17%) 0 (0%)
3 1 2 0 0 0
(20%) (1%) (2%) (0%) (0%) (0%)
IMD (deprivation index) 1 2 3 4 5
192 132 107 126 122
24 19 18 22 21
(9%) (10%) (11%) (13%) (11%)
13 (5%) 5 (3%) 7 (4%) 3 (2%) 12 (7%)
51 30 23 25 27
2 2 1 0 1
(1%) (1%) (1%) (0%) (1%)
Vaccine type Unvaccinated LAIV IIV Vaccinated (unknown)
391 (64%) 194 (74%) 87 (87%) 7 (78%)
78 (13%) 19 (7%) 6 (6%) 1 (11%)
28 (5%) 10 (4%) 2 (2%) 0 (0%)
112 (18%) 38 (15%) 5 (5%) 1 (11%)
5 1 0 0
(1%) (0%) (0%) (0%)
0.723
0.275
<0.001
0.871 (22%) (17%) (18%) (5%)
(1%) (0%) (0%) (50%) P < 0.001
P < 0.001
P < 0.001
P = 0.862 (68%) (70%) (68%) (72%) (67%)
(18%) (16%) (15%) (14%) (15%)
P < 0.001
p-values relate to a Chi-square test on all influenza positives (A and B) and influenza negatives.
Table 2 Samples positive (cases N = 307) and negative (controls N = 679) for influenza A and B according to vaccination status and VE estimates, England, 1 October 2018–8 April 2019. Influenza
A and B A A/H1N1pdm09 A/H3N2
Cases
Controls
Vac
Unvac
Vac
Unvac
83 82 26 12
224 219 78 28
288 288 288 288
391 391 391 391
Crude VE (95%CI)
Adjusteda VE (95% CI)
48.7 49.0 54.7 41.8
53.0 55.2 63.5 31.1
(31.2, 61.8) (31.4, 62.1) (27.6, 71.7) ( 16.4, 70.9)
(33.3, 66.8) (36.2, 68.5) (34.4, 79.7) ( 53.9, 69.2)
CI: confidence interval; VE: vaccine effectiveness; Vac: vaccinated; Unvac: unvaccinated. a Adjusted for age-group, month, region, IMD and risk group.
have similar severity of illness in order to be tested and hospitalised. A relatively large proportion of influenza A detections remain unsubtyped, which reflects local NHS laboratory practice in many instances. Clinicians will not know whether a person admitted to hospital with acute respiratory illness has H1N1pdm09 or H3N2, and so likelihood of sampling will be similar. Flu A sub-typing is undertaken locally regardless of factors such as vaccine history and dependent upon test platforms in use in the laboratory. We have also looked at aVE for flu A unsubtyped,
with the estimate lying, as anticipated between that seen for H3N2 and H1N1pdm09. Finally, although primary care is recommended to keep vaccine history up to date for primary school agechildren, it is possible that vaccine history may not have been captured on some children vaccinated in schools. The missing vaccine history will result in non-differential exposure misclassification resulting in bias to the null. We found that influenza vaccine in children 2–17 years of age provided moderate levels of protection in preventing influenza
Please cite this article as: R. G. Pebody, H. Zhao, H. J. Whitaker et al., Effectiveness of influenza vaccine in children in preventing influenza associated hospitalisation, 2018/19, England, Vaccine, https://doi.org/10.1016/j.vaccine.2019.10.035
5
R.G. Pebody et al. / Vaccine xxx (xxxx) xxx Table 3 Adjusted vaccine effectiveness estimates for influenza associated hospitalisation by sub-type, age-group and vaccine type, England, 1 Oct 2018–8 April 2019. Cases
Adjusteda VE
Controls
Vac
Unvac
Vac
Unvac
(95% CI)
A&B By age-group 2–9 10–17
61 11
147 62
212 61
245 114
52.3 (29.4, 67.8) 62.6 (12.9, 84.0)
By risk group None In risk group
41 31
162 47
100 173
199 160
65.4 (43.8, 78.7) 35.9 ( 9.0, 62.3)
By vaccine type LAIV IIV
58 12
209 209
183 83
359 359
49.1 (25.9, 65.0) 64.4 (29.4, 82.0)
A By age-group 2–9 10–17
60 11
145 61
212 61
245 114
55.6 (33.8, 70.2) 61.2 (8.9, 83.5)
By risk group None In risk group
40 31
161 45
100 173
199 160
67.3 (46.5, 80.0) 38.2 ( 5.8, 64.0)
By vaccine type LAIV IIV
57 12
206 206
183 83
359 359
52.2 (30.0, 67.3) 64.2 (28.7, 82.0)
A/(H1N1)pdm09 By age-group 2–9 10–17
19 2
52 21
212 61
245 114
61.4 (26.2, 79.8) NR
By risk group None In risk group
10 11
61 12
100 173
199 160
81.6 (57.3, 92.1) 4.8 ( 176.8, 60.3)
By vaccine type LAIV IIV
14 6
73 73
183 83
359 359
70.7 (41.8, 85.3) 44.4 ( 51.9, 79.6)
A/H3N2 By age-group 2–9 10–17
8 3
13 14
212 61
245 114
14.5 ( 137.1, 69.2) NR
By risk group None In risk group
6 5
18 9
100 173
199 160
NR NR
By vaccine type LAIV IIV
9 2
27 27
183 83
359 359
20.1 ( 87.3, 65.9) NR
CI: confidence interval; VE: vaccine effectiveness; Vac: vaccinated; Unvac: unvaccinated; LAIV – live attenuated vaccine, aTIV – adjuvanted trivalent inactivated vaccine, QIV – quadrivalent inactivated vaccine; NR – not reported. a Adjusted for age-group, month, region, IMD and risk group.
Table 4 Adjusted vaccine effectiveness estimates for influenza hospitalisation by subtype and prior vaccine history, England in 2–17 year olds, 1 Oct 2018–8 April 2019. Casesa
Controlsa
Adjustedb VE
Vac
Unvac
Vac
Unvac
(95% CI)
A and B 2017/18 only 2018/19 only Both seasons
22 35 33
185 185 185
53 95 171
292 292 292
11.3 ( 58.1, 50.3) 44.5 (10.5, 65.5) 64.8 (44.5, 77.7)
A/(H1N1)pdm09 2017/18 only 2018/19 only Both seasons
6 10 9
65 65 65
53 95 171
292 292 292
10.8 ( 138.0, 66.6) 64.9 (20.0, 84.6) 71.0 (35.2, 87.0)
A/H3N2 2017/18 only 2018/19 only Both seasons
5 3 7
22 22 22
53 95 171
292 292 292
NR NR 21.7 ( 111.1, 70.9)
CI: confidence interval; VE: vaccine effectiveness; Vac: vaccinated; Unvac: unvaccinated; NR not reported. a Frequencies were calculated substituting the median vaccination date where unknown and by excluding those with unknown risk status. b Adjusted for age-group, month, region, IMD and risk group.
Please cite this article as: R. G. Pebody, H. Zhao, H. J. Whitaker et al., Effectiveness of influenza vaccine in children in preventing influenza associated hospitalisation, 2018/19, England, Vaccine, https://doi.org/10.1016/j.vaccine.2019.10.035
6
R.G. Pebody et al. / Vaccine xxx (xxxx) xxx
confirmed hospitalisation during a season with co-circulation of A (H1N1)pdm09 and A(H3N2). The aVE figures we describe are similar to those reported elsewhere in earlier seasons [12,13], and provide on-going encouragement that influenza vaccination of children provides important direct protection against severe disease. The level of aVE was high against A(H1N1)pdm09, though we found no significant protection against A(H3N2). This apparent lower VE against A(H3N2) was also seen in 2017/18 and 2018/19 for both QIV and LAIV for laboratory confirmed infection in primary care [14,15]. It has been hypothesised that this may be related to egg adaption of A(H3N2) vaccine viruses during the vaccine production process [19]. We report that the overall effectiveness of LAIV and QIV in preventing influenza confirmed hospitalisation were similar, though there were some potential differences by influenza A subtype, with a higher, non-significant, effectiveness of LAIV against A(H1N1) pdm09 compared to QIV. It is important to note though that those children receiving QIV will be those contraindicated LAIV for medical reasons – including severe asthma or immunosuppression. These findings are still encouraging and support the update in the vaccine composition of A(H1N1)pdm09 from A/Bolivia/559/2013 in 2016/17 to A/Slovenia/2903/2015 in 2017/18. This latter strain was intended to replicate more efficiently compared to the antecedent strain – and our findings of likely effectiveness are supported by early season aVE reports from elsewhere in the Northern hemisphere, such as Hong Kong [16]. In 2019/20, the A (H1N1)dm09 vaccine strain will be updated again to a A/Brisbane/02/2018 (H1N1)pdm09-like virus. The reasons for the apparent divergent A(H1N1)pdm09 findings in 2015/16 between the United States and elsewhere, including the UK, remain unclear, though early life priming with inactivated influenza vaccine in the USA, could be one hypothesis [20]. It will be important to ensure vaccine effectiveness surveillance continues to provide assurance of on-going clinical protection. We found no evidence of a significant difference in effectiveness by age-group. However, there was a suggestion of higher aVE for healthy children compared to those with an underlying clinical risk factor. Lower VE seemed particularly apparent for A(H1N1)pdm09 infection in at-risk children. Our results of significant VE against influenza in secondary care are supported by similar observations in primary care [17] and emphasises the important protection that is provided by influenza vaccine for healthy children, but that vaccine protection may be sub-optimal for groups who are at elevated risk of poor outcome following infection. It is not clear what is driving this, for example LAIV induces immunity mainly in the upper-respiratory tract, which may not be as protective against infection in the lower respiratory tract. Further work is required to understand the specific risk factors for vaccine failure in these groups. Finally, our findings continue to support the need for annual revaccination, with no evidence of residual protection against hospitalisation from prior season vaccination or that repeat vaccination had impaired vaccine induced protection the following season in 2018/19 [18]. Some investigators have suggested that repeat vaccination may impair vaccine related immunity, particularly when A(H3N2) is circulating. Further epidemiological and immunological studies over several seasons are required to better understand these phenomena and their policy implications. In summary, our study provides evidence of important protection that annual influenza vaccination provides against hospitalisation in eligible children. There is strong evidence that protection remains good against A(H1N1)pdm09, with reduced effectiveness against A(H3N2) for both live attenuated and inactivated vaccines.
Declaration of Competing Interest The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: MD received lecturing fee from Sanofi Pasteur MSD; SpeeDx provided partial financial support for an educational meeting and UK Clinical Virology Network (UK CVN) which he chairs is a registered charity which includes a number of commercial partners. No other co-authors had conflicts to declare.
Appendix A. Supplementary material Supplementary data to this article can be found online at https://doi.org/10.1016/j.vaccine.2019.10.035.
References [1] Joint Committee on Vaccination and Immunisation. Meeting minute.5 Oct 2011. London. Available from: http://webarchive.nationalarchives.gov.uk/ 20120907090205/http://www.dh.gov.uk/prod_consum_dh/groups/dh_ digitalassets/@dh/@ab/documents/digitalasset/dh_133598.pdf. [2] The national flu immunisation programme 2018/19 https://assets.publishing. service.gov.uk/government/uploads/system/uploads/attachment_data/file/ 694779/Annual_national_flu_programme_2018-2019.pdf. [3] Chung JR, Flannery B, Thompson MG, Gaglani M, Jackson ML, Monto AS, et al. Seasonal effectiveness of live attenuated and inactivated influenza vaccine. Pediatrics 2016;137(2):e20153279. [4] Pebody R, McMenamin J, Nohynek H. Live attenuated influenza vaccine (LAIV): recent effectiveness results from the USA and implications for LAIV programmes elsewhere. Arch Dis Child 2018;103(1):101–5. https://doi.org/ 10.1136/archdischild-2016-312165. [5] ACIP votes down use of LAIV for 2016-2017 flu season http://www. cdc.gov/media/releases/2016/s0622-laiv-flu.html. [6] Penttinen PM, Friede MH. Decreased effectiveness of the influenza A(H1N1) pdm09 strain in live attenuated influenza vaccines: an observational bias or a technical challenge?. Euro Surveill 2016;22(21):38. https://doi.org/10.2807/ 1560-7917.ES.2016.21.38.30350. [7] Belongia EA, McLean HQ. Influenza vaccine effectiveness: defining the H3N2 problem. Clin Infect Dis 2019. https://doi.org/10.1093/cid/ciz411. [8] Pebody R, Sile B, Warburton F, Sinnathamby M, Tsang C, Zhao H, Ellis J, Andrews N. Live attenuated influenza vaccine effectiveness against hospitalisation due to laboratory-confirmed influenza in children two to six years of age in England in the 2015/16 season. Euro Surveill 2017;22(4):30450. https://doi.org/10.2807/1560-7917.ES.2017.22.4.30450. [9] Jefferson T, Rivetti A, Di Pietrantonj C, Demicheli V. Vaccines for preventing influenza in healthy children. Cochrane Database Syst Rev 2018(2). CD004879. [10] Surveillance of influenza and other respiratory viruses in the UK Winter 2018 to 2019 https://assets.publishing.service.gov.uk/government/uploads/system/ uploads/attachment_data/file/807472/Surveillance_of_influenza_and_other_ respiratory_viruses_in_the_UK_2018_to_2019-FINAL.pdf. [11] Zhao H, Green HK, Lackenby A, Donati M, Ellis J, Thompson C, et al. A new laboratory-based surveillance system (Respiratory DataMart System) for influenza and other respiratory viruses in England: results and experience from 2009-2012. Euro Surveill 2014;19(3):20680. [12] El Omeiri N, Azziz-Baumgartner E, Thompson MG. REVELAC-i network participants, Clará W, Cerpa M, Palekar R, Mirza S, Ropero-Álvarez AM. Seasonal influenza vaccine effectiveness against laboratory-confirmed influenza hospitalizations - Latin America, 2013. Vaccine 2018;36 (24):3555–66. [13] Buchan SA, Chung H, Campitelli MA, Crowcroft NS, Gubbay JB, Karnauchow T, et al. Vaccine effectiveness against laboratory-confirmed influenza hospitalizations among young children during the 2010–11 to 2013–14 influenza seasons in Ontario, Canada. PLoS One 2017;12(11):e0187834. [14] PHE Surveillance of influenza and other respiratory viruses in the UK: winter 2017 to 2018. https://assets.publishing.service.gov.uk/government/uploads/ system/uploads/attachment_data/file/740606/Surveillance_of_influenza_and_ other_respiratory_viruses_in_the_UK_2017_to_2018.pdf. [15] PHE surveillance of influenza and other respiratory viruses in the UK: winter 2018 to 2019. https://assets.publishing.service.gov.uk/government/uploads/ system/uploads/attachment_data/file/807472/Surveillance_of_influenza_and_ other_respiratory_viruses_in_the_UK_2018_to_2019-FINAL.pdf. [16] Chiu SS, Kwan MY, Feng S, Chan EL, Chua H, Wong JS, et al. Early season estimate of influenza vaccination effectiveness against influenza hospitalisation in children, Hong Kong, winter influenza season 2018/19. Euro Surveill 2019;24(5). https://doi.org/10.2807/1560-7917. [17] Shang M, Chung JR, Jackson ML, Jackson LA, Monto AS, Martin ET, et al. Influenza vaccine effectiveness among patients with high-risk medical
Please cite this article as: R. G. Pebody, H. Zhao, H. J. Whitaker et al., Effectiveness of influenza vaccine in children in preventing influenza associated hospitalisation, 2018/19, England, Vaccine, https://doi.org/10.1016/j.vaccine.2019.10.035
R.G. Pebody et al. / Vaccine xxx (xxxx) xxx conditions in the United States, 2012–2016. Vaccine 2018;36(52):8047–53. https://doi.org/10.1016/j.vaccine. [18] Belongia EA, Skowronski DM, McLean HQ, Chambers C, Sundaram ME, De Serres G. Repeated annual influenza vaccination and vaccine effectiveness: review of evidence. Expert Rev Vaccines 2017;16(7):1–14. Epub 2017 Jun 9. Review.
7
[19] https://www.pnas.org/content/114/47/12578.long. [20] Ilyushina NA, Haynes BC, Hoen AG, Khalenkov AM, Housman ML, Brown EP, et al. Live attenuated and inactivated influenza vaccines in children. J Infect Dis 2015;211(3):352–60.
Please cite this article as: R. G. Pebody, H. Zhao, H. J. Whitaker et al., Effectiveness of influenza vaccine in children in preventing influenza associated hospitalisation, 2018/19, England, Vaccine, https://doi.org/10.1016/j.vaccine.2019.10.035