Effect of Vaccination on Preventing Influenza-Associated Hospitalizations Among Children During a Severe Season Associated With B/Victoria Viruses, 2019–2020

Abstract

Background

The 2019–2020 influenza season was characterized by early onset with B/Victoria followed by A(H1N1)pdm09 viruses. Emergence of new B/Victoria viruses raised concerns about possible vaccine mismatch. We estimated vaccine effectiveness (VE) against influenza-associated hospitalizations and emergency department (ED) visits among children in the United States.

Methods

We assessed VE among children aged 6 months–17 years with acute respiratory illness and ≤10 days of symptoms enrolled at 7 pediatric medical centers in the New Vaccine Surveillance Network. Combined midturbinate/throat swabs were tested for influenza virus using molecular assays. Vaccination history was collected from parental report, state immunization information systems, and/or provider records. We estimated VE from a test-negative design using logistic regression to compare odds of vaccination among children testing positive vs negative for influenza.

Results

Among 2029 inpatients, 335 (17%) were influenza positive: 37% with influenza B/Victoria alone and 44% with influenza A(H1N1)pdm09 alone. VE was 62% (95% confidence interval [CI], 52%–71%) for influenza-related hospitalizations, 54% (95% CI, 33%–69%) for B/Victoria viruses, and 64% (95% CI, 49%–75%) for A(H1N1)pdm09. Among 2102 ED patients, 671 (32%) were influenza positive: 47% with influenza B/Victoria alone and 42% with influenza A(H1N1)pdm09 alone. VE was 56% (95% CI, 46%–65%) for an influenza-related ED visit, 55% (95% CI, 40%–66%) for B/Victoria viruses, and 53% (95% CI, 37%–65%) for A(H1N1)pdm09.

Conclusions

Influenza vaccination provided significant protection against laboratory-confirmed influenza-associated hospitalizations and ED visits associated with the 2 predominantly circulating influenza viruses among children, including against the emerging B/Victoria virus subclade.

influenza, vaccination, test-negative design, children

The 2019–2020 influenza season caused substantial morbidity and mortality among children. The season was characterized by the early onset of influenza B/Victoria viruses followed by influenza A(H1N1)pdm09 virus circulation [1, 2]. Cumulative 2019–2020 United States (US) hospitalization rates among pediatric and young adult populations were the highest reported for the last 10 influenza seasons [3]. One hundred eighty-nine pediatric influenza deaths were reported in the US during 2019–2020, which just exceeded the 188 deaths reported in the 2017–2018 season and represented a peak in a nonpandemic season since reporting began in 2004 [3]. Of those, 116 (62%) were associated with influenza B infections, and B/Victoria was the predominant B lineage detected in the US (>98%) [3].

Over the 2019–2020 season, influenza B/Victoria predominated in the age group 5–24 years, while children 0–4 years of age had nearly equal detection of B/Victoria and A(H1N1)pdm09 influenza viruses [3]. More than 95% of circulating influenza B/Victoria viruses were subclade V1A.3, which differed antigenically and genetically from the Northern Hemisphere influenza B/Victoria vaccine strain V1A.1, indicating that antigenic drift had occurred [1]. Among A(H1N1)pdm09 viruses, the majority belonged to 5A subclades, with increasing numbers of subclade viruses emerging over the season with an N156K amino acid substitution in the hemagglutinin gene, which also differed from the vaccine virus strain (54% of sequenced A(H1N1)pdm09 viruses had this substitution over the entire season) [4]. Circulation of these viruses raised questions about vaccine effectiveness (VE).

Using data from the New Vaccine Surveillance Network (NVSN), we estimated VE among children aged 6 months–17 years who were hospitalized in 7 pediatric medical centers or treated in associated EDs during the 2019–2020 influenza season. To assess protection against circulating seasonal viruses including the new B/Victoria viruses, we evaluated VE against laboratory-confirmed influenza hospitalizations and laboratory-confirmed influenza emergency department (ED) visits by circulating influenza type.

METHODS

Study Population

The NVSN is comprised of 7 sites that perform prospective surveillance for acute respiratory illness among children who are hospitalized or seek care in the ED [5–7]. All sites enrolled hospitalized children <18 years of age. Among ED visits, 5 sites enrolled all children <18 years of age (Rochester, New York; Cincinnati, Ohio; Nashville, Tennessee; Houston, Texas; Kansas City, Missouri), and 2 sites enrolled children primarily <5 years of age (Pittsburgh, Pennsylvania; Seattle, Washington). For our VE analysis, we included participants enrolled during each site-specific influenza season, defined as date of first through last influenza-positive case in enrolled subjects at each site. Eligibility criteria for NVSN enrollment and information regarding data collection through parent/guardian interview with standardized medical record review have been previously published [5, 6].

Sample Collection and Influenza Testing

At enrollment, we collected and combined midturbinate nasal and throat swabs in viral transport medium. A comparable respiratory specimen (eg, tracheal aspirate, nasal wash, nasopharyngeal swab, or midturbinate nasal swab) collected for clinically indicated testing independent of this study was an acceptable alternative to swabs collected by research staff.

Research testing was performed for influenza viruses using site-specific molecular assays: Luminex NxTAG Respiratory Pathogen Panel (Kansas City, Cincinnati) [8], BioFire FilmArray Respiratory Panel (Seattle) [9], Applied Biosystems TaqMan Array microfluidic card (Rochester) [10, 11], and in-house real-time reverse-transcription polymerase chain reaction assays using primers, probes, and testing protocol developed by the Centers for Disease Control and Prevention (CDC) (Pittsburgh, Nashville, Houston) [12, 13].

In the spring of 2020, enrollment in NVSN with dedicated research specimen collection and testing was diminished because of increasing circulation of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) in the US [7]. Thus, we included influenza test results from standard-of-care molecular testing performed in the hospital clinical laboratories in our analyses and for determination of site-specific influenza seasons. The study protocol was approved by the CDC, and participating institutional review boards and informed consent and assent were obtained.

Influenza Vaccination History

For all children, we asked the parent or guardian about their child’s receipt of 2019–2020 influenza vaccine during the enrollment interview to determine whether a vaccine had been received for the current season (defined as since 1 July 2019). Parental verbal report was verified using documentation from the state immunization information systems, the electronic medical record, and/or via phone call or fax from providers [5, 6]. Per protocol, the process of verifying parental report for influenza vaccination was required for all hospitalized children and optional for children enrolled in the ED. Five sites (Cincinnati, Houston, Nashville, Pittsburgh, and Rochester) verified parental report for ED patients. Complete data regarding influenza vaccination status for past seasons were not yet available; therefore, we defined children as vaccinated or unvaccinated for the current season and did not evaluate full and partial vaccination status. When using documented vaccination status, we classified a child as vaccinated if they received a current season vaccine at least 2 weeks prior to symptom onset. For children missing vaccine records or enrolled from ED sites that did not verify parental report, vaccination history was determined from parental/guardian report alone. When using parental report, a child was considered vaccinated for the current season if they were said to have received 1 current season influenza vaccine prior to the hospitalization or ED visit. We presented VE results based on vaccination ascertainment by documented verification supplemented with parental verbal report, because differences in VE estimates using documented verification alone vs VE estimates using parental report alone were ≤5% with widely overlapping confidence limits.

Statistical Analysis

We compared characteristics among influenza-positive cases and influenza-negative controls, and vaccinated and unvaccinated children, using descriptive statistics (χ ² or Wilcoxon rank-sum tests). We estimated the effectiveness of influenza vaccination against influenza-associated hospitalizations and ED visits associated with any influenza virus and for B/Victoria lineage and A(H1N1)pdm09 viruses using the test-negative study design, comparing the odds of current season influenza vaccination in influenza cases compared with test-negative controls [14–18]. VE was calculated as (1 – adjusted odds ratio) × 100%. Exclusions from analysis have been previously described [5] and are shown in Supplementary Figures 1 and 2.

A priori, we included study site, age, and calendar time (month of enrollment) in logistic regression models [19]. Additional potential covariates including sex, time since symptom onset (0–2, 3–4, 5–7, and 8–10 days), race/ethnicity, underlying medical conditions (0 or ≥1), insurance status (public/self-pay/unknown compared with private/both), and number of hospitalizations in the past year were evaluated [5]. Models were built using a manual forward stepwise procedure. Covariates were retained in the final model and all adjusted models if they were a priori confounders, remained significant at a P value <.05, or influenced the VE estimate by >5%. For all analyses, we conducted a subgroup analysis to estimate VE by age (<5 years and 5–17 years) and evaluated P values from an interaction term to determine if estimates between age groups were statistically significantly different. All analyses were completed in SAS version 9.4 software.

RESULTS

During the 7 site-specific dates for influenza season (encompassing 15 September 2019 to 13 April 2020), there was an early predominance of enrolled participants with B/Victoria influenza viruses followed by a peak of A(H1N1)pdm09 viruses (Figure 1). We enrolled 3133 hospitalized children and 2571 children with ED visits that did not result in hospitalization (Supplementary Figures 1 and 2).

Figure 1.

Epidemiologic curve of influenza virus infections in New Vaccine Surveillance Network enrollees, including hospitalized children and children presenting to the emergency department for acute respiratory illness during influenza season, 2019–2020.

Hospitalized Participants

Characteristics

After exclusions (mostly for children <6 months of age who are not vaccine eligible), 2029 (65%) hospitalized children were eligible for the primary VE analysis. Among the hospitalized group, 335 (17%) children were positive for influenza virus infection: 149 (44%) for A(H1N1)pdm09, 17 (5%) A not subtyped, 2 (1%) A(H3N2), 124 (37%) B/Victoria, 35 (11%) B lineage unknown, 3 (1%) B/Yamagata, and 5 (2%) with >1 influenza virus (Supplementary Figure 1, Supplementary Table 1). Among the 2029 hospitalized children in the VE analysis, 1649 (81%) had confirmed documentation of vaccination status available, and 380 (19%) had vaccination status by parental report alone.

Influenza-positive hospitalized cases were more likely to be aged 5–17 years (45%) and male (64%) compared with influenza-negative hospitalized controls (26% and 55%, respectively) (Table 1). A higher percentage of vaccinated children were <5 years of age (76%) compared with unvaccinated children (66%). Vaccinated children were more likely than unvaccinated children to be non-Hispanic white (48% vs 35%) and less likely to be non-Hispanic black (17% vs 31%). Overall, 53% (1084/2029) of hospitalized participants received influenza vaccination (Table 1), 35% (118/335) among influenza cases and 57% (966/1694) among influenza-negative controls (Table 2). The percentage of hospitalized children vaccinated by site ranged from 47% to 66% (Supplementary Table 2).

Table 1.

Characteristics of Children Hospitalized With Acute Respiratory Illness, Overall and by Influenza Positivity and Influenza Vaccination Status, New Vaccine Surveillance Network, 2019–2020

Characteristic	No. Enrolled	Influenza Positive	Influenza Negative	P Value	Vaccinated (≥1 Dose)	Not Vaccinated	P Value
Overall	2029	335	1694		1084	945
Sex				.002			.60
Male	1149 (57)	216 (64)	933 (55)		608 (56)	541 (57)
Female	880 (43)	119 (36)	761 (45)		476 (44)	404 (43)
Age category				<.001			<.001
<5 y	1438 (71)	185 (55)	1253 (74)		819 (76)	619 (66)
5–17 y	591 (29)	150 (45)	441 (26)		265 (24)	326 (34)
Race/ethnicity				.089			<.001
White, non-Hispanic/Latino	852 (42)	133 (40)	719 (42)		522 (48)	330 (35)
Black, non-Hispanic/Latino	478 (24)	94 (28)	384 (23)		189 (17)	289 (31)
Other, non-Hispanic/Latino	199 (10)	37 (11)	162 (10)		109 (10)	90 (10)
Hispanic/Latino	500 (25)	71 (21)	429 (25)		264 (24)	236 (25)
Underlying medical conditionsa				.90			.033
≥1	1114 (55)	185 (55)	929 (55)		619 (57)	495 (52)
None	915 (45)	150 (45)	765 (45)		465 (43)	450 (48)
Hospitalizations in past year				.098			<.001
0	1264 (62)	226 (67)	1038 (61)		623 (57)	641 (68)
≥1	748 (37)	107 (32)	641 (38)		453 (42)	295 (31)
Unknown	17 (1)	2 (1)	15 (1)		8 (1)	9 (1)
Surveillance site				<.001			.006
Cincinnati	186 (9)	37 (11)	149 (9)		91 (8)	95 (10)
Houston	508 (25)	62 (19)	446 (26)		268 (25)	240 (25)
Kansas City	124 (6)	28 (8)	96 (6)		58 (5)	66 (7)
Nashville	294 (14)	48 (14)	246 (15)		148 (14)	146 (15)
Pittsburgh	527 (26)	102 (30)	425 (25)		280 (26)	247 (26)
Rochester	183 (9)	39 (12)	144 (9)		103 (10)	80 (8)
Seattle	207 (10)	19 (6)	188 (11)		136 (13)	71 (8)
Insurance status				.29			<.001
Public (or none/self-pay)	1362 (67)	237 (71)	1125 (66)		656 (61)	706 (75)
Private or both	644 (32)	94 (28)	550 (32)		415 (38)	229 (24)
Unknown	23 (1)	4 (1)	19 (1)		13 (1)	10 (1)
Days of symptoms since illness onset				.008			.19
0–2	732 (36)	100 (30)	632 (37)		388 (36)	344 (36)
3–4	681 (34)	109 (33)	572 (34)		381 (35)	300 (32)
5–7	490 (24)	103 (31)	387 (23)		257 (24)	233 (25)
8–10	126 (6)	23 (7)	103 (6)		58 (5)	68 (7)
Received MV	70 (3)	14 (4)	56 (3)	.43	41 (4)	29 (3)	.38
Admitted to ICUb	369 (18)	52 (16)	317 (19)	.16	212 (20)	157 (17)	.091
Death	2 (<1)	1 (<1)	1 (<1)	.20	0 (0)	2 (<1)	.32

Data are presented as No. (column %) unless otherwise indicated.

Abbreviations: ICU, intensive care unit; MV, mechanical ventilation.

^aUnderlying medical conditions obtained from medical record review included chronic pulmonary/airway, cardiac, gastrointestinal, kidney, endocrine, neurologic/neuromuscular, developmental, hematologic/oncologic, genetic/metabolic, immunocompromised conditions, pregnancy, and prematurity.

^bFour children had unknown status for ICU admission.

Table 2.

Influenza Vaccine Effectiveness for Prevention of Influenza A and B–Associated Hospitalizations Among Children, New Vaccine Surveillance Network, 2019–2020

Virus Type	Influenza Positive (Cases)		Influenza Negative (Controls)		Vaccine Effectiveness
	No. Vaccinated/Total	%	No. Vaccinated/Total (%)	%	VE, % (95% CI)	Adjusteda VE, % (95% CI)
All influenza viruses	118/335	35	966/1694	57	59 (48–68)	62 (52–71)
<5 y	74/185	40	745/1253	59	54 (38–67)	60 (44–71)
5–17 y	44/150	30	221/441	50	58 (38–72)	65 (47–77)
A(H1N1)pdm09 viruses	56/152	37	966/1694	57	56 (38–69)	64 (49–75)
B/Victoria viruses	46/127	36	966/1694	57	57 (38–70)	54 (33–69)

Abbreviations: CI, confidence interval; VE, vaccine effectiveness.

^aAdjusted for study site, age as a continuous variable, and calendar time (monthly intervals).

Vaccine Effectiveness

Overall, adjusted VE against any influenza-associated hospitalization in children was 62% (95% confidence interval [CI], 52%–71%) (Table 2). When stratified by virus subtype, VE was 64% (95% CI, 49%–75%) against A(H1N1)pdm09 viruses and 54% (95% CI, 33%–69%) against B/Victoria lineage viruses. By age group for hospitalized patients for all influenza viruses, VE was 60% (95% CI, 44%–71%) among children 6 months to <5 years of age and 65% (95% CI, 47%–77%) among children aged 5–17 years (P value for differences between age groups = .72).

Emergency Department Participants

Characteristics

Among ED patients, 2102 of 2571 (82%) were eligible after exclusions. Of these, 671 (32%) were positive for influenza virus infection: 285 (42%) for A(H1N1)pdm09, 16 (2%) A not subtyped, 3 (<1%) A(H3N2), 315 (47%) B/Victoria, 41 (6%) B lineage unknown, 4 (1%) B/Yamagata, and 7 (1%) for >1 influenza virus (Supplementary Figure 2, Supplementary Table 1). Among the 2102 ED children in the VE analysis, 1218 (58%) had confirmed documentation of vaccination status available, and 884 (42%) had vaccination status by parental report alone.

Among 2102 enrolled children with ED visits, 266 of 671 (40%) influenza-positive cases and 350 of 1431 (24%) influenza-negative ED controls were 5–17 years of age (Table 3). Although race/ethnicity did not differ significantly by influenza positivity in hospitalized children, influenza-positive children in the ED were more likely than influenza-negative children to be non-Hispanic black (52% vs 40%, respectively) and to have public or no insurance (87% vs 79%, respectively). Similar to hospitalized children, a higher percentage of vaccinated children were <5 years of age (80%) compared with unvaccinated children (66%). Vaccinated children were more likely than unvaccinated children to be non-Hispanic white (28% vs 16%) and less likely to be non-Hispanic black (30% vs 53%). Overall, 42% (881/2102) of ED participants received influenza vaccine (Table 3), 28% (191/671) among influenza cases and 48% (690/1431) among influenza-negative controls (Table 4). The range of percentage of ED children who had been vaccinated by site was from 29% to 73%.

Table 3.

Characteristics of Children in the Emergency Department With Acute Respiratory Illness, Overall and by Influenza Positivity and Influenza Vaccination Status, New Vaccine Surveillance Network, 2019–2020

Characteristic	No. Enrolled	Influenza Positive	Influenza Negative	P Value	Vaccinated (≥1 Dose)	Not Vaccinated	P Value
Overall	2102	671	1431		881	1221
Sex				.74			.33
Male	1126 (54)	363 (54)	763 (53)		483 (55)	643 (53)
Female	976 (46)	308 (46)	668 (47)		398 (45)	578 (47)
Age category				<.001			<.001
<5 y	1486 (71)	405 (60)	1081 (76)		702 (80)	784 (64)
5–17 y	616 (29)	266 (40)	350 (24)		179 (20)	437 (36)
Race/ethnicity				<.001			<.001
White, non-Hispanic/Latino	445 (21)	91 (14)	354 (25)		250 (28)	195 (16)
Black, non-Hispanic/Latino	919 (44)	348 (52)	571 (40)		267 (30)	652 (53)
Other, non-Hispanic/Latino	159 (8)	41 (6)	118 (8)		87 (10)	72 (6)
Hispanic/Latino	579 (28)	191 (28)	388 (27)		277 (31)	302 (25)
Underlying medical conditionsa				.025			.012
≥1	558 (27)	157 (23)	401 (28)		259 (29)	299 (24)
None	1544 (73)	514 (77)	1030 (72)		622 (71)	922 (76)
Hospitalization in past year				<.001			<.001
0	1806 (86)	612 (91)	1194 (83)		721 (82)	1085 (89)
≥1	293 (14)	59 (9)	234 (16)		158 (18)	135 (11)
Unknown	3 (<1)	0 (0)	3 (<1)		2 (<1)	1 (<1)
Surveillance site				<.001			<.001
Cincinnati	296 (14)	124 (18)	172 (12)		87 (10)	209 (17)
Houston	160 (8)	32 (5)	128 (9)		73 (8)	87 (7)
Kansas City	363 (17)	167 (25)	196 (14)		108 (12)	255 (21)
Nashville	497 (24)	153 (23)	344 (24)		191 (22)	306 (25)
Pittsburgh	302 (14)	71 (11)	231 (16)		127 (14)	175 (14)
Rochester	218 (10)	77 (11)	141 (10)		101 (11)	117 (10)
Seattle	266 (13)	47 (7)	219 (15)		194 (22)	72 (6)
Insurance status				<.001			<.001
Public (or none/self-pay)	1716 (82)	587 (87)	1129 (79)		661 (75)	1055 (86)
Private or both	373 (18)	81 (12)	292 (20)		213 (24)	160 (13)
Unknown	13 (1)	3 (<1)	10 (1)		7 (1)	6 (1)
Days of symptoms since illness onset				<.001			.44
0–2	1115 (53)	400 (60)	715 (50)		485 (55)	630 (52)
3–4	557 (26)	156 (23)	401 (28)		223 (25)	334 (27)
5–7	361 (17)	102 (15)	259 (18)		147 (17)	214 (18)
8–10	69 (3)	13 (2)	56 (4)		26 (3)	43 (4)

Data are presented as No. (column %) unless otherwise indicated.

^aUnderlying medical conditions obtained from medical record review included chronic pulmonary/airway, cardiac, gastrointestinal, kidney, endocrine, neurologic/neuromuscular, developmental, hematologic/oncologic, genetic/metabolic, immunocompromised conditions, pregnancy, and prematurity.

Table 4.

Influenza Vaccine Effectiveness for Prevention of Influenza A and B–Associated Emergency Department Visits Among Children, New Vaccine Surveillance Network, 2019–2020

	Influenza Positive (Cases)		Influenza Negative (Controls)		Vaccine Effectiveness
Influenza Type	No. Vaccinated/Total	%	No. Vaccinated/Total (%)	%	VE, % (95% CI)	Adjusteda VE, % (95% CI)
All influenza viruses	191/671	28	690/1431	48	57 (48–65)	56 (46–65)
<5 y	126/405	31	576/1081	53	60 (50–69)	56 (42–66)
5–17 yb	65/266	24	114/350	33	33 (4–53)	40 (11–59)
A(H1N1)pdm09 viruses	93/288	32	690/1431	48	49 (33–61)	53 (37–65)
B/Victoria viruses	80/322	25	690/1431	48	64 (53–73)	55 (40–66)

Abbreviations: CI, confidence interval; VE, vaccine effectiveness.

^aAdjusted for study site, age as a continuous variable, and calendar time (monthly intervals).

^bChildren <18 years of age were enrolled in the emergency department primarily in 5 of 7 New Vaccine Surveillance Network sites: Cincinnati, Houston, Kansas City, Nashville, and Rochester.

Vaccine Effectiveness

Among ED patients, adjusted VE was 56% (95% CI, 46%–65%) against any influenza virus infection (Table 4). When stratified for virus subtype, VE was 53% (95% CI, 37%–65%) against A(H1N1)pdm09 viruses and 55% (95% CI, 40%–66%) for B/Victoria lineage viruses. VE for all viruses was 56% (95% CI, 42%–66%) among children aged 6 months to <5 years and 40% (95% CI, 11%–59%) among those aged 5–17 years (P value for differences between age groups = .08).

DISCUSSION

During the 2019–2020 influenza season, our data showed that influenza vaccination significantly reduced laboratory-confirmed influenza hospitalization among US children enrolled in NVSN by 62%. Notably, vaccination halved the odds of hospitalizations and ED visits associated with circulating B/Victoria viruses that had genetic and antigenic differences compared with the B/Victoria component of the 2019–2020 influenza vaccine. Past annual estimates of influenza VE in children that typically measure outpatient medically attended laboratory-confirmed influenza have shown 40%–60% effectiveness when vaccine viruses and circulating viruses are similar [19–25]. Protection against new circulating B/Victoria viruses suggests cross-reactivity to antigenically drifted viruses within the same lineage of the vaccine antigen, although the contribution of potential cross-lineage immunity due to the B/Yamagata vaccine antigen cannot be excluded [19, 26, 27]. Our data indicate that the unusually intense and early disease that occurred nationally in children due to the emergence of these new B/Victoria viruses, the first time B viruses had predominated since the 1992–1993 season [2], was likely related to failure to vaccinate rather than vaccine failure. Moreover, the second half of the season in the US was predominated by A(H1N1)pdm09 viruses, with majority subclades that were antigenically distinct from the vaccine component [4]. Despite these antigenic distinctions, we also found that influenza vaccination reduced laboratory-confirmed influenza hospitalizations and ED visits related to A/H1N1pdm09 by 53%–64%.

Overall, 53% of hospitalized participants and 42% of ED participants received influenza vaccination. This is far below the desired goal for pediatric influenza vaccination, despite the recommendation that all people 6 months and older without contraindication be vaccinated against influenza [28]. Furthermore, we noted racial disparities with a higher percentage of non-Hispanic black children unvaccinated than non-Hispanic white children in our study population. All available US pediatric vaccine formulations for the 2019–2020 season were quadrivalent and contained A(H3N2), A(H1N1)pdm09, B/Victoria virus, and influenza B/Yamagata virus strains [28]. Our VE point estimates are higher than the US Flu VE Network estimates against medically attended pediatric outpatient influenza illness related to B/Victoria viruses for the 2019–2020 season (final estimate, 41% [95% CI, 28%–51%] for children aged 6–17 years) [4]. Our estimate is also consistent with VE estimates against influenza B/Victoria (range, 49%–56%) during seasons when the B/Victoria component of the vaccine was well-matched to circulating viruses [1].

Our study has strengths and limitations. The prospective multisite enrollment across both the inpatient and ED settings in geographically diverse locations is a strength of the NVSN. Not all verification of vaccination was complete for inpatients and not routinely performed for all ED children, which might result in misclassification of vaccination status for some participants. However, we have demonstrated in previous seasons that VE against influenza-associated hospitalizations in NVSN does not differ substantially when using parental report compared with documented records as the source for vaccination history [5, 29]. A limitation of this study was our inability to differentiate complete vaccination with 2 vaccine doses in a previous or current season compared with only 1 dose in children <9 years of age. We used an observational study design and may not have controlled for all confounders. However, the test-negative design is widely used nationally and internationally for observational studies of influenza VE and has been extensively validated, including against findings from randomized trials [14–18, 30]. Last, we did not do genetic characterization of viruses from our study and therefore are not certain which influenza subclades circulated in our network sites. This is unlikely to be a limitation for B/Victoria viruses because >95% of circulating influenza B/Victoria viruses nationally were subclade V1A.3 [1].

In conclusion, our data showing relatively high VE for preventing both hospitalizations and ED visits emphasize that while influenza virus antigenic drift is unpredictable, vaccination can still be protective in children even when the circulating influenza viruses do not match the vaccine antigens. This is reassuring and further illustrates that genetic and antigenic characterization of circulating influenza viruses for the annual vaccine strain selection, which typically focuses on the evolving receptor binding domain of the hemagglutinin surface glycoprotein, may not consistently predict clinical protection conferred by vaccination. While immunity against the head domain of the hemagglutinin is an important correlate of protection, other components of immunity that contribute to protection may include conserved hemagglutinin epitopes, neuraminidase surface protein, other viral proteins, or cross-protective T-cell responses [31]. Timely annual studies of effectiveness can provide supporting evidence for vaccine strain selection, address gaps, and offer research directions for improving influenza vaccines. In our study, influenza vaccines provided significant protection against both B/Victoria and A/H1N1pdm09 viruses, and vaccination reduced the likelihood of pediatric hospitalizations and ED visits by approximately 53%–65%. However, only about half of study participants were vaccinated, indicating missed opportunities for reducing the burden of severe influenza in children. With the potential for influenza and SARS-CoV-2 viruses both circulating during the 2020–2021 influenza season, influenza vaccination is more important than ever to decrease the overall burden of febrile and respiratory disease and complications among children.

Supplementary Data

Supplementary materials are available at Clinical Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

Notes

New Vaccine Surveillance Network Author Group. The additional authors for this study are: Robert W. Hickey, MD (Department of Pediatrics, University of Pittsburgh School of Medicine and University of Pittsburgh Medical Center Children’s Hospital of Pittsburgh, Pennsylvania); Pedro A. Piedra, MD and Flor M. Munoz, MD (Baylor College of Medicine and Texas Children’s Hospital, Houston, Texas); Vasanthi Avadhanula (Baylor College of Medicine, Houston, Texas); Jennifer E. Schuster, Barbara A. Pahud, Gina Weddle, and Mary E. Moffatt (Children’s Mercy Hospital, Kansas City, Missouri); Christina Albertin and Wende Fregoe (University of Rochester School of Medicine and Dentistry, Rochester, New York); and Elizabeth Schlaudecker (Department of Pediatrics, University of Cincinnati College of Medicine, and Division of Infectious Disease, Cincinnati Children’s Hospital Medical Center, Cincinnati, Ohio).

Acknowledgments. The authors acknowledge Rebekkah Varjabedia and Yesenia Romero (Nashville, Tennessee); Sahle Amsalu, Monica Asman, Maureen Carolin, Rose Davis, Skylar Deaton, Carley Dunkman, Nicole Knab, Nicholas Leahy, Julia Majchszak, Chelsea Rohlfs, Kevin Schmidt, Senora Schwab, Meghan Sweeney, Marilyn Rice, Nicole Meyer, Amy Singh, Christina Quigley, Samantha Thomas, Tara Keller, and Amy Ostrow (Cincinnati, Ohio); Samar Musa, Noreen Jeffrey, and Monika Johnson (Pittsburgh, Pennsylvania); Lynne Shelley, Jenelle Putzig, Jennifer Carnahan, Miranda Marchand, Theodore Pristash, and Joshua Aldred (Rochester, New York); Kirsten Lacombe, Bonnie Strelitz, Ashley Akramoff, Jennifer Baxter, Rachel Buchmeier, Kaitlin Cappetto, Jennifer Jensen, Sarah Korkowski, Katarina Ost, Hannah Schlaack, Hannah Smith, Sarah Steele, Ariundari Tsogoo, and Emily Walter (Seattle, Washington); Paushpala Sen, Anne Kleinwolterink, Dithi Banerjee, Anjana Sasidharan (Kansas City, Missouri); Yingtao Zhou, Iddrisu Sulemana, Roberto Mejia, LaShondra Berman, and John Barnes (Atlanta, Georgia); and all the members of the New Vaccine Surveillance Network.

Disclaimer. The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the US Centers for Disease Control and Prevention (CDC).

Financial support. This work was supported by the US Centers for Disease Control and Prevention (cooperative agreement CDC-RFA-IP16-004).

Potential conflicts of interest. N. B. H. receives grant support from Sanofi and Quidel, received an educational grant for speaker fees from Genentech, received vaccine donations from Sanofi, and was a consultant for Moderna and Karius. J. V. W. serves on a scientific advisory board for Quidel, a scientific advisory board for ID Connect, and an independent data monitoring committee for GlaxoSmithKline (GSK). Children’s Mercy Hospital–Kansas City receives grant funding from GSK, Merck, and Pfizer for pneumococcal, meningococcal, and rotavirus vaccine studies on which C. J. H. and B. A. P. are investigators, and receives grant funding from Merck (for J. E. S. and R. S.). B. A. P. was a consultant for Merck, Pfizer, GSK, and Sanofi, and reports grants from Merck, Seqirus, GSK, Pfizer, and Sanofi, outside the submitted work. J. A. E. receives grant support from GSK, AstraZeneca, Merck, and Novavax, and is a consultant for Sanofi Pasteur and Meissa Vaccines. F. M. M. receives grant support from Novavax and Janssen for respiratory syncytial virus research, Gilead for COVID-19 research, and Merck for pneumococcal vaccine research, and is a member of the data and safety monitoring board for various vaccines for Pfizer, Moderna, Meissa, and Virometrix. M. M. reports grants to Children’s Mercy Hospital - Kansas City from Luminex Corporation, outside the submitted work. P. A. P. reports scientific advisory fees from Roche and grants from Shionogi, outside the submitted work. G. A. W. reports reimbursement for writing chapters from Merck and consulting fees from Reviral, outside the submitted work. All other authors report no potential conflicts of interest.

All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.

Author notes

Additional authors of the NVSN are listed in the Notes.