Epidemiology and severity of pediatric influenza hospitalizations before and during the COVID-19 pandemic: a surveillance study of the Canadian immunization monitoring program (IMPACT), 2012–2023

Summary

Background

The COVID-19 pandemic disrupted respiratory virus circulation, including influenza. We described trends in seasonality, demographics, and severity of pediatric influenza hospitalizations in Canada before and during the pandemic.

Methods

We conducted national prospective surveillance through the Canadian Immunization Monitoring Program, ACTive (IMPACT), across 12 tertiary-care pediatric centers. Children <17 years hospitalized with laboratory-confirmed influenza between August 2012–August 2023 were included. Periods were categorized as pre-pandemic (2012–March 2020), and pandemic (2021–2022, 2022–2023). Severe disease was defined as intensive care unit (ICU) admission, mechanical ventilation, extracorporeal membrane oxygenation, and/or death. Adjusted odds ratios (aOR) were estimated using multivariable logistic regression.

Findings

Among 10,091 hospitalizations, 7938 occurred pre-pandemic (annual average = 955), 335 in 2021–2022, and 1810 in 2022–2023. The median age was 3.0 (IQR 1.2–6.4) in pre-pandemic, 3.3 (IQR 1.4–7.3) in 2021–2022, and 3.8 years (IQR 1.7–7.2, p < 0.001) in 2022–2023. ICU admissions declined in 2021–2022 (11.3%) and rose in 2022–2023 (15.9%), though remaining below pre-pandemic levels (17.3%). Mortality was stable across periods (0.6%). The odds of severe disease were lower in 2021–2022 (aOR 0.63, 95% CI 0.44–0.89) and 2022–2023 (aOR 0.85, 95% CI 0.74–0.98) compared to pre-pandemic.

Interpretation

Pediatric influenza hospitalizations nearly doubled in 2022–2023, with a modest shift in age distribution toward older children. Although severity indicators rose compared to 2021–2022, they remained slightly reduced compared to pre-pandemic seasons, suggesting suggest a partial return toward pre-pandemic epidemiologic patterns.

Funding

IMPACT is managed by the Canadian Paediatric Society and conducted on behalf of the Public Health Agency of Canada.

Research in context.

Evidence before this study

We searched PubMed and Google Scholar between January 15, 2023, and January 18, 2026, for studies evaluating pediatric influenza hospitalizations and severity before and during the COVID-19 pandemic. Search terms included combinations of (“influenza” OR “seasonal influenza”) AND (“children” OR “pediatric”) AND (“hospitalization” OR “severity” OR “intensive care unit”) AND (“COVID-19 pandemic” OR “post-pandemic”). We also reviewed surveillance reports from the US Centers for Disease Control and Prevention (CDC), FluWatch Canada, and the European Centre for Disease Prevention and Control (ECDC). Previous studies consistently reported major disruptions in influenza circulation during the COVID-19 pandemic, followed by altered seasonality and resurgence of pediatric influenza hospitalizations after relaxation of public health measures. However, most Canadian studies were limited to single centers, provincial datasets, or few pre-pandemic comparator seasons. Few studies evaluated severity across multiple pandemic seasons using long-term national active surveillance data.

Added value of this study

This study provides a national description of pediatric influenza hospitalizations in Canada before and during the COVID-19 pandemic using 10 influenza seasons of prospective active surveillance from the Canadian Immunization Monitoring Program, ACTive (IMPACT), a network of 12 tertiary-care children's hospitals capturing more than 90% of Canada's pediatric tertiary-care admissions. After an unusually mild 2021–2022 season, influenza activity rebounded sharply in 2022–2023, exceeding all pre-pandemic seasons. Despite this surge, the odds of severe outcomes remained slightly lower than in the pre-pandemic period. We also identified a modest shift toward older hospitalized children, although children younger than five years consistently accounted for more than half of admissions.

Implications of all the available evidence

Available evidence indicates that pandemic-related public health measures substantially altered pediatric influenza epidemiology, but severity patterns largely returned toward historical norms by 2022–2023. The substantial increase in hospitalizations despite slightly lower per-case severity highlights the need for sustained pediatric influenza surveillance and preparedness planning. Children younger than five years continue to account for most influenza hospitalizations and should remain a priority for preventive strategies, including seasonal vaccination.

Introduction

During the first year of the COVID-19 pandemic, several countries, including Canada, demonstrated a remarkable decrease in the spread of influenza and other seasonal respiratory viruses.1, 2, 3, 4 Following the relaxation of the nonpharmaceutical interventions (NPIs) adopted to mitigate the pandemic,⁵ there was a re-emergence of respiratory viruses worldwide.6, 7, 8 This resurgence led to unprecedented hospital admissions for acute respiratory tract infections–especially influenza–in children and substantial health system pressures in Canada and elsewhere in late 2022.^9,10

Infants and young children typically have the highest incidence of influenza infection and, along with the elderly, are the age groups most affected by severe disease.¹¹ Although most influenza cases in the very young are mild and self-limited,¹² children under five years of age and older children with comorbidities are at higher risk of developing severe illness including hospitalization, intensive care unit (ICU) admission, or death.12, 13, 14, 15

Understanding the epidemiology and severity of influenza hospitalizations among children during the COVID-19 pandemic has important implications for public health, clinical care, and health system planning for future respiratory virus outbreaks where NPIs may be implemented.^10,16 Recent studies have begun to examine influenza epidemiology in the context of the COVID-19 pandemic, reporting increased pediatric hospitalizations during the 2022–2023 season, often with lower severity compared to pre-pandemic periods.17, 18, 19 However, these findings are based on limited data, including studies restricted to a small number of pre-pandemic seasons or single-center analyses, highlighting the need for comprehensive, population-based assessments across multiple seasons. In this study, we used national surveillance data of children hospitalized for influenza to characterize the pre-pandemic baseline epidemiologic features, seasonality, and severity of pediatric influenza in Canada and compared these characteristics to those observed during the period of influenza and COVID-19 co-circulation between 2021 and 2023.

Methods

Study design and setting

This study was conducted through sentinel hospital-based, national, active surveillance at 12 tertiary-care pediatric academic centers as part of the Canadian Immunization Monitoring Program, ACTive (IMPACT).²⁰ IMPACT collects individual-level data on children hospitalized for microbiologically-confirmed influenza and other vaccine-preventable diseases.²¹ IMPACT centres admit more than 75,000 children annually, account for more than 90% of Canadian pediatric tertiary care beds, and receive referrals from all provinces and territories. Institutional ethics or hospital approval for participation in IMPACT was obtained at each center and data were collected from hospital charts with a waiver of consent.²¹

Study population and timeframe

Children aged <17 years hospitalized for laboratory-confirmed, community-acquired influenza between August 26, 2012, and August 26, 2023, were identified through year-round active screening of hospital admission lists and laboratory results by trained research nurses, under the supervision of pediatric infectious diseases physicians. The study period encompassed ten influenza seasons (2012–2013 to 2022–2023), excluding the 2020–2021 season due to minimal influenza circulation and insufficient case numbers.² Seasons were defined according to the Public Health Agency of Canada, which designates each season as beginning in epidemiologic week 35.²² To ensure complete case ascertainment, annual audits of hospital discharge abstract databases were performed.

Across all centers, children admitted with an acute respiratory infection were tested for influenza in accordance with centers’ clinical testing policies. Microbiologically-confirmed influenza was defined as a positive result by polymerase chain reaction, immunofluorescence, or viral culture. Both influenza A and B lineages were included; however, viral subtyping was not routinely performed. Incidental cases, defined as children with laboratory-confirmed influenza identified during hospitalization for reasons unrelated to acute influenza illness, were determined at the center level and were not reported to IMPACT.

The study period was divided into four distinct periods: (1) the pre-pandemic period, encompassing the 2012–2013 influenza season through March 31, 2020; (2) the influenza suppression period (April 1, 2020 to August 28, 2021) (i.e., beginning of the 2021–2022 influenza season), during which influenza virus circulation was nearly absent due to COVID-19 mitigation measures¹; (3) the first influenza re-emergent season (2021–2022), beginning on August 29, 2021; and (4) the second influenza re-emergent season (2022–2023), beginning on August 28, 2022.

Outcomes and variables

Influenza severity was evaluated and compared among children hospitalized during the pre-pandemic period, the 2021–2022 season, and the 2022–2023 season. The influenza suppression period was excluded (n = 8 cases). Severity outcomes of interest included ICU admission, mechanical ventilation, extracorporeal membrane oxygenation (ECMO), hospital and ICU length of stay, and in-hospital death. These outcomes were not mutually exclusive, as patients could contribute to more than one category. For the assessment of overall disease severity, we derived a composite binary variable defined as the occurrence of any of the following: ICU admission, mechanical ventilation, ECMO, or death. This composite severity outcome was used as the dependent variable in multivariable logistic regression models to estimate adjusted odds ratios (aORs) for severe influenza across time periods.

Variables analyzed included demographic and clinical characteristics used to describe the study population, as well as covariates selected a priori for adjustment in multivariable analyses. These included sex, age (as both continuous and categorical), influenza vaccination status during the current season (among children aged ≥6 months), presence of comorbidities (including respiratory, neurological, hematological, oncologic and other immunocompromising conditions, metabolic, cardiovascular, and renal), and length of hospital stay. Vaccination status referred to influenza vaccination status at the time of hospital admission during the current season, and was categorized as vaccinated, not vaccinated, or unknown/missing. Comorbidities (as described in medical charts) were classified using the Medical Dictionary for Regulatory Activities (MedDRA),²³ according to their “system organ class” (SOC).

Statistical analysis

Descriptive statistics were used to summarize demographic and clinical characteristics. Continuous variables were reported as medians with interquartile ranges (IQR), and categorical variables as frequencies and percentages. Comparisons between the pre-pandemic period and each of the 2021–2022 and 2022–2023 seasons were conducted using Wilcoxon rank-sum tests, χ² tests, and Fisher's exact tests. We also expressed pre-pandemic statistics as annual averages, considering only the 2012–2013 to 2018–2019 seasons (excluding 2019–2020 cases given this was a partial season prior to the emergence of COVID-19 in early 2020). To understand potential changes in seasonality, time series plots were analyzed using the absolute number of influenza hospitalizations and daily rate of hospitalizations per month.

Multivariable mixed-effects logistic regression was used to estimate the odds of severe influenza outcomes between the pre-pandemic period and each of the 2021–2022 and 2022–2023 seasons, adjusting for age category, sex, presence of chronic comorbidities (any vs. none), and influenza virus type (A, B, or A/B coinfection), with IMPACT centre included as a random effect. Age was modeled as a categorical variable (<6 months, 6–23 months, 24–59 months, 5–9 years, and 10–16 years) to reflect established clinical risk categories for severe influenza in children.^13,24 Covariate adjustment was defined prior to analyzing the data, based on literature review and expert knowledge of pediatric influenza risk factors. In two secondary models, we assessed effect modification using categorical interaction terms between time period and each of age group (<5 vs. 5–16 years) and influenza lineage (A vs. B). Finally, to assess whether severity in the 2021–2022 and 2022–2023 seasons was similar to an individual pre-pandemic season, we modeled influenza season as a categorical variable, replacing the broader time period grouping to allow direct comparison with each pre-pandemic season. Model fit was assessed using a calibration plot by assessing the intercept and slope of observed and predicted probabilities by deciles of predicted risk, while multicollinearity was checked using variance inflation factors.

All tests were two-tailed and used a statistical significance threshold of α = 0.05. We accounted for multiplicity using the Bonferroni correction when multiple categories (e.g., comorbidities), outcomes, or influenza seasons were analyzed. All analyses were conducted as complete case and without imputation of missing data. Missing data were explicitly reported using the missing indicator method in tables (e.g., vaccination status) or footnotes. Analyses were conducted using Stata version 19.5.²⁵ Frequencies between 1 and 4 were reported as <5 according to IMPACT privacy protection policies.

Ethical approval

Research ethics and/or hospital approvals for the Immunization Monitoring Program Active (IMPACT) data collection were obtained from each IMPACT site.

Role of the funding source

IMPACT influenza surveillance is conducted by the IMPACT network of pediatric investigators on behalf of the Public Health Agency of Canada's Centre for Immunization and Respiratory Infectious Diseases. The funder supported the surveillance infrastructure but had no role in the study design; data analysis; results interpretation; manuscript preparation, review, or approval; or decision to submit the manuscript for publication.

Results

Epidemiological trends and virus circulation

A total of 10,091 influenza hospitalizations were identified across 12 IMPACT centres between the 2012–2013 and 2022–2023 seasons. Among these, 7938 hospitalizations occurred during the pre-pandemic period, 335 in the first re-emergent season (2021–2022), and 1810 in the second re-emergent season (2022–2023) (Table 1). The 2022–2023 season recorded more influenza hospitalizations than both the average annual cases pre-pandemic (955 per year) and the highest pre-pandemic peak (1354 cases in 2018–2019) (Fig. 1A–B).

Table 1.

Demographic and baseline health characteristics of children hospitalized for influenza at IMPACT centres from 2012 to 2023, overall and by time period.

Characteristic	All influenza	Time perioda
		Pre-pandemic	2021–2022 season	p-value vs. pre-pandemic	2022–2023 season	p-value vs. pre-pandemic
Number of cases, N	10,091	7938	335	–	1810	–
Age (years), median (IQR)	3.2 (1.2–6.6)	3.0 (1.2–6.4)	3.3 (1.4–7.3)	0.09	3.8 (1.7–7.2)	<0.001
Age category, n (%)				0.002		<0.001
<6 months	1293 (12.8)	1050 (13.2)	46 (13.7)		197 (10.9)
6–23 months	2365 (23.4)	1965 (24.8)	65 (19.4)		332 (18.3)
24–59 months	2826 (28.0)	2178 (27.4)	97 (29.0)		550 (30.4)
5–9 years	2330 (23.1)	1799 (22.7)	65 (19.4)		464 (25.6)
10–16 years	1277 (12.7)	946 (11.9)	62 (18.5)		267 (14.8)
Sex, n (%)				0.91		0.29
Female	4393 (43.5)	3434 (43.3)	146 (43.6)		808 (44.6)
Male	5698 (56.5)	4504 (56.7)	189 (56.4)		1002 (55.4)
Any comorbid conditions, n (%)	4912 (48.7)	3895 (49.1)	154 (46.0)	0.27	857 (47.4)	0.19
Blood/lymphatic disorders	582 (5.8)	432 (5.4)	31 (9.3)	0.003c	119 (6.6)	0.06
Cardiovascular disorders	546 (5.4)	445 (5.6)	19 (5.7)	0.96	81 (4.5)	0.05
Congenital disorders, not otherwise classified	486 (4.8)	363 (4.6)	13 (3.9)	0.55	108 (6.0)	0.01c
Endocrine/metabolic disorders	592 (5.9)	440 (5.5)	21 (6.3)	0.57	129 (7.1)	0.009c
Gastrointestinal/hepatobiliary disorders	624 (6.2)	493 (6.2)	19 (5.7)	0.69	108 (6.0)	0.70
Immunocompromiseb	655 (6.5)	502 (6.3)	31 (9.3)	0.03c	122 (6.7)	0.51
Musculoskeletal disorders	302 (3.0)	214 (2.7)	16 (4.8)	0.02c	72 (4.0)	0.004c
Neoplasms	458 (4.5)	387 (4.9)	17 (5.1)	0.87	54 (3.0)	<0.001
Nervous system disorders	1285 (12.7)	1040 (13.1)	37 (11.0)	0.27	204 (11.3)	0.04c
Renal disorders	341 (3.4)	280 (3.5)	11 (3.3)	0.81	50 (2.8)	0.10
Respiratory disorders	1920 (19.0)	1492 (18.8)	63 (18.8)	1.00	361 (19.9)	0.26
Obesity, n/N > 2 years (%)	91/6433 (1.4)	61/4923 (1.2)	6/224 (2.7)	0.07	23/1281 (1.8)	0.12
Gestational age ≤35 weeks, n/N < 1 year (%)	166/2113 (7.9)	138/1737 (7.9)	7/71 (9.9)	0.56	21/304 (6.9)	0.53
Influenza virus type, n (%)d				<0.001		<0.001
A	7548 (74.8)	5583 (70.3)	331–334 (98.8–99.7)		1628 (89.9)
B	2499 (24.8)	2311 (29.1)	<5 (<1.5)		182 (10.1)
A and B	44 (0.4)	44 (0.6)	0 (0.0)		0 (0.0)
Patients eligible for influenza vaccinatione, N	8798	6888	289	–	1613	–
Influenza vaccination status, n/N (%)f				<0.001		<0.001
Vaccinated	937/8798 (10.7)	765/6888 (11.1)	42/289 (14.5)		129/1613 (8.0)
Not vaccinated	5448/8798 (61.9)	3924/6888 (57.0)	221/289 (76.5)		1299/1613 (80.5)
Unknown	2413/8798 (27.4)	2199/6888 (31.9)	26/289 (9.0)		185/1613 (11.5)

Statistically significant values are indicated in bold.

^aAnalysis of n = 8 cases from the influenza suppression period (i.e., April 1, 2020 to August 28, 2021) is not included.

^bIncludes immune system disorders (n = 180) and immunocompromising medications (n = 491).

^cP-values were no longer statistically significant after Bonferroni adjustment for multiple testing (adjusted threshold of p = 0.0019 for n = 26 tests; 2 comparisons for 13 conditions).

^dIncludes n = 44 influenza A/B coinfections, all in the pre-pandemic period.

^ePatients aged <6 months were considered ineligible for influenza vaccination.

^fVaccination status during the current influenza season.

Fig. 1.

(A) Number of influenza hospitalizations per month from the 2012–2013 to 2022–2023 seasons. (B) Daily rate of influenza hospitalizations per month in 2021–2022 and 2022–2023, compared to the pre-pandemic (2012–2019) average. In panel B, pre-pandemic data excludes hospitalizations from 2019–2020 given this season partially overlapped with the onset of the COVID-19 pandemic. The number of hospitalizations per month we divided by the number of days per month, to generate the daily rate of influenza-related hospitalizations.

Influenza seasonality also shifted, with a delayed peak in May 2021–2022 followed by an early peak in November 2022–2023, compared with the typical January peak observed pre-pandemic (Fig. 1B and 2).

Fig. 2.

(A) Number of influenza hospitalizations per month from the 2012–2013 to 2022–2023 seasons, by influenza virus type. (B/C) Daily rate of influenza hospitalizations per month in 2021–2022 and 2022–2023 by influenza virus type, compared to the pre-pandemic (2012–2019) average. In panels B and C, pre-pandemic data excludes hospitalizations from 2019–2020 given this season partially overlapped with the onset of the COVID-19 pandemic. The number of hospitalizations per month we divided by the number of days per month, to generate the daily rate of influenza-related hospitalizations. N = 44 cases with influenza A and B coinfection were not included in these figures.

Across the full study period, 74.8% (n = 7548) of cases were attributable to influenza A and 24.8% (n = 2499) to influenza B (0.4% [n = 44] with influenza A/B coinfection during pre-pandemic period). Influenza A predominance increased in the re-emergent seasons, accounting for nearly all cases in 2021–2022 and 89.9% (n = 1628) in 2022–2023, compared to an average of 70.3% (n = 5583) in pre-pandemic seasons (Table 1, Fig. 2).

Demographic and clinical characteristics

Among 10,091 pediatric influenza hospitalizations, demographic characteristics were broadly similar across seasons, although children hospitalized in 2022–2023 were slightly older compared with the pre-pandemic period (p < 0.001). Sex distribution remained consistent across periods (Table 1). Chronic comorbidities were reported in 48.7% (n = 4393) of cases overall, most commonly respiratory (19.0%, n = 1920) and neurological disorders (12.7%, n = 1285), with a consistent distribution across pre-pandemic, 2021–2022, and 2022–2023 seasons (Table 1).

Severity outcomes

Overall, 1700 children (16.9%) were admitted to ICU, and 621 (6.2%) required mechanical ventilation (Table 2). ICU admission was lower in 2021–2022 compared with the pre-pandemic period (11.3% vs. 17.3%, p = 0.005), but similar in 2022–2023 (15.9% vs. 17.3%, p = 0.16) (Table 2, Supplemental Figure S1). Similar patterns were observed when stratified by influenza type (A and B), with reduced ICU admission in 2021–2022 (11.7% vs. 17.8%, p = 0.003) and comparable proportions in 2022–2023 (15.9% vs. 17.8%, p = 0.08) similar to those observed in pre-pandemic seasons for each influenza type (Table 2).

Table 2.

Disease severity outcomes of children hospitalized for influenza at IMPACT centres from 2012 to 2023, overall and by time period.

Characteristic	All influenza	Time perioda
		Pre-pandemic	2021–22 season	p-value vs. pre-pandemic	2022–23 season	p-value vs. pre-pandemic
Influenza hospitalizations, N	10,091	7938	335	–	1810	–
Severe influenza, n (%)	1712 (17.0)	1381 (17.4)	39 (11.6)	0.006c	289 (16.0)	0.94
ICU admission	1700 (16.8)	1371 (17.3)	38 (11.3)	0.005	288 (15.9)	0.16
Mechanical ventilation	621 (6.2)	492 (6.2)	13 (3.9)	0.08	114 (6.3)	0.87
Hospital stay (days), median (IQR)b	3 (2–5)	3 (2–5)	2 (1–4)	<0.001	3 (2–5)	0.95
ICU stay (days), median (IQR)	3 (1–6)	3 (1–6)	2 (1–5)	0.13	2 (1–5)	0.01c
Ages <5 years, N	6484	5193	208	–	1079	–
Severe influenza, n (%)	992 (15.3)	801 (15.4)	21 (10.1)	0.04c	168 (15.6)	0.90
ICU admission	987 (15.2)	796 (15.3)	21 (10.1)	0.04c	168 (15.6)	0.84
Mechanical ventilation	405 (6.2)	323 (6.2)	7 (3.4)	0.09	73 (6.8)	0.50
Hospital stay (days), median (IQR)	3 (2–5)	3 (2–5)	2 (1–4)	0.001	3 (2–5)	0.98
ICU stay (days), median (IQR)	3 (1–6)	3 (1–6)	2 (2–5)	0.81	2 (1–5)	0.07
Ages 5–16 years, N	3607	2745	127	–	731	–
Severe influenza, n (%)	720 (20.0)	580 (21.1)	18 (14.2)	0.06	121 (16.6)	0.006c
ICU admission	713 (19.8)	575 (20.9)	17 (13.4)	0.04c	120 (16.4)	0.006c
Mechanical ventilation	216 (6.0)	169 (6.2)	6 (4.7)	0.51	41 (5.6)	0.58
Hospital stay (days), median (IQR)	3 (2–6)	3 (2–6)	3 (2–5)	0.05c	3 (2–5)	0.60
ICU stay (days), median (IQR)	3 (1–6)	3 (1–7)	1 (1–5)	0.05	3 (1–5)	0.07
Influenza A, Nd	7548	5583	332	–	1628	–
Severe influenza, n (%)	1296 (17.2)	996 (17.8)	39 (11.7)	0.005	260 (16.0)	0.08
ICU admission	1289 (17.1)	991 (17.8)	38 (11.4)	0.003	259 (15.9)	0.08
Mechanical ventilation	465 (6.2)	352 (6.3)	13 (3.9)	0.08	100 (6.1)	0.81
Hospital stay (days), median (IQR)	3 (2–5)	3 (2–5)	2 (1–4)	<0.001	3 (2–5)	0.86
ICU stay (days), median (IQR)	3 (1–6)	3 (1–6)	2 (1–5)	0.14	2 (1–5)	0.02c
Influenza B, Nd	2499	2311	–	–	182	–
Severe influenza, n (%)	411 (16.4)	380 (16.4)	–	–	29 (15.9)	0.86
ICU admission	406 (16.2)	375 (16.2)	–	–	29 (15.9)	0.92
Mechanical ventilation	156 (6.2)	140 (6.1)	–	–	14 (7.7)	0.38
Hospital stay (days), median (IQR)	3 (2–5)	3 (2–5)	–	–	3 (2–5)	0.42
ICU stay (days), median (IQR)	3 (1–7)	3 (1–7)	–	–	2 (1–7)	0.44

ICU = Intensive care unit; IQR = Interquartile range.

Comparisons were conducted within influenza type: influenza A cases in each evaluated season were compared with influenza A cases in the pre-pandemic period, and similarly for influenza B. Subgroup analysis for influenza B cases in the 2021–22 season are not presented given n < 5 cases. Statistically significant values are indicated in bold.

^aSubgroup of n = 8 cases from the early pandemic (i.e., April 1, 2020 to August 28, 2021) is excluded.

^bLength of stay was missing for n = 8 cases (n = 6 pre-pandemic and n = 2 in the 2022–23 season).

^cP-values were no longer statistically significant after Bonferroni adjustment for multiple testing (adjusted threshold of p = 0.005 for n = 10 tests per stratum; 2 comparisons for 5 outcomes).

^dAnalyses exclude n = 44 cases with influenza A and B coinfection.

Overall, 35 patients (0.4%) received extracorporeal membrane oxygenation, and 56 patients died, corresponding to an in-hospital case fatality rate of 0.6% (n = 29, 0.4% among ages <5 years; n = 27, 0.7% among ages 5–16 years).

Median hospital length of stay decreased in 2021–2022 (2.0 days [IQR = 1.0–4.0]) compared with pre-pandemic seasons (3.0 days [IQR = 2.0–5.0], p < 0.001), returning to baseline in 2022–2023 (3.0 days [IQR = 2.0–5.0]). ICU length of stay did not differ across periods after adjusting for multiple testing (3.0 days [IQR = 1.0–6.0]) (Table 2).

Patterns were consistent across age groups (<5 and 5–16 years), while children with comorbidities had consistently higher severity across all periods (Table 2, Supplemental Table S1).

Risk of severe influenza

Overall, 7938 children (17.4%) experienced severe influenza (i.e., ICU admission, ventilation, and/or death) during the pre-pandemic period compared with 39 (11.6%) in 2021–2022 and 289 (16.0%) in 2022–2023. In adjusted analyses, the odds of severe influenza were lower in both 2021–2022 (adjusted OR = 0.63, 95% CI = 0.44–0.89, p = 0.008) and 2022–2023 (aOR = 0.85, 95% CI = 0.74–0.98, p = 0.03) compared to the pre-pandemic period (Fig. 3).

Fig. 3.

Odds of severe influenza (i.e., ICU admission, ventilation, ECMO, and/or death), between the pre-pandemic period vs. 2021–2022 and 2022–2023 seasons. Models adjust for age category (<6 months, 6–23 months, 24–59 months, 5–9 years, 10–16 years), sex, chronic comorbid conditions (any vs. none), and influenza virus type (A, B, A/B coinfection), with IMPACT site (n = 12) included as a random effect. In subgroup analyses of age category and influenza virus type, interaction terms between time period and the subgroup of interest were analyzed in separate models. ¹Model excludes n = 44 cases with influenza A and B coinfection. ²Subgroup analysis for influenza B cases in the 2021–22 season are not presented given n < 5 cases.

Age-stratified analyses showed reduced odds of severe influenza among children aged 5–16 years in 2022–2023 (aOR = 0.71, 95% CI = 0.57–0.88, p = 0.002), with a similar trend in 2021–2022 (aOR = 0.62, 95% CI = 0.37–1.04, p = 0.07) (Fig. 3, Supplemental Table S2). Among children aged <5 years, estimates were lower in 2021–2022 but not significantly different, and no difference was observed in 2022–2023 (Fig. 3, Supplemental Table S2).

In subgroup analyses, influenza A showed consistently lower odds of severe disease in both 2021–2022 (aOR = 0.63, 95% CI = 0.44–0.89, p = 0.008) and 2022–2023 (aOR = 0.83, 95% CI = 0.71–0.97, p = 0.02) compared with the pre-pandemic period (Fig. 3).

Across pre-pandemic seasons, the odds of severe influenza varied, reflecting expected heterogeneity. The 2021–2022 season was consistently less severe than pre-pandemic seasons, whereas the 2022–2023 season fell within the range of severity observed across pre-pandemic seasons. Notably, the 2022–2023 season showed comparable, or slightly lower, severity than pre-pandemic seasons with known H3N2 predominance (Supplemental Figure S2).

Discussion

We describe the impact of the COVID-19 pandemic on pediatric influenza hospitalizations in Canada, comparing the pre-pandemic period with the first two seasons of influenza re-emergence during the pandemic, using national active surveillance data from the IMPACT network. Overall, influenza re-emerged with altered seasonality and increased hospitalization burden, while per-case severity remained lower or comparable to pre-pandemic seasons.

Following the near absence of influenza hospitalizations after the onset of the COVID-19 pandemic,² influenza re-emerged during the 2021–2022 season with atypical seasonal patterns, characterized by a delayed and lower peak in hospitalizations in May 2022, followed by an early and intense peak in November 2022, exceeding those observed in any pre-pandemic season in this study. These shifts deviated from the typical January peaks seen pre-pandemic and are consistent with international influenza surveillance reports from the United States (Centers for Disease Control, CDC) and Europe (European Centre for Disease Prevention and Control, ECDC).^3,4 The relatively low burden observed in 2021–2022 and the unprecedented resurgence in pediatric hospitalizations in 2022–2023 likely reflect the prolonged impact and subsequent relaxation of NPIs, which remained in place to varying degrees across Canada into 2022 and were largely lifted by early 2023.²⁶

Despite the increase in hospitalizations, per-case severity remained lower or comparable to pre-pandemic seasons. Multivariable analyses confirmed a lower likelihood of severe influenza in 2021–2022 and a modest reduction in 2022–2023 compared to the pre-pandemic period. This finding aligns with previous studies reporting increased influenza hospitalization rates following the COVID-19 pandemic, without a corresponding increase in severity at the individual level in both northern and southern hemispheres.^3,27, 28, 29 Furthermore, this pattern is consistent with North American surveillance data. In the United States, CDC reports showed that, although the 2022–2023 season had an earlier and different temporal pattern, the cumulative all-age hospitalization rate was not substantially higher than in several pre-pandemic seasons.³⁰ In Canada, national surveillance data (FluWatch/CCDR) documented a marked increase in pediatric hospitalizations during the same season, without a corresponding increase in overall population-level hospitalization rates.³¹ Together, these observations support the interpretation that the marked increase observed in our study reflects a pediatric-specific increase in hospitalization burden, rather than a generalized increase in influenza severity across all age groups.

This patter characterized by higher pediatric influenza hospitalizations with lower severity contrasts with that observed for other respiratory viruses, such as RSV, which showed marked and atypical surges in pediatric hospitalizations, including out-of-season peaks in 2021 across North America and Europe.32, 33, 34, 35 Similarly, IMPACT surveillance data comparing RSV hospitalizations in 2022–2023 with those observed during the pre-pandemic period (2017–2018 through 2019–2020) showed that RSV activity was characterized by an earlier onset, substantially higher peak, and more prolonged duration.³⁶ These differing temporal patterns likely contributed to overlapping circulation and increased cumulative healthcare demand during the 2022–2023 season.

The dissociation between influenza burden and severity suggests that different mechanisms may underlie infection incidence and clinical outcomes. Reduced influenza circulation during the pandemic likely resulted in a larger pool of immunologically naïve children, increasing susceptibility to infection without necessarily increasing disease severity.^37,38 In addition, changes in healthcare-seeking behavior and hospital admission thresholds may have broadened the clinical spectrum of hospitalized cases.¹⁸ Third, the progressive relaxation of NPIs likely altered exposure dynamics and timing of infection, influencing both transmission and clinical presentation.²⁶ Overall, these patterns are consistent with a return toward more typical influenza epidemiology rather than a sustained shift in disease severity.

Although differences in circulating influenza subtypes have historically been associated with variability in severity,³⁹ our findings do not support subtype as the primary driver of the observed patterns. Results were consistent across subtype-stratified and season-specific analyses, indicating that findings were not explained by virus type or by the choice of pre-pandemic reference. Notably, the 2022–2023 season, characterized by A (H3N2) predominance, showed severity comparable to or lower than previous H3N2-dominant seasons (e.g., 2012–2013, 2014–2015, 2016–2017, and 2017–2018),^30,40 further supporting a multifactorial explanation.

We observed a modest shift toward older age among hospitalized children during the re-emergent seasons, although children <5 years continued to account for more than half of all hospitalizations. These findings may reflect evolving patterns of exposure and susceptibility following pandemic-related disruptions.^27,28 Importantly, children <5 years remain a key vaccination priority in Canada. Age-stratified analyses suggested a partial return toward pre-pandemic severity among younger children, while older children (5–16 years) continued to experience lower severity.

Recent Canadian surveillance data (FluWatch/CCDR) further support the interpretation of 2022–2023 as a transient post-pandemic perturbation. The subsequent 2023–2024 season showed a return to more typical epidemiological patterns, and preliminary data from the ongoing 2025–2026 season do not suggest a comparable surge.^41,42 Together, these observations indicate that the elevated burden observed in 2022–2023 is unlikely to represent a sustained shift in influenza epidemiology.

Our study has several limitations. Changes in microbiologic testing strategies across IMPACT centres during the COVID-19 pandemic may have influenced case ascertainment. Reduced testing early in the pandemic may have led to an underestimation of influenza hospitalizations, whereas increased and more systematic testing in later periods may have identified more mild cases. However, these factors are less likely to have affected the identification of patients requiring ICU care. More comprehensive testing in later periods would be expected to preferentially capture milder cases, potentially leading to an underestimation rather than an overestimation of disease severity. Furthermore, the lack of inclusion of subsequent influenza seasons limits the ability to determine whether the increased hospitalizations observed in 2022–2023 represent an anomalous event or the beginning of a sustained upward trend in Canada. However, national surveillance data suggest a return to more typical patterns in 2023–2024, supporting the interpretation of a transient surge.⁴¹ The absence of influenza A subtyping is an additional limitation, although consistency across influenza type and season-specific analyses supports the robustness of our findings. Moreover, influenza vaccination status was missing for a substantial percentage of cases (27.4%) and was therefore not adjusted for in the multivariable models. We observed differences in influenza vaccination between time periods, potentially resulting in residual confounding of our modeling results. However, among hospitalized children in the 2022–2023 season, only a small proportion were vaccinated (9% among cases with known vaccination status). In contrast, available national survey data suggest that approximately 30–35% of children in Canada received seasonal influenza vaccination during 2022 and 2023, with variability across provinces.⁴³ Given the known protective effect of vaccination against severe influenza outcomes, the lower proportion of vaccinated individuals among hospitalized cases is expected. Additionally, IMPACT does not capture all pediatric admissions in Canada and is based in tertiary care hospitals, potentially limiting generalizability. However, the network's long-term, standardize surveillance across centers representing almost all of the country's tertiary care pediatric beds is a key strength, allowing robust comparisons over time.

In conclusion, national surveillance data from the IMPACT network provide a comprehensive view of pediatric influenza hospitalizations in Canada across ten seasons. The COVID-19 co-circulation era was marked by an atypical, mild 2021–2022 season followed by a 2022–2023 season with high hospitalization burden and a partial return to typical severity. These data serve as an important benchmark for future surveillance and can help determine whether changes in pediatric influenza epidemiology are transient or enduring. Ongoing monitoring is essential to identify children at highest risk of severe disease and to guide planning for clinical care and prevention strategies in the years to come.

Contributors

CDC: conceptualization, interpretation of data, contributing to the analysis, writing—original draft; DSF: statistical analysis, interpretation of data, figures, writing—review & editing; JP, CBu, SH, TJ, KK, and MS: data collection, conceptualization, writing—review & editing; JAB: curation, interpretation of data, conceptualization, writing—review & editing; WW: conceptualization, writing—review & editing; SKM: data collection, conceptualization, overall supervision, writing—review & editing. JAB, JP, and SKM had full access to all the data in the study. CDC, JAB, DSF, and SKM have directly accessed and verified the underlying data reported in the manuscript.

All authors had full access to all the data in the study, approved the final manuscript as submitted, and accepted the responsibility to submit it for publication.

Data sharing statement

Due to the nature of ethics approval and data collection, the IMPACT team cannot share data directly. However, influenza hospitalization data reported by IMPACT network is publicly available from the Government of Canada website (https://www.canada.ca/en/public-health/services/diseases/flu-influenza/influenza-surveillance.html). Details of the conditions for the use of this open dataset are available on the Government of Canada website.

Declaration of interests

SKM has received honoraria for lectures from GlaxoSmithKline, Pfizer, and Sanofi-Pasteur was a member of ad hoc advisory boards for GlaxoSmithKline, Pfizer, and Sanofi Pasteur, all unrelated to this study. SAH has served on ad hoc advisor boards for Pfizer, Moderna, Merck, AstraZeneca, GSK, and Aramis, and has received funding paid to his institution for clinical trials from Pfizer, Merck, Sanofi, GSK, and Moderna. In the last 3 years, MS has been an investigator on projects funded by GlaxoSmithKline, Merck, Moderna, Pfizer, and Sanofi-Pasteur. All funds have been paid to his institute, and he has not received any personal payments. JP reports research grants to his institution from Merck and MedImmune and personal fees from Enanta, unrelated to this study. JAB: none exist. CBu is a site co-investigator on a study funded by Moderna, unrelated to this study.