Nirsevimab effectiveness, number needed to immunize and impact on severe RSV outcomes, a test-negative study in Quebec, Canada

Summary

Background

In 2024–25, Quebec implemented a universally publicly funded infant nirsevimab program. We estimated effectiveness, number-needed-to-immunize (NNI) and impact against severe respiratory syncytial virus (RSV) outcomes.

Methods

A test-negative study estimated adjusted effectiveness among nirsevimab-eligible children RSV-tested during emergency room (ER) consultation or hospitalization between October 2024 and March 2025. Eligible children were born during the RSV season (at-birth group) or <6- to <19-month-olds healthy, preterm, with chronic conditions or living in remote regions (catch-up groups). We used 2023–24 respiratory hospitalization rates, RSV-positivity from a systematic hospital-based surveillance network, and Quebec 2024–25 coverage estimates to derive the theoretical NNI and averted hospitalizations.

Findings

Effectiveness analyses included 3172 specimens linked to ER consultations (1775 [56·0%] from male and 1397 [44·0%] from female participants; 668 [21·1%] RSV-positive) and 1758 linked to hospitalizations (988 [56·2%] from male and 770 [43·8%] from female participants; 549 [31·2%] RSV-positive). Effectiveness was 86% (95% CI: 82–90) against ER consultation, 89% (95% CI: 84–92) against hospitalization (79% (95% CI: 60–89) for those with chronic diseases and 93% (95% CI: 83–97) for preterm infants), and 88% (95% CI: 58–97) against ICU admission. We estimated 41 at-birth and 58 catch-up immunizations needed to avert one RSV-associated hospitalization. We estimated that 72% (at-birth group) and 59% (catch-up group with 64% coverage by November) of RSV-associated hospitalizations were prevented, potentially increasing to 82% for a hypothetical coverage of 90% by October.

Interpretation

Nirsevimab is highly effective and could more substantially impact the RSV burden through broad and timely administration to healthy-term and high-risk infants.

Funding

Ministère de la santé et des services sociaux du Québec.

Research in context.

Evidence before this study

We searched PUBMED using the terms “nirsevimab” and “effectiveness”, to obtain meta-analyses evaluating real-world nirsevimab effectiveness against severe respiratory syncytial virus (RSV) outcomes and also reviewed primary studies from March (last publication date of studies included in the meta-analyses) to November 24, 2025. Two recent meta-analyses summarized 37 studies from the USA, France, Spain and Luxembourg, showing an estimated protection of 81%–83% against RSV hospitalizations and of 76%–81% against intensive care unit admissions. Additional primary studies from the USA, Europe, Chile and Australia published since March 2025 have likewise reported effectiveness against hospitalization ranging from 75 to 98%.

Few studies examined nirsevimab effectiveness among high-risk infants. Two studies reported effectiveness against hospitalization for infants with comorbidities at 62% and 65%, comparable to their estimates among those without comorbidities. Among preterm infants, two studies reported contrasting results: effectiveness of 70% versus 65% among full-term infants in a French large cohort, and 62% versus 82%, respectively, in a Spanish study with few preterm participants.

In summary, although real-world effectiveness studies conducted in Europe, the USA, and Australia corroborate clinical trial findings, data from Canada remain unavailable. Furthermore, evidence regarding effectiveness in high-risk populations is very limited.

Added value of this study

Using provincial population-level data from the province of Quebec, Canada, we estimated that nirsevimab provides over 85% effectiveness against RSV-associated emergency room consultation, hospitalizations and intensive care admissions. Our results showed high effectiveness also among children with comorbidities and preterm infants, confirming initial but scarce evidence that suggested strong protection in these high-risk populations. Our estimates of the impact of Quebec's first season of universal infant immunization provide key inputs for cost-effectiveness analyses and for program planning, particularly in comparable settings such as other Canadian provinces.

Implications of all the available evidence

Nirsevimab is highly effective for preventing RSV severe disease among heathy and at-risk infants. Nirsevimab programs substantially reduce RSV burden, and their impact could increase through widespread and timely administration.

Introduction

Respiratory syncytial virus (RSV) is a leading cause of emergency room (ER) visits and hospitalizations among young children in North America, with the highest hospitalization risk during the first three months of life.¹ While certain underlying health conditions predispose children to severe RSV illness, most pediatric RSV hospitalizations occur in healthy full-term infants.² Nirsevimab is a long-acting monoclonal antibody, which reduced the risk of hospitalization due to RSV by approximately 80% for 150 days in clinical trials,^3,4 as well as in studies evaluating nirsevimab programs implemented during the 2023–24 RSV season in Europe and the USA.^5,6 However, data on effectiveness among preterm infants and high-risk populations, and evidence on health systems impact and number needed to immunize (NNI) remain sparse.^7,8

Nirsevimab was first authorized in Canada on April 2023.⁹ In May 2024, Canada's National Advisory Committee on Immunization recommended building towards a universal RSV immunization program for all infants during their first RSV season.¹⁰ Quebec, with a population of 9 million and about 86,000 births annually, was one of only two Canadian provinces to implement a publicly-funded universal infant nirsevimab program during the 2024–25 season.11, 12, 13 Nirsevimab was offered to all newborns during the RSV season, with a catch-up program for infants <6 months at the start of the season. However, the catch-up phase was delayed and early limitations in nirsevimab supply, along with low awareness among both the population and healthcare providers, likely precluded optimal nirsevimab uptake during the launch. The maternal RSVpreF vaccine was not publicly funded but was available through the private market.

This study among Quebec infants eligible for the 2024–25 nirsevimab immunization program aimed to estimate the protection (effectiveness) against RSV-associated ER consultations, hospitalizations and intensive care unit (ICU) admissions associated with nirsevimab administration, overall, by eligible group and by time since immunization; and NNI and number of hospitalizations and ICU admissions averted.

Methods

Study design and population

We estimated nirsevimab effectiveness using a test-negative design. The study population included all Quebec children eligible for nirsevimab immunization, assessed for RSV by nucleic acid amplification test (NAAT) during the 2024–25 RSV season (from October 1st, 2024 (week 2024–40) to March 31st, 2025 (2025–14)), and having an ER consultation or hospitalization between 7 days after and 2 days before testing. In Quebec there is no provincial policy guiding RSV testing in emergency departments, which is usually performed at the clinician's discretion. Children born during the 2024–25 RSV season (at-birth group) were eligible as were the following catch-up groups by category and age at the start of the RSV season: (1) healthy full-term children <6 months old; (2) children <8 months old born prematurely; (3) children <8 months old living in the remote regions of Nunavik and Terres-Cries-de-la-Baie-James; and (4) children <19 months old with one of the following chronic conditions increasing the risk of severe RSV disease: bronchopulmonary dysplasia, chronic lung disease, congenital heart disease or cardiomyopathy, pulmonary hypertension, Down syndrome, cystic fibrosis, neuromuscular disorder or congenital anomaly that impairs clearance of upper airway secretions, and bone marrow, stem cell, or solid organ transplantation.¹¹ Comorbidities were identified using ICD-10 codes specified in the main or secondary diagnosis of any hospitalization or ER consultation since birth (Appendix p 2). Prematurity was identified by the encoded duration of pregnancy in the birth hospitalization and the ICD-10 code P07·3. RSVpreF vaccine was authorized for use in pregnancy in Canada in January 2024 but was not publicly funded in Quebec during the 2024–25 season.¹⁴ Regardless, we excluded any children born to mothers vaccinated with RSVpreF vaccine. We also excluded children who received nirsevimab <7 days before consultation or admission, despite the immediate protection conferred by passive immunization, to account for the RSV incubation period and delay between symptom onset and consultation, and those without the health insurance number necessary for data linkage, usually representing infants born near the date of data extraction or recently arrived in Quebec. We conducted NNI analyses using the most recent RSV season (from October 1st, 2023 (week 2023–40) to March 31st, 2024 (2024–13)) preceding the nirsevimab availability in Quebec to estimate RSV hospitalization rates among non-immunized children. As nirsevimab was not available in Québec before fall 2024, we assumed all children during this period were non-immunized. Based on 2024–25 nirsevimab eligibility, we estimated NNI for children born during the 2023–24 season (at-birth group) or <6 months of age on October 1st, 2023 (universally eligible catch-up group). Hospitalization rates could not be estimated for catch-up groups based on high-risk conditions.

The research proposal was examined by the Research Ethics Board of the Centre Hospitalier Universitaire de Québec-Université Laval and considered a program evaluation exempt from ethics review.

Data sources

For effectiveness analyses we used the health insurance number to combine four administrative databases: the Quebec provincial immunization registry; the Quebec sentinel laboratory registry; the common ER database; and the administrative hospitalization database. The immunization registry includes all Quebec residents and all administered immunizations, with each child being linked to the mother. It was the source for demographic variables including biological sex. The laboratory registry includes approximately 80 laboratories in Quebec contributing RSV NAAT results, but RSV-subtyping data was not available. The ER and hospitalization databases contain administrative and clinical information, including reason for ER consultation and ICD-10 codes for discharge diagnosis. By May 30, 2025, data extraction date, information on discharge diagnosis was complete for 98% of ER consultations and 75% of hospitalizations. Participants were included in the primary analysis regardless data completeness on discharge diagnosis.

For NNI analyses, we used: the administrative hospitalization database; the Quebec sentinel surveillance network of six hospitals that conduct systematic testing of hospital admissions for acute respiratory infection using a multiplex PCR panel for respiratory viruses¹⁵; and birth data from the Institut de la statistique du Québec.¹⁶

Exposure and outcome definitions

Immunized children received nirsevimab ≥7 days before ER consultation or hospitalization.

For each outcome, cases were ER consultations not leading to hospitalization or hospitalizations lasting >24 h with RSV NAAT-positive result between 7 days before and 2 days after consultation. ICU admissions were defined as RSV-associated if occurring during an RSV-associated hospitalization. Controls were NAAT-negative for RSV 7 days before and 2 days after an ER consultation or a hospitalization, respectively. First post-test ER consultation (without hospitalization) was included, although hospitalization could occur during subsequent ER visits. Our main outcome definition was based only upon NAAT results and date of admission, regardless of the clinical diagnosis at discharge in order to include participants with incomplete discharge diagnostic data. We performed sensitivity analyses restricted to NAAT-positive RSV consultations or hospitalizations with reason for ER consultation or main discharge diagnosis being a respiratory presentation, or bronchiolitis/bronchitis (Appendix p 3).

Effectiveness analyses

In univariate analyses, standardized differences in means of proportions of covariates were calculated by nirsevimab immunization status. For each outcome, multivariate logistic regression models compared the odds of immunization between test-positive cases and test-negative controls. Under the standard test-negative design assumption that nirsevimab has no impact on other respiratory infections, the estimated odds ratios are equivalent to risk ratios.¹⁷ Models were adjusted for a-priori defined risk factors for severe RSV outcomes such as sex, age in months at testing (<6, 6–11, 12–17 and ≥18), region of residence (seven categories), social and material deprivation indices (three categories), presence of any comorbidity (dichotomous variable), prematurity (dichotomous variable), and calendar time (two-week periods of admission). Other potential risk factors, such as race/ethnicity or low birth weight were not available. Each child could contribute several ER consultations or hospitalizations during the study period if associated with negative RSV NAAT but was censored after the first positive test. The same hospitalized test-negative controls were used to estimate effectiveness against hospitalization and ICU admission. Nirsevimab effectiveness (%) and 95% confidence intervals (95% CI) were calculated as $\left(1-adjustedoddsratio\right)\times 100$. Stratified analyses evaluated effectiveness separately for each eligible group (healthy term, preterm, with chronic conditions or living in remote regions), and for universally eligible at-birth and catch-up groups, and for sex without statistical comparison across groups. Trends of effectiveness by time (four-week intervals) since nirsevimab administration were statistically tested using a linear contrast with coefficients reflecting the ordinal spacing of the levels.

Statistical analyses to estimate NNI and hospitalizations averted

We used administrative hospitalization data to estimate the number of respiratory and bronchiolitis admissions, as previously defined, among all newborns during the 2023–24 RSV season and infants <6 months in October 2023. We inferred the number of RSV-attributable hospitalizations by applying the RSV positivity among respiratory or bronchiolitis hospitalizations within the Quebec hospital-based surveillance network during the 2023–24 season by the specified age groups. Using the proportion of ICU admissions among RSV-positive hospitalizations reported by the Quebec surveillance network, we estimated the number of RSV-attributable ICU admissions, defined as admissions occurring as a direct consequence of RSV infection. We calculated the RSV hospitalization rate (per 100,000 infants) for each age group and diagnosis group (respiratory admissions or bronchiolitis) as $\frac{Numberofhospitalizations\times RSV\%positivity}{Quebecinfantpopulation}\times 100,000$. We calculated NNI as $\frac{1}{RSVhospitalizationrate\left(amongnon-immunized\right)\times nirsevimabeffectiveness}$, where rates and effectiveness are parameterized as proportions, and nirsevimab effectiveness is based on our estimations and 95% CIs. Nirsevimab coverage was defined as the proportion of test-negative participants who had received nirsevimab in 2024–25. We estimated the number and proportion of averted hospitalizations among at-birth and catch-up groups considering two theoretical situations: catch-up immunization completed by November 1st, 2023 (week 2023–44), or, in sensitivity analyses, by October 1st, 2023 (2023–40). Similar calculations were used to estimate NNI to avert one ICU admission.

Statistical analyses were done using SAS (version 9·4)

Role of the funding source

The funder of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report.

Results

Population

During the study period, 13,306 specimens between October 1st, 2024 (week 2024–40) and March 31st, 2025 (2025–14), collected from newborns or children <19 months old in October 2024, were subjected to RSV NAAT (Fig. 1). Of these, 4645 and 2008 tests performed among nirsevimab-eligible children were associated, respectively, with an ER consultation or a hospitalization. After exclusion of repeat tests performed for a single admission (2·4% and 6·4%, respectively) and admissions occurring within 7 days of nirsevimab administration (0·3% and 0·1%, respectively), 3172 RSV-tested specimens (from 2966 children) were included in the estimation of nirsevimab effectiveness against ER consultations (without hospitalization) and 1758 tests (from 1583 children) against hospitalizations. Most children contributed to a single specimen (2782, 93·8% and 1450, 91·6%, respectively, for each analysis), while 184 (6·2%) and 133 (8·4%), respectively, contributed to 390 and 308 specimens (Appendix p 4).

Fig. 1.

Population flowchart. Abbreviations: ER, emergency room; NAAT, nucleic acid amplification tests; RSV, respiratory syncytial virus.

Over half of test-participants were male (2763 [56·0%] of 4930), and aged <6 months at RSV testing (2700 [54·8%]). Standardized differences comparing immunized versus non-immunized children showed imbalance (absolute value > 0·20) in age, presence of comorbidities and month of testing (Appendix p5). Most ER consultations (2339 [71·3%]) and hospitalizations (1052 [59·8%]) occurred in healthy-term children (Table 1). While most participants belonged to the catch-up group, the at-birth group showed a higher proportion of test-negative controls (e.g. among tests performed in ER, 92·4% in the at-birth group versus 75·3% in the catch-up group were negative), likely reflecting a greater tendency for testing during the first months of life.

Table 1.

Characteristics of children included in the effectiveness analyses.

	ER consultations		Hospitalizations		ICU admissions
	Test-positive	Test-negative	Test-positive	Test-negativea	Test-positive
	N (%)	N (%)	N (%)	N (%)	N (%)
N	668	2504	549	1209	34
Nirsevimab status
Immunized	69 (10·3)	1265 (50·5)	55 (10·0)	627 (51·9)	3 (8·8)
Days since immunization, median (IQR)	53 (34–77)	71 (49–98)	49 (31–76)	67 (40–97)	41 (40–49)
Sex
Female	293 (43·9)	1104 (44·1)	208 (37·9)	562 (46·5)	12 (35·3)
Male	375 (56·1)	1400 (55·9)	341 (62·1)	647 (53·5)	22 (64·7)
Age at testing, months
Median (IQR)	5·7 (3·3–7·5)	5·4 (2·8–8·5)	4·8 (2·8–7·6)	5·7 (3·1–9·1)	3·6 (2·1–6·0)
<6 months	355 (53·1)	1383 (55·2)	339 (61·7)	623 (51·5)	25 (73·5)
6–11 months	263 (39·4)	949 (37·9)	160 (29·1)	420 (34·7)	5 (14·7)
12–17 months	47 (7·0)	139 (5·6)	45 (8·2)	124 (10·3)	3 (8·8)
18+ months	3 (0·4)	33 (1·3)	5 (0·9)	42 (3·5)	1 (2·9)
Age on October 1st, 2024, months
0 (born after)	51 (7·6)	623 (24·9)	48 (8·7)	250 (20·7)	1 (2·9)
0–<6 months	514 (76·9)	1577 (63·0)	406 (74·0)	706 (58·4)	27 (79·4)
6–18 months	103 (15·4)	304 (12·1)	95 (17·3)	253 (20·9)	6 (17·6)
Delay test – admission, days, median (IQR)	0 (0–0)	0 (0–0)	0 (−1 to 0)	0 (−1 to 0)	0 (−1 to 0)
Eligible groups, days since immunizationb
Healthy-term children	476 (71·3)	1863 (74·4)	351 (63·9)	701 (58·0)	23 (67·6)
Days, median (IQR)	54 (34–77)	69 (49–94)	38 (23–74)	64 (42–91)	40 (40–55)
Children with comorbidities	135 (20·2)	450 (18·0)	125 (22·8)	402 (33·3)	9 (26·5)
Days, median (IQR)	45 (16–70)	75·5 (46–108)	52 (34–82)	68 (35–103)	42 (42–42)
Preterm children	65 (9·7)	248 (9·9)	90 (16·4)	184 (15·2)	4 (11·8)
Days, median (IQR)	45 (42–55)	81 (51–115)	50 (40·5–65·5)	74 (38–112)	NE
Children of remote regions	1 (0·1)	4 (0·2)	3 (0·5)	16 (1·3)	0 (0·0)
Days, median (IQR)	NE	43 (7–51)	163 (163–163)	57 (22–73)	NE
Main diagnosis
Missing	9 (1·3)	44 (1·8)	75 (13·7)	432 (35·7)	0 (0·0)
Respiratory diagnosisc	655 (98·1)	2336 (93·3)	451 (82·1)	491 (40·6)	32 (94·1)
Bronchiolitis/bronchitis diagnosis	196 (29·3)	169 (6·7)	383 (69·8)	166 (13·7)	26 (76·5)

Abbreviations: ER, emergency room; ICU, intensive care unit; IQR, interquartile range; NE, not estimable.

Note: Numbers are n and column percentages except otherwise specified.

^aTest-negative hospitalizations were controls for both hospitalization and ICU effectiveness analyses.

^bGroups of children with comorbidities and preterm children are not mutually exclusive (100 hospitalizations and 139 ER consultations belong to both eligible subpopulations).

^cFor ER consultations, respiratory diagnosis is based on reason for consultation and discharge diagnosis.

Among nirsevimab-eligible children during the 2024–25 season, RSV-positive NAAT included 668 ER consultations without hospitalization, 549 hospitalizations and 34 ICU admissions (Table 1). Males accounted for a higher proportion of ER consultations (375 [56·1%]) or hospitalizations (341 [62·1%]) compared with females. Most had a respiratory diagnosis (98%, 82% and 94%, respectively, for each outcome), while bronchiolitis/bronchitis was the main diagnosis for 70% of hospitalizations and 77% of ICU admissions. RSV-associated ER consultations and hospitalizations increased steadily starting in week 2024–40 and peaked at weeks 2024–46 to 2025–01, in line with Quebec sentinel laboratory network RSV positivity and with a similar profile in 2023–24 (Fig. 2). The number of admissions among infants from remote regions was insufficient to estimate nirsevimab effectiveness.

Fig. 2.

RSV circulation in Quebec, nirsevimab vaccination and RSV cases during the 2024–25 season. (a) RSV positivity in Quebec sentinel laboratory network, (b) doses of nirsevimab administered, (c) proportion of controls immunized by eligibility criteria, and (d) number of severe RSV outcomes. Abbreviations: ER, emergency room; ICU, intensive care unit; M, month; RSV, respiratory syncytial virus.

Due to a supply shortage in October 2024, only high-risk infants were prioritized that month. Most doses of nirsevimab were administered in November and December 2024 (Fig. 2). Among test-negative children, the proportion immunized with nirsevimab by the end of the season was 79·2% for the at-birth group, 64·4% for the catch-up group, 66·8% for healthy-term children, 51·1% for children with chronic diseases and 76·3% for preterm children. Among ER consultations, 50·5% (n = 1265) of controls versus 10·3% (n = 69) of cases were immunized with nirsevimab at least 7 days before testing (Table 1). Among hospitalizations, 51·9% (n = 627) of controls versus 10·0% (n = 55) of cases were so immunized. Median time since immunization was longer for high-risk (52 days for children with chronic diseases) than for healthy-term children (38 days).

Nirsevimab effectiveness

Nirsevimab effectiveness overall was 86% (95% CI: 82–90) against ER consultation without hospitalization, 89% (95% CI: 84–92) against hospitalization and 88% (95% CI: 58–97) against ICU admission (Fig. 3, Appendix pp 7–9). Protection against ER consultations ranged from 86% to 89% regardless of diagnosis (any, respiratory or bronchiolitis admission) (Fig. 3). Similar effectiveness were estimated when stratified by age, sex or eligible group, except for slightly lower estimate of 83% (95% CI: 65–92) for the at-birth group and higher at 97% (95% CI: 91–99) for preterm children, albeit with overlapping CIs (Appendix p 7). Effectiveness against hospitalizations ranged from 88% to 93% for all groups except the at-birth group (84%; 95% CI: 58–94) and children with chronic diseases (79%; 95% CI: 60–89) (Fig. 3, Appendix p 9). Effectiveness against hospitalizations was similar when restricted to respiratory diagnosis, suggesting limited misclassification when defining RSV-hospitalization only by a positive NAAT. Few ICU admissions precluded meaningful stratification (Appendix p 11). Point estimates of effectiveness against ER consultations and hospitalizations were lower for those immunized ≥12 weeks prior, but with wide and overlapping 95% CIs following the drop in RSV infections in February 2025 (Figs. 2 and 3). Trend tests of estimates stratified by time since immunization were not significant for any of the outcomes (Fig. 3).

Fig. 3.

Nirsevimab effectiveness against severe outcomes by outcome definition, eligible subpopulation and time since immunization. (a) Effectiveness against RSV+ emergency room consultations, (b) effectiveness against RSV+ hospitalizations, and (c) effectiveness against ICU admissions during RSV+ hospitalizations. Note 1: P-values for trend of estimates by time since immunization were p = 0·12 for hospitalizations and p = 0·93 for ER consultations. Note 2: Panel c shows only the estimates that could be calculated considering sample size. Abbreviations: Bronch dx, bronchiolitis or bronchitis diagnosis; CI, confidence interval; ChrDis, children with chronic diseases; ER, emergency room; ICU, intensive care unit; M, month; Nborn, newborn; Preterm, preterm children; Respirat dx, respiratory diagnosis; RSV, respiratory syncytial virus; W, week.

NNI and averted hospitalizations/ICU admissions

Based on the 2023–24 RSV season, between 41 and 59 nirsevimab at-birth immunizations would be required to avert one RSV hospitalization or one RSV hospitalization with bronchiolitis/bronchitis diagnosis, respectively (Table 2). Among the catch-up group, corresponding NNIs were 76 and 79 if immunization started in November, as per the Quebec 2024–25 roll-out. Assuming nirsevimab coverage of 79% and 64% within the at-birth and catch-up groups, respectively (as per the 2024–25 season) was reached by 1st November, we estimate that 734 (72%) and 342 (45%) hospitalizations, respectively, could have been prevented within the 2023–24 scenario. Increasing nirsevimab coverage to 90% by 1st October would avert up to 82% of hospitalizations in both subpopulations. In sensitivity analyses, if the catch-up group was immunized by 1st October, the NNI would decrease from 76 to 58, consequently increasing averted hospitalizations from to 45% to 59%, while minimal changes were observed in the at-birth group (Appendix p 12).

Table 2.

Number needed to immunize to prevent an RSV hospitalization and an ICU admission and potentially averted severe outcomes.

Parameter	Data sources and calculationsa	Birth cohort
		Born during 2023–24 season		Aged <6 months on October 1st, 2023
		Respiratory diagnosis	Bronchiolitis/bronchitis	Respiratory diagnosis	Bronchiolitis/bronchitis
Number of respiratory hospitalizations (Nov–Mar)	Quebec hospitalization administrative database	1548	851	1150	803
RSV positivity (%)	Quebec hospital-based surveillance network	65·4	81·3	50·4	66·5
Proportion of RSV hospitalizations with ICU admission (%)	Quebec hospital-based surveillance network	8·0	6·6	10·8	11·3
RSV hospitalizations	Nb of hospitalizations ∗ RSV positivity	1012	692	580	534
RSV ICU admissions	RSV hospitalizations ∗ % of ICU admissions	81	46	62	60
Births cohort (by eligible subpopulation)	Institut de la statistique du Québec	38,250	38,250	40,050	40,050
RSV hospitalization rate per 100,000	RSV hospitalizations/birth cohort	2646·8	1808·8	1447·2	1333·3
RSV ICU admission rate per 100,000	RSV ICU admissions/birth cohort	211·7	119·4	155·9	150·7
Effectiveness against hospitalization (95% CI)	Quebec 2024–25 effectiveness	91·5 (85·7–95·0)	95·4 (89·4–98·0)	91·5 (85·7–95·0)	95·4 (89·4–98·0)
Effectiveness against ICU admission (95% CI)	Quebec 2024–25 effectiveness	83·9 (39·3–95·7)	83·9 (39·3–95·7)	83·9 (39·3–95·7)	83·9 (39·3–95·7)
Nirsevimab coverage (%)	2024–25 immunization among test-neg controls	79·2	79·2	64·4	64·4
NNI to avert one hospitalizationb	1/(hospitalization rate ∗ effectiveness and 95%CI)	41 (40–44)	59 (56–62)	76 (73–81)	79 (77–84)
(Potentially) averted hospitalizationsc	(Coverage ∗ Birth cohort)/NNI	734	518	342	325
Proportion of hospitalizations averted (%)	Averted hospitalizations/RSV hospitalizations Nov–March	72·5	74·9	58·9	61·4
	Averted hospitalizations/RSV hospitalizations Oct–Marchd	71·7	73·8	44·9	48·4
NNI to avert one ICU admission	1/(ICU admission rate ∗ effectiveness and 95%CI)	563 (493–1202)	998 (875–2131)	765 (670–1633)	791 (694–1689)
(Potentially) averted ICU admissions	(Coverage ∗ Birth cohort)/NNI	54	30	34	33
Proportion of ICU admissions averted (%)	Averted ICU admissions/RSV ICU admissions Nov–March	66·5	66·5	54·0	54·0
	Averted ICU admissions/RSV ICU admissions Oct–Marchd	65·7	65·5	38·0	38·8

Abbreviations: CI, confidence interval; ICU, intensive care unit; Nb, number; NNI, number needed to immunize; RSV, respiratory syncytial virus; test-neg, test-negative.

^aScenario considering 2023–24 season between October 1st, 2023, and March 31st, 2024, and infants aged <6 months on October 1st, 2023 (catch-up group), immunized by November 1st.

^bCalculation of NNI as $\frac{1}{RSVhospitalizationrate\left(amongnon-immunized\right)\times nirsevimabeffectiveness}$, where rates and effectiveness are parameterized as proportions. Example: NNI to avert one hospitalization with respiratory diagnosis among infants born during the 2023–24 season = $\frac{1}{0·026468\times 0·915}$ = 41·3.

^cCalculation of averted hospitalizations as $\frac{Nirsevimabcoverage\times birthcohort}{NNItoavertonehospitalization}$, where coverage is parameterized as a proportion. Example: Averted hospitalizations with respiratory diagnosis among infants born during the 2023–24 season = $\frac{0·792\times 38250}{41·3}$ = 733·5.

^dProportion of averted hospitalizations or ICU admissions considering the number of hospitalizations or ICU admissions from October 1st, 2023, to March 31st, 2024, calculated in the Appendix p7.

Compared to hospitalization, about ten times as many nirsevimab immunizations are required to prevent one ICU admission at 563 and 765 for the at-birth and catch-up groups, respectively. However, based upon 2024–25 nirsevimab coverage, up to 60% of ICU admissions could be prevented and increasing to 90% coverage could prevent up to 76% (Table 2).

Discussion

In this population-based evaluation of universal nirsevimab immunization from Canada, we report substantial protection with effectiveness exceeding 85% against RSV-associated ER consultation, hospitalization and ICU admission. We estimate between 41 and 76 immunizations among newborns during the season and children <6 months at the start of the season, respectively, to prevent one hospitalization. Given such high effectiveness, ultimate program impact depends upon the level and timeliness of nirsevimab coverage achieved. We estimate two-thirds of RSV hospitalizations and 60% of ICU admissions could be prevented based upon 2024–25 coverage in Quebec and catch-up immunization completed by November, but increasing the latter to 90% could prevent more than three-quarters of both outcomes. Conversely, delaying catch-up immunization by even a month would decrease averted hospitalizations by about 15%.

Our observed effectiveness of 89% against RSV-associated hospitalization is consistent with published data from observational studies and from randomized controlled trials (RCT). A systematic review summarizing data from the USA, France, Italy, Luxembourg and Spain reported a pooled estimate of nirsevimab effectiveness of 83% against hospitalization,⁵ whereas pooled efficacy from RCTs is 81%.¹⁸ Our estimated effectiveness of 88% against RSV-associated ICU admissions and 86% against RSV-associated ER consultations also aligns with published pooled effectiveness (81%) and efficacy (90%) data.^5,18 These results suggest potentially greater protection against hospitalization offered by nirsevimab versus maternal RSVpreF vaccination, with effectiveness of the latter estimated in the UK and Argentina to be 58% and 71%, respectively.^19,20 Although preterm infants were included in RCTs,²¹ this is one of the few studies to report nirsevimab effectiveness stratified for preterm infants (93%) and other high-risk groups such as children with chronic diseases (79%).⁶

Data on the duration of nirsevimab-induced protection are sparse. A few studies to date have suggested progressive decline over the first 5 months after administration.^22,23 A systematic review reported decrease in effectiveness with the study follow-up time, lower if ≥ 150 days.²⁴ Our results suggest that nirsevimab effectiveness is maintained over the course of a season, or at least during the typical peak period in the northern hemisphere, but we also observed lower point estimates beyond four months post-administration. The low number of events during the last 10 weeks of the study period led to imprecision in effectiveness estimates among children immunized 12–20 weeks earlier. Bias due to the depletion of susceptibles could also explain an apparent pattern of waning: given high infant attack rates (exceeding 50%),²⁵ the proportion of non-immunized children being protected through infection-induced immunity will increase over the course of the season. If protection waning at the tail end of the season is confirmed, this potential downside of early administration should be considered when establishing program launch dates.

RSV exerts important pressure on Canadian healthcare systems during the winter season, with the burden of hospitalizations and ICU admissions being especially high in infants <6 months.²⁶ In this first roll-out season in Quebec, nirsevimab potentially reduced 2024–2025 hospital admissions for RSV-associated respiratory diagnosis by 60% (72% in the at-birth and 45% in the catch-up groups). This resulted in a potential absolute reduction of 1075 respiratory hospitalizations among a birth cohort comprising just 86,000 infants (or 1250 per 100,000 infants). The proportion of averted RSV-associated hospitalizations depends on timing of the immunization campaign and the start of the RSV season, with RSV seasonality suggesting that, optimally, catch-up immunization should be completed by October for maximal impact. Our estimated NNI of 41 at-birth and 76 catch-up immunizations to prevent one hospitalization are consistent with two Spanish studies that reported NNI of 15 (from October to December) and 41 (from October to March) for at-birth immunization and of 71 and 90 for catch-up immunization.^27,28 Similarly, NNI estimates ranged 25–128 according to a recent review summarizing six studies.²⁹ Previous Canadian studies have reported that a universal infant nirsevimab program would not be cost-effective at the current list price.³⁰ Our findings of higher hospitalization rates and slightly greater effectiveness than assumed in those models indicate improved cost-effectiveness, though this remains highly dependent on nirsevimab's price.

Our study has some limitations. At the time of our analysis, administrative databases for the 2024–25 season were incomplete. Some hospitalizations lacked a documented diagnosis and nearly 20% of recently born infants did not have a unique identifier required for database linkage. Although it was unrelated to nirsevimab administration or RSV testing and should not have compromised the validity of our effectiveness estimates, it reduced the available sample size contributing to imprecision, particularly in the at-birth group. Also, some eligibility criteria for nirsevimab included a notion of clinical impact (i.e. hemodynamically significant chronic heart disease) which could not be adequately captured by ICD codes. Some included children may then have been ineligible for nirsevimab. Due to low numbers of ICU admissions among test-negative controls, we used hospitalized controls to estimate effectiveness against ICU admissions. Although controls with less severe disease may be used when the outcome is rare,³¹ this approach diverges from test-negative design assumptions and could potentially lead to effectiveness underestimation. Delays in catch-up immunization due to supply shortage may result in overestimation of effectiveness if children were infected with RSV before receiving nirsevimab. However, this potential bias was limited in our analysis, as effectiveness was similar between the at-birth versus catch-up groups. The absence of race/ethnicity data prevented stratified analyses, limiting the evaluation of generalizability across diverse ethnic groups. Estimations of NNI and averted hospitalizations depend on the RSV season, the cumulative hospitalization rates and the timing of the immunization campaign. We used the 2023–24 season, as data completeness allowed estimation of provincial rates and nirsevimab was not yet available in Quebec. Although following similar seasonality, Quebec hospitalization surveillance showed that the 2024–25 RSV season was more intense than 2023–24 among age-groups not eligible for RSV immunization.³² Our estimates of averted hospitalizations may therefore be conservative for the at-birth group. On the other hand, we assumed that catch-up immunization was completed two months earlier that in the launch campaign of 2024–25, leading to a potential overestimation of averted hospitalizations. Effectiveness may be affected by the predominance of RSV-subtype circulation that could be different in 2023–24 versus 2024–25, however recent publications suggest similar effectiveness for each subtype.³³ RSV burden may be underestimated when based on administrative databases.³⁴ We overcame this limitation using RSV positivity from a surveillance network with systematic testing for respiratory admissions. NNI to avert one ICU admission may be underestimated because two out of the four tertiary pediatric hospitals in the province, with higher expected proportion of ICU admissions than regional hospitals, participate to the surveillance network, and we did not estimate uncertainty around numbers of cases. Generalizing our estimated averted outcomes based on the 2024–25 campaign would be limited due to the delayed nirsevimab administration during this launch season. Instead, we relied on theoretical assumptions about campaign timing, which should be considered when extrapolating our findings to other jurisdictions.

In conclusion, nirsevimab immunization provides substantial benefit against RSV-associated severe outcomes to both healthy-term infants and high-risk groups. This finding reinforces the impact of a universal program inclusive of healthy-term children, the latter estimated responsible for more than 80% of RSV-associated hospitalizations notwithstanding the highest individual risk among preterm infants and/or those with comorbidity.² Maximizing coverage and impact through early-season program launch must consider end-of-season durability of nirsevimab protection, the latter warranting further investigation.

Contributors

All authors had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. All authors have read and approved the final version of the manuscript. Concept and design: S.C., R.G. and D.T. Acquisition, analysis, or interpretation of data: All authors. Drafting of the manuscript: S.C, J.P. and M.P. Critical revision of the manuscript for important intellectual content: All authors. Statistical analyses: M.O. and S.C. Supervision: J.P. and S.C.

Data sharing statement

Unpublished individual data are not available because the provincial registries used in this study are a property of the “Ministère de la Santé et des Services sociaux du Québec”. Data access to researchers who fulfill local requirements was authorized under the Health and Social Services Information Act. Aggregate data are available within the manuscript and the Supplementary Material.

Declaration of interests

SC and MO report funding from the Ministère de la santé et des services sociaux du Québec to conduct this work, paid to their institution. SC reports funding from the Public Health Agency of Canada for unrelated work, paid to her institution. DMS reports grants from the Public Health Agency of Canada, the Pacific Public Health Foundation, the Canadian Institutes of Health Research and the Michael Smith Foundation for Health Research for unrelated work, paid to her institution. JP reports a grant from Merck for unrelated work, paid to his institution, and consulting fees from Enanta. Other authors have no conflict of interest to declare.