Population immunity to measles in Canada using Canadian Health Measures survey data - A Canadian Immunization Research Network (CIRN) study.

We aimed to determine population immunity to measles in Canada, and to assess the risk of future outbreaks. We tested 11,176 sera from Cycles 2 (2009-2011) and 3 (2011-2013) cohorts from the biobank of Statistics Canada's Canadian Health Measures Survey (CHMS) using the BioPlex 2220 MMRV IgG assay. We then tested all BioPlex negative and equivocal samples using a more sensitive Plaque Reduction Neutralization Test (PRNT). We determined the weighted proportion of positive, equivocal, and negative samples by age, sex, region and whether individuals were born in Canada. We found that 90.0% (95% confidence interval (CI): 88.2, 91.9) of samples were positive, 4.5% (95% CI: 3.4, 5.5) were equivocal and 5.5% (95% CI: 4.3, 6.7) were negative. Individuals in the 12-19 year age band had the lowest proportion positive at 78.7% (95% CI: 74.2, 83.2) and the highest proportion of positive samples was found in those 60-79 years (99.6%, 95% CI: 99.3, 99.9). Seropositivity was consistently <90% across a broad range of pediatric and adult age bands (6-39 years). We found that a slightly higher proportion of females were positive (91.9%, 95% CI: 90.1, 93.6) compared to males (88.3%, 95% CI: 85.8, 90.7). When taking into account interaction between age and born in Canada status, we found individuals born in Canada aged 19 and under were less susceptible (OR = 0.6 (95% CI: 0.4, 0.95)) compared to those born outside Canada whereas, those aged 20 and over were more susceptible (OR = 1.7 (95% CI: 1.1, 2.8)). Our findings indicate that measles immunity in Canada is below the 95% immunity threshold required to sustain measles elimination, underscoring the importance of maintaining high vaccine coverage to prevent future measles outbreaks and sustain Canada's elimination status.

Introduction

Measles is one of the most infectious diseases in humans, with a basic reproductive number (R0) generally reported as 12–18 [1]. Canada eliminated measles in 1998 [2], with elimination defined as an interruption of endemic measles transmission for ≥12 months [3]. Since it is so infectious, a level of population immunity of 90–95% [4], [5] is thought to be required to meet the herd immunity threshold to achieve and sustain measles elimination.

Measles vaccine was approved in Canada in 1963 [6]. Measles vaccination programs were implemented in the 1970s, with varying dates of introduction by province/territory [7]. A second dose of a measles containing vaccine was introduced in 1996 [6]. In most jurisdictions, vaccination programs for measles over the last 40 years have used combined formulations of measles-mumps-rubella (MMR) and more recently measles-mumps-rubella-varicella (MMRV) [2]. Children receive their first dose at 12 months, and their second dose at either 18 months, 36 months, or 4–6 years depending on the provincial or territorial immunization schedule [8]. Despite vaccine approval, measles continued to circulate in Canada during the 1960s so individuals born before 1970 are considered to be immune through previous infection [2].

Despite Canada’s sustained elimination status, measles vaccine coverage estimates in children fall below the Canadian target of 95% coverage for all recommended childhood vaccines [9]. The latest published survey reported that in 2017, 90.2% of two year olds in Canada had at least one dose of measles containing vaccine and 87.0% of 7-year olds had at least two doses [10]. During this same period, there have been increasing reports of measles importation into Canada from abroad [11], [12], causing local outbreaks. Furthermore, there is now compelling evidence that vaccine mediated immunity is less robust than immunity from previous infection [13] and can wane over time [14], [15], [16], [17], [18]. Waning immunity (i.e. secondary vaccine failure) is particularly problematic in settings that have eliminated measles [14], [15], [16], [17], [18] due to the lack of immunological boosting from circulating virus. The combination of sub-optimal vaccine uptake, waning vaccine-mediated immunity and ongoing importations from measles endemic jurisdictions abroad presents a critical challenge to sustaining measles control. In this context, the objective of this study was to assess measles population immunity in Canada using sero-epidemiology on specimens collected through Statistics Canada’s Canadian Health Measures Survey (CHMS), in order to assess the risk of future sustained outbreaks and loss of elimination status.

Methods

Sera source and accompanying data

Serum samples used in this study were collected as part of the CHMS, which is a cross-sectional, nationally representative survey conducted by Statistics Canada. The CHMS utilized a multistage sampling strategy, using the Canadian Labour Force Survey as a sampling frame, which uses clusters that were combined to construct collection sites [19], [20]. The CHMS collects health data from individuals aged 3–79 years old, through household interviews and physical measurements as well as biological specimen collection [21], [22]. The survey did not collect data on individuals living on reserves and other Indigenous settlements, individuals who serve full time in the Canadian armed forces, institutionalized persons, and those living in specific remote regions [19], [20]. For this study, we used CHMS samples collected from 2009 to 2011 (Cycle 2) and from 2011 to 2013 (Cycle 3). Survey respondents were first contacted by mail, and underwent a health interview and visited a mobile examination centre to conduct physical measurements and provide biological samples [21], [22]. We tested samples that had sufficient serum volume and for which consent was provided by participants for further testing and research use. We accessed CHMS demographic variables included in the survey, including age group, sex, region in Canada, and whether or not the individual was born in Canada.

Laboratory testing

All testing was performed at the National Microbiology Laboratory (NML) in Winnipeg, Canada. Samples were first tested using the BioPlex 2200 measles, mumps, rubella and varicella-zoster virus (MMRV) IgG assay, which was previously validated by our Network for measles sero-epidemiology work [23]. In our validation study, we determined the cut-offs that would result in a 100% sensitivity (95% CI = 93.5–100%) and 100% specificity (95% CI = 83.2–100%), which were then used in our study [23]. Samples with a BioPlex result ≥ 1.1 AU/mL were considered positive, those 0.13≤ and <1.1 AU/mL were equivocal, and <0.13 AU/mL were negative [23]. We retested samples with negative or equivocal BioPlex results using a Plaque Reduction Neutralization Test (PRNT), which is the gold-standard test for measuring measles immunity [24]. The PRNT was performed at the NML as previously described [25]. Samples with a PRNT result ≥ 192 mIU/mL were considered to be positive, 112≤ and <192 mIU/mL were equivocal, and <112 mIU/mL were considered non-immune [23]. We used PRNT results to determine the final immunity status of the BioPlex negative and equivocal samples.

Statistical analyses

We conducted descriptive and regression analyses to investigate the association between the proportion of samples overall and within various sub-groups, and measles antibody concentrations above a threshold thought to confer protection. We weighted prevalence estimates using survey weights, to estimate seroprevalence in the Canadian population from 2009 to 2013, and used bootstrap weights to estimate their variances [21], [22]. The CHMS corrected for possible non-response bias using weighting adjustments [19], [20]. To assess the quality of an estimate we used the Coefficient of Variation (CV). Estimates with a CV of >16.6 and ≤33.3 were considered to have high sampling variability and estimates with a CV > 33.3 were considered unreliable [19], [20]. We also assessed the characteristics of missing data to determine whether survey respondents who were not tested (i.e. sample wasn’t provided or did not have consent) were statistically different from those that were tested.

We determined the weighted proportions of samples that were above the threshold of protection, as well as negative and equivocal samples by age group, sex, region, and whether individuals were born in Canada. Age was categorized into 6 groups: 3–5, 6–11, 12–19, 20–39, 40–59, and 60–79 years old. Due to sample size considerations, provincial level analysis was feasible only for Ontario and Quebec. Therefore, the region variable was categorized into Ontario, Quebec, and others (Newfoundland and Labrador, Nova Scotia, New Brunswick, Manitoba, Alberta, British Columbia). Weighted geometric mean concentrations (GMC) based on the BioPlex result only (since BioPlex and PRNT units differed we were not able to combine results) were calculated by age and any zero values were imputed using random sampling from a beta-distribution (α = 5, β = 1). To assess whether there was significant difference in GMCs by age, we conducted the Wald test. We performed Chi-square tests to identify whether there were differences in immune (positive samples) and non-immune (negative and equivocal samples) samples, by age group, sex, region, and by whether an individual was born in Canada. We assessed whether there was significant difference in susceptibility to measles infection (defined as samples with final immunity result of negative or equivocal) by sex when adjusting for age by conducting logistic regression. To assess the odds of susceptibility to measles infection by whether an individual was born in Canada, we performed logistic regression analyses un-adjusted and adjusted for age and sex. As per Statistics Canada dissemination guidelines, weighted estimates were rounded to the nearest hundred. We conducted analyses using R versions 3.5.3 and 3.6.2.

We received ethics approval from the Public Health Ontario Ethics Review Board and the University of Manitoba Research Ethics board.

Results

There were 12,180 survey respondents included in Cycles 2 (2009–2011) and 3 (2011–2013) of the CHMS, of whom 12,157 provided blood samples. Of these, samples from 938 individuals were not tested for measles antibody levels for various reasons, including insufficient serum volume and lack of consent for further testing. Results were unavailable for an additional 43 samples from Cycle 2 due to instrument error, bringing the total number of samples tested to 11,176 (weighted N = 29,570,500) (Table 1). We found the highest proportion of missing samples in the youngest age group, females, and individuals born in Canada compared to those born outside Canada (p < 0.01) (Supplemental Table 1).

Table 1

Characteristics of the study population, from the Canadian Health Measures Survey (CHMS) Cycles 2 and 3, 2009–2013.

Variable	Unweighted (N = 11,176)	Weighted (N = 29,570,500)	Proportion of the Canadian population (%) (N = 33,476,688)**
	n (%)	n (%)
Age group (years)
3–5	879 (7.9)	823,100 (2.8)	3.3
6–11	1,811 (16.2)	1,830,100 (6.2)	6.5
12–19	1,901 (17.0)	2,980,300 (10.1)	10.0
20–39	2,269 (20.3)	8,821,800 (29.8)	26.0
40–59	2,241 (20.1)	9,698,200 (32.8)	29.9
60–79	2,075 (18.6)	5,416,900 (18.3)	16.9
Sex
Male	5,513 (49.3)	14,888,800 (50.4)	49.0
Female	5,663 (50.7)	14,681,600 (49.6)	51.0
Region
Ontario	3,733 (33.4)	11,525,000 (39.0)	38.4
Quebec	2,532 (22.7)	6,902,100 (23.3)	23.6
Others*	4,911 (43.9)	11,143,300 (37.7)	34.2
Born in Canada
No	2,224 (19.9)	7,663,800 (26.0)	20.6
Yes	8,950 (80.1)	21,856,200 (74.0)	79.4
NA	<10	–

Others include: Newfoundland and Labrador, Nova Scotia, New Brunswick, Manitoba, Alberta, British Columbia.

Data from Statistics Canada, 2011 Census [48].

The proportions of samples collected by age group, sex, and region are provided in Table 1. The proportion of samples collected from individuals born outside Canada (19.9%) was similar to the foreign-born Canadian population (20.6%) according to the 2011 Statistics Canada Census [26].

All samples were first tested using the BioPlex MMRV IgG assay. Of 11,176 samples, 8,893 (79.6%) were positive, 1,915 (17.1%) were equivocal, and 368 (3.3%) were negative (Table 2). When BioPlex equivocal and negative specimens were tested using PRNT, 970/2,283 (42.5%) were positive, 622/2,283 (27.2%) were equivocal, and 691/2,283 (30.3%) were negative (Table 2). The use of the PRNT testing for non-positive samples increased the proportion of samples above the threshold of protection from 79.6% to 88.3%.

Table 2

BioPlex and PRNT test results for 11,176 samples from CHMS Cycles 2 and 3 (2009–2013).

BioPlex (n)	PRNT (n)
	Positive	Equivocal	Negative	NA	Total, n (%)
Positive	NA	NA	NA	8,893	8,893 (79.6%)
Equivocal	960	587	368	NA	1,915 (17.1%)
Negative	10	35	323	NA	368 (3.3%)
Total n (%)	970 (8.7%)	622 (5.6%)	691 (6.2%)	8,893 (79.6%)	11,176 (100%)

Overall the weighted proportion of samples above the threshold of immunity in cycle 2 (2009–2011) were 90.0% (95% confidence interval (CI): 87.1, 92.8) and 90.4% (95% CI: 88.2, 92.6) in cycle 3 (2011–2013). When the cycles were combined, we found that the weighted proportion of positive samples was 90.0% (95% CI: 88.2, 91.9). A further 4.5% were equivocal (95% CI: 3.4, 5.5), and 5.5% (95% CI: 4.3, 6.7) were negative (Table 3). The proportion of seropositive samples varied by age group (p < 0.0001). The proportion of positive samples in children decreased as age increased, with highest proportion of positive samples in 3–5 years old (92.4%, 95% CI: 89.4, 95.4) and the lowest proportion in the 12–19 years age group (78.7%, 95% CI: 74.2, 83.2). In adults, the proportion of positive samples increased with age, peaking in those aged 60–79 years old (99.6%, 95% CI: 99.3, 99.9). Samples from individuals aged 12–19 years old had the highest proportion of equivocal samples at 11.5% (95% CI: 8.7, 14.3), while adults aged 40 – 79 years had small proportions of equivocal samples with high CV values, thus deemed unreliable results. Assessing antibody concentration, we found that the GMC was lowest in sera from children 12–19 years old, at 1.7 AU/mL (95% CI: 1.4, 1.9), and highest in sera from individuals 60–79 years old, at 19.3 AU/mL (95% CI: 18.2, 20.4), with a significant association between GMC and age (p < 0.0001) (Fig. 1).

Table 3

Weighted proportion of samples that were positive¶, equivocal and negative for measles antibody by age group, sex, region and whether born in Canada.

Variable	% Positive (95% CI)	% Equivocal (95% CI)	% Negative (95% CI)	P values***
Age group (years)
3–5	92.4 (89.4, 95.4)	1.6 (0.5, 2.7)*	6.0 (3.2, 8.8)*	<0.0001
6–11	84.6 (80.2, 88.9)	7.2 (5.1, 9.3)	8.2 (5.2, 11.2)*
12–19	78.7 (74.2, 83.2)	11.5 (8.7, 14.3)	9.7 (6.3, 13.2)*
20–39	83.7 (80.2, 87.2)	7.3 (4.8, 9.8)	9.0 (6.6, 11.4)
40–59	94.8 (92.8, 96.8)	1.8 (0.9, 2.7)*	3.4 (2.0, 4.7)*
60–79	99.6 (99.3, 99.9)	0.2 (0.0, 0.4)**	0.2 (0.0, 0.5)**
Sex
Male	88.3 (85.8, 90.7)	4.9 (3.5, 6.3)	6.8 (5.2, 8.4)	<0.001
Female	91.9 (90.1, 93.6)	4.0 (2.9, 5.1)	4.2 (2.9, 5.4)
Region
Ontario	90.6 (88.2, 93.0)	4.0 (2.3, 5.7)*	5.4 (3.9, 6.9)	0.24
Quebec	92.1 (88.6, 95.6)	4.1 (2.2, 6.1)*	3.8 (1.9, 5.6)*
Others	88.2 (84.5, 91.9)	5.1 (3.3, 6.9)*	6.7 (4.0, 9.4)*
Born in Canada
No	92.9 (90.6, 95.3)	3.4 (1.7, 5.1)*	3.7 (1.8, 5.7)*	0.01
Yes	89.0 (86.9, 91.1)	4.9 (3.6, 6.1)	6.1 (4.9, 7.4)
Total	90.0 (88.2, 91.9)	4.5 (3.4, 5.5)	5.5 (4.3, 6.7)

Interpret with caution due to high sampling variability, coefficient of variation (CV) > 16.6% and ≤ 33.3%.

Interpret with caution due to extreme sampling variability, CV > 33.3%.

Chi2 p-value is assessing statistically significant differences between the proportion of immune (positive samples) and non-immune (negative and equivocal samples) samples.

Final immunity status: All samples were tested in BioPlex. Negative and equivocal samples were re-tested in PRNT.

Fig. 1

BioPlex geometric mean titre by age group (years).

The point estimate of the proportion of samples above the threshold of protection was higher in females (91.9%, 95% CI: 90.1, 93.6) compared to males (88.3%, 95% CI: 85.8, 90.7) (p < 0.001) (Table 3). To assess the effect of age, we modelled susceptibility by age-group and sex, and found that adjusted for age-group, males were 1.5 times (95% CI 1.2, 1.9) more likely to be susceptible than females. We found that male children (3–5, 6–11, and 12–19 years old) had a higher proportion of positive samples compared to females. Conversely, adult females had a higher proportion of positive samples compared to males (Fig. 2).

Fig. 2

Proportion of positive samples by age group (years) and sex, presented along with the measles herd immunity threshold of 95% [5].

The variation in the proportion of positive samples by region was not statistically significant (Table 3) (p = 0.24). The lowest proportion of positive samples was found in the combined category of other regions (88.2%, 95% CI: 84.5, 91.9). In Ontario, 90.6% (95% CI: 88.2, 93.0) of the samples were positive, and the highest proportion of positive samples was found in Quebec (92.1%, 95% CI: 88.6, 95.6).

The point estimate of the proportion of positive samples was higher in those born outside of Canada (92.9%, 95% CI: 90.6, 95.3) compared to those born in Canada (89.0%, 95% CI: 86.9, 91.1) (p = 0.01) (Table 3). We found that samples from children (3–5, 6–11, 12–19 years old) who were born in Canada had a higher proportion of positive samples compared to samples from those that were born outside Canada (Fig. 3). However, in samples from older age groups (20–39 and 40–59 years old) those born outside Canada had a higher proportion of positive samples. For individuals 60–79 years of age, the proportion of samples above the threshold of protection was equal for those born in and outside of Canada, 99.6% (95% CI: 99.1, 100.0) and 99.6% (95% CI: 98.9, 100), respectively. To explore this further, we used logistic regression to compare predicted susceptibility to measles infection in individuals born in Canada compared to those born outside of Canada. Immunity status was categorized into immune (positive samples) and susceptible (negative and equivocal samples). We found that the unadjusted odds of being susceptible to measles infection were 1.6 (95% CI: 1.1, 2.3) times greater for those born in Canada compared to individuals born outside of Canada. When adjusted for age and sex, the odds ratio (OR) did not change substantially but was no longer significant (OR = 1.4 (95% CI: 0.9, 2.0)). As a result we assessed for interaction between age and whether individuals were born in Canada, and found an interaction between these variables (p < 0.01). Individuals aged 19 and under that were born in Canada had a 0.6 (95% CI: 0.4, 0.95) odds of being susceptible, whereas those aged 20 and over had 1.7 (95% CI: 1.1, 2.8) times greater odds of being susceptible when compared to those born outside Canada (Supplementary Table 2).

Fig. 3

Proportion of positive samples by age group (years) and whether the sample was from an individual born in Canada.

Discussion

In the current work, we measured measles population immunity in Canada using samples collected between 2009 and 2013 using a nationally representative cohort from a survey performed by Statistics Canada. We found a weighted population estimate of measles immunity of 90.0%, which is below the 95% required to meet herd immunity [1]. However, 4.5% of the weighted study population were found to be equivocal. Although we do not know whether individuals with equivocal samples are protected from measles, we know that they represent individuals who have either been previously infected, or more likely, vaccinated against measles. Using a combination of serologic and cell-based assays, it has been shown that some individuals with negative or equivocal serology results retain some degree of immune memory for measles antigens [27], [28], [29]. If the samples with positive and equivocal tests results in the current study are combined, the proportion of the population likely to be protected increases to 94.5%, which is very close to herd immunity threshold [1]. By age group, we found seroprevalence ranging between 78.7% and 92.4% in samples from individuals aged 3–39 years old (estimated to be born between 1970–2008 for Cycle 2 and 1972–2010 for Cycle 3), who were eligible for either one or two measles vaccine doses, and much higher seroprevalence of 94.8–99.6% in samples from individuals 40 years and older, who were born before 1970 and are assumed in Canadian guidelines to be immune through previous infection[2].

We observed the highest proportion of measles susceptibility in those 12–19 years of age, with only 78.7% of the samples in this group having antibody levels above the threshold of protection. This age group also had the highest proportion of equivocal samples, at 11.5%.The high proportion of equivocal samples in this age-group likely reflects vaccinated individuals and may suggest waning immunity in this group, since measles immunity through vaccination can result in lower antibody levels than those developed through previous infection [13], [30]. These results are concordant with previous findings related to measles immunity and vaccine failure in Canada. In 2019, 57% of measles cases reported in Canada occurred in individuals 15–24 years of age who had received at least one dose of measles containing vaccine, much higher than the reported rate of two dose vaccine failure from other age-groups [31]. Seroprevalence reported in our study from children aged 12–19 in Ontario (80.6% seropositive, Supplemental Fig. 1) is consistent with a previous study of measles seroprevalence in Ontario (Canada’s most populous province) performed by our group, which found that 79.4% of samples in the 12–19 age-group were positive (Supplemental Fig. 2) [14]. For specimens obtained from Quebec, seroprevalence for this age-group in the current study was 76.5%, and may explain the large measles outbreak that occurred in Quebec in 2011. Of over 700 cases associated with this outbreak, the majority were children age 10–19 years old, of whom 20% had received two doses of measles vaccine [32].

Other jurisdictions that have eliminated measles have also reported lower than expected seroprevalence in younger age groups, suggesting waning immunity. For example, a study from Australia, where measles was eliminated in 2014 [18], compared measles seroprevalence estimates from multiple serosurveys conducted between 1996 and 2013, and found that the proportion of seropositive specimens decreased over time in vaccine-eligible groups. The lowest proportion of positive specimens were observed in 2012–2013 in specimens collected from 10 to 19 year olds. This age-group also exhibited a high proportion of equivocal results [18]. In Korea, which interrupted endemic measles transmission in 2006 [33] and eliminated measles in 2014 [16], [34], serosurveys of vaccine eligible individuals born from 1992 to 2008 were conducted in 2010 and again in 2014, revealing an overall seroprevalence decline of 16.4% between the two surveys. In 2014, only 48.5% of specimens collected in 2014 from youth aged 16–19 were seropositive [16].

Despite the growing evidence suggesting waning immunity in vaccine-eligible individuals in Canada, it should be noted that low vaccine coverage may also contribute to the lower than desired seroprevalence estimates observed in 12–19 year olds. Our findings are concordant with a 2013 national measles vaccine coverage survey conducted in Canada during this time period, which reported that only 83.6% of 17 years old in Canada had received two doses of MMR/MMRV [10]. Assuming field vaccine effectiveness of approximately 94% [35], [36], this coverage estimate is within the seroprevalence estimates found in our study for the 12–19 age group.

We found that individuals 40 years (born between 1969–1971 for Cycle 2 and 1971–1973 for Cycle 3) and older had a population immunity of 95% or greater, reflecting immunity predominantly through previous infection [2]. This is consistent with seroprevalence estimates from this birth cohort in other jurisdictions that have eliminated measles, including Australia, Brazil, Japan, Korea, and the United States [16], [18], [37], [38], [39].

Overall, we found a higher proportion of positive samples in females compared to males, even when adjusting for age-group. When sex differences in measles immunity following vaccination have been assessed, seroprevalence has generally been higher in females or equal among the sexes, with contributing factors possibly including genetic, hormonal, and environmental factors [13], [40]. In addition, postpartum MMR vaccination following prenatal screening for rubella antibodies, could also have contributed to higher measles immunity in females.

Our study found that individuals born in Canada had greater odds of being susceptible to measles compared to those born outside Canada. However, when adjusted for age and sex, the odds ratio was not statistically significant. We found that adults 20–59 years of age who were born outside Canada had a higher proportion of seropositive specimens compared to those born in Canada, while seroprevalence in older individuals was similar between the two groups. Although measles incidence and vaccine coverage rates vary by country [41], we hypothesize that the difference in seroprevalence in those 20–39 years of age relates to a higher burden of measles and lower vaccine coverage in the countries of origin of these individuals. In adults aged 40–79, there was little difference in immunity by place of birth, reflecting childhood infection prior to the availability of vaccine [2]. Conversely, children and young adults (3–19 years old) born outside of Canada had a lower proportion of specimens above the threshold of protection. Despite the existence of clinical guidelines recommending that, in the absence of vaccination records, immigrant children should be vaccinated against measles [42], the lower level of seroprotection in this group could be due to barriers to vaccination access, vaccine hesitancy, or suboptimal vaccine coverage in their countries of origin. These results are consistent with another Canadian seroprevalence study investigating measles, mumps and rubella immunity in recent Montreal immigrants and refugees, which observed that adult newcomers under the age of 35 years were more susceptible to at least one of these diseases compared to recent immigrants and refugees above the age of 35 [43]. There are conflicting study results on whether newcomer children in Canada have lower vaccine coverage. Using data from the from the 2013 Childhood National Immunization Coverage Survey (cNICS) Perinet et al., found that two year old newcomer children were more likely to have vaccination delays for one-dose of MMR/MMRV [44], while Gilbert et al. found that measles non-vaccination was not significantly different between newcomer and two year old children born in Canada [45].

The strengths of our study include the large sample size and the nationally representative set of sera, as well as our reference testing approach [23]. Our study has added knowledge on immunity to measles in those born outside Canada. This is a topic on which there are scant data, despite the fact that 21.9% of Canadian residents are newcomers, rising to 51.1% in Ontario [26]. Therefore, our finding that those born outside Canada have higher seropositivity than those born in Canada is noteworthy, although further work is required to understand the lower seropositivity in newcomer children. The limitations of our study include the lack of vaccination data for study participants, hampering our ability to determine whether gaps in immunity in vaccine-eligible cohorts were due to waning immunity or low vaccine coverage in these samples. Since seroprevalence studies do not measure cellular immunity, it is possible that our estimates underreport measles protection, which have both humoral and cellular immunity components [13].

Our study is the first national measles serosurvey in Canada, and demonstrates that population immunity to measles in Canada may not be sufficient to sustain elimination, particularly with ongoing measles importations from abroad. Even if all equivocal samples in our serosurvey represent protected individuals, pockets of susceptibility in vaccine-eligible cohorts may result in outbreaks, particularly in school and university settings [32]. Furthermore, over time the proportion of individuals in the Canadian population protected only by vaccination will almost certainly continue to increase, as infants are born and are vaccinated while the proportion of those immune through previous infection or combined vaccination and natural exposure will decrease as these older individuals die. This demographic shift has the potential to further lower population immunity. In addition, due to the COVID-19 pandemic, associated disruptions in primary care services and the likely decrease in routine vaccine coverage, future serosurveys are a priority in order to understand additional immunity gaps due to the pandemic [46], [47]. Strategies to maintain high vaccination coverage and adequate immunity in all age groups are a priority to ensure that the chains of transmission following inevitable introductions from abroad remain short so that large measles outbreaks do not occur in the future.

Contributors

NC, SB, AS, TH conceived the study. EM, AS tested samples at NML. SO wrote the manuscript. SO, CI, EJ, JW performed data analyses. All authors reviewed and edited the manuscript.

Funding

This work was supported by the Canadian Immunization Research Network (CIRN).

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.