Measles is one of the leading causes of vaccine-preventable childhood morbidity and death, causing an estimated 107 500 deaths in 2023.¹ Measles-related incidence and mortality are particularly high among infants,^2,3 many of whom are below the age of vaccination recommended by national immunisation programmes for measles-containing vaccine (MCV).^2,3 About 12% of reported measles cases between 2013 and 2017 fell into this group of infants,² and cases among young infants have been documented in several outbreaks.^4–6 WHO recommends two doses of MCV for all children. The first dose of MCV (MCV1) is typically recommended at 9 months of age in countries with endemic measles transmission and at 12 months of age in countries with low measles incidence. The second dose (MCV2) is administered at a later age, typically in the second year of life.³ During large measles outbreaks and in other settings of increased circulation or increased risk, WHO recommends using a MCV starting at 6 months of age to protect young infants from measles infection. This dose of MCV is considered a zero dose (MCV0); therefore, the infants still require both regularly scheduled doses and parents are advised that infants who receive an early dose of MCV should continue to receive two doses of MCV according to the national schedule.³

Most previous studies of MCV immunogenicity in infants younger than 9 months used a monovalent measles vaccine in settings with lower vaccination rates and higher rates of circulating measles virus.^7–10 However, almost all countries now use a combined measles–rubella (MR) vaccine, which is sometimes also combined with mumps (MMR) or mumps and varicella (MMRV). Data on the immunogenicity of MR vaccine administered at age younger than 9 months and its effect on immunogenicity of subsequent MR vaccination routinely administered at 9 months are scarce. Furthermore, a higher proportion of infants are now born to mothers whose measles immunity derives from vaccination rather than measles infection. The decline of maternal antibodies below protective levels occurs earlier in infants of vaccinated mothers than in infants of mothers with immunity from measles infection.^11–13 As more countries replace monovalent measles vaccine with MR vaccine in their routine immunisation programmes and infants become increasingly susceptible to measles, the need for additional data on immunogenicity of MR vaccine in younger infants becomes greater.

We aimed to assess the immunogenicity and safety of a first dose of MR vaccine administered at age 6 months and its subsequent effect on the immunogenicity of a second dose of MR vaccine administered at age 9 months.

637 infants and their mothers were screened between March 9, 2017, and March 18, 2018. 17 (3%) participants were found ineligible, and 620 infants were randomly assigned to the two-dose group (n=312) and the one-dose group (n=308; figure 1). Following all exclusions, 301 infants were included in two-dose group and 292 infants were included in the one-dose group for the primary and secondary immunogenicity analyses.

Baseline characteristics measured at age 6 months were similar for infants in the study groups (table 1). The distributions of measles and rubella IgG concentrations in infants at baseline and throughout the study are shown in figure 2 and appendix (pp 1–2). The proportions of mothers with undetectable, detectable, and protective concentrations of measles and rubella IgG were similar between groups at baseline (figure 2; table 1; appendix pp 1–2). We found a strong correlation in infant and maternal IgG concentrations for both measles and rubella (appendix p 3). However, there was limited variation in infant IgG concentrations at age 6 months for both measles and rubella (table 1); and the decay of antibodies in infants from 6 months to 9 months was seen only in rubella antibodies in infants in the one-dose group (vaccination at 9 months) whose mothers were antibody positive at baseline (appendix pp 4–5).

The proportions of infants who seroconverted for measles and rubella are shown in table 2. Seroconversions for measles and rubella after the first dose of MR vaccine at age 6 months (two-dose group) were lower than in infants receiving their first MR vaccine dose at 9 months (one-dose group). However, by final follow up at age 11 months, almost all infants seroconverted for measles and rubella regardless of whether they had received two or one MR vaccine doses.

Median measles IgG and rubella IgG concentrations among infants who seroconverted after receiving each vaccine dose are shown in figure 3. The median measles IgG concentration among infants who seroconverted at 11 months was lower in the one-dose group than in the two-dose group (1311 mIU/mL [IQR 881–1976] vs 1453 mIU/mL [1057–2094]; p=0·029). In the two-dose group, the median rubella IgG antibody concentration among seroconverting infants was higher after the first dose at age 6 months than after two doses (102 IU/mL [IQR 59–166] vs 94 IU/mL [60–137]), although this was not compared statistically.

123 adverse events were observed or reported over the duration of the study; 72 in the two-dose group and 51 in the one-dose group (table 3). There were no statistically significant differences between study groups in severity or outcome of adverse events or their relation to the MR vaccine. 12 (17%) events in the two-dose group and seven (14%) in the one-dose group were severe; most events were mild, resolved without sequelae, and were unrelated to the MR vaccine. The ongoing events were classified as mild, and the death in the one-dose group was deemed unrelated to the MR vaccine. 11 (4%) infants receiving the MR dose at age 6 months experienced adverse symptoms within 48 h, the most prevalent symptom being fever. There were three reports of redness at the site of injection reported within 30 min of vaccination; these were all in infants in the two-dose group receiving their first vaccine at age 6 months. No adverse events were reported within 30 min of vaccination among infants receiving MR vaccine at 9 months.

This study showed the immunogenicity and safety of the MR vaccine administered at age 6 months. For the population enrolled in this trial, less than 5% of infants were seropositive for either measles or rubella antibodies at age 6 months, and 94% of infants vaccinated with MR at age 6 months seroconverted after a single dose, with minimal adverse events. Although seroconversion after a first dose at age 6 months (94%) was lower than when a first dose was given at age 9 months (100%), seroconversion at 11 months was equivalent (99% in the two-dose group vs 100% in the one-dose group). Nearly all infants had seroconverted by age 11 months, whether they received one or two doses. We detected a significant difference in rubella antibody concentrations at 11 months between the two-dose and one-dose groups; however, the clinical significance is uncertain because the median antibody concentrations exceeded clinical protection thresholds in both groups. The MR vaccine was safe, both when administered early at 6 months, before routine vaccination, and at the age of the routine first dose (9 months).

The immunogenicity of measles vaccine depends on several factors, including the presence and potential interference of maternally derived antibodies and the maturity of the immune system of the vaccine recipient. A systematic review of previous studies reported a higher immunogenicity of measles vaccine when administered at a later age, with 76% seroconversion at age 6 months.¹⁰ Although some studies have directly demonstrated a negative association between maternal antibody concentrations and vaccine immunogenicity,^7,16 declining maternal antibody concentrations in the vaccine era might limit such concerns.

A history of vaccination or measles or rubella infection among the mothers enrolled in this trial were not solicited due to challenges with verification. More than 80% of mothers in both study groups had seroprotective concentrations of measles and rubella antibodies at baseline. It is likely that for most of the mothers, measles seroprotection derived from vaccination rather than previous infection. Bangladesh introduced MCV1 into the EPI schedule in 1989, with periodic supplementary immunisation activities since that time.¹⁷ MCV2 was introduced in 2012. Bangladesh’s estimated coverage was 97% for MCV1 and 93% for MCV2 in 2018 according to WHO and UNICEF estimates of national immunisation coverage (WUENIC).¹⁸ Measles has been well controlled in Bangladesh over the 30 years before the enrolment period for this study. Maternal rubella seroprotection could have derived from vaccination or infection; a nationwide MR catch-up campaign was conducted in 2014 with a target age of 9 months–15 years.¹⁹ The epidemiological context seen in Bangladesh is representative of many countries where measles and rubella are eliminated or well controlled.

Although most mothers were seropositive for both measles and rubella antibodies in this study population, infants at age 6 months had low concentrations of maternally derived antibody concentrations. Low levels of seroprotection at age 6 months have been documented previously.^20–23 Results of one study in Matlab, reported in 1998, showed that 95% of infants were seronegative by age 8 months.²⁴ This observation is consistent with the more rapid decline of maternal antibodies in infants born to mothers with vaccine-induced immunity.^11,12 As the number of infants born to mothers with vaccine-induced immunity continues to increase, pockets of susceptibility to measles and rubella will continue to accumulate among the vulnerable age group of infants too young to be vaccinated and therefore ineligible for the MR vaccine, posing a challenge for global immunisation programmes, particularly those still endemic for measles with ongoing transmission.

Measles and rubella seroconversion rates after a first dose of MR vaccine at age 6 months were unexpectedly high among the infants enrolled in the two-dose group. This seropositivity appears to be sustained after receiving a second dose at age 9 months, giving no indication of immunological blunting. Although few studies have investigated immunological blunting, most have shown high seropositivity and vaccine effectiveness when MCV doses are administered early.²⁵ However, some studies have suggested longer-term effects on the durability of the immune response, which need to be further investigated. For example, studies from the Netherlands have showed that early receipt of MCV at age 6–12 months results in lower measles antibody concentrations after subsequent doses.^13,26 One such study showed that 11% of children receiving MCV before age 9 months ultimately lost protective immunity to measles later in life.²⁶ Our study does not assess immunogenicity across a longer observation period because no long-term follow-up was planned.

The findings of this trial support the use of a supplementary dose of the MR vaccine for infants younger than 9 months during periods of intense transmission.³ However, the trial was conducted in a setting with high vaccination rates among infants. Results might therefore differ in settings where vaccination rates are lower or with different measles and rubella epidemiology. Infants are more likely to receive a higher concentration of passively transferred antibodies from mothers who have been infected, and these might interact with the MR vaccine to reduce immunogenicity. Further studies in similar and different environments are needed to ensure the robustness of the findings. Future studies might also assess cell-mediated immunity markers and long-term humoral immunity. Any considerations to vaccinate infants routinely at an age younger than 9 months should balance the potential risk of primary vaccine failure and long-term effects on the immune response against the risk of measles infection and measles-related complications, including death, during periods of intense transmission.

Our study had some limitations. First, the study only followed up participants up to age 11 months, and therefore we were unable to assess the longer-term effects of MR vaccination at 6 months on the routine second dose of MCV at age 9 months and beyond. Second, data were collected in 2017–18, but delays in transfer and testing of samples and the COVID-19 pandemic significantly delayed data analysis and dissemination. Further limitations are associated with the open-label study design, such as potential bias in reporting of adverse events, which could affect interpretation of results. However, the laboratory and data analysts were not involved in implementation of the trial and did not have access to the data until the trial had ended, which might have reduced biases.

In summary, the MR vaccine was well tolerated and had high immunogenicity in infants aged 6 months. At age 11 months, after receiving their second dose of the MR vaccine at age 9 months, their immunogenicity was equivalent to those infants aged 11 months who received their first MR vaccine at age 9 months. We thus recommend additional studies to evaluate antibody concentrations over time to assess longer-term immunity based on age at vaccination.