Measles is a highly contagious disease that requires high population immunity (93–95%) for elimination. Vaccinating all eligible children with two doses of measles vaccine is the standard for national immunisation programmes aimed at eliminating measles, with on-time delivery of the first dose of measles-containing vaccine (MCV) being the highest priority.¹ Rubella is an infectious disease with similar—although milder—symptoms to measles. Rubella virus is highly teratogenic, especially during the first trimester of pregnancy, when it causes congenital rubella syndrome, a constellation of severe birth defects of the eyes, ears, heart, and brain.² In 2005, the second dose of measles vaccine was moved to 18–24 months of age. In 2008, rubella vaccination was included in China’s National Immunisation Programme, with a schedule of measles-rubella vaccine given at 8 months of age and measles-mumps-rubella vaccine given at 18 months of age.

Japanese encephalitis is the most important vaccine-preventable cause of encephalitis in many Asian countries, with an estimated 68000 clinical cases every year, primarily among children.^3,4 Approximately 50% of these cases occur in mainland China. Vaccination is considered essential for Japanese encephalitis control in endemic areas,⁵ and WHO recommends integration of Japanese encephalitis vaccine into national immunisation schedules in all areas where Japanese encephalitis is recognised as a public health priority.⁶ A live, attenuated, SA 14–14–2 Japanese encephalitis vaccine (LJEV) is manufactured in China and has been prequalified by WHO since October, 2013. This vaccine has been used in China and other Asian countries, such as Nepal, South Korea, Sri Lanka, India, Thailand, Cambodia, Vietnam, and Laos.^7,8 China’s National Immunisation Programme has included LJEV since 2008, with one dose given at 8 months and a second dose given at 2 years of age. In China, LJEV and measles-rubella vaccine are recommended for co-administration at 8 months, which is the earliest administration age of these vaccines across the countries using Japanese encephalitis vaccine. Available evidence supports the immunogenicity and safety of co-administration of LJEV with single-antigen measles vaccine among infants, but there are no studies of co-administration of LJEV and measles-rubella vaccine at 8 months of age.⁶ Furthermore, WHO’s position on measles vaccine is that MCV administered before 9 months of age should be considered supplementary (ie, should not count as one of the two routine MCV doses) unless the country has data showing high seroconversion after vaccination of infants younger than 9 months.⁹

In this Article, we report a study to evaluate seroconversion and possible LJEV interference with measles-rubella vaccination when co-administered at 8 months of age.

1173 infants were assessed for eligibility between Aug 13, 2015, and June 10, 2016, and 80 were excluded (figure 1). Of 1093 infants enrolled, 545 were randomly assigned to the measles-rubella plus LJEV group and 548 to the measles-rubella group. 1013 (93%) enrolled infants completed the study, 507 in the measles-rubella plus LJEV group and 506 in the measles-rubella group (figure 1); 504 infants were from the Hebei province and 509 from the Zhejiang province. Characteristics of the participants and their mothers are shown in table 1. There was little difference between the two groups in infant weight at enrolment or in maternal characteristics (age, and measles and rubella disease and vaccination history; appendix).

Before vaccination, six (1%) of 507 infants in the measles-rubella plus LJEV group and one (<1%) of 506 in the measles-rubella group were seropositive for measles; eight (2%) infants in the measles-rubella plus LJEV group and two (<1%) in the measles-rubella group were seropositive for rubella antibody. 6 weeks after vaccination, seropositivity for measles antibody increased to 99% in both groups (501 of 507 infants in the measles-rubella plus LJEV group; 500 of 506 infants in the measles-rubella group; p=0·9973), yielding a non-inferior proportion of seroconversion of 98% in the measles-rubella plus LJEV group (496 of 507 infants) and of 99% in the measles-rubella group (499 of 506 infants; difference −0·8% [90% CI −2·6 to 1·1], p_{non-inferiority}<0·0001; figure 2 and table 2). 6 weeks after vaccination, 483 (95%) of 507 infants in the measles-rubella plus LJEV group and 475 (94%) of 506 infants in the measles-rubella group were seropositivie for the rubella antibody (p=0·33), yielding a non-inferior proportion of seroconversion of 94·3% in the measles-rubella plus LJEV group (478 of 507 infants) and of 93·5% in the measles-rubella group (473 of 506 infants; difference 0·8% [90% CI −1·8 to 3·4], p_{non-inferiority}=0·0039; figure 2 and table 2).

The median concentration of measles IgG antibody was 25 mIU/mL in both groups before vaccination (the Virion test kit lower limit of detection for measles antibody was 50 mIU/mL, and values below 50 mIU/mL were assigned to 25 mIU/mL); after vaccination, the median concentration of measles IgG antibody increased to 1830·9 mIU/mL in the measles-rubella plus LJEV group and to 1739·7 mIU/mL in the measles-rubella group, with no significant difference between groups (table 3). The median concentration of rubella IgG antibody was 2·7 IU/mL in both groups before vaccination; after vaccination, it increased to 56·4 IU/mL in the measles-rubella plus LJEV group and 58·9 IU/mL in the measles-rubella group, with no significant difference between groups (p=0·63; table 3).

Among the seven infants who were seropositive for measles before the measles-rubella vaccination at 8 months of age, prevaccination IgG concentrations ranged from 411·4 mIU/mL to 1604·7 mIU/mL. Among these infants, only the infant with the lowest prevaccination IgG antibody concentration seroconverted, showing a four-time increase from 411·4 mIU/mL to 1810·2 mIU/mL. The other six did not seroconvert—ie, none had a four-time rise in measles IgG antibody titres (appendix).

Among the ten infants who were seropositive for rubella antibody before the measles-rubella vaccination at 8 months of age, prevaccination IgG concentrations ranged from 10·4 IU/mL to 357·6 IU/mL. Three of these infants seroconverted, with 5·0-times to 6·5-times increases in IgG antibody concentrations after vaccination. The other seven infants did not seroconvert after vaccination; one infant’s IgG antibody concentration increased by 3·8-times, near the threshold for seroconversion. Four of the infants included in this study were seropositive for measles and rubella antibody before the measles-rubella vaccination; none of these four seroconverted to either vaccine component (appendix).

AEFI information was collected for all infants who completed the study. Mild local reactions included redness, swelling, and pain; systemic reactions included fever, diarrhoea, and vomiting. There were no serious AEFIs (table 4; appendix). In the measles-rubella plus LJEV group, seven (1%) of 507 infants were reported to have one or more local AEFI; in the measles-rubella group, 14 (3%) of 506 infants were reported to have one or more local AEFI. There was no evidence of a difference in the incidence of any local AEFI between the two groups (χ²=2·40, p=0·12). Systemic AEFI were reported in the 6 weeks following vaccination for 144 (28%) infants in the measles-rubella plus LJEV group and 162 (32%) infants in the measles-rubella group. There was no evidence of a difference in the incidence of any systemic AEFI between the two groups (χ²=1·57, p=0·21; table 4). Fever was the most common systemic AEFI. All infants with reported AEFIs fully recovered.

In this study, we found that measles and rubella IgG antibody seroconversion following measles-rubella vaccination was not adversely affected by co-administration with LJEV at 8 months of age. The high rates of seroconversion and high antibody concentration against measles and rubella viruses found in both groups provide reassurance that infants will be protected from both measles and rubella, whether the first dose of measles-rubella is given alone or co-administered with LJEV. These data support the integration of LJEV into routine childhood immunisation schedules in countries where Japanese encephalitis vaccine is part of the national immunisation schedule or where Japanese encephalitis is endemic and introduction of the vaccine is being considered. Co-administration of the two vaccines might be particularly useful in areas with scarce resources or access to care, as it would avoid the extra costs and inconvenience associated with an additional clinic visit, possibly leading to greater uptake.

The safety profile for simultaneous administration of LJEV with measles-rubella vaccine was similar to that of measles-rubella vaccine alone. Both vaccines have been in use for decades in China and are considered safe. Passive postmarketing surveillance of Japanese encephalitis vaccination in China between 2008 and 2013 has found no safety concerns, with 214·4 AEFIs reported per million doses of administered LJEV.¹⁵ During these years, 39·0% of the time Japanese encephalitis vaccine was co-administered with other vaccines—of which MCV was the most common (62·7%).¹⁵ As expected, the active adverse-event monitoring we used in our study found a higher prevalence of AEFIs than were identified in passive, postmarketing surveillance.

Our results are consistent with a study done in Taiwan,¹⁶ which showed co-administration of live attenuated Japanese chimeric virus vaccine with measles-mumps-rubella at 9 months did not affect measles and rubella antibody seroconversion proportions compared with a measles-mumps-rubella only group. Although Japanese chimeric virus vaccine is a different class of Japanese encephalitis vaccine than LJEV, the consistency of findings is reassuring.

LJEV co-administration and measles seroconversion were assessed in a randomised study¹⁷ done in 2005–06 of infants aged 9 months in the Philippines. The investigators showed the proportion of measles seroprotection (defined as IgG antibody concentration ≥340 mIU/mL) in infants co-administered a single antigen measles vaccine and LJEV was not inferior to the proportion of seroprotection in infants receiving the measles vaccine and LJEV separately (95·5% [95% CI 91·9–97·8] vs 100% [98·0–100]). In addition, the proportion of Japanese encephalitis seroprotection in the group that was co-administered measles vaccine plus LJEV was non-inferior to that in the LJEV-alone group (90·5% [85·9–94·1] vs 92·1% [84·3–96·7]). The geometric mean concentration for measles antibody titres¹⁸ in the measles vaccine and LJEV co-administration group (302 mIU/mL) was higher than in the measles vaccine-alone group (263 mIU/mL), and the proportion of measles seroprotection in the measles vaccine and LJEV co-administration group was non-inferior to the proportion in the measles vaccine-alone group (91·8% [87·3–95·1] vs 86·5 [80·6–91·2]). In our study, we used the same LJEV vaccine as in the study^17,18 done in the Philippines and our results provide additional assurance that LJEV does not have an adverse effect on the immune response to measles vaccine.

A non-randomised, single-group study¹⁹ of 257 Sri Lankan infants who received simultaneous doses of LJEV and measles vaccine reinforced the safety profile of LJEV vaccine, but because there was no control group, the effect of LJEV on measles antibody seroconversion could not be measured.

LJEV has been introduced or is being considered by many countries for routine infant immunisation, but because high proportions of seroconversion are needed for protective population immunity to measles virus, immunogenicity and safety data on the possible effects of co-administering LJEV on the effectiveness of measles vaccine are crucial.¹ The non-inferiority and safety of co-administered LJEV and measles-rubella vaccine shown in this randomised trial support the concomitant delivery of these vaccines in infants—a finding of importance to the global priorities of measles eradication and of expansion of Japanese encephalitis vaccination in endemic countries.⁸ This study also showed no reduction in rubella antibody seroconversion with concomitant administration of LJEV and measles-rubella vaccine. This information is particularly useful given the widespread introduction of rubella vaccination, with 152 countries having included rubella-containing vaccines (usually as measles-rubella vaccine) in their national childhood immunisation schedules as of December, 2016.²⁰

An additional finding of specific relevance for China is that the infants aged 8 months in this study had low measles and rubella antibody titres before vaccination, whereas after measles-rubella vaccination, the prevalence of seropositivity to measles and rubella antibody was high (98%). This result countervails the concern that maternal antibodies interfere with the immune response to measles-rubella vaccine in infants younger than 9 months. WHO recommends the first dose of MCV to be administered at 9 months of age, and that MCV administered before this age should be considered a supplementary dose and not count as one of the two routine MCV doses unless the country has seroconversion data for MCV administered before age 9 months.⁹

Our study is consistent with previous observations that pre-existing measles and rubella antibody will prevent seroconversion, as only 14% and 30% of infants with pre-existing antibodies to measles and rubella, respectively, seroconverted. The numbers of infants in these categories were small, but only about 1% of infants aged 8 months in our study were seropositive before vaccination. This finding suggests that maternal measles antibody had waned by this age, providing further support for the administration of measles-rubella vaccine to infants aged 8 months in China.

A limitation is that comparison of results between different ELISA tests is problematic because of the different sources and concentrations of antigens, and thresholds for determining protective antibody concentrations have not been standardised.¹⁰ Thus, it is difficult to make a direct comparison of antibody concentrations in our study with the concentrations measured in studies that used other ELISA kits. A second potential limitation is the loss to follow-up of 80 (7%) of 1093 enrolled infants who did not complete the study. However, proportion of participants lost to follow-up was similar in each group and most losses to follow-up were due to parents refusing to continue until the end of the 6 weeks of the study or being unreachable. Only six infants did not complete the study because of illness after enrolment. Because the proportion of loss to follow-up did not differ substantially between study groups, and among infants who completed the study, the two groups were not substantially different from each other in height, weight, sex, prevalence of low birthweight, and feeding pattern (appendix), we believe it is unlikely that losses to follow-up biased our seroconversion and safety results. Loss to follow-up was approximately half of the proportion accounted for in the sample-size calculation, indicating that study power was not compromised.

Overall, we found that infants aged 8 months seroconverted equally well to measles and rubella when measles-rubella vaccine was administered alone and simultaneously with LJEV. Proportions of seroconversion were high and consistent with the proportion necessary to eradicate measles and rubella. No safety concerns were identified for either of the vaccines used in this study, regardless of whether measles-rubella vaccine was co-administered with LJEV. Co-administration of these two vaccines can reduce the number of needed immunisation visits and the time and effort dedicated by health workers, parents, and caregivers. Measles-rubella and LJEV co-administration has, therefore, the potential of increasing the coverage of Japanese encephalitis, measles, and rubella vaccination and of protecting an increased number of young infants from the morbidity and mortality associated with these three serious diseases.