Exploiting Real-Time Genomic Surveillance Data To Assess 4CMenB Meningococcal Vaccine Performance in Scotland, 2015 to 2022

ABSTRACT The United Kingdom implemented the first national infant immunization schedule for the meningococcal vaccine 4CMenB (Bexsero) in September 2015, targeting serogroup B invasive meningococcal disease (IMD). Bexsero contains four variable subcapsular proteins, and postimplementation IMD surveillance was necessary, as nonhomologous protein variants can evade Bexsero-elicited protection. We investigated postimplementation IMD cases reported in Scotland from 1 September 2015 to 30 June 2022. Patient demographics and vaccination status were combined with genotypic data from the causative meningococci, which were used to assess vaccine coverage with the meningococcal deduced vaccine antigen reactivity (MenDeVAR) index. Eighty-two serogroup B IMD cases occurred in children >5 years of age, 48 (58.5%) of which were in unvaccinated children and 34 (41%) of which were in children who had received ≥1 Bexsero dose. Fifteen of the 34 vaccinated children had received one dose, 17 had received two doses, and two had received three doses. For 39 cases, meningococcal sequence data were available, enabling MenDeVAR index deductions of vaccine-preventable (M-VP) and non-vaccine-preventable (M-NVP) meningococci. Notably, none of the 19 of the children immunized ≥2 times had IMD caused by M-VP meningococci, with 2 cases of NVP meningococci, and no deduction possible for 17. Among the 15 children partially vaccinated according to schedule (1 dose), 7 were infected by M-VP meningococci and 2 with M-NVP meningococci, with 6 for which deductions were not possible. Of the unvaccinated children with IMD, 40/48 were ineligible for vaccination and 20/48 had IMD caused by M-VP meningococci, with deductions not being possible for 14 meningococci.

Of the 82 children with serogroup B IMD, 48 (58.5%) had not received any doses of Bexsero (referred to here as unvaccinated), of whom 36 (75.0%) were born before implementation and too old for the catch-up program, 4 (8.3%) were born after July 2015 but developed IMD before the first 8-week dose, and 8 (16.7%) were eligible but not vaccinated. Of the IMD patients, 34/82 (41.5%) had received $1 dose of Bexsero, 15  . For the 28 patients born after 1 September 2015 who had received $1 dose of Bexsero, median age at first dose was 9 weeks (range, 8 to 12 weeks) and median age at second dose was 18 weeks (range, 17 to 20 weeks); only one child received a booster dose at 13 months.
The 82 IMD cases occurred across all seven epidemiological years, 2015/16 to 2021/22 ( Fig. 1a and c), but incidence decreased over time, with most occurring in the first three epidemiological years (Table 1) Table 1).
There were 38/82 (46.3%) IMD cases among infants, the highest-risk group (Table 1). These were distributed among infants who were unvaccinated (14/38 [36.8%]), partially   Fig. 2). Among the 10 children .1 year old who had received $1 vaccine dose and developed IMD, 8 had not received their 12-month booster dose, with 3 being only days past their first birthday, and 2 developed IMD at 3 to 4 years old, despite completing a 2 1 1 Bexsero schedule. The time between the last dose of vaccine and IMD onset was highly variable: the time after the first dose was 16 to 110 days, that after the second dose was 7 to 1,029 days, and that after the third dose was 875 to 1,235 days.
Fifteen IMD cases occurred in children who were partially vaccinated according to schedule, of which 11 were culture confirmed and four PCR confirmed with vaccine antigen typing. For 12 cases, MLST was determined: CC41/44, 6 isolates; CC35, 1; C32, 2; CC213, 2; and CC269, 1 (Fig. 3). The fHbp peptide was present in all isolates and categorized in Novartis subfamilies (subfamily 1, 7 isolates; subfamily 2, 4; and subfamily 3, 4) and Pfizer subfamily (subfamily A, 8 isolates, and subfamily B, 7). The most frequent fHbp Cases are shown with dots by increasing age at onset of IMD, measured in weeks. The dots are colored by vaccination status (yellow, fully vaccinated; purple, partially vaccinated; blue, unvaccinated). For each case, the MenDeVAR index for the invasive meningococcal isolate is shown by the arrow above the dot, colored green for an exact match to vaccine variants, amber for a cross-reactive match to vaccine variants, red for no match to vaccine variants, gray where there were insufficient data to interpret reactivity, or blue where there was inadequate DNA to determine an antigenic profile for PCR-confirmed cases. The timing of the 2 1 1 Bexsero dosing schedule is shown with the yellow dotted line, with the first dose at 8 weeks, the second dose at 16 weeks, and the booster dose at 12 months. Children were deemed fully or partially vaccinated according to the UK immunization schedule.
Assessment of Meningococcal Vaccine Performance mBio peptide was peptide 4, found in 4 meningococci, and vaccine variant peptide 1 was not present. PorA VR1 and PorA VR2 were present in all organisms, with two (18.2%) possessing vaccine variant P1.4. NHBA was present in all 11 meningococci with WGS data, with vaccine variant NHBA peptide 2 in 3/11 (27.2%) organisms. NadA was present in 1/11 (9.1%). The MenDeVAR index predicted no cross-reactivity for two isolates, both children received 1 dose (Fig. 4b). Seven children with IMD were infected by meningococci considered to be vaccine-preventable, five meningococci with exact matches (4 received 1 dose, and 1 received 2 doses) and two with cross-reactive (both received 2 doses) antigen variants. Six specimens had insufficient data to facilitate prediction of phenotype. Nineteen IMD cases occurred across all seven epidemiological years in children who were fully vaccinated according to schedule: 12 in infants, five in toddlers 1 to 2 years old, and two in 3-to 5-year-olds (Table 1). Of these, 11 were culture confirmed and eight were PCR confirmed (Table 2). For 13 cases, MLST profiles of the invasive meningococci were determined (CC213, 4 isolates; CC461, 3; CC32, 2; CC269, 2; CC35, 1; and CC41/44, 1) (Fig. 3). The fHbp peptide was determined for 17 meningococci and categorized into Novartis subfamilies (subfamily 1, 4 isolates; subfamily 2, 4; and subfamily 3,9) and Pfizer subfamily (subfamily A, 4 isolates, and subfamily B, 13). The commonest fHbp peptides were 13 (3 isolates) and 47 (3 isolates). PorA VR1 and PorA VR2 were identified in 15/17 specimens. NHBA was present in 11 isolates with WGS data and NadA in 3 of these ( Table 2). The MenDeVAR index predicted no cross-reactivity for two meningococci; both children received 2 doses (Fig. 4c). No meningococci had an exact or cross-reactive match with Bexsero antigens, and 15 isolates had insufficient data for phenotype prediction. Among these were variants for which insufficient MATS data were available, including fHbp peptides 13 and 321 and NHBA peptides 3, 21, 115, and 118. Children who were fully

DISCUSSION
These data demonstrate how genomic surveillance of a variable bacterial pathogen, which can be achieved in real time or nearly real time, can support vaccination implementation. In 2015, the United Kingdom was the first country to introduce the proteinbased meningococcal vaccine Bexsero into a national infant immunization program, although its efficacy against all meningococcal variants causing IMD was uncertain at that time. Vaccine uptake was high in Scotland, at 94.5 to 96.9% for the 2-dose priming course and 91.4 to 95.0% for the booster dose. Routine real-time genomic surveillance of IMD in Scotland (2015 to 2022) enabled the MenDeVAR index to be used. This demonstrated the effectiveness of Bexsero in fully vaccinated children, who experienced no IMD cases caused by vaccine-preventable meningococci, i.e., those with antigen variants that either were identical to those in the vaccine or had been shown in in vitro assays to be reactive to vaccinee sera (cross-reactive variants) (12,16). However, some individuals who received a single dose of Bexsero did develop vaccine-preventable IMD.
There were 19 IMD cases in children who were fully vaccinated according to schedule, two of whom developed IMD caused by meningococci regarded as not preventable and therefore not anticipated to be covered by the Bexsero immunization program. The remaining 17 cases in the group who were fully vaccinated according to schedule (89%) had antigenic variants not tested by the MATS assay at the time of writing, precluding the determination of their vaccine-preventable status using the MenDeVAR index. A lasting protective effect of the booster dose was observed in children who were fully vaccinated according to schedule, with low numbers of cases in 3-and 4-year-old children who had received three doses and developed IMD (n = 2), supporting previous efficacy studies (17). These data therefore support the 2 1 1 schedule for priming advised by the UK Joint Committee of Vaccination and Immunization, rather than the 3 1 1 schedule from clinical trials (18,19). Further postimplementation immunogenicity data may explain these observations in vaccinees, with .97% Where the isolate was culture confirmed, the meningococci were subjected to whole-genome sequencing, and so a genome record is present on pubmlst.org and can be found using the PubMLST ID. For each isolate, culture confirmed or PCR confirmed, the typing data for ST, CC, fHbp, NHBA, NadA, and PorA VR1 and VR2 are listed where available. NHBA and NadA testing was not routinely performed as part of NHS workflow at the time of the study.
Assessment of Meningococcal Vaccine Performance mBio of infants who received a 2 1 1 schedule achieving a .4-fold rise in human SBA titers after both primary and booster doses (20). Bexsero protects children through the period of highest risk, from infancy to 5 years old, against certain meningococcal variants, and this should be communicated to parents and health care professionals; however, the occurrence of breakthrough cases with nonpreventable meningococci indicates that those caring for vaccinated children should remain alert for the signs and symptoms of suspected IMD and manage them with appropriate urgency. IMD occurred most frequently in unvaccinated and partially vaccinated children (76.8%), with 17.1% of children who developed IMD not receiving all or any of the doses for which they were eligible. For 37.5% (3/8) of eligible but unvaccinated children and 46.7% (7/15) of children partially vaccinated according to schedule, their IMD was caused by a vaccine-preventable meningococcus. Although these numbers of individuals are small, the impact of IMD on children and families is often profound and long-lasting (3). These data suggest that vaccination is highly protective against diverse meningococcal variants, perhaps as many as 88% of meningococci (10,21,22); therefore, if universally applied, the existing schedule might have prevented up to half of IMD cases in eligible children ,5 years old. Most partially vaccinated children were infants, of whom five had not received their second dose (due at 16 weeks) by 18 to 26 weeks of age, and three of these developed IMD with meningococci characterized as vaccine preventable. This supports the need for at least two priming doses, as one is likely insufficient to generate protective immunity. For the five children .1 year old who had received their priming course but not their boosters by 20 to 48 months, it was not possible to determine whether their meningococcal variants could have been prevented by adequate vaccination, as their antigens had not been tested in the MATS assay. Childhood vaccinations can be delayed for many reasons, including concurrent febrile illness, vaccination appointment delays, concern over side effects, and vaccine hesitancy or refusal. This was compounded by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pandemic, resulting in reduced vaccine uptake in England during the first UK lockdown compared to prepandemic levels (23), although Scotland had an increase in timely vaccine uptake (24,25). Refocusing efforts on positive and sustained messaging to carers and encouraging proactive identification and contact of unvaccinated children would help to prevent such cases.
Known hyperinvasive meningococcal lineages were responsible for IMD among Scottish children, reflecting the circulating meningococcal population causing disease in unvaccinated children and adults (see Fig. S1 in the supplemental material). There was no evidence of CC41/44 causing IMD in fully vaccinated children, consistent with the fact that the OMV component in Bexsero originates from the MeNZB vaccine, which is based on a CC41/44 meningococcus and therefore contains the characteristic PorA P1.7-2,4 variant as a dominant antigen. In post-MeNZB introduction surveillance in New Zealand in 2008, 34 vaccine breakthrough cases occurred in persons ,19 years old caused by the epidemic PorA variant P1.7-2,4 meningococcus, with half occurring in children ,5 years old (26). We did not observe any breakthrough cases with the PorA variant P1.7-2,4 or any CC41/44 meningococci. Of the CC41/44 isolates that did occur in unvaccinated and partially vaccinated Scottish children, 81% possessed at least two reactive fHbp, NHBA, or PorA VR2 antigens, suggesting that in Bexsero vaccinees, multiple antigens may result in enhanced protection rather than relying on the PorA P1.7-2,4 antigen alone for CC41/44.
Postimplementation vaccine surveillance is required to monitor ongoing trends in disease-causing meningococci, including secular changes in lineages and vaccine antigen prevalence. Until 2016, this was performed through phenotypic assays, but with the development and application of genomic typing, particularly WGS, in public health laboratories, genomic surveillance has become achievable. Genomic data allow assessment of genetic population structure and detailed characterization of certain bacterial features, including capsular type and protein vaccine antigens. Furthermore, deduction of peptide variants from genomic data enables a functional assessment, although at the time of writing, no methods exist for reliably assessing protein expression and cross-reactivity from genomic data. Determining cross-reactivity is key to estimating how broadly Bexsero can cover diverse meningococcal variants, as there are only four variants that exactly match the vaccine components. These real-world data suggest that the MATS assay, a surrogate marker of the correlate of protection against N. meningitidis infection (SBA assay), predicts crossreactivity reliably: none of the meningococci that occurred in individuals who were fully vaccinated according to schedule were predicted to be reactive. These data also support the thresholds used in the development of the MenDeVAR index (testing of $5 isolates by MATS and at least three-fourths being cross-reactive or not reactive). Variants with insufficient phenotyping testing are problematic, and availability of more phenotypic data will improve genotype-phenotype predictions through the MenDeVAR index. Identifying more meningococci that are known to be vaccine preventable in vitro and inferred in vivo would enable global regions to rigorously assess the breadth of vaccine coverage in their population, providing more detailed vaccine effectiveness and cost-effectiveness estimates for vaccine policy decisions.
The limitations of this study included the incomplete availability of meningococcal isolates, required for WGS analysis, although the rate observed (47.6%) was comparable to that in other studies (15). Attempts were made to optimize data from PCR-confirmed cases, but in many specimens the amount of meningococcal DNA was too small for antigen typing. Underlying immunodeficiencies among cases were not considered, as this information is not routinely collected during public health surveillance. There was a notable reduction in IMD cases across all age groups in Scotland from early 2020 continuing into 2022, compared to previous years, which was likely secondary to social distancing measures and other restrictions implemented from March 2020 in response to the coronavirus disease 2019 (COVID-19) pandemic, interrupting meningococcal transmission.
Effective and timely public health interventions are necessary for the prevention of infectious diseases, with vaccines against N. meningitidis being a highly effective intervention. The evidence obtained in the 7 years following introduction of the Bexsero vaccine for infants in Scotland provides a real-world demonstration of vaccine efficacy assessed using real-time, integrated WGS in public health. This study also demonstrates the need to characterize additional, diverse meningococci as vaccine preventable to understand the impact of the vaccine and possible modification of vaccine formulations in the future.

MATERIALS AND METHODS
The Public Health Scotland Order 2019 in Article 9(2)(i) places an obligation on Public Health Scotland (PHS) to engage in the control of spread of infectious diseases in accordance with section 43 of the National Health Service (Scotland) Act 1978. In accordance with sections 15, 16(5), and 21(2) of the Public Health etc. (Scotland) Act 2008, PHS is obliged to process data in relation to notifiable diseases, health risk states of patients, notifiable organisms, and carrying out public health investigations, and therefore, individual patient consent is not required.
At the time of Bexsero introduction, the population of Scotland was ;5.47 million, with an annual birth cohort of ;55,000. Data for all Scottish IMD cases were captured by the Meningococcal Invasive Disease Augmented Surveillance (MIDAS) scheme, managed by Public Health Scotland and SMiRL-G. Records for children with IMD from 1 September 2015 to 30 June 2022 were retrieved with demographic information, including date of birth, number of Bexsero doses received, date of last Bexsero dose, age at IMD presentation, date of IMD diagnosis, clinical presentation, and laboratory sample identifiers. For each case, microbiological specimens were identified, mode of diagnosis was established (i.e., microbiological culture or meningococcal ctrA PCR), and sequence data were retrieved, where available. The PubMLST database (pubmlst.org) held isolate records, WGS data, and associated provenance information. To assess vaccination status, the number of vaccine doses received and the interval between the last vaccine dose and IMD onset was determined (Table 3). Fourteen days postvaccination was considered sufficient for vaccine-induced immunity to develop. The term "vaccine breakthrough cases" was used to describe IMD cases that occurred in children who were fully vaccinated according to schedule.
For non-culture-confirmed IMD cases, diagnosis was by molecular methods. Briefly, DNA extracts from clinical specimens (EDTA-blood or CSF) were screened for meningococcal DNA using an in-house ctrA quantitative reverse transcription-PCR (RT-PCR) (27). Serogroup was determined with siaD RT-PCR (28). Nested PCR was used for amplification of porA (29) and fHbp (30) directly from clinical samples, allowing the deduction of PorA VR1 and VR2 and the fHbp peptide. Nested PCR success was correlated with positivity of the ctrA RT-PCR, with extracts with high cycle threshold values (weakly positive) often failing subsequent MLST or porA/fHbp typing. NHBA and NadA typing were not routinely performed.
Where possible, typing information was deduced from WGS data, including species, strain designation, MLST and ST, CC, and BAST (14). The MenDeVAR index (16) was determined for Bexsero and Trumenba: for a given isolate, each vaccine was assigned a green, amber, red, or gray status. Green indicated that the isolate possesses an antigenic variant identical to the vaccine component. For Bexsero, Assessment of Meningococcal Vaccine Performance mBio this was one of fHbp1, NHBA 2, NadA 8, or PorA VR2 P1.4. For Trumenba, it was fHbp peptide 45 or 55. Amber indicated that the isolate possesses an antigenic variant found experimentally to cross-react with vaccine-derived antibodies. Red indicated that the vaccine antigen variants present in the meningococcus had all been shown to be nonreactive experimentally with vaccine-derived immune responses. Gray indicated that the meningococcus possessed antigen variants for which insufficient information was available from experimental studies to determine potential cross-reactivity. The putative cross-reactive or nonreactive antigenic variants were previously determined using MATS (21) for Bexsero-induced immune responses and MEASURE or SBA assay data for Trumenba-induced immune responses.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only. FIG S1, TIF file, 0.1 MB.