Recent studies suggest a measles-induced immune amnesia that could have long-term immunosuppressive effects via preferential depletion of memory CD150+ lymphocytes.1–3 An association with a 2–3-year period of increased mortality and morbidity from infectious diseases other than measles has been shown in children from wealthy countries,4–6 and increased morbidity of clinical signs of infectious disease was shown in children 9–59 months of age in the Democratic Republic of the Congo (DRC) who had been classified as measles cases within the past 2 years.7

Vaccine-induced immunologic memory, as with natural immunity, is mediated by B and T memory cells. While immunity against many pathogens, such as measles virus, can be achieved either via natural infection or immunization, this is not the case with tetanus. Tetanus, a disease caused by a neurotoxin from the bacterium Clostridium tetani, can be prevented via immunization, but natural infection with tetani toxin will not result in protective antibody against future infection.8,9 Three doses of tetanus toxoid are thought to confer protection against tetanus for at least 5 years, and the accepted minimal level of toxoid-induced antitoxin antibody required to protect against disease is 0.01 IU/mL, as measured via the “gold standard” in vivo neutralization assay.8 While tetanus is uncommon in wealthy countries, it is a greater concern in middle- and low-income countries, particularly those whose healthcare delivery is disrupted by armed conflict, with 79% of deaths due to tetanus estimated to occur in south Asia and sub-Saharan Africa (44,612 and 56,743, respectively).9,10

In 2015, there was a mortality rate of 1.34 per 100,000(confidence interval [CI]: 0.24-6.35) due to non-neonatal tetanus in the DRC.10 While this is an approximately 50% decrease from the 1990 mortality estimate, it is still too high for this vaccine-preventable disease. In 2018 and 2019, the DRC had a reported incidence rate (IR) for total tetanus (including neonatal tetanus) of 9.4 and 9.3, respectively. By comparison, the African Region had IRs of 5 and 3.7 for the same years, and the global IRs were 3.1 and 2.7 for 2018 and 2019, respectively. DRCs total tetanus IR for 2020 and 2021 was only 0.5,11 but this may have been due to limited surveillance during the severe acute respiratory syndrome coronavirus 2 pandemic. Immunization against tetanus in DRC is recommended at 6, 10 and 14 weeks via the Diphtheria and Tetanus and Pertussis and Haemophilus influenzae and Hepatitis B vaccine (DTwPHibHepB),11 yet findings from the 2013-2014 DRC Demographic and Health Survey (DHS) reveal that only 60% of children 6–59 months of age received the full 3-dose tetanus toxoid-containing series, and only 40% had seroprotective antibody titres.12

To further examine the associations previous measles virus infection may have on immunologic memory among children in the DRC, we examined tetanus antitoxin levels among children fully vaccinated against tetanus, with and without a history of measles. Because natural infection with tetanus toxin does not induce seroprotection, tetanus antibody is ideal for assessment of measles virus’ potential impact on vaccine-induced immunologic memory in this population. Previously, we presented a statistically significant difference in vaccine-induced immunity to tetanus among children with past measles disease versus children who had not had measles.13 Here, we provide additional analyses assessing this association and include examination of seroprotective tetanus antibody levels as an additional endpoint.

METHODS

We examined the association between past measles cases and tetanus antibody levels among children participating in the 2013–2014 DRC DHS. Our sample consisted of 711 children 9–59 months old.

Data Source and Study Population

The 2013–2014 DHS, a nationally representative survey based on a stratified 2-stage cluster design, occurred from November 2013 to February 2014. The first stage of the survey consisted of Enumeration Area formation in which a stratified sample of geographic locations, or clusters (n = 540), was selected with proportional probability according to size. The second stage involved sampling households from each Enumeration Area; complete listings of households were created within each cluster, and households (n = 9000) were selected with equal probability.14–17 Within the selected households, individuals interviewed included 18,827 women 15–49 years old (all selected households) and 8656 men 15–59 years old (50% of selected households). The DHS collected biomarker data only on children from households in which men were interviewed. In our dataset, there were 812 children 9–59 months old with written (and dated) record of receiving the full 3-dose tetanus vaccine series and who also had anthropometric, measles and tetanus IgG serology data available, and of these, 713 children had data on measles that occurred following DTwPHibHepB vaccination as well as all other covariates of interest. Four of these children who met measles case criteria had had measles less than 2 months before dried blood spot (DBS) collection and so were removed from analyses to prevent associations due to acute measles-induced immune suppression in this study, bringing the final total to 711 children (this total incorporates complex survey weighting methods utilized by the 2013–2014 DRC DHS).

Information was collected on weight, health outcomes, vaccination history and vaccine-preventable disease serology. After parental consent, DBSs were collected from participants to assess immunity to vaccine-preventable diseases and processed at the University of California, Los Angeles (UCLA)—DRC laboratory at the National Laboratory for Biomedical Research in Kinshasa. All survey data were transferred from paper questionnaires to an electronic format using the Census and Survey Processing System (U.S. Census Bureau, ICF Macro). Data were double entered and verified by comparison of both datasets.

At the time of interview, if mothers possessed a health care worker-provided vaccination card indicating the date the child was vaccinated, this was considered a “dated card” report. Since “dated card” reports are considered most reliable, we utilized only this type of report when categorizing a child as vaccinated, whether the vaccination was against tetanus or measles. Children classified as “unvaccinated” in this study were those reported as such in the 2013–2014 DHS. Tetanus vaccination status was obtained via dated vaccination card and limited to children receiving the complete 3-dose vaccination series (all other children were removed from analyses). Due to concern of live versus killed vaccines resulting in nonspecific effects on vaccinees,18 we examined whether the DTwPHibHepB versus measles vaccine was given last within the sample of participating children. Of the 649 children who received both DTwPHibHepB and measles vaccines, 5 (0.1%) children received DTwPHibHepB after the measles vaccine (indicating an incorrect vaccination schedule), 22 (3.3%) children received both vaccines within the same month and 622 (96%) received the measles vaccine after DTwPHibHepB. We did not examine other vaccines a child may have received.

Laboratory Analysis

DBS samples were extracted using a modified extraction protocol19 and processed at the UCLA-DRC laboratory at the National Laboratory for Biomedical Research. The DBS extraction and assay protocol and multiplex technology have been described elsewhere.20,21 Briefly, a 0.25“ DBS punch was extracted, shaking at room temperature, in 1 mL phosphate buffered saline, 0.05% Tween-20 and 5% dried milk, which represents a 1:143-fold dilution assuming 7 µL of serum per punch. Polystyrene beads coated separately with antigen were immobilized within 54-well SmartPLEX assay strips with 10 beads per well and processed using a prototype DYNEX Multiplier chemiluminescent automated immunoassay platform with a research use only Smart-PLEX assay for measles, mumps, rubella, varicella-zoster virus and tetanus (DYNEX Technologies, Chantilly, VA).21 Assay Score (AS) for each sample was calculated as a ratio of the signal to a 5-donor, pooled positive control included in each run. Given that it was outside the scope of this project, this research use only immunoassay panel was not correlated with International Units for measles antibody via the plaque reduction neutralization test. In previous work,22 we have reported antibody according to the AS, but in an effort to keep units consistent in this study, we converted the measles AS to IU/mL. For tetanus IgG antibody, based on validation (R. Budd, personal communication, 15 January 2020) and epidemiologic studies,8 the seroprotective cutoff was set to a value of 0.69 AS, which corresponds to 0.2 IU/mL.

As part of our measles case definition, we utilized the limit of quantitation (LOQ) for measles antibody, which was calculated to be 12.06 IU/mL, and this value, in supplement to maternal report of measles, was used as an additional inclusion criterion for classifying previous measles disease among children with maternally reported history of measles. Maternal report captured whether the child had ever had measles in the past, and most maternal reports of measles had an associated month and year date. Requiring maternally reported measles cases to have a minimum measles antibody level to be included as measles cases in this analysis was done to decrease misclassification of maternal reports of measles cases, yet avoid missing cases of infants who may not have displayed a robust immune response despite previous infection. Among the measles-vaccinated children (n = 649), 42 children had a maternal report of measles. Of these 42, 40 (95%) met the LOQ antibody threshold while 2 (5%) were below this threshold and so excluded as measles cases (coding them as noncases for measles). Among children not vaccinated against measles (n = 63), 2 children had a maternal report of measles, and of these, 2 (100%) met the LOQ antibody threshold. To further examine the possibility of misclassification of measles, we included a sensitivity analysis, described in the statistical analysis section.

Covariate Selection

A priori confounders, such as age, sex and geographic location, were included in descriptive and regression analyses. Other potential confounding variables were identified in the literature and based largely on variables representing social determinants of health and found by others to capture health disparities, such as chronic malnutrition (stunting), maternal education, wealth index and birth order.16,23,24 Malaria status (via blood smear) was included due to its association with host immunosuppression,25 and the authors previously found that rural versus urban residence impacted measles vaccination coverage.26 Finally, due to the history of conflict in the DRC, we created a dichotomous variable of children living in provinces experiencing high levels of conflict versus all other provinces. The 5 (old) provinces (Katanga, Maniema, Nord-Kivu, Orientale and Sud-Kivu) with the highest levels of human displacement in 2014 were classified as provinces experiencing high levels of conflict, as most of the human displacement (97%) was conflict-related.27,28

Due to the challenges inherent with directly measuring a household’s wealth, the DHS assumes that the economic state of a household is related to whether the household has access to certain amenities or services. Such variables can easily be included into a questionnaire or directly observed by the interviewer. Within the DHS, households are placed into quintiles according to household ownership of previously selected assets and via principal components analysis. Asset scores are standardized relative to a standard normal distribution and each household receives a summed score based on the standardized score for each asset. Individuals receive a score reflective of their household score, and the resulting quintiles make up the wealth index for the DHS survey.29 In this study, we dichotomized wealth index to compare the most disadvantaged individuals to all others in the sample, resulting in a wealth index that compares the poorest children to all others, or the lowest quintile children to those in the combined group of second lowest, middle, fourth lowest and highest quintile.

Collinearity diagnostic statistics were performed on all explanatory variables included in the final regression models, with acceptable results (tolerance for all variables was ≥ 0.75, variance inflation factor for all variables was ≤ 1.33) for all variables of interest.

Statistical Analyses

Descriptive statistics were performed, examining associations of sociodemographic covariates with both tetanus antibody seroprotection and measles case status via Wald χ2 test. Because our sample was limited to children with 3 doses of DTwPHibHepB reported via dated vaccination card, we observed some differences in descriptive data for variables such as wealth index, rural versus urban residence and measles vaccination status, relative to the authors’ previous work examining these variables among children with vaccinations reported via maternal recall or marked card. One such difference was that the overall measles attack rate (AR) was higher for children who were measles vaccinated versus unvaccinated in this analysis. To further explore this finding, we provided an additional table stratifying AR by age and time of initial measles vaccination, as the subgroup ARs are more reflective of findings from the larger sample of children in DRC including those not fully vaccinated against tetanus.22

Multiple linear regression was used to examine the linear relationship between tetanus antibody and history of measles, adjusting for potential confounding variables. We examined the relationship between our log transformed continuous tetanus antibody variable and measles case status, adjusting for covariates. Multiple logistic regression was used to examine predictors of tetanus seroprotection among this fully vaccinated group of children. Additionally, we examined differences in antibody among subgroups by calculating geometric mean concentration (GMC) of tetanus IgG by age, measles status and other covariates of interest. To control for variability in probability of selection and interview, DHS stratum, cluster and individual sampling weights were applied to analyses and SAS 9.4 PROC SURVEY procedures were utilized.

Because laboratory confirmation of maternally reported measles cases was beyond the scope of this study, to examine the potential effects of misclassification of measles cases (such as incorrectly classifying a child as being a measles case in the past when in fact the child had not been a measles case), we performed additional analyses to examine GMC of tetanus antibody by subgroup. A limitation in categorizing a measles case via positive maternal + LOQ is that children who have been measles vaccinated are more likely than unvaccinated children to meet LOQ criterion even if the maternal report is falsely positive. We have no way of differentiating whether children who meet the LOQ criterion do so due to vaccination or prior infection, but as a mitigating measure, we have provided a closer examination of measles antibody among the following 4 groups of children: (1) measles-unvaccinated + no maternal measles report, (2) measles-unvaccinated + maternal measles report, (3) measles-vaccinated + no maternal measles report and (4) measles-vaccinated + maternal measles report.

To examine the impact that varying levels of misclassification might have on odds ratio (OR) estimates in our logistic regression model, we conducted a supplementary sensitivity analysis. Our concern was that we might be more likely to include false positive measles cases (children categorized as having had measles in the past due to positive maternal report and qualifying LOQ that were in fact not measles cases) among children who were vaccinated against measles versus children who had not received measles vaccine. While the authors think it is plausible that probability of misclassification of measles case status (exposure) is independent of probability of misclassification of tetanus seroprotection status (outcome), due to tetanus seroprotection status being measured via biosample, there is concern of nondifferential dependent misclassification between measles case status (exposure) and measles vaccination (covariate).30

Utilizing unweighted data, we demonstrated via 2 × 2 tables the potential impact of nondifferential dependent misclassification between measles case status and measles vaccination, with 10%, 25% and 50% of measles cases among vaccinated children considered to be false positives and so reclassified as measles noncases. Using updated values for exposure (measles status) and outcome (tetanus seroprotection status) variables, we acquired ORs and 95% Wald confidence limits via univariate regression.

All analyses were performed using SAS software, Version 9.4 (SAS Institute, Cary, NC). Ethical approval was obtained at UCLA Fielding School of Public Health, the Kinshasa School of Public Health and the U.S. Centers for Disease Control and Prevention. As children were younger than the standard age of assent, the parent or guardian of participating children provided consent on the child’s behalf.

RESULTS Descriptive Analyses

Descriptive statistics (Table 1) show that, among children receiving 3 doses of tetanus vaccine, seroprotection for tetanus varied by age (Wald χ2P = 0.0057), maternal education, wealth index and severe stunting. Measles case status varied by age (Wald χ2P = 0.0253) and wealth index (Wald χ2P = 0.0111).

TABLE 1. - Descriptive Data in Children 9–59 Months of Age, by Tetanus Seroprotective Status and Measles Report (n = 711) Variable Total Seroprotection for Tetanus* Measles Case n n % n % Age (months)  9–11 75 31 41 1 1  12–23 233 167 72 4 2  24–35 179 109 61 7 4  36–47 105 53 50 16 15  48–59 119 60 50 14 12 P † 0.0057 0.0253 Sex  Male 330 195 59 22 7  Female 381 226 59 20 5 P 0.9682 0.5104 Measles vaccination  Yes 649 393 61 40 6  No 63 28 44 2 3 P 0.0821 0.4815 Maternal education  < 7 years 352 181 51 23 7  ≥ 7 years 360 240 67 19 5 P 0.003 0.6118 Wealth index‡  Poorest 75 30 40 1 1  All others 636 391 61 41 6 P 0.0052 0.0111 Rural vs. urban residence  Urban 347 227 65 19 5  Rural 365 194 53 23 6 P 0.0584 0.7667 Severely stunted§  Yes 133 57 43 8 6  No 579 364 63 34 6 P 0.0032 0.9599 Firstborn vs. later born  Firstborn 186 124 67 12 6  Later born 525 297 57 31 6 P 0.097 0.8538 Malaria (blood smear) result  Positive 82 44 54 7 9  Negative 630 377 60 35 6 P 0.3046 0.5604 Conflict  High conflict province¶ 307 171 56 22 7  Other province 405 250 62 21 5 P 0.3419 0.4485 Province  Kinshasa 109 78 72 4 3  Bandundu 60 29 48 4 7  Bas-Congo 43 24 56 0 1  Equateur 70 42 60 6 9  Kasai-Occidental 78 47 60 1 1  Kasai-Oriental 46 29 63 5 11  Katanga 46 27 59 4 9  Maniema 6 5 83 — —  Nord-Kivu 170 91 54 14 8  Orientale 13 7 54 1 8  Sud-Kivu 72 43 60 2 3 P 0.5588 — Total 711 421 59 42 6

Bold entries indicate statistically significant estimates via the Wald X2 test.

The — indicates that no measles cases were reported.

*Seroprotection defined as 0.2 IU/mL.

†Wald χ2 test for independence of measles status and row variables.

‡Wealth index is the Demographic and Health Survey composite measure of a household’s cumulative living standard. Based on household ownership of previously selected assets, and utilizing principal components analysis, households are placed within 1 of 5 quintiles. The dichotomized variable compares the poorest children to all others, or the lowest quintile children to those in the combined group of second lowest, middle, fourth lowest and highest quintile.

§Calculated according to National Center for Health Statistics/CDC/WHO international reference standard for height/age, dichotomized as –2.0 to ≤ –3.0 SD below the mean (chronically malnourished, or stunted) and normal to ≥ 3.0 SD above the mean (normal/overnourished). Here, we compare the most severely stunted children with all others.

¶Dichotomous variable of children living in provinces experiencing high levels of conflict (Katanga, Maniema, Nord-Kivu, Orientale and Sud-Kivu) vs. all other provinces. Classified according to the United Nations Office for the Coordination of Humanitarian Affairs.

CDC indicates Centers for Disease Control and Prevention; WHO, World Health Organization.

Examination of measles AR stratified by age (Table 2) revealed among both vaccinated and unvaccinated children, most measles reports occurred during the 24–59 months of age range. Most unvaccinated children fell within the 9–11 and 12–23 months of age groups (n = 53); there were few measles reports (n = 5) overall in these age groups, with none being in unvaccinated children. Only 10 unvaccinated children were in the 24–59 months of age group. The overall AR among the vaccinated (ARv) was 6.2 and among the unvaccinated (ARU), 3.2. Among the 34 vaccinated children 24–59 months of age in our study who reported measles, 32 (94%) received measles vaccination at less than 12 months of age. There were no cases of measles among unvaccinated children in the 9–11 month and 12–23 months of age groups, and the ARU findings within the 24–59 months of age group was 20.0.

TABLE 2. - Measles Attack Rate Stratified by Age in Children 9–59 Months of Age (n = 711) Age (Months) No History of Measles Measles Case Total (n) Attack Rate (per 100)*,† 9–11  Unvaccinated 26 0 26 0.0  Vaccinated 48 1 49 2.0 12–23  Unvaccinated 27 0 27 0.0  Vaccinated 202 4 206 1.9   < 12 months of age‡ 182 4 186 2.2   ≥ 12 months of age 20 0 24–59  Unvaccinated 8 2 10 20.0  Vaccinated 359 34 393 8.7   < 12 months of age 308 32   ≥ 12 months of age 51 2 Total  Unvaccinated 60 2 62 3.2  Vaccinated 609 40 649 6.2

*Attack rate in the unvaccinated calculated as: (cases among the unvaccinated/total unvaccinated) × 100.

†Attack rate in the vaccinated calculated as (cases among the vaccinated/total vaccinated) × 100.

‡Denotes age at which the child was vaccinated against measles.

Regression Analyses

A linear regression model (Table 3) was used to examine relationship between