Integration of microbiological, epidemiological and statistical methods for estimating protection level within populations against diphtheria
Posted on July 17, 2023 by Dr Noriko Kitamura
Dr Noriko Kitamura takes us behind the scenes of their latest publication 'Evaluation and validation of a commercial ELISA versus the in vitro toxin neutralisation assay for determination of diphtheria antitoxin in human serum’ in Journal of Medical Microbiology.
My name is Noriko Kitamura, I’ve just graduated a joint PhD programme between the London School of Hygiene and Tropical Medicine, UK and Nagasaki University, Japan. This article, published in Journal of Medical Microbiology, is the last piece of work in a series of publications in my PhD related to diphtheria serology.
The research started when I visited small rural villages in Vietnam for a diphtheria outbreak investigation in 2017. The diphtheria outbreak was shocking news for me as at the time I thought diphtheria was a rare infectious disease. I’ve subsequently found several instances of recent diphtheria outbreaks in neighbouring countries after the investigation.
A few months after my outbreak investigation in Vietnam, the largest diphtheria outbreak in the 21st century started in Rohingya refugee camps at the boarder of Bangladesh and Myanmar. This event encouraged me to conduct research on this disease. I wanted to find out whether the booster dose of the vaccine is necessary for controlling diphtheria in Vietnam and other low- and middle-income countries.
I planned a few seroprevalence surveys with the Pasteur Institute in Nha Trang, the designated government health authority in central Vietnam where the outbreak occurred. The first part of my research focused on the waning rate and protective duration of vaccine-derived immunity among children. One study used two-point IgG measurements in sera among 500 individuals in Vietnam which were recruited in two years apart, in 2017 and 2019, and another study used existing cross-sectional serology data collected in European countries between 1995 and 2003. The protective duration after four doses of vaccine were 5 years in the first study and 10 years in the second.
The next part of my research resulted in a paper which described the carriage prevalence of C.diphtheria and seroprevalence anti-diphtheria toxin antibody in the epidemic prone setting. The article clearly showed the reverse correlation between carriage prevalence and seroprevalence at diphtheria epidemic prone areas. This article also compared the seroprevalence in the community where the vaccination coverage was supposed to be high in the last 20 years to an epidemic-prone community where vaccine coverage was not optimal. The difference in seroprevalence by age between these two settings explains the background population immunity when diphtheria outbreaks occurs. However, one question remained.
All these serological surveys were conducted using ELISA (enzyme-linked immunosorbent assay) to detect anti-diphtheria IgG antibodies in study participants’ sera. The international standard cut-off value of anti-diphtheria IgG was 0.1 IU/ml when IgG was quantified by ELISA. However, this cut-off value does not necessarily correlate with the protection from disease. Therefore, the seroprevalence I measured in my paper did not accurately describe the protection level in the population. To identify the accurate seroprevalence, the ELISA values were compared with toxin neutralization assay, which is the most reliable method to identify individuals who have protective level of antitoxin. The UK Health Security Agency kindly measured anti-diphtheria antitoxin levels in the subset of sample collected in the serosurvey by neutralization assay.
In addition to the parallel comparison of ELISA and neutralization assay, I tried to estimate the accurate proportion of protected individuals against diphtheria in Vietnamese community. This information is really important for government officials to be able to judge the needs of booster dose introduction. I therefore used two statistical methods based on receiver operating characteristic (ROC) curve analysis and multiple imputation approach to estimate the protective level in the population. I then found that the estimated protective level in the population was much higher than the crude seroprevalence measured by ELISA.
ELISA is an easy and cheap alternative to other methods and the only option for serosurveys in rural areas in low- and middle- income countries where the neutralization assay is not plausible due to the limited resources. These are also areas where a diphtheria epidemic is most likely to occur. This paper suggested methods to identify the protection level in the target population from ELISA measurements. Therefore, I believe this paper is especially useful when serosurvey will be conducted in those settings.
Unexpectedly high numbers of diphtheria cases have been recently identified in asylum seekers in Europe. Nigeria have also recently reported a large number of cases. The needs for serological assessment in various populations will be necessary in the near future.
Thumbnail Credit: Pasteur Institute Vietnam