Seroprevalence of pertussis in Madagascar and implications for vaccination

Abstract

Pertussis is a highly contagious infectious disease and remains an important cause of mortality and morbidity worldwide. Over the last decade, vaccination has greatly reduced the burden of pertussis. Yet, uncertainty in individual vaccination coverage and ineffective case surveillance systems make it difficult to estimate burden and the related quantity of population-level susceptibility, which determines population risk. These issues are more pronounced in low-income settings where coverage is often overestimated, and case numbers are under-reported. Serological data provide a direct characterisation of the landscape of susceptibility to infection; and can be combined with vaccination coverage and basic theory to estimate rates of exposure to natural infection. Here, we analysed cross-sectional data on seropositivity against pertussis to identify spatial and age patterns of susceptibility in children in Madagascar. A large proportion of individuals surveyed were seronegative; however, there were patterns suggestive of natural infection in all the regions analysed. Improvements in vaccination coverage are needed to help prevent additional burden of pertussis in the country.

Keywords: Madagascar; pertussis; serology; vaccination.

Fig. 1.

Epidemiology of pertussis and DTP vaccination in Madagascar. (a) Nationally reported pertussis cases from the WHO indicate consistently low case numbers after 2004; vertical dashed lines indicate the time-range experienced by individuals within our cross-sectional seroepidemiological sample that occurred in 2016 and reached children aged between 9 months and 15 years (2001–2016). (b) Location of the five districts where the study took place (shown in green) across Madagascar, with sample sizes of each shown in the insets. (c) Vaccination coverage (number of doses given over the estimated target population) for DTP 1–3 nationally and for the five districts studied. Mean coverage decreased by the dose number and was highly variable amongst the districts (DTP1 = 83%, min = 30%, max = 124%; DTP2 = 75%, min = 28%, max = 98%; DTP3 = 74%, min = 34%, max = 95%).

Fig. 2.

Age and geographic seropositivity results. Serum samples were tested for the presence of IgG anti-pertussis toxin by (a) age profile with a blue dashed line indicating <5 IU/ml and red dashed line for 100 IU/ml, (b) location, (c) distribution by seroprotection group per location, and (d) the proportion who are seronegative by age and location.

Fig. 3.

Determinants of seropositivity estimates of the probability of becoming seropositive by vaccination (y-axis, first panel) in each of the five communes; and associated force of infection (y-axis, panels 2–6) across years (x-axis) for each of the five locations (panel titles); showing 95% (light grey) and 80% (dark grey) credible intervals from the posterior distributions of associated rates. See text for model assumptions; estimates are robust to altering the prior distributions. The pattern of decay in the force of infection broadly matches the reported pattern from national scale reporting (Fig. 1a).

Fig. 4.

Individual-level analysis. For a small number of individuals who possess a vaccination card, the majority were not serologically protected (titre values below 5). We analysed the age profile of these individuals relative to others who also possessed a vaccination card and the study population. Those who had a low titre level and a vaccination card were on average 1.47 years whereas those who had a low titre level but did not possess a vaccination card were on average 6 years old (P < 0.001).