An update of the global burden of pertussis in children younger than 5 years: a modelling study

Summary

Background

Since the publication in 2003 of a model to estimate the disease burden of pertussis, new evidence of the protective effect of incomplete pertussis vaccination against severe pertussis has been reported. We revised the model to provide new estimates of regional and global pertussis cases and deaths for children younger than 5 years.

Methods

We developed a revised model with data from 2014 to estimate pertussis cases and deaths. Pertussis cases were defined according to the WHO clinical case definition, as a coughing illness lasting at least 2 weeks with paroxysms of coughing, inspiratory whooping, or post-tussive vomiting. We used UN population estimates and WHO and UNICEF data on national pertussis immunisation coverage. Estimates were made for vaccine effectiveness against pertussis cases and deaths for one, two, and three doses of vaccination, probability of infection in low and high coverage countries, and case fatality ratios in low and high mortality countries in two age groups: infants younger than 1 year and children aged 1–4 years. We did sensitivity analyses with a range of input parameters to assess the effect of uncertainty of the input parameters on the model outputs.

Findings

We estimated that there were 24·1 million pertussis cases and 160 700 deaths from pertussis in children younger than 5 years in 2014, with the African region contributing the largest proportions (7·8 million [33%] cases and 92 500 [58%] deaths). 5·1 million (21%) estimated pertussis cases and 85 900 (53%) estimated deaths were in infants younger than 1 year. In the sensitivity analyses, the estimated number of cases ranged from 7 million to 40 million and deaths from 38 000 to 670 000.

Interpretation

Our estimates suggest that, compared with the 1999 estimates published in 2003 (30·6 million pertussis cases and 390 000 deaths from pertussis in children younger than 5 years), the numbers of cases and deaths of pertussis have fallen substantially. Model sensitivity emphasised the importance of better surveillance to improve country-level decision making and pertussis control.

Funding

None.

Introduction

Pertussis, caused by the bacterium Bordetella pertussis, results in substantial morbidity and mortality in infants and children and remains a public health concern despite an increase in vaccination coverage during the past decade. The disease is endemic in all countries and epidemic cycles occur every 2–5 years, even after the implementation of effective vaccination programmes and the achievement of high vaccination coverage.¹

All available pertussis vaccines are produced in combination with other antigens such as diphtheria and tetanus toxoid. Two types of pertussis vaccine are available: acellular vaccines based on one or more highly purified individual pertussis antigens; and whole-cell vaccines based on killed B pertussis organisms. The 2015 WHO position paper¹ recommended that every country should seek to achieve early and timely vaccination initiated at 6 weeks of age and no later than 8 weeks, and maintain high coverage (≥90%) with at least three doses of quality assured pertussis vaccine.

In 2003, Crowcroft and colleagues² developed a model to estimate regional and global pertussis disease burden in children younger than 15 years with data up to 1999. Their estimates have helped WHO and national authorities make decisions about pertussis control, particularly in countries with poor disease surveillance. The model estimated 48·5 million pertussis cases worldwide (9·4 million in infants aged <1 year, 21·2 million in children aged 1–4 years, 17·9 million in children aged 5–14 years) and 390 000 deaths from pertussis (all in children <5 years) in 1999, of which the African region contributed the largest burden (12·8 million cases and 200 000 deaths). Since the publication of the Crowcroft model, new evidence of the protective effect of incomplete pertussis vaccination has become available: 50% protection against severe pertussis with a single dose and at least 80% protection with two doses.³ However, modelling disease burden in children aged 5–14 years is considered more challenging; although there is evidence of waning of immunity, there is a paucity of data on the duration of vaccine protection, particularly in the context of many different schedules for providing booster doses. We aimed to estimate pertussis cases and deaths in children younger than 5 years and to estimate regional and global pertussis burden using data for 2014.

Research in context

Evidence before this study

Because several countries, particularly low-income countries, do not have good surveillance systems to monitor pertussis cases and deaths, Crowcroft and colleagues developed a model in 2003 to estimate the disease burden in children younger than 15 years. With data for 1999, this model estimated a global total of 48·5 million pertussis cases (30·6 million in children <5 years) and 390 000 deaths from pertussis (all in children <5 years). We searched PubMed for studies published between May 1, 2003, and April 30, 2016, with the keywords “pertussis”, “whooping cough”, “pertussis vaccine”, and “burden” for any updated information. We identified three recent estimations of pertussis deaths: 60 000 (95% uncertainty interval 43 000–94 000) from WHO vital registration data for 2013, 56 400 (20 700–127 000) from the Global Burden of Disease Study 2013, and 54 500 (18 800–117 000) from the Global Burden of Disease Study 2015. Protection provided by acellular pertussis vaccine has been shown to wane faster than that induced by whole-cell pertussis vaccine.

Added value of this study

This refined pertussis disease burden model estimates the number of pertussis cases and deaths in children younger than 5 years in 2014. Because of the paucity of data for the duration of vaccine protection beyond 5 years of age, estimates of cases and deaths from 5 to 14 years were no longer included in the revised model. We used updated population and coverage data, and incorporated the protective effects of incomplete pertussis vaccination. New data show that pertussis vaccination provides 50% protection against severe disease with a single dose of vaccine and at least 80% protection with two doses. By contrast with the original model, we classified countries into low or high mortality groups according to mortality in children younger than 5 years. Countries with under-5 mortality lower than the mean of all countries were assigned to the low mortality group and the rest to the high mortality group. Because of the improved estimations of population by age group and vaccine coverage in countries, we expect our estimates of cases and deaths to be more accurate than those of the previous model. We also did sensitivity analyses with a range of input parameters to assess the effect of uncertainty of the input parameters on the model outputs.

Implications of all the available evidence

Since the 1999 estimates were reported by Crowcroft and colleagues, the global number of pertussis cases in children younger than 5 years has fallen from 30·6 million to 24·1 million and the number of pertussis deaths from 390 000 to 160 700. These differences reflect an improvement in vaccination coverage, a protective effect of incomplete vaccination, and a more realistic mortality classification grouping for countries. However, pertussis burden in the African region remains high. The sensitivity of the model is driven by a small number of parameters, emphasising the importance of better surveillance. Improved surveillance, resulting in more robust disease burden estimates, will improve country-level decision making and enhance pertussis control.