Dynamics of Pertussis Transmission in the United States

F. M. G. Magpantay; P. Rohani


Am J Epidemiol. 2015;181(12):921-931. 

In This Article

Abstract and Introduction


Past patterns of infectious disease transmission set the stage on which modern epidemiologic dynamics are played out. Here, we present a comprehensive account of pertussis (whooping cough) transmission in the United States during the early vaccine era. We analyzed recently digitized weekly incidence records from Morbidity and Mortality Weekly Reports from 1938 to 1955, when the whole-cell pertussis vaccine was rolled out, and related them to contemporary patterns of transmission and resurgence documented in monthly incidence data from the National Notifiable Diseases Surveillance System. We found that, during the early vaccine era, pertussis epidemics in US states could be categorized as 1) annual, 2) initially annual and later multiennial, or 3) multiennial. States with predominantly annual cycles tended to have higher per capita birth rates, more household crowding, more children per family, and lower rates of school attendance than the states with multiennial cycles. Additionally, states that exhibited annual epidemics during 1938–1955 have had the highest recent (2001–2010) incidence, while those states that transitioned from annual cycles to multiennial cycles have had relatively low recent incidence. Our study provides an extensive picture of pertussis epidemiology in the United States dating back to the onset of vaccination, a back-story that could aid epidemiologists in understanding contemporary transmission patterns.


Pertussis, commonly known as whooping cough, is a contagious disease caused by the bacterium Bordetella pertussis. Historically, pertussis was responsible for substantial morbidity and mortality in children. In the prevaccine era in the United States, it is estimated that pertussis accounted for 200,000 cases and 4,000 deaths annually.[1] The widespread rollout of infant vaccination in the 1940s and 1950s[2] dramatically reduced pertussis incidence in much of the developed world.[3] However, over the past decade or so, a clear rise in pertussis incidence has been observed in some countries with consistently high vaccine coverage.[4–8] The absence of a clear explanation for these events highlights gaps in our understanding of pertussis epidemiology, particularly the nature of the immunity rendered by infection and vaccination.[9–13]

While some authors have touted improvements in surveillance and diagnostic methods[14] as an important factor in rising incidence, much of the recent resurgence literature has focused on explanations rooted in biology, attempting to identify changes in the epidemiologic landscape that have led to increasing pertussis transmission. Candidate explanations include loss of vaccine-derived immunity,[15] evolutionary changes in bacterial virulence[7] and antigenic escape,[16] and the lower protection afforded by acellular vaccines.[17] Intriguingly, it was recently suggested that in populations with a long-standing history of incomplete vaccination with an imperfect vaccine, pertussis resurgence may be inevitable even in the absence of changes in underlying transmission biology.[18] These authors demonstrated that a rebound in transmission, especially among older age groups, can arise as a natural consequence of demographic changes and changes in the immune profile of the population.

In this paper, our intention is to examine the epidemiologic dynamics of pertussis during the transition to mass infant immunization, partly with a view toward exploring the link between past patterns of transmission (and immunity) and contemporary pertussis epidemiology in the United States. Specifically, we present results from analyses of newly digitized records on pertussis incidence[19] in the 48 contiguous states and Washington, DC, from 1938 to 1955 (Figure 1). One of the key challenges to understanding how the profile of pertussis immunity in the United States has changed over the past few decades is the absence of quantitative information regarding vaccine uptake, with the earliest national estimates dating back only to 1962.[20] As shown in Table 1, while the combined diphtheria-tetanus-pertussis whole-cell vaccine became widely available in the United States in 1948,[21] other pertussis vaccines were already accessible and in sporadic use as early as the 1930s.[22–24] Indeed, some states, such as Michigan, were known to have already started routine infant vaccination ahead of the national immunization program.[2] Thus, we sought an explanation for contrasting dynamics in different states by examining potential predictors of vaccine uptake, demographic information, and insights from epidemiologic theory. Finally, we investigated the extent to which historical transmission dynamics in each state were predictive of modern observed resurgence patterns.

Figure 1.

Pertussis incidence in the United States from 1938 to 1955. The figure shows normalized square-rooted incidence rates for each continental state and Washington, DC, ordered by longitude. Yellow indicates high levels of incidence, and red indicates low levels of incidence. Missing data were interpolated for all states except Mississippi and Nevada.