Disease Risk – Pertussis


Risk to a child from Pertussis if not vaccinated until the age of 5:

Pertussis, also known as whooping cough, is a highly contagious bacterial disease caused by Bordetella pertussis.  It’s characterized by severe coughing spells that end in a “whooping” sound when the person breathes in.  Historically, pertussis caused significant morbidity and mortality in the world.  In the 1940’s in the US, pertussis was responsible for more infant deaths than measles, scarlet fever, diphtheria, polio, and meningitis combined.[1]

Routine pertussis immunization got underway in the United Stated in the 1940’s and mass pediatric vaccination has been successful in significantly reducing both the morbidity and mortality of Pertussis.  Unfortunately, pertussis vaccination was also associated with high rates of adverse reactions related to severe neurologic disease and death.[2]  Although pediatric vaccination for pertussis has demonstrated the ability to protect unvaccinated infants < 6 months of age through herd immunity, pertussis is still endemic in highly vaccinated populations worldwide.  Despite high rates of immunization, cyclical outbreaks of pertussis in children < 5 still occur as documented by Fine and Clarkson [3].  Cherry suggests that this indicates that immunization controls disease but does not control the prevalence of the organism in the population.[4]

Due to a variety of reasons, both Japan and Sweden eliminated their pediatric pertussis vaccination programs in relatively recent times (Japan between 1975-1981, Sweden between 1979-1995).  Both countries re-instituted pediatric vaccination using newer vaccines with less adverse reactions.

Factors Not Considered: There is considerable evidence that the risk from severe pertussis disease is highly variable depending on factors influenced by economic and living conditions.  Morbidity and mortality due to pertussis is far higher in the developing world.  In a study from the UK, Maclure found that the risk of hospitalization from pertussis in children living in deprived households was almost 10 times higher than in areas where households were not deprived.[5]  For pertussis, overcrowding and unemployment were more correlated with pertussis hospitalization than vaccination rates.[5]  In the United States today, pertussis rates are highly correlated to race/ethnicity.  The reasons for the correlation are unknown although several hypotheses include the factors of living conditions and access to health care [6-8].

Case Fatality Rate: Medical treatment and modern living conditions have drastically reduced the rates of both morbidity and mortality from pertussis over time.  As described by Romanus [9], the fatality rate in major outbreaks in both Swedish and England and Wales outbreaks (1977-1979 for England and Wales, 1981-1983 for Sweden) remained low.  Cherry [2] makes a case that the England and Wales deaths were underreported during the 1977-1979 epidemic.  However, the experience from Sweden provides the most relevant information because they had the longest period of an unvaccinated population (17 years) in recent times.  The tracking of the disease in Sweden has been quite rigorous compared to other countries and during their period of non-vaccination, they instituted modern medical protocols to treat the disease including aggressive antibiotic treatment and post-exposure prophylaxis for infants below 6 months. [10]  Romanus documents only 3 fatalities over a 3 year period in Sweden (1981-1983 a period of highly endemic pertussis in Sweden).  Two of the cases were in children older than 5 years with severe congenital disease and the third was only 4 months old.[9]  Due to the extremely low number of fatalities in Sweden during the period of vaccination, this analysis can’t derive a usable case fatality rate from Sweden.

In the US, all of the recent deaths (9) from the epidemic in the highly vaccinated California population occurred in infants < 2 months and thus were not vaccine preventable.[6]  Kanai documents that in 1977 (during the period when Japan stopped vaccinating), 14 of 19  deaths (74%) occurred in infants 2 months old or less,  17 of 19 were < 6 months old (89%) and the remaining deaths all occurred in infants < 1 year old.[11]  These statistics illustrate that in modern times (in populations that have both high or low levels of vaccination), pertussis vaccination does not directly provide a significant reduction to risk of death for the individual since the vast majority of the risks to infants are at an age prior to vaccination.  The vast majority of the reduction in pertussis fatalities in highly vaccinated populations is due to reduced pediatric disease circulation and the resulting herd immunity protecting the vulnerable infant population.  This analysis is focusing on the individual incremental risk of death due to lack of vaccination.  The statistics from Kanai show us that only 11% of the deaths in infants < 1 year of age were vaccine preventable given the age distribution of deaths.  This is consistent with the notion that infants < 6 months account for the vast majority of serious and subsequently fatal cases of pertussis in both vaccinated and unvaccinated populations.  The risk to infants < 6 months has changed from historical times because prior to universal vaccination, infants < 6 months would have been maternally protected if they were breast fed.  However, it would likely require at least two decades of non-vaccination for this lost pattern to re-emerge – the length of time for unvaccinated females to bear children.  This analysis will base the incremental fatality ratio on the average US case fatality rate of 1% from Cherry [2]. The vaccine preventable case fatality rate used will be 0.11% of incidence from 0-1 year and 0 for all other ages.  The case fatality rate for children between 6-12 months of age is therefore estimated to be 0.22%.

Rate of Long Term Sequelae: According to Cherry [12], pertussis is a unique illness in that systemic manifestations are rare.  Cherry hypothesizes that the neurological symptoms of convulsions and encephalopathy are most likely caused by anoxia resulting from respiratory damage.  Pertussis epidemics can cause significant morbidity and hospitalizations in children < 5 years of age.  The Romanus study[9] documented around 18% of the hospitalized cases suffered neurologic or serious secondary sequelae or infection.  Although long term follow-ups were not directly studied by Romanus in Sweden[9], the study makes the following comment about the hospitalized cases:  “However, it is on record that all the surviving patients were apparently healthy on leaving the hospital and at the next visit.”[9]  An analysis of hospitalizations in Sweden (unvaccinated) illustrates again the vast majority of severe cases of pertussis (90%) occur in children < 5 years, with almost half (48%) occurring in infants < 1 year of age.[9]    Cherry [2] paints a more skewed picture in the US (a vaccinated population) with 67% of hospitalizations occurring in infants < 6 months of age.  None of the large studies of pertussis reviewed in this analysis provide any statistics of long term sequelae.  The few references found were all in infants < 6 months of age which is consistent with the repeated observation that severe disease is associated with the children in that age group.  Cherry indicates that long term sequelae does occur in encephalitis cases, but the reference is not from recent times.  The data from Cherry et al [12] also indicates that the majority of encephalitis cases were in children < 6 months of age.  Given the lack of evidence for long term sequelae in children over 6 months of age in the modern age, this analysis assumes an incremental risk for non-vaccination of 0.

Incremental Risk in Population with Low Rates of Vaccination (if a child does not vaccinate by age 5): The most modern statistics from an unvaccinated population comes from Sweden between 1979-1995.  From the analysis above, the mortality risk from Pertussis occurs in infants < 1 year of age.  This analysis takes the average incidence over 5 years around a peak year 1991 (1989-1993).[10]  Applying the case fatality rate of 0.11% against an average incidence of 914 per 100,000 yields a fatality rate of 1.006 per 100,000 or 1 in 99,375. 

Note: The theoretical case fatality rate for the US in unvaccinated populations is higher than it should be.  The data from Sweden [9] documents a total case fatality rate of 1 death per year or 1 per 90,000 over three years of outbreaks.  However, none of those deaths in Sweden were vaccine preventable.  This means that the vaccine preventable case fatality rate in unvaccinated Sweden was actually far lower.  This could be due to several factors including:  living conditions in the US differ from Sweden, medical protocol or care is better in Sweden than the US, or the 1% case fatality rate for the US is an overestimate in the modern era.  This analysis will use the calculated rate from above and it should be treated as an upper bound.

Incremental Risk in Highly Vaccinated Population (if a child does not vaccinate by age 5): Calculations of vaccine preventable risk are very difficult in highly vaccinated populations due to the lack of good statistics regarding vaccinated population counts, and percentage of vaccinated incidence.  As described above, in the case of pertussis, the reduction of deaths is almost entirely due to herd immunity.  If we review the historical epidemiological data from Cherry[12], the case fatality rate has dropped to approximately 1% of its previous historical value, largely paralleling the drop in pertussis incidence.  Taking 1% of the risk from populations with low rates of vaccination, yields a fatality rate of 0.01006 per 100,000 or 1 in 9,930,000.



[1] Gordon JE, Hood HI: Whooping cough and its epidemiological anomalies.  AmJ Med Sci 1951; 222:333-361

[2] Cherry JD, Baraff LJ, Hewlett E: The past, present, and future of pertussis. The role of adults in epidemiology and future control.  West J Med. 1989 Mar;150(3):319-28.

[3] Fine PEM, Clarkson JA.  The recurrence of whooping cough:  possible implications for assessment of vaccine efficacy.  Lancet 1982;1:666-9.

[4] Cherry JD:  Historical Review of Pertussis and the Classical Vaccine.  J Infect Dis 1996 Nov;174 Suppl 3:S259-63

[5] Maclure A, Stewart GT.  Admission of children to hospitals in Glasgow: relation to unemployment and other deprivation variables.  Lancet 1984 Sep 22;2(8404):682-5.

[6] Pertussis Report, January 7, 2011.  California Department of Public Health

[7] Melnick M.  California’s Whooping Cough Epidemic Hits Latino Babies Hardest.  Time Healthland, Friday October 29, 2010.  http://healthland.time.com/2010/10/29/californias-whooping-cough-epidemic-hits latino-babies-hardest/

[8] Ostrov BF.  Whooping Cough (Pertussis) Epidemic in California:  Tips for Convering the Story.  Barbara Feder Ostrov’s Helath Jounalism Blog.  http://www.reportingonhealth.org/blogs/whooping-cough-pertussis-epidemic-california-tips-covering-story

[9] Romanus V, Jonsell R, Berqquist SO.  Pertussis in Sweden after the cessation of general immunization in 1979.  Pediatr Infect Dis J. 1987 Apr;6(4):364-71.

[10] ELEVEN YEAR REPORT Pertussis surveillance in Sweden Progress Report October 1, 1997 with an executive summary.  SMITTSKYDDSINSTITUTETS RAPPORTSERIE NR 6:2009

[11] Kanai K.  Japan’s experience in pertussis epidemiology and vaccination in the past thirty years.  Jpn J Med Sci Biol. 1980 Jun;33(3):107-43.

[12] Cherry JD, Brunell PA, Golden GS, Darzon DT. Report of the task force on pertussis and pertussis immunization–1988. Pediatrics 1988; 81(suppl):939-84.