A Critical Analysis of the Pertussis Vaccine

In a recent post, we showed that the measles vaccine helped bring down deaths from 400 to 20 in the immediate years after the vaccine’s licensure and we painted a rather optimistic outlook by examining changes in trends, and now I’d like to use the same technique to examine the Pertussis Vaccine for Whooping Cough.

The key insight was that one can compare post-vaccine deaths to various points in time in the past and the further back you go, the more of that decline will be attributable to non-vaccine factors like nutrition and sanitation. For example, Kinlay et al conclude that less that 3 percent of the decline in mortality from infectious diseases, was attributable to vaccines. Meanwhile, government agencies like comparing the 400 Measles deaths in 1962 to the 20 deaths about 5 years after, reporting a 95% decline. There is a valid case to be made for both, but w.r.t. the latter one cannot simultaneously assert that vaccines save millions of lives. By looking at changes in trends rather than changes in levels, we can remove the impact that the timing of vaccine introduction has on its overall effectiveness. Please see the previous post for a further discussion on the issue.

The following graph shows the log-rate of Pertussis mortality rates from the CDC. In 1920, there were 13,308 deaths, which when divided by the population yields a rate of 12.5 per 100,000. The Y-axis below shows the logarithm of this quantity, which for this year is around 4. Taking the logarithm helps to straighten out the exponentially declining mortality rates and allows us to zoom in on fluctuations in the post-vaccine era, which would otherwise be too small to see.

Left Y-Axis Shows Log Mortality Rates. Right Y-Axis Shows Coverage Rates

Left Y-Axis Shows Log Mortality Rates. Right Y-Axis Shows Coverage Rates

By the mid 1940’s there were about 1,700 deaths (1.2 per 100K) and in the 1980’s there were usually less than 10 deaths (around 2 per 100M), which indicates a dramatic change in mortality rates 30 years following the introduction of the Pertussis Vaccine.

Note that vaccine coverage rates ranged from 65 to 75% from 1960 to 1980 and then increased dramatically after the introduction of the acellular vaccine in 1991 [3]. We will comment more on this later.

As with most of my analyses, I exclude the 1900’s to 1920’s from my analysis because that time-period experienced unusual economic hardships like the Great Depression and World War I. The world and outlook was vastly different during that time period.

Licensing Dates

The first pertussis vaccine was developed in the 1930s and was in widespread use by the mid-1940s, when pertussis was combined with diphtheria and tetanus toxoids to make the combination DTP vaccine. The acellular vaccine replaced the whole-cell vaccine in 1991 because of safety concerns.

Characterizing the Trends

By comparing slopes on the curve, we can compare the relative impact that vaccines have on fast-forwarding the decline in mortality rates.

Period Slope T-Stat Diff
1920-1946 -0.05 N/A
1946-1991 -0.07 4.84
1992-2007 +0.03 -10.0

The slope from 1920 to 1946 is -0.05, while the slope from 1946 to 1990 is -0.07, so there is an improvement in the elimination of Pertussis, and the difference in slopes is statistically significant (t-stat = 4.84, p-val = 7.9E-06). However, when the acellular vaccine was introduced, the slope flattens out to +0.03, and the t-stat on the difference in slopes between the following periods is -10.0 (p-val 1.45E-14), which means that Pertussis elimination came to a grinding halt.

While it does not mean causation, the post 1991 period suggests that higher coverage rates are correlated with higher mortality rates.

A Turn of History

There are two types of pertussis vaccines, whole-cell (wP) and acellular (aP). Whole-cell pertussis vaccines contain a whole-cell preparation, which means they contain killed, but complete, B. pertussis bacteria. The acellular pertussis vaccine is more purified and uses only selected portions of the pertussis bacteria to stimulate an immune response in an individual.

While the whole-cell vaccine was able to demonstrate efficacy, 90% of the children developed systemic reactions such as fever, drowsiness, vomiting, and anorexia [2]. Between 1 and 2 out of 1000 children became unresponsive or experienced serious convulsions, and concerns about the safety led to the development of the more purified acellular vaccine that had less side effects. These complications must have contributed to the declining coverage rate, which fell from 79% to 65% between 1971 and 1985.

Since aP uses only selected portions, one might naturally expect reduced efficacy after the 1990s, and that is what we observe in the mortality data above, but controlled trials have also been conducted to re-enforce this idea.

Clinical Evidence That the aCellular Vaccine Fails To Prevent Transmission

A new study, conducted by the FDA, was released in 2013, which helps to provide a better understanding of vaccines for whooping cough. Based on an animal model with baboons, the study shows that acellular pertussis vaccines may not prevent infection from the bacteria that causes whooping cough in those vaccinated or its spread to other people, including those who may not be vaccinated. So, this study that vaccinated infant baboons at 2, 4, and 6 months with aP or wP and then challenged them with B. pertussis at 7 months, provides proof that it does not provide herd immunity and that it furthermore endangers immuno-compromised individuals by spreading infections to others. The study concluded:

Baboons vaccinated with aP were protected from severe pertussis-associated symptoms but not from colonization, did not clear the infection faster than naive animals, and readily transmitted B. Pertussis to unvaccinated contacts…

While all groups possessed robust antibody responses, key differences in T-cell memory suggest that aP vaccination induces suboptimal immune response that is unable to prevent infection. These data provide a plausible explanation for pertussis resurgence.

Another recent study by Smallridge that published in the Journal of Infectious Diseases arrived at similar results and conclusions in their study on mice [4]:

whole-cell vaccination was able to prevent transmission, but an acellular vaccine that effectively controls disease failed to control shedding and transmission.

Smallridge further points out that the acellular vaccine “fails to confer the full benefits of herd immunity” and highlights additional risks to immuno-compromised individuals.

It is possible that the effects of acellular vaccination could mask symptoms to allow infected individuals to act as unsuspecting resevoirs for potential spread to more susceptible individuals

The good news is that the vaccinated groups had milder symptoms, showed no signs of reduction of activity, loss of appetite, or other outward signs of disease. Moreover, white blood cell counts did not change, which indicated that they were more resistant to severe signs of disease and morbidity. However, this latter finding contradicts the rising mortality rates seen in the graphs above. Their study was quite short-term (and they were challenged only one month after vaccination) and the FDA study does emphasize issues with waning immunity:

Recent cohort and case-control studies concluded that 5 y following the fifth aP dose, children are fourfold to 15-fold more likely to acquire pertussis compared with within the first year, consistent with waning aP immunity

The results are worth reiterating: even after 5 doses of aP, you are about 10 times more likely to acquire Pertussis than after the first year. One of the 3 referenced studies can be found here.

Pertussis Outbreaks in Fully Vaccinated Populations

Whooping cough outbreaks in fully vaccinated populations are more common than might be imagined. Here are a few recent outbreaks worth mentioning:

  • Summit, Co., Mar 2015: 19 Cases of Whooping Cough among students who were fully up-to-date on their vaccinations.
  • Montgomery, Aug 26, 2014: 15 Cases of Whooping Cough. Health officials confirmed that “all of the campers who developed whooping cough were vaccinated against it.”
  • California Outbreak 2010: Vaccine effectiveness was as low as 24% in children aged 8-12 years.

A more complete list of various outbreaks can be found here.

However, one thing to note, which vaccine critics nearly always get wrong, is that the expected proportion of reported cases that are vaccinated can be quite high. Mathematically, this has to do with Bayes Law, which is counter-intuitive to human nature. Vaccinologists calculate vaccine effectiveness using the screening method:


where PCV is defined as the proportion of cases vaccinated, PPV is defined as the proportion of population vaccinated, and VE is defined as vaccine effectiveness. If you assume 85% effectiveness (VE) and a 95% coverage rate (PPV), then you can solve for PCV. The result is that in any given outbreak you should expect 74% of individuals to have been vaccinated. Hence, if less than 3/4th of the outbreak constitutes vaccinated individuals, the vaccine should be deemed effective, otherwise it should be viewed as a failure. “Anti-vaxxers” will nearly always mistakenly believe that the reference point should be a low number. The model may miss some subtle points about the networked nature of transmission and more complicated models may be warranted, but that will most likely be the subject of a future blog post.


The whole-cell Pertussis vaccine significantly increased the rate at which whooping cough deaths were falling. 10 years following the introduction of the Pertussis vaccine, deaths declined 72% from 1,679 to 467. However, the statistical significance may not be that economically meaningful in the near-term because extrapolating pre-vaccine trends over the same 10-year period would have achieved roughly the same 72% decline. A meaningful impact could not be seen until coverage rates for the whole-cell vaccine reached nearly 80% in 1971, but complications and adverse events caused low participation rates and the need to rethink the vaccine. While the acellular vaccine was pursued aggressively in the 1990s and as coverage rates succeeded in pushing over 95% in 1996, the incidence of deaths primarily associated with B. Pertussis have continued to rise. Both the clinical evidence and the mortality statistics seem to refute claims that a resurgence in Pertussis is the result of a rise in religious and/or personal belief exemptions.

  1. Warfel, Jason; Zimmerman, Lindsey; Merkel, Tod; Acellular pertussis vaccines protect against disease but fail to prevent infection and transmission in nonhuman primate model; PNAS (Jan 2014) vol. 111 no. 2
  2. Cody CL, Baraff LJ, Cherry JD, Marcy SM, Manclarck CR. The nature and rate of adverse reactions associated with DTP and DT immunization in infants and children. Pediatrics 1981;68:650-60
  3. Simpson, D., et al.; Forty Years and Four Surveys; Am J Prev Med 2001, 20
  4. Smallridge, W., et al. Different Effects of Whole-Cell and Acellular Vaccines on Bordatella Transmission; Journal of Infectious Diseases (2014)
  5. Klein, N.; Waning Protection after Fifth Dose of Acellular Pertussis Vaccine in Children; The New England Journal of Medicine (2012)
  6. Witt, MA., et al.; Unexpectedly limited durability of immunity following acellular pertussis vaccination in preadolescents in a North American outbreak.; Clin Infect Dis 2012 Jun; 54(12)
  7. Farrington, CP; Estimation of Vaccine Effectiveness Using the Screening Method; Int. J. Epidemiol. (1993) 22 (4)

About PD

PD is passionate about applying his background in math, statistics, and economics to apply new and interesting ideas about health, nutrition, and the incentives that drive products and the policies that surround them.