Apparently there was a lot of interest in a vaccine against
whooping cough around this time. Makes sense; it sounded pretty serious. People
had recently discovered that when growing the pertussis bacteria in the lab,
they started off with smooth-looking colonies that were virulent, which over
time sometimes got more rough and less harmful. Some graded these from Phase I
(smoothest, most virulent) to Phase IV (roughest, least virulent). Sauer and
others thought the Phase I would make the best vaccine.
Studying vaccines of pertussis was tricky: the way the
vaccine was made could influence things, obviously (as this study itself seems
to show), and diagnosing the disease could be uncertain, and they weren’t even
sure if having the bacteria in your body was even enough to make you sick; some
might have the bacteria and not be sick, though there seemed to be some
evidence against that idea.
Based on previous experience, James Doull, Gerald Shibley,
and Joseph McClelland decided that a study of 200-300 children in each group
(vaccinated and unvaccinated) could detect a difference in pertussis cases
decently. But just to be safe, they decided to go with a higher number, around
500 per group.1
They recruited these children at health stations in
Cleveland that gave out free milk. The requirements were that the children must
not have had pertussis before, and needed to have an older sibling who also had
not had pertussis (to increase the likelihood of exposure), and had to be
between 6-15 months old.
They started out planning to give every other child the
vaccine they made, but a lot of parents refused to let their children be
vaccinated, so they ended up counting some refusers as controls instead. So
there ended up being 483 vaccinated (most with all three doses, a few with only
one or two), 247 selected as controls, and 249 that refused. So almost 1000
total.
The vaccine was made with 5 recently isolated (and thus smooth
and virulent) strains of bacteria, grown on agar, scraped off, washed with
distilled water, and suspended in saline with 0.5% phenol. Standardized doses
were given once a week for three weeks, subcutaneously in the buttocks. Sounds
like fun.
They noticed some reactions to it, but not many. Some had a
slight fever on the same day, or a local reaction lasting up to 3 days. It was
claimed that 3 had convulsions, one of which was observed by a nurse.
After inoculation, the authors kept tabs on their subjects,
sometimes sending someone to check on the families. If any reported a case, an
epidemiologist or pediatrician obtained records and tests to confirm.
The distribution of race and sex in the groups seemed pretty
even, though there might’ve been more white people refusing the vaccine than
minorities. The age distribution was fairly even though too.
Ok, results: not great. 61 out of 483 vaccinated got
whooping cough, or 12.6%. Of those not vaccinated, either from randomness or
refusal, 71 of 496 (14.3%) got whooping cough. So at most, the vaccine was
11.6% effective, or prevented the disease in 11.6% of those who would’ve gotten
it. Pretty awful.
They did mention that some thought the vaccinated cases
might’ve been milder, but it seemed like a rather subjective judgment:
"The opinion of physicians who have seen representative attacks in both groups is that those in the inoculated have been milder. This is a difficult question to settle. There has been only 1 death, and that in a control child."
Overall, seems like pretty clear negative results. It wasn’t
a great study (no placebo, no blinding, not even very good randomization), but
usually one expects those problems to give more positive results, not less
positive. So it’s somewhat interesting, considering the positive results others
had observed with similar vaccines around that time.
Some people in later articles had some ideas about what the
difference might be:
"In three field studies...favorable results were reported. In another very thorough study [this one] in which the method of preparation of vaccine was slightly different unequivocal evidence of immunization was not obtained."2
"The Cleveland vaccine [this study] was also a Phase I vaccine [freshly isolated organisms] containing 10 billion organisms per ml., but in its preparation the organisms had been washed once with distilled water, while the Michigan vaccine [showing a positive result] had been washed once with saline."3
"These varying estimates of vaccine performance had at least two possible explanations. They might have reflected either differing manufacturing processes, or variations in the methodological rigour of the studies."4
"In his pertussis studies, Doull relied on alternation, which in his view was sufficient to prevent bias in selecting which children would receive the vaccine. However, the actual allocations departed from this 'ideal' somewhat: in one recruiting location, Doull used children whose parents refused inoculation as the controls, while at a later point in the study, he assigned the vaccine in a 2:1 ratio, with every third child serving as a control."5
Citations:
1. Doull, J. A., Shibley, G. S. & McClelland,
J. E. Active Immunization Against Whooping Cough: Interim Report of the
Cleveland Experience. Am J Public Health Nations Health 26, 1097
(1936).
2. Singer-Brooks,
C. & Miller, J. J. The Opsono-Cytophagic Test in Children with Pertussis
and in Children Vaccinated with H. pertussis Antigens. J Clin Invest 16,
749–761 (1937).
3. Perkins, J. E.,
Stebbins, E. L., Silverman, H. F., Lembcke, P. A. & Blum, B. M. Field Study
of the Prophylactic Value of Pertussis Vaccine. Am J Public Health Nations
Health 32, 63–72 (1942).
4. Jefferson, T.
Why the MRC randomized trials of whooping cough (pertussis) vaccines remain
important more than half a century after they were done. J R Soc Med 100,
343–345 (2007).
5. Marks, H. M.
James Angus Doull and the well-controlled common cold. J R Soc Med 101,
517–519 (2008).
