046 - Results of Field Studies With Poliomyelitis Vaccine

One of the vaccines I'm excited to learn about is polio, because apparently it works well enough that we've almost eradicated it from the world. But apparently it took a while to develop, probably because it took a while to figure out what caused it. (Something way smaller than bacteria? Whoa)

But in today's study, Gilliam and Onstott do a pretty good trial to test a vaccine against polio!¹ This version was developed by someone named Brodie, so it's not one of the ones in use today (which are Salk and Sabin), but I wonder why not. Apparently it was made of an emulsion of virus from monkey spinal cords, treated with formaldehyde long enough to inactivate the virus.² It seemed safe enough in preliminary tests, so the authors went ahead with a field trial.

They selected several communities in North Carolina and Virginia, near where polio outbreaks were occurring but not near enough to be in the midst of an outbreak (they wanted to make sure there was enough time to establish immunity before exposure). To parents of young children in these communities, they offered the vaccine as experimental and reasonably safe, and got 1452 volunteers for the trial in total.

Of these, they split them into those who would receive the vaccine, 766, and the control group, 686, based on the place of their surnames in the alphabet. Nice and random, supposedly. Each patient's physician would be the one doing the vaccination though.

Brodie cooperated with the study by providing the vaccine and also by withholding it from anyone else in the area who might want it, so that the researchers could be sure the controls really were controls.

When it came time to actually get the shots, only 458 volunteers actually showed up; a few of the controls managed to come upon a vaccine in other ways too, so they were excluded from the study.

There were two doses of vaccine, about two weeks apart. At first, they tried giving a shot of novocaine along with the vaccine to make it less painful, but it seemed just to complicate things without helping much. So in the end, 422 subjects got both doses, mostly at the right interval, and 36 got only one.

Gilliam and Onstott were closely involved in following up with the patients, asking about side effects or symptoms they may have had, since apparently the physicians missed some things. They observed that 50% of the subjects had some sort of reaction, though mostly mild and temporary: redness, swelling, etc. Most of the reactions seemed to be in patients that had received the novocaine. Hmmmm.

However, there were 15 local reactions that were more serious. Fourteen patients had an abscess or similar at the injection site, all of which healed, sometimes slowly.

Seventeen patients had general reactions (like fever or headache or nausea), which isn't bad, and these were usually associated with the abscesses. But some of the reactions were disturbing. For example, one girl fainted and stopped breathing after the shot, though she was fine after they revived her. Another had leg cramps at the injection site, though they got less frequent over time.

So it's pretty clear that this vaccine was not as safe as they thought, or at least not as safe as it should be. Better than being vulnerable to polio? Perhaps, but still not ideal.

What were the results other than safety? Well, of people who weren't involved in the study at all, about 6 in 10000 had polio, so the number of people you'd expect to have the disease in a similar group of a thousand is about 1, plus or minus 1. So between zero and two. In actuality, they didn't see any cases in either the vaccinated or controls. The authors calculated that even if the vaccine were 100% effective, they'd have to have about 10000 patients in each group to see a significant difference between them.

So that's unfortunate; the study was pretty much useless for judging effectiveness. As a later publication says,

"They were able to show, however, that Brodie's vaccine was relatively harmless, although they encountered a few very disturbing reactions."²

I had some other issues with the design of the trial as well. First, a minor issue: randomizing by alphabet? I don't think it's safe to assume that there isn't some bias inherent in people's family names, such that a group of people with names from the first half of the alphabet would be effectively identical to a group of people with names from the second half. But maybe.

Second, there's no placebo, which means there's almost no blinding (everyone knows who gets the vaccine and who doesn't), so bias in diagnosis could be introduced (like assuming people who got the vaccine are less likely to get polio, so being more likely to attribute symptoms to some other disease).
It also means that people dropping out of the study may not have been random; certain types of people might drop out more often in the vaccinated group but not in the control group (because there's no shot or other commitment involved), so the groups wouldn't be equivalent.

It didn't seem like there was much monitoring of rates of adverse events in the controls, to see if the observed side effects could be reasonably attributed to the vaccine or not, but perhaps that would need a different kind of study.

So overall, it's not great.

Citations:

1.

Gilliam, A. G. & Onstott, R. H. Results of Field Studies With Poliomyelitis Vaccine. Am J Public Health Nations Health 26, 113–118 (1936).

2.

Maxcy, K. F. Changing Conceptions of the Prevention of Acute Anterior Poliomyelitis. The American Journal of the Medical Sciences 192, 436–444 (1936).