The polio vaccine: friend or foe?

Before I launch into this entry, let me add the following disclaimer:

You and you alone are responsible for your health. This entry, and any other medical advice you read or get (even from doctors) with an honest, open-minded skepticism. Get the facts, and evaluate the evidence. Even doctors, and Ph.D. biostatisticians, make mistakes.

This entry continues a review of the book Vaccines: Are They Really Safe and Effective? by Neil Miller (2002 edition). I’ve already covered the tone of this book, and you might be able to guess it by its mere existence. It claims that vaccines, in general, are not safe and effective. This is a huge claim, but what makes this book different from other things you might read (especially forum postings and blog entries) is that the book is well-referenced. You can go to the source, at least in theory.

First, Miller tackles the polio vaccine. This vaccine is pointed to as one of the major successes of vaccination programs, taking polio from epidemic status to nonexistent. We even have the apocryphal story of Jonas Salk, leader of the research team that developed the first, inactivated virus vaccine, which was reported said to his celebrating team after their first victory with the vaccine in a trial: “We have to do this again.”

More below the fold.

Of course, Miller has a different take on the effects of the polio vaccine. First, he claims, it did not really take polio from epidemic status to virtually nonexistent. He deconstructs this claim as follows:

1. Polio is usually not devastating

His claim is that 95 percent of people exposed to polio virus will not develop symptoms, and, of the 5 remaining percent, most will experience only mild symptoms. Only 0.1 percent develop the serious scary symptoms. Of the sources cited to support these assertions, we have everything from the Physician’s Desk Reference to Mothering. It’s worth noting that Mothering tends to be friendly to alternative medicine and unfriendly to conventional medicine. While that doesn’t rule it out as a good source of information (a view I’m sure some of the “skeptics” will take issue with), it’s good to keep the biases in mind.

However, the Physician’s Desk Reference isn’t so unfriendly. Since I do have free access, I checked out some of these statements.

Of note, from the Pediarix (a combined vaccine that includes inactivated polio) cites CDC sources and claims that up to 2% get neurological symptoms.

The CDC’s source on vaccinations, of course, is the pink book. The “clinical features” section of the poliomyelitis chapter pretty much agrees with Miller’s assertions about the relationship between fraction of cases to severity of polio.

However, Miller does not cover the postpolio syndrome, which occurs 30-40 years after an episode of poliomyelitis which leads to paralysis. Postpolio syndrome is a non-infectious condition that leads to new muscle weakness, and is thought to have to do with how the body copes with the paralytic polio.

2. Cases were already falling when the vaccine was introduced

Miller presents something similar to the following graph in his book (click on images to enlarge):

Polio cases (Neil Miller)

I actually pulled this graph out of Miller’s article in Medical Veritas. Strangely enough, this is not exactly the same graph from the book. The only change I can see is that the x-axis begins at 1920 in the book, but begins at 1923 in the Medical Veritas article. However, the time series plots are not shifted. I find this a bizarre change.

The CDC pink book shows the following graph on polio cases:

Polio cases (CDC)

The story from this graph is different. This shows a strong temporal association between inactivated polio virus(IPV) and decline in polio cases. The cases for the subsequent polio vaccines (live virus oral and “enhanced potency”) are of course much weaker.

So here we have the case of the dueling graphs, and the dueling stories. And I think it will be an interesting digression to talk about the manipulation of graphs. Of course, I recommend anyone get How to Lie With Statistics (preferably from this link, so I get a kickback 😉 ), which talks about the different ways in which charts can be manipulated to tell any story the designer requires. So, on with the comparison of the two graphs.

A closer look at Miller’s graph shows the y-axis as “decreasing percentages.” The y-axis for the CDC graph appears to be raw number of cases. Miller’s graph goes from 1923 to 1978 (the one I show above goes from 1920 to 1975) (time on x-axis), and the CDC graph goes from 1950 to 2004. Both show 1955 as the year that the Salk IPV(inactivated polio virus) was introduced. Miller’s graph shows the vaccine introduced during a decline in polio; the CDC’s shows the vaccine introduced right before a huge drop-off. Miller’s graph includes both US cases and UK cases (with UK on a steeper decline before the introduction of IPV(inactivated polio vaccine)), and the CDC shows only US cases.

The next step, of course, is to validate Miller’s graph (and the CDC’s, but let’s check out Miller’s graph first). This is harder than you might guess, at least on the internet. A search for “International Mortality Statistics” polio returns many results that reference Miller’s graph. The other option is to try to scale the x-axis to be the same on both graphs. So what I ended up doing was reverse engineering both Miller’s and CDC’s graphs (using Datathief), and rescaling Miller’s graph to actual cases by making Miller’s and CDC’s graph intersect at 1950. What I got was the following result:

Neil Miller and CDC on polio cases

It’s worth noting that the graphs agree at 1950, 1955, and close to 1960 (though I might be suffering from minimum distance bias, rather than vertical distance). Miller’s graph tapers off much more slowly than the CDC’s graph after 1960, indicating that the graphs are probably measuring different things (or one is wrong, or I need to shift Miller’s down even more). The CDC gives a more detailed view of cases between 1950 and 1955, which is important information because there was a spike in cases then.

I noted the most bizarre thing about the CDC graph when reverse engineering it, though. The CDC claims the IPV(inactivated polio virus) was introduced in 1955. However, the arrow on the CDC graph in the pink book points to 1954, which was a timewise local peak in cases. I tried to account for the poor resolution of the graph, but the graph simply doesn’t agree with the text.

So for an overall evaluation, I don’t think that the CDC or Miller is telling the whole story. The question of whether the polio vaccine resulted in the eradication of indigenous polio (and indigenous/wild-type polio vs. vaccine-induced polio is a topic for the next blog entry in this series) is not confirmed or denied in this retrospective look at the data. Of course, it is not the only data available, as there is a global campaign to eradicate polio.

I’ve only covered a part of this question. I haven’t yet looked into Miller’s claims that public health officials declared the vaccine worthless and refused to vaccinate their own children. I also haven’t addressed the vaccine-induced polio, which the CDC in the pink book does tacitly admit is a problem with the live-virus vaccine (and claims that IPV does not have the problem). I have also not addressed redefinition of polio, which Miller claims occurred in 1955 and contributed a lot to the huge drop-off we see in the graphs. These are important topics to take up later.

Previous articles in this series:

Technorati Tags:


One Response

  1. […] from the title and circumstances, the book is very skeptical about the utility of vaccines. In the last examination of this book, I tackled Miller’s assertion that the polio vaccine was not responsible for the […]

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: