August 2010: Onset intervals

This month I will explore onset intervals, the number of days between taking a vaccination and getting adverse symptoms. I will examine VAERS data by asking this question:

What is the typical onset interval, and does it vary with each vaccine?

As my January 2010 report showed, examining the database by looking at all of the 70 different vaccines shows way too much information to examine clearly. The solution is to graph vaccine types instead. This groups the vaccines according to the diseases they are attempting to prevent. Since there are only 27 different diseases, the graph is easier to analyze.

So for each vaccine type, we want a plot of the number of VAERS events that have a particular onset interval. To do this, check “Show Graph” (in part 1 of the search form) and under it graph “Onset Interval” and “Vaccine Type”. You can uncheck “Show Table” and “Show Event Details”.

Here is the graph (ignoring the “unknown” onset intervals):

The graph clearly shows that most adverse events happen in the first few days. There seems to be a spike from 10 to 30 days, but this is really a graphing side effect caused by the fact that it starts off going one day at a time and then switches to groups of days (10-14, 15-30, 31-60, etc.) If we redo the graph so that days are never grouped, and then look more closely at the first month, it shows a smooth tapering-off:

From these two graphs, some trends can be seen. The second graph, which zooms closely into the first month, shows a period from 7 to 10 days after the vaccination when some vaccines have a resurgence. What this tells us is that for most vaccines, if there is no adverse reaction in the first 6 days, there probably won’t be one at all. The exceptions are the rubella (German measles) and varicella (chicken pox) vaccines, which may still cause reactions up to 10 days.

But the first graph shows another spike that happens years after the vaccine has been given, and it happens only for varicella. This vaccine seems to have extremely long onset intervals, and the graph shows that there are thousands of reports of damage that happen years after the innoculation.

What is going on?

A look at the reports shows two unfortunate reactions:

  • Varicella. In some cases, the vaccine simply does not work, and the patient develops chicken pox years later. A search of the VAERS data for late onset of chicken pox (more than 120 days) after getting the varicella vaccine turns up 922 reports.

  • Zoster. In some cases, the the patient develops zoster (also called “shingles”) years later. A search for late onset of shingles after getting the Varicella vaccine turns up 775 reports.

Can the varicella vaccine cause shingles? When I did my original research for this this article, just a few months ago, I found a document on the CDC website ( which had this in it:

How often does zoster (shingles) occur following varicella vaccination?

Varicella vaccine is a live virus vaccine, and may result in a latent infection, similar to that caused by wild varicella virus. Consequently, zoster in vaccinated persons has been reported. Not all of these cases have been confirmed as having been caused by vaccine virus. The risk of zoster following vaccination appears to be less than that following infection with wild-type virus. However, longer follow-up is needed to assess this risk over time.

This statement is no longer available on the CDC website, but it can be found at the Internet ‘wayback machine’ which saves old copies of web pages. A 2005 New York Times article also discusses this problem.

So even if the CDC no longer believes that the varicella vaccine can cause shingles, it still seems to be the case from the VAERS data.