Flu vaccine blunder

I got my flu shot - did you get yours? Well, you should.

There, now that I've said that, I can explain why the shot might not do you any good. Unfortunately, the CDC (Centers for Disease Control) and WHO (World Health Organization), who jointly decide what flu strains to put into the vaccine each year, may have made a serious mistake this year. I don't want to criticize these organizations, which do wonderful work each year that benefits public health in countless ways, but if I'm going to be consistent, I can't keep silent when I see a mistake, even if I like the organization(s) behind it.

The short version of the problem is that the CDC/WHO panel decided to put the same strain of H3N2 (the most common type of human flu) into the vaccine this year as we had last year. They made this decision despite the fact that the data clearly showed that a new H3N2 was emerging. The consequence is likely to be - I fear - that those of us who get the flu shot will have almost no protection against this year's flu. Not good.

Some background: the flu shot has 3 strains in it: H3N2, H1N1, and influenza B. Last year was a mild season dominated by H1N1, which is usually mild. The CDC/WHO committee decided to replace the H1N1 in the shot, using a recent isolate - so if you get the shot, then at least you're protected against H1N1. However, H3N2 is dominant in most years, and "bad" flu years are always H3N2-dominant. Why do I say a new strain was emerging? Well, here is some of the data:In this chart, which shows flu cases throughout last season (2006-7), the blue bars indicate H1N1 cases and the red bars show H3N2. (Yellow bars are untyped, so we don't know what those are.) Notice that through the middle of the season, which was in mid- to late-February, H1 was dominant. But in the later part of the season, from March on, H3 became dominant, and by late March almost all the cases were H3N2. Clearly, H3N2 was taking over again.

The WHO's own report on this data admits that a new strain was emerging. So why did they choose an OLD strain - from 2005, no less! - for this year's vaccine? Here is what their report says: "An increasing proportion of isolates was distinguishable both antigenically and genetically from the vaccine strains; however, antigenic analysis did not reveal the emergence of a sufficiently well characterized antigenically variant group." In other words, the new isolates were clearly different from the vaccine strain (A/Wisconsin/67/2005), but they weren't sure which of the new isolates was going to be the new one. They also reported that "the lack of egg isolates precluded the selection of a new vaccine candidate"; in other words, they weren't sure if they could grow the new isolates in eggs, which is how we make vaccines here in the U.S. (Using eggs is an outdated system that must be changed, but I'll have to save that topic for another blog.)

So the bottom line is that: (a) they knew that a new strain was emerging, (b) they weren't sure if we could grow it in eggs, and (c) they weren't sure which of the new isolates was the best choice for the vaccine. So they decided to put the old strain in the vaccine, even though they knew it was ineffective. In fact, their own data show that it was only effective in about 34-58% of cases last season - a number that is likely to be much lower this year.

Part of the problem is that the panel that chooses the flu vaccine strains continues to rely heavily on traditional serotyping methods rather than the newer, more precise genotyping methods available through influenza genome sequencing. A related problem is that the CDC doesn't release its influenza genome data, as I and others have pointed out in the past. (See our letter to Nature on this topic from March 30, 2006 - despite numerous calls to release data immediately, the CDC and WHO still sit on their flu data, often holding it for years.)

Why doesn't the CDC release flu genome data immediately? Part of the problem, though they won't admit it, is that they don't want to be second-guessed by "outsiders" on what strain to put into the vaccine next year. The data used to make this decision should have been available last winter, but the CDC only shares it with three other WHO collaborating centers (more info here). Of course they are trying to do the right thing - to choose the right vaccine strain - but they shouldn't be afraid of hearing the advice and opinions of other scientists who might disagree with them.

Still, get your flu shot. Maybe we'll get lucky and we'll have another H1N1 flu season. I hope so.


  1. It seems that most (or at least a plurality) of the strains are untyped. Does that mean that whatever shot we get, it only protects against one-half the viruses?

  2. Good question - actually, the yellow bars are "unsubtyped", meaning they were identified as influenza A, but the CDC didn't do the additional tests required to distinguish H3N2 from H1N1. It is likely that nearly all of these influenza A's are one or the other: in a typical year, over 99% of the influenza A viruses belong to one of these two types.

    The shot protects against H1N1, H3N2, and influenza B. In most years when H3N2 is dominant, 80-90% of the viruses will belong to that flu type. So if this year's vaccine has a strain of H3N2 that doesn't match the H3N2 viruses in the population, then we will have little or no protection against that virus. The shot will still protect you against flu B and H1N1.

  3. The Washington Post today (Feb. 10, 2008) has a feature article by David Brown entitled, "This Season's Flu Strains Are Not a Good Match for Vaccine".

    I was right. Somehow I doubt the CDC will admit they were wrong, though.


Markup Key:
- <b>bold</b> = bold
- <i>italic</i> = italic
- <a href="http://www.fieldofscience.com/">FoS</a> = FoS

Note: Only a member of this blog may post a comment.