Adam Cifu, Churnalist
I liked my previous post on why I don’t think the shingles vaccine (actually our previous shingles vaccine, Zostavax, rather than the present one, Shingrix) does not prevent dementia, despite a Nature article and coverage pretty much everywhere to the contrary. I stand by my conclusions.
However, a couple of astute readers, one in the comments and one email, called me out for a bit of an unintentional sleight of hand in the math. It was unintentional, but I want to set the record straight.
One of the reasons that I argued that the results of this study was misleading was that the vaccine seemed to reduce the risk of dementia by more than it did shingles. This was kind of a mistake. I’ll quote the important lines from the article.
With regard to shingles:
When calculating the effect of actually receiving the zoster vaccine, we find a reduction in the probability of having at least one shingles diagnosis of 2.3 (95% CI = 0.5–3.9; P = 0.011) percentage points over the seven-year follow-up period; an effect (37.2% (95% CI = 19.7–54.7) in relative terms) that is similar in size to that observed in clinical trials of the live-attenuated zoster vaccine (Zostavax).
With regard to dementia:
…we find that actually receiving the zoster vaccine reduced the probability of a new dementia diagnosis by 3.5 (95% CI = 0.6–7.1; P = 0.019) percentage points, corresponding to a relative reduction of 20.0% (95% CI = 6.5–33.4).
When I made my statement that the vaccine was more effective at preventing dementia than shingles, I was comparing absolute risk reductions (ARR), 3.5% for dementia vs. 2.3 for shingles.
Clinically, the ARR is generally more important than the relative risk reduction (RRR), so I usually focus on that number.
Here, however, because the baseline rates of shingles and dementia are different, I should have highlighted the RRR, a number that argues that the vaccine is more effective at preventing shingles than dementia.
My conclusions don’t change, given all the other reasons I listed, but I wanted to make sure I was not guilty of anything like what I was blaming others for. Kudos to the readers who called me on this; I’ll be reaching out to send you one of our coveted t-shirts.
With regard to the recommendation to take the shingles vaccine, it appears that in the NEJM article, the incidence of post-herpetic neuralgia in those over 70 was 0.01% in the vaccine group vs 0.09% in the placebo group over 3.8 yrs. That's an NNT of 1,250. Pretty slim benefit.
The numbers for shingles itself are somewhat better, but there you're preventing a nuisance, not a lot-term morbidity.
Not sure why Dr. Cifu is so enthusiastic about pt's over 50 all taking the shingles vaccine, "full stop."
Keep in mind that they made an active effort not to use the actual receipt of Zostavax in their analyses. At no point in their analysis do they separate groups based on actual vaccine receipt. Both in the case of dementia and in the case of zoster, they note a reduction in the overall incidence based on the magnitude of the regression discontinuity, and then divide by the fractional overall uptake of the vaccine to multiply the "scaled" effect to account for the incomplete uptake.
As I wrote before, this leads to the counter-intuitive result that the lower the vaccine uptake, the higher the scaled effect becomes. If something else caused the effect - chance or an unsuspected outside influence - this technique would not only fail to find that, but would be utterly insensitive to it.
This isn't classical epidemiology at all, so ARR and RRR aren't meaningful in the usual sense. This is handwaving algebra dressed in fancy doctor clothes and pretending to belong at the party.