Letters Response

“Antivaccine Lobby” replies to the BMJ

BMJ 2010; 341 doi: https://doi.org/10.1136/bmj.c4001 (Published 27 July 2010) Cite this as: BMJ 2010;341:c4001

Re: “Antivaccine Lobby” replies to the BMJ

I am afraid that Dr Puliyel misunderstands my comments, and he makes a bad fist of the comparative mortality data. To make a comparison of the rates of infant deaths following pentavalent vaccine in Kerala to the SIDS rate in the USA is entirely spurious; his assumption is that someone (myself perhaps?) is attributing deaths that coincidentally follow pentavalent vaccination to SIDS, but I have never made this claim.

The fact is that when a child dies on the same day as a health intervention such as vaccination, the natural reaction is to blame the intervention, invoking a cause and effect relationship. However, this scenario does not even play out the way Dr Puliyel suggests.

To start with, we need evidence that the claimed deaths in Kerala occurred on the same day as vaccination - this testimony is lacking (and if I have missed it, I stand corrected, and would be grateful if Dr Puliyel could provide the evidence). Even so, we can still make some form of comparison if for the sake of debate we assume all deaths took place within 24 hours of vaccination.

However, why should we choose SIDS as the comparator? Surely any child who dies the day of vaccination is likely to have the vaccine raised as a probable cause of the death, whether that child actually died of something else such as SIDS, or as the case was in Thailand other conditions such as respiratory infection and congenital heart disease.

So we need to consider all infant deaths. The under one-year-old IMR in India is around 50 per thousand, meaning that in a sample population of 40,000 vaccine recipients, we would expect on average 50x40/365 deaths every day. That is over 5 deaths, which would happen each day, whether vaccine was administered or not. However I note that Dr Puliyel refers to recent Kerala data showing an IMR of 14 per 1000. Using these data, the expected number of deaths on any day would be 14x40/365 (1.5).

To find that there were 1.5 deaths the day "following" vaccination in Kerala is really quite meaningless in statistical terms (recall that Dr Puliyel gave the death rate after vaccination as 1 per 10,000, so the reported number was four deaths). We also do not know if we are just counting deaths within 24 hours of vaccination, or within the first few days following vaccination (which would add a multiplier to the number of expected deaths).

I am afraid that we must stop clinging to prejudices and not attribute in a post hoc ergo propter hoc manner isolated events that might clearly be purely coincidental.

Unless we have firmer evidence of course... Dr Puliyel seems to suggest that because diagnosis of vaccine reactions may be difficult, that we should just presume that the deaths are due to vaccination. That seems perverse. It is unclear what pathological mechanism Dr Puliyel is invoking here. If it is a form of hypersensitivity/anaphylaxis, as he first suggested, then there usually will be typical symptoms and clear signs which make the diagnosis fairly straightforward (people can see last week's BMJ for a good refresher). But there is no reported narrative here to indicate any anaphylactic process. Now Dr Puliyel has also mentioned SIDS, which implies a different type of presentation. It would help if we could see reliable medical reports pertaining to the deaths; until then all we can do is speculate.

Finally, I must say I am well aware that penicillin allergies can be detected by skin testing; however in normal medical practice skin tests are not done on each and every infant prior to giving a penicillin-containing antibiotic, as Dr Puliyel stated.

Competing interests: No competing interests

20 February 2013
Peter J Flegg
Victoria Hospital
Blackpool, FY3 8NR