Cont: The One Covid-19 Science and Medicine Thread Part 5

Association Between Guillain-Barré Syndrome and COVID-19 Infection and Vaccination: A Population-Based Nested Case-Control Study

ETA: The study was conducted in Israel, which used almost exclusively the BioNTech vaccine (aka Pfizer vaccine) during the study period.

Guillain-Barré Syndrome is quite rare so you need to study millions of people to get enough cases to be statistically significant.
Rare is relative. I've seen 3 cases (one in the hospital and 2 friends) which doesn't seem rare to me.

From Wiki:
Guillain–Barré syndrome is rare, at 1 or 2 cases per 100,000 people every year.[3][4] Both sexes and all parts of the world have similar rates of disease.[1][3]
You don't need to look at "millions" in this case. And there are millions of COVID cases and vaccinations.

They looked at over 3 million cases. From your link:
We conducted a nested-case control study in a cohort of 3,193,951 patients aged≥16 years, without a diagnosis of prior GBS, from the largest healthcare provider in Israel.
GBS occurs in 3 months or less from the offending cause (IIRC, it may be 6 weeks) and they followed these folks for a year.

IOW I find the study credible.
 
Last edited:
Then how did they get statistically significant results with only 836 subjects?

They studied a cohort of 3,193,951 patients. In other words, "millions of people", like I said.

Of those, 76 patients were diagnosed with GBS during the 18-month study period. Each of the 76 patients was matched with 10 randomly selected people of the same age and sex from the same cohort who were not diagnosed with GBS.

So, yes, only 836 became subjects for analysis, but they started out with millions so that the number of cases would be large enough to be meaningful.
 
They studied a cohort of 3,193,951 patients. In other words, "millions of people", like I said.

No. That's the population from which the subjects for the study were drawn. It's like when you do a survey among eligible voters in the US. The population is millions, but you only interview a thousand, say. Your sample size isn't the population of millions from which the 1000 were drawn, it's the N=1000 that were actually studied.

In the Guillain-Barré study, 76 cases were compared to 760 controls. That's it. The remaining members of the cohort were not studied. Nothing about them entered into any of the study results. The sample size, N, was 836, and all the calculated statistics—the effect sizes, the confidence intervals, the p-values, etc. reflect that sample size.

ETA: To put it another way, if the sample size were 3,193,951, then 76 cases would have been compered with 3,193,875 controls, and the results and the computed statisitics would have reflected that comparison instead. That would have been another way to conduct the study, but that is not what they did.
 
Last edited:
Let me reword what I wrote earlier since some people want to nitpick.

Guillain-Barré Syndrome is quite rare so you need a population of millions of people to get enough cases to put together a statistically significant study.

Which is obviously what I meant.

If, at the beginning of the study you say we're going to study 1000 people for the next 2 years, you might not get any cases at all in just 1000 people.

GBS occurs worldwide with an overall incidence rate of 1–2 cases per 100,000 people per year,
(source: Google search)
 
Let me reword what I wrote earlier since some people want to nitpick.

Guillain-Barré Syndrome is quite rare so you need a population of millions of people to get enough cases to put together a statistically significant study.

Which is obviously what I meant.


And it isn't true. One way that studies of rare diseases are done is to select patients from a medical practice that specializes in that disease. Then you don't have to screen millions of patients. Depending on how specialized the clinic is, up to 100% of their patients could qualify.
 
Last edited:
And it isn't true. One way that studies of rare diseases are done is to select patients from a medical practice that specializes in that disease. Then you don't have to screen millions of patients. Depending on how specialized the clinic is, up to 100% of their patients could qualify.

Is there a medical practice in the world that only treats GBS and no other medical condition?

 
Is there a medical practice in the world that only treats GBS and no other medical condition?


Did I say there needs to be a medical practice in the world that only treats GBS and no other medical condition? Use your brain. Would there need to be? Should we pretend you're a student and make it an essay question?
 
One way that studies of rare diseases are done is to select patients from a medical practice that specializes in that disease.

Which could introduce some sort of inadvertent selection bias.

Perhaps you should write the authors of this study and explain to them that they're doing it wrong.

From the study:
Methods: We conducted a nested-case control study in a cohort of 3,193,951 patients aged≥16 years, without a diagnosis of prior GBS, from the largest healthcare provider in Israel. Subjects were followed from January 1st, 2021, until June 30th, 2022, for the occurrence of GBS.

If you want to target in your study people who haven't already been diagnosed with GBS, you will need to start with a large population because the rate of occurrence is only around one to two cases per 100,000 per year.

And if you want to study the incidence of GBS in people without a diagnosis of prior GBS after COVID infection and after the particular vaccine, which only began to be distributed in Israel from 19 December 2020, 12 days before the start of this study, the more people you have to study, the better.
 
Which could introduce some sort of inadvertent selection bias.


Yes, it could. That's the first the correct thing you've said on the subject.

Perhaps you should write the authors of this study and explain to them that they're doing it wrong.


I didn't say the authors were wrong. I said you were. That's why I'm writing to you and not to them.

If you want to target in your study people who haven't already been diagnosed with GBS, you will need to start with a large population because the rate of occurrence is only around one to two cases per 100,000 per year.


Yeah, but that's not really what they did, despite how they worded that statement. Their study is retrospective. They took patients in their database who had been diagnosed with GBS and matched them to patients in their database who were not, and compared the COVID and vaccination rates in each group. That is not conceptually different than taking patients from a GBS clinic and matching them to community controls and comparing their respective COVID and vaccination rates. Basically, because they had a large database, they were able to find enough cases. But you don't need a large database to study a rare disease; you need a population of cases.

And if you want to study the incidence of GBS in people without a diagnosis of prior GBS after COVID infection and after the particular vaccine, which only began to be distributed in Israel from 19 December 2020, 12 days before the start of this study, the more people you have to study, the better.


I don't know what you are talking about because their study design does not allow them to estimate the incidence of GBS after COVID infection. It allows them to estimate the inverse: the incidence of COVID infection in those with a GBS diagnosis.
 
Last edited:
Basically, because they had a large database, they were able to find enough cases.
That's the only point I was making in that statement, which you just had to nitpick for the sake of argument.

But you don't need a large database to study a rare disease; you need a population of cases.
To get a large enough population of cases for a rare disease, you need a big population to generate those cases. Because it's rare.

The interesting finding here is that vaccination is associated with lower risk of GBS and SARS-CoV-2 infection is associated with lower risk of GBS.
Conclusions: Our study suggests that SARS-CoV-2 infection is associated with increased risk of GBS, while Pfizer-BioNTech COVID-19 vaccine is associated with decreased risk of GBS.
 
The interesting finding here is that vaccination is associated with lower risk of GBS and SARS-CoV-2 infection is associated with lower risk of GBS.

Typo? SARS-CoV-2 infection is associated with higher risk of GBS.
 
That's the only point I was making in that statement, which you just had to nitpick for the sake of argument.

No, you said you need a large population. You don't.

To get a large enough population of cases for a rare disease, you need a big population to generate those cases. Because it's rare.

Try and remember more than one post back. You need a population of cases. One such population, if you recall, is the population of cases from clinics that get a lot of such cases. This, in fact, is the more common source of cases for studies of rare diseases.
 
Typo? SARS-CoV-2 infection is associated with higher risk of GBS.

The abstract says, "Multivariable conditional logistic regression models showed that the OR for GBS associated with SARS-CoV-2 infection and COVID-19 vaccine administration was 6.30 (95% CI 2.55-15.56) and 0.41 (95% CI 0.17-0.96), respectively."

It's not clear to me whether that means the OR for vaccination is controlled for SARS-CoV-2 or not. It should be clearer in the full paper, but I won't have access to it until it is officially published in an issue of the journal (for now, it's just on the journal's website). If not, then vaccination would likely be protective for GBS by reducing COVID-19 incidence or severity.
 
Last edited:

ISF - Join now!

Every member here is approved by hand. No bots, no spam, just people who care about evidence and honest debate.

Membership is free!

Create your free account

Back
Top Bottom