Hacker News new | past | comments | ask | show | jobs | submit login
Should individuals who already had a SARS-CoV-2 infection receive 1 or 2 shots? [pdf] (medrxiv.org)
23 points by tosh 32 days ago | hide | past | favorite | 42 comments

While "zero" obviously isn't the answer long term, I do think it's a reasonable one in the short term: The re-infection rate is clearly less than 1% given we only have a few anecdotal edge-cases there, so "being infected" clearly establishes solid protection. The first peak of the pandemic was 9 months ago, and we're still not seeing re-infections rise, so there's no reason to think this infection-based immunity is short-lived.

Obviously there might be long-term benefits to still vaccinating such people, but right now we're dealing with a very limited resource. Giving the vaccine to someone who hasn't been infected is a huge gain (say, 0% to 60% protection), while the second shot is a smaller gain (say, 60% to 90%), and someone who already has natural immunity is probably even smaller yet (say, 90% to 99%)

(Numbers are off-the-top-of-my-head and meant to indicate magnitude not an exact value)

What is your source for reinfection rate less than 1%? It seems much higher based on the recent second wave in Manaus [0].

[0] https://www.cidrap.umn.edu/news-perspective/2021/01/variant-...

Here's a source for the reinfection rate. I've seen a few studies that all came to the < 1% conclusion:


The Manaus data is interesting. The basis for assuming there's reinfection happening seems to be a seroprevalence survey done over the summer which estimated 76% of the population had antibodies to the virus. The accuracy of that survey has been called into question for a variety of reasons. The survey was done through blood donors and apparently they incentivized blood donation by offering people the results of their antibodies test as part of donation. Since testing capacity during peak infection was limited, this would give people who thought they might have had the virus (and likely did) incentive to donate blood and made the sample not nearly as random as it was presented to be.

> The basis for assuming there's reinfection happening seems to be a seroprevalence survey done over the summer which estimated 76% of the population had antibodies to the virus. The accuracy of that survey has been called into question for a variety of reasons.

This is exactly correct. Extrapolating from the Manaus seroprevalence paper is scientifically tenuous, at best. They say so in the Lancet commentary cited in the CIDRAP piece cited by OP:

"In the Lancet report, experts said four factors may be at play, some possibly related. First, scientists might have overestimated the attack rate for the first surge, and infections might have been below the herd immunity threshold"

> The survey was done through blood donors and apparently they incentivized blood donation by offering people the results of their antibodies test as part of donation. Since testing capacity during peak infection was limited, this would give people who thought they might have had the virus (and likely did) incentive to donate blood and made the sample not nearly as random as it was presented to be.

Not just that, but if you read the paper carefully, you'll see that they've applied a number of fairly arbitrary "corrections" to the data, which pull the seroprevalence from an observed 20% range up to the cited "76%" number (Figure 2A in this link):


While any of these adjustments might be individually merited, when you see a >300% aggregate "adjustment" to the raw data, it should make you very skeptical about any claims made with that data. Particularly when the raw data is below the likely herd-immunity threshold, and has been "corrected" to be above it.

More simply: the most likely conclusion is that the original paper was wrong.

This is perfect example of bad science being propagated by journal title and headline. Whomever wrote that (IMHO, terrible, editorialized) CIDRAP piece didn't bother to read the original data, and just took the first paper's claim at face value. Journalists have been (and continue to be) insufficiently skeptical when it comes to Covid science -- except when it suits the narrative they're trying to create. If a paper supports their desired narrative, any half-baked, highly implausible claim is treated as worth serious consideration. If a paper conflicts with their claim, no counterargument is too implausible to be considered.

Isn’t that an antibody-escaping variant? Existing vaccines aren’t very efficacious for that variant either.

My comment is being downvoted by an emotional crowd, it seems, but read this: https://www.biorxiv.org/content/10.1101/2021.01.15.426911v1

"Our experiments indicate that these variants, and potentially others that carry K417N/T, E484K and N501Y mutations, can reduce the neutralization potency of vaccinee plasma."

Preprint, not peer-reviewed, but still concerning.

TL;DR: It doesn't currently seem like a problem, but it is definitely a topic worth keeping an eye on.

"The vaccine generated a weaker immune response against the South African strain, but the antibodies remained above levels that are expected to be protective against the virus, the company said, adding the findings may suggest “a potential risk of earlier waning of immunity to the new B.1.351 strains.”

“Out of an abundance of caution and leveraging the flexibility of our mRNA platform, we are advancing an emerging variant booster candidate against the variant first identified in the Republic of South Africa into the clinic to determine if it will be more effective to boost titers against this and potentially future variants,” Moderna CEO Stephane Bancel said in a statement."

Source: https://www.cnbc.com/2021/01/25/covid-vaccine-moderna-workin...

Source for my source, because he posts a pretty solid round-up of recent research and well-cited sources every Friday: https://thezvi.wordpress.com/2021/01/28/covid-1-28-muddling-...

Your body throws the kitchen sink at SARS-CoV-2, it may generate spike-based antibodies, but it also might generate core-shell-based antibodies, which might mutate more (and therefore bypass acquired immunity in the future).

That’s one reason why spike-based mRNA vaccines are so exciting...

It is certainly of scientific interest to know these things. But I suspect the logistics of finding out who already have been infected, and only giving them one/zero doses, is not going to be worth it. And even then, only a small portion of the population have a confirmed infection.

In the coming months a very large amount of vaccines is going to arrive in most (western/rich) countries. It is then no longer going to be important to optimize exactly who gets the vaccine, but instead to increase the vaccination rate.

Yes. "Using quantitative serological assays that measure antibodies to the spike protein could be used to screen individuals prior to vaccination if the infection history is unknown." So it's possible, with enough testing, to see if someone really needs a second shot of vaccine. Then, of course, follow up after a month, and probably again later, to see how it worked.

That is the logistics I'm talking about. Testing was incredibly slow to scale up last spring, and scaling up tests for antibodies as well seems like unnecessary complication. We need fewer moving parts, not more.

Titers are measured before, after first jab and before second, so logistically this is pretty easy. The methods are described in the paper. It’s basically a couple of blood draws and quite simple to get a cohort based on questionnaire given that there’s 20M confirmed cases in the US.

This seems like an efficient study that could increase vaccination rate since people are more likely to get one dose rather than two. For many people, traveling to a vaccination appointment is a big deal.

This also serves to help convince people who have had COVID the value in getting a vaccine.

the gist

> In this short report, we show that the antibody response to the first vaccine dose in individuals with pre-existing immunity is equal to or even exceeds the titers found in naïve individuals after the second dose. We also show that the reactogenicity is significantly higher in individuals who have been infected with SARS-CoV-2 in the past. Changing the policy to give these individuals only one dose of vaccine would not negatively impact on their antibody titers, spare them from unnecessary pain and free up many urgently needed vaccine doses.


There is nothing obvious about anything like that.

Why not?

Vaccines don't actually protect you against the virus, they just activate your immune system (your body's natural protection). Guess what else also activates the immune system? The actual virus!

At the beginning of the pandemic, there was some talk that anti-bodies might not last longer than 3 months, but that was obviously just FUD (there was literally no data). I had antibodies 6 months after being infected. Two additional things I heard (but don't have the source), your body can restart production of antibodies (B-cells remember), and people had SARS(-CoV-1) antibodies years after infection.

Finally, this: https://www.ft.com/content/929ef3cd-8611-49b2-9f23-918dc3470... Covid infection shown to provide as much immunity as vaccines - I haven't checked the source but FT isn't really known for fake news.

Are there any credible studies that say that infection doesn't confer immunity?

> Why not?

Something can feel instinctively correct and not be correct. For example, the sun going around the earth, or the earth being flat, for that matter.

I'm not saying it should be 0, 1, 2 or 10 shots; I'm saying the answer is absolutely not obvious, even if you instinctively want to say zero.

In a world where we're lacking information, the default should probably be zero. But given that this is a worldwide pandemic, we're not going to be running out of information any time soon, so it's good it's being studied.

It's not FUD. It happens. It might be a small percentage but it's real.

There's a case of someone close to me that works in the ER that was positive around November with very little symptoms, got cured and now has it again but with stronger symptoms.

And from the antibodies tests done to the health personal in that hospital, it seems some people get good antibody response and others not so much.

Trying to figure out why a perfectly logical train of thought like this is getting downvoted

Because nature (and by extension reality) has a bad habit of not taking our brilliant logic into account.

> Are there any credible studies that say that infection doesn't confer immunity?

Of course! Economic studies said that vaccine manufacturers might not make as much of their investment back if infection confers immunity, so we can conclude that infection doesn't confer immunity.

Not obvious, and probably incorrect.

The immunity induced by the vaccines is said to be potentially more effective and/or longer-lasting than the immunity obtained from a prior infection.

There is also an open question about whether people who are sufficiently immune from prior infection that they won't get serious symptoms are still able to pass the virus to others during a "fighting it off again successfully" phase.

Immunity is not a simple yes-no boolean. It's not even a single variable, because there are different parts of the immune system memory.

Long-term, I agree, vaccines are generally (depends on the vaccine) more effective, but when we're dealing with a vaccine shortage, deprioritizing people who already have antibodies isn't necessarily crazy. I had doubts about prioritizing hospital staff who work with covid patients because they might have already been exposed (I haven't seen any data either way on this), but they also arguably "earned" it by being on the front line, so I'm not complaining too much, either.

> There is also an open question about whether people who are sufficiently immune from prior infection that they won't get serious symptoms are still able to pass the virus to others during a "fighting it off again successfully" phase.

Isn't this just as much an open question with vaccines? AFAIK the main vaccine trials were only checking for symptoms (so you could still be asymptomatic & transmitting the virus).

FWIW, US Rep. Stephen Lynch (D-MA) tested positive on Friday for COVID after being exposed to an infected staffer. He had received the 2nd vaccine shot prior to the inauguration:



It's worth basically nothing. We already know it's 95% effective; someone has to be the 5%, and there's a publication bias towards reporting on those.

The immune response from a vaccine can differ considerably from an actual infection.

At this point isn't the immune response from an infection pretty well known? I'd say more well known than the response from a vaccine.

I fail to any disadvantage of "0 shots" other than people not getting the vaccine because they think they had covid but didn't.

But then again the entire public policy around covid has been to treat the entire population as morons that wouldn't make the decisions thought to be correct on their own.

If the previous infection had only mild symptoms then probably any vaccine would indeed be a waste. After all, the vaccine manufacturers only claim to lower the severity of the disease but not the infectiousness.

Vaccine manufacturers only claim this, because that is the only claim that they have the data to make, beyond a reasonable doubt.

It's entirely possible that the vaccine also lowers the infectiousness. We don't know either way, though, because the data from the studies isn't in yet.


The number of anti-science misinformation red flags in this are remarkable.

Yes, there is strong evidence that vaccine-derived immunity is quite a bit more protective (as evidenced by antibody titers) than disease-acquired immunity.

The PCR cycle number misinformation is the current tinfoil being used to try to claim that there is no "real" epidemic, just false positive testing for some nefarious purpose. It's nonsense. The case number (ie, positive tests) strongly correlate to hospital admissions, which are strongly correlated with deaths, which are not f*cking false positives.

So just stop this nonsense please.

Unfortunately, putting your fingers in your ears and saying “it’s just misinformation, it’s just misinformation” over and over doesn’t make this RT cycle issue go away. This is coming from Fauci himself:

“…If you get a cycle threshold of 35 or more…the chances of it being replication-competent are miniscule…you almost never can culture virus from a 37 threshold cycle…even 36…it’s just dead nucleoids, period.”

Anything 35 and over gives results that can’t be trusted. False negatives but way more important are the false positives. The FDA has recommended the PCR cycle threshold up to 40. I would be happy to provide citations if you’re too lazy to do it yourself.

Sorry, inconvenient truths like this will continue. No, claiming something is nonsense and attempting to link it to actual tinfoil conspiracy theories will not work. You and your ilk will have to come up with some new talking points to fool people who can’t do basic research or read a medical study.

Have yourself a wonderful day.

Tell us, learned friend, what is the average CT recorded on positive tests today? And what is the false positive rate of those tests? I'll give you a hint - it's less than 30, and about 0.005%. So why, pray tell, is CT such an important attack point against testing?

Bringing up CT is just a smokescreen to try to discredit the idea that there is a pandemic. So please stop.

I’ll respectfully ask for more than a hint instead of blindly taking your word as fact. What I need is a source that backs up the statistics you just dropped. Unless of course you pulled both directly out of your rear end.

Go ahead, I’ll wait.

I know you didn’t just use an image from a Twitter post as a source. This is coming from the guy who starts off his first rebuttal calling me “anti science”. Hopefully you can see the irony.

No, I used an image from a Medrxiv preprint from a world-famous immunology institution. The attribution is listed in the image. There's no easy way to embed images in HN, and that one was available.

Go ahead, keep deflecting. It's obvious you don't have a leg to stand on.

I have verified that indeed case counts by PCR correlate with hospitalizations and death. It’s a good indicator.

I am, however, puzzled by the refusal of testing facilities to actually disclose what cycle count they are using. It could be too high, it could be too low, it’s likely “just right”. But “trust me I am doing it right” is the wrong answer to someone who hasn’t yet made up their mind.

Because science is complex, there is no universal answer to "what is the right CT" for PCT. Each assay/test configuration and mechanism has its own characteristics and constraints. The right number can vary, even from lab to lab using "the same test", due to variance in sample technique, storage, source of materials, etc.

And besides, the average CT for positive tests has been below 30 for the vast majority of tests, with something like a 0.005% false positive rate [1]. The people who are barking about CT are misinformation specialists, not infectious disease specialists

[1] https://twitter.com/LaymansScience/status/134538112170665164...

So to the anti-vaxers who want to score points over CT values, just ask them what CT threshold they use in their own lab, and how they justify it vs 1 higher, or 2 lower, or double, or half.

They don't have answers because it is complex and they don't know what the specific lab(s) are doing.

With all due respect, this is how religion is done, not how science is done.

I am not an antivaxer, I am up to date on all my vaccines. I was reading the Pfizer report, and I noticed that, w.r.t to severe disease, the placebo arm had 4 people, and the vaccine arm had 1 (please don’t trust me and check for yourself). That is 75% RRR with (any reasonable) confidence interval (0,100). I have mentioned it and got attacked for being an anti vaxxer, mostly by people who had not read the Pfizer report, would not understand it if they did, but were adamant that “it is 95%. Pfizer said so, and the FDA approved”. Statistics people I reached out too, however confirmed my reading of the results.

Now, you may very well be right, but your response is indistinguishable from that (provably wrong) attack I received, except perhaps to PCR experts. Similarly, almost everyone I’ve met, including many biologists and doctors, claim that two weeks after the 1st vaccine does protection is 50% and only gets to 95% after the 2nd one, which is very wrong (but only people who looked at Pfizer’s data are aware of that). Do you actually know what you said for sure, or just repeating what you’ve been told?

Furthermore, I am at this second stuck in Israel, a country with 35% vaccinated population, 6 weeks of lockdown (of which 3 have been relatively tight). I had believed the infectious disease experts in the beginning, but every single falsifiable prediction they made since March 2020 has been falsified. And in the times I looked at their data and methodology, it was abysmal 3rd grade statistics and politically motivated.

So, my trust for “leave it to the experts” right now is at absolute zero.

(And... the only data I was able to find here in Israel is unofficial, but supposedly they use 35 cycles - which every PCR using biologist I’ve talked to told me was bonkers, although some said “well, perhaps they have a good reason to”. This is not science. It’s religion)


Page 25.

Actual infectious disease scientists. There are two conclusions you can draw from these graphs:

1) The vast majority of 'positive' PCR tests do not need > 30 cycles to demonstrate presence of virus.

2) Actual, employed, doing-the-work scientists do indeed use > 30 cycles, and publish the work, to demonstrate facts with actual scientific merit.

Of course, your buddies who tell you that "it's bonkers" are a reliable source, right?

My buddies do PCRs every day, some of them for decades. Non of them work on human tissues, but they all have done PhDs and postdocs in prestigious labs, and some of them are already tenured and considered rising stars. So, yes, they are a reliable source.

Thank you for the medrxiv pointer, I will read it later today when I have the time.

The conclusions mins you, which one should never trust without actually reading the paper (and I admit I have not, especially not pg 25 you pointed to) are: “ Conclusions Community SARS-CoV-2 infections show marked variation in viral load. Ct values could be a useful epidemiological early-warning indicator.”

At face value, this reads “Ct value should be published”. But I will only adopt this interpretation (or disown it) after reading the entire paper.

Yes, if you only look at pg 25, you’d get these results. But if you actually read the paper (which actually supports the conclusions they reaxh). Median Ct>34.9 group seems to be essentially false positives, and even Ct>33 is not strong, whereas median Ct=~24 is a very strong positive.

Turns out, my buddies seem to be right and you sit, by your own reference, are wrong in claiming that there’s no reason to publish Ct.

You think you are part of the scientific solution, but you have just proved yourself to be part of the cargo cult religion.

Applications are open for YC Summer 2021

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact