1. Last 7 days
    1. On 2021-04-25 13:30:44, user Robert Saunders wrote:

      Clery and colleagues state that “evidenced based treatments are available” for chronic fatigue syndrome. These are listed as Cognitive Behavioural Therapy-for-fatigue (CBT-f), Activity Management (AM) and Graded Exercise Therapy (GET).

      In 2017 the US Centers for Disease Control and Prevention concluded that there are no effective treatments for CFS, after it re-examined the scientific evidence and removed CBT and GET as recommended treatments [1].

      Similarly, the 2020 draft NICE guideline for ME/CFS specifically warns against the prescription of CBT and GET as treatments due to the evidence that they are ineffective and potentially harmful [2]. 89% of outcomes in studies of non-pharmacological interventions for ME/CFS have been graded as “very low quality” with a high or very high risk of bias by NICE’s independent experts. And no outcomes in any studies of CBT or GET are graded as better than “low quality” [3].

      Clery and colleagues cite Nijhof et al (FITNET) [4] for their claim that “at least 15% of children with CFS/ME [sic] remain symptomatic after one year of treatment”. It should be noted that Nijhof et al used the 1994 CDC Fukada diagnostic criteria [5], which is less specific than other criteria as it does not require post-exertion malaise (PEM) as a symptom.

      Evidence suggests that most people with fatigue and other persistent symptoms following viral infection will recover within 2 years with no treatment, but a minority with ME/CFS will not recover [6,7]. There is no reliable evidence to suggest that long term outcomes are any better for those who have been prescribed CBT or GET and there is good evidence to suggest that these interventions are harmful [8].

      There is undoubtedly a need for children and adults with post-viral fatigue syndromes and ME/CFS to be given appropriate advice and support to manage and cope with the effects of their illnesses. However, acknowledgement of the very low quality of past studies and the evidence that CBT and GET are neither safe nor effective treatments for ME/CFS should be considered a prerequisite for any research pertaining to the provision of such services.

      References:

      1. https://meassociation.org.u...

      2. https://www.nice.org.uk/gui...

      3. https://www.nice.org.uk/gui...

      4. https://www.thelancet.com/j...

      5. https://pubmed.ncbi.nlm.nih...

      6. https://pubmed.ncbi.nlm.nih...

      7. https://pubmed.ncbi.nlm.nih...

      8. https://www.bmj.com/content...

    1. On 2021-05-04 02:06:41, user Uri Kartoun wrote:

      Ref 9 actually does rely on combining structured and unstructured data elements. The paper is one of the earliest to identify NAFLD patients using EMRs - indeed it is limited, but I wouldn't write "fail to provide the full clinical picture of NAFLD" because it is not true.

    1. On 2021-05-05 01:00:46, user Andre Boca Ribas Freitas wrote:

      Unfortunately, the drop of proportion of elderly people among total of deaths is due in large part to the increase in deaths among young people!<br /> This is due to the characteristics of variant P.1, which leads to more serious cases among young people.

    1. On 2021-05-06 12:22:51, user Steeve Asselin wrote:

      The old adage I feel applies here: It is not because we can do it that we should do it...Has thoughts ever been given to the potential of such innovative process to be misused by Life Insurance Companies to increase or worse, deny life insurance to a person because that innovation "estimated" (because it is an estimation NOT a calculation) that the probability of this person to die is above 50% in the coming years...

    1. On 2021-05-10 02:20:37, user Jogen ( G12 Student) wrote:

      Good day, may I request for the questionnaire because we're currently conducting the same study and it would be a big help for us, thankyou in advance.

    2. On 2021-05-12 06:41:10, user Mary b wrote:

      may I request for the questionnaire? We are currently conducting our data gathering for the barriers in online learning that is related to your studies. Thank you in advance

    1. On 2021-09-16 10:11:35, user kdrl nakle wrote:

      The paper claims significant increase of virulence yet many epidemiologists in the US would claim that is not the case with any of the variants. Which is true?

    1. On 2021-08-21 14:43:23, user Paul-Olivier Dehaye wrote:

      This paper has now been published (and presumably peer-reviewed).

      New title: Adherence and Association of Digital Proximity Tracing App Notifications With Earlier Time to Quarantine: Results From the Zurich SARS-CoV-2 Cohort Study

      https://www.ssph-journal.or...

      It still includes in the introduction, despite showing the opposite:

      These findings provide evidence that DPT may reach exposed contacts faster than MCT, with earlier quarantine and potential interruption of SARS-CoV-2 transmission chains.

    1. On 2021-08-23 10:32:30, user ingokeck wrote:

      Dear authors, I have a problem following your comparison between vaccinated cases and non-vaccinated cases. I understand how you select the vaccinated cases with your flowchart (thank you for providing one, this really helps to understand!), but I don't understand how you create the non-vaccinated controls. If you simply add up all non-vaccinated cases, you will get a huge bias towards non-vaccinated cases, as the vaccination campaign was still ongoing during your analysis period. So you will need to account for the differences in exposure, i.e. for a vaccinated case in week 10 which got infected in week 8 you need to look at the dose2 percentage even 2 weeks earlier, i.e. week 6 to normalize the exposure risk between vaccinated and non-vaccinated persons. It would be interesting to see the result. If you can share the anonymized raw data, I may be able to help.

    1. On 2021-08-25 12:07:48, user Prof. W Meier-Augenstein, FRSC wrote:

      What other than the difference in antibody titer post-vaccination and post-infection is the take-home message of this study? Surely, the decline in antibody titer per se months after vaccination or primary infection is not a surprising finding but could be expected? Antibodies have a finite life-span given by their Ig specific half-life (for example 21 days for IgGs). In the absence of a subsequent challenge (e.g. by a secondary infection) antibodies formed in response to the challenge posed by vaccination or primary infection will have all but cleared from serum after 6+ months. Furthermore, the difference in antibody titer between mRNA vaccinated and SARS-CoV-2 infected could not have come as a big surprise either considering mRNA vaccination results in expression of spike-protein “only” which means in contrast to a viral infection host cells are actually not infected and do not reproduce copious amounts of the virus which will take longer to fight and clear from the body than the spike protein. For the same reason, macrophages (phagocytes) are unlikely to be involved in the mRNA vaccinated group to the same degree as they are in the group infected by the virus. The natural decline of IG antibodies produced in response to the mRNA vaccine does not offer an exclusive explanation for breakthrough infection. Instead, breakthrough infection occurring 146+ days post vaccination are most likely the result of a “perfect storm”, an unfortunate coincidence of the higher virulence of the Delta variant of <<7 days incubation time, the associated higher viral load produced, and the fact the production of neutralising antibodies by B-memory cells takes up to 4-5 days to reach its peak.

    1. On 2021-08-26 03:53:13, user IMO wrote:

      Interesting study. Funny how little discussion there is from the media or the public health machine about immunity conferred by infection. No mention from on high about extending "passport" privileges to those who have been infected and then chosen not to get vaccinated. What's up with that?

    2. On 2021-08-28 11:49:17, user Doug Truitt wrote:

      So if one lives in Kentucky previous infection is less protective than vaccination - https://www.cdc.gov/mmwr/vo... - and if one lives in Israel previous infection is more protective than vaccination (this study). I'd be interested in discussion as to why these two studies are at odds with one another.

    3. On 2021-08-29 05:35:10, user some_guy wrote:

      Is there any variable control?

      The first question that pops into my mind is: are we comparing the same populations?

      People who were first vaccinated are disproportionately old people with weaker immune systems, more likely to be infected anyways. I wonder how much does this impact the results.

    4. On 2021-08-30 07:37:30, user 4qmmt wrote:

      This paper has one major flaw that is not discussed: We know for certain who was vaccinated and of those, 199 had symptoms. Though 8 of the matched recovered had symptoms, they were assumed to have been infected because of a + PCR test in the past, which as we all know produces tons of false positives (Israel's Ct is 35-40). In fact, the paper's own data shows that of the 238 PCR+ in the vaccinated cohort, only199 were actually positive, i.e., symptomatic = ill, and in the recovered cohort, of 19 PCR+, only 8 were symptomatic, suggesting at least 60% false positives. Thus, while the number of vaccinated is certain, the true number of previously infected is at most 8, but could in fact be 0, as the Cleveland Clinic study found.

      In fact, last weeks' MoH data in Israel shows that 73% of PCR + tests are on people with no symptoms, i.e., not infected. So this paper is actually saying that the risk of infection for vaccinated vs. recovered is between 27X to Infinite X.

    5. On 2021-09-01 03:04:45, user bgoo2 wrote:

      To all the people commenting on this article who have no idea how to critically evaluate study methodology.... let's just skip over the lack of control variables and omission of addressing the Santa's naughty list of biases.

      Let's just get right to the bones of it... while you are searching for your "natural immunity" as opposed to vaccination (how did that go with Polio, by the way? And Tetanus? And Rubella? And Hepatitis? ... hint: those vaccines came out before the Twitter era) ... where the FAHQK do you intend to keep the millions of symptomatic infected people who require hospitalization while they recover?

      How did that work out in 2020? How is that working out now? Every whack job shouting about natural immunity doesn't seem to acknowledge that to get there... you have to sacrifice a whole lot of people who would've otherwise been protected by a vaccine.

      Alabama USA has less than 5 million people. 38% of the population is double vaccinated as of August 31. Nearly 2500 people are in hospital with mild to severe Covid symptoms.

      Ontario Canada has 15 million people. Triple the population. 67% double vaccinated. Less than 350 hospitalized.

      We can repeat these comparisons across the globe. The results of this years' vaccination program is completely undeniable.

      The results in an under-vaccinated population is undeniable.

      Thankfully, each day, the percentage of unvaccinated grows smaller and smaller.

      And it is very quickly isolating who the remaining lunatics are in your community. (hint: they're ironically usually the ones who clearly have trouble saying NO to refined carbohydrates)

      Literally, that's the end of the story.

    6. On 2021-09-04 14:10:39, user criticalscientist wrote:

      It is likely that emergence of the Delta variant, which has 10 mutations in the spike protein, is due to leakiness of the mRNA vaccines based on the single spike antigen. I more worry about the Mu variant that appears to be completely resistant to monoclonal antibodies. Time to develop new vaccines using multiple viral proteins or develop an "attenuated" virus mimicking natural immunity.

    7. On 2021-09-09 04:36:47, user Salvatore Chirumbolo wrote:

      Dear Colleagues, my congratulations for this valuable and excellent paper. With a Colleague of mine, Sergio Pandolfi (Rome) we too addressed the topic of natural immunized subjects against SARS-CoV2 (see Pandolfi S, Chirumbolo S. On reaching herd immunity during COVID-19 <br /> pandemic and further issues. J Med Virol. 2021 Sep 7. doi: <br /> 10.1002/jmv.27322. Epub ahead of print.) but we highlighted also the controversial issue of the immunization route, i.e. mucosal versus intramuscular, I mean sIgA-B cell mediated respect to a DC-mediated immunity towards the IgGs production, with obviously different B-cell memory. Do you think that this is one of the major differences between immunized vs vaccinated people? And if serum anti-RBD IgGs are the only clue for evaluating immunized people, why a "certain" discriminating attitude exists towards natural immunized subjects respect to vaccinated ones?

      Many thanks in advance<br /> Regards<br /> SC

    8. On 2021-09-10 07:18:00, user Jim Ayers wrote:

      I hate to ask a dumb question but did the study include people with long haul covid and people who died? Quoting the study '46,035 persons in each of the groups.' The fraction of a percent who died or the few percent who have long covid who may feel too sick to participate in a study could invert the study if not accounted for since they have been weeded out of the cohort and need to be adjusted for. This study could be 100 percent wrong if the up to 20% who have long haul covid aren't participating.

    1. On 2021-08-26 10:17:10, user Ollie wrote:

      This might be an interesting approach. However, something is worrying me:<br /> 1/ The first equation in this paper "r*dr/dt = ..." is not derived, just presented as a citation from a book consisting of 304 pages. A book that is not readily accessible lest one borrows or buys it. The reader thus cannot understand the validity of this equation. The number of open and close brackets is not equal, which implies that the citation is incorrect. Further, it is a pity that parameter e_s(T_a) in the equation is explained in a slightly sloppy manner by omitting the subscript a for T in the text.<br /> 2/ The second equation is stated by the authors, rather than derived from hypotheses. A derivation seems relevant here, as the intuition of the reader (at least mine) tells him or her that the relationship between evaporated volume and surface area reduction of spherical drops is only linear for evaporation that causes very little radius decrease, or in other words: only for evaporation (dV) where dR<<r, where="" dv="the" evaporated="" volume,="" r="the" initial="" radius="" of="" a="" droplet,="" and="" dr="" the="" change="" of="" r.="" if="" this="" intuition="" is="" correct,="" it="" should="" be="" evaluated="" why="" the="" indicator="" air="" drying="" capacity="" is="" indeed="" relevant,="" as="" it="" is="" likely="" that="" in="" a="" given="" timeframe="" for="" some="" drops="" who="" evaporate="" only="" slightly="" dr<<r="" indeed,="" but="" for="" other="" droplets="" dr="R" (complete="" evaporation).="">

    1. On 2021-08-26 18:46:13, user vicweast wrote:

      If a vaccinated person contracts covid-19 (breakthrough infection) and their symptoms do not include symptoms like coughing/sneezing... are they as contagious as a person who has coughing/sneezing symptoms? This is what I am not reading anywhere, and yet it seems to be exactly the point.

    2. On 2021-09-05 11:40:00, user OverSpun wrote:

      The term "infected" appears to translate in this article to the presence of virus (detectable SARS-Cov2 at Ct<25) rather than the presence of the disease (i.e clinical manisfestions of COVID-19). Chronic carriers of other pathogens such as Staphylococcus are sometimes referred to as "not infected".

      Beyond the cellular pathophysiology of bacteria vs virus, this is more than linguistic trivia because it challeges the assumption that SAR-Cov-2 asymptomatic carriers are in a transient subclinical state rather than chronic carriers. If such chronic carriers exist, are they more likely to have been vaccinated or obtained their partial immunity from a live virus infection, i.e. an acute case of COVID-19?

      The following quote from this article highlights the importance of determining if mRNA vaccines have the potential to create chronic carriers: <br /> "Notably, 68% of individuals infected despite vaccination tested positive with Ct <25, including at least 8 who were asymptomatic at the time of testing."

      Clinical management and public health policy require confirmation that all asymptomatic carriers are eventually clear of SARS-Cov-2 and any causative relationship between the vaccine and such carrier state is well undestood.

    3. On 2021-09-30 02:46:32, user Fiona Weir wrote:

      70% of people in Dane County are vaccinated [6-your source]. If representative, you would expect a similar profile in your study; however only 38% of your cohort is vaccinated. This raises crucial questions about selection/inclusion... Was self-reporting reliable? Were cases excluded when vaccination status was not self-reported? Were unvaccinated people much more likely to test? Were vaccinated people likely to test only if they were actually ill, and therefore likely to have a higher virus load in any case? These issues plus the low cohort size may make your findings unreliable.

    1. On 2021-08-29 16:06:28, user Paul Wolf wrote:

      I wonder if the infectiousness of the delta variant could be a blessing in disguise, if it dominates over other, potentially more dangerous variants.

    1. On 2021-08-29 19:50:51, user Yvonne wrote:

      Based on Pfizer (6 month) study, the vulnerable started getting vaccinated on a great scale in February (USA) if you add 6 months that would put that vaccinated population at August, therefore the question would be asked, once the 6 month time frame of those vaccinated within a specific period, be deemed “unvaccinated” come August? With increasing spike cases/hospitalization in August (USA) and the term “unvaccinated” being used, who are within the description of “unvaccinated”, those never getting a vaccine? Or would the term include those who were vaccinated and now have passed the 6 months? I think August would be more of a complete study, if the term “unvaccinated” group is clearly defined. That would require maintaining that data, tracking the expiration of the 6 months when those vaccinated are spiking in cases. While this is helpful, the public should be shown how the spikes are increasing in August for full transparency and even comparison.

      The next spike of population vaccinated in April 2021 (USA) will hit the 6 month cycle in October. Therefore December will show if the spikes in November are from that vaccinated group.

    1. On 2021-09-01 16:46:01, user WGardner wrote:

      I would like to know more about how you controlled for fidelity of mask usage? This should consider whether or not people consistently wore masks, and wore them correctly. Having a mask mandate is a poor marker for whether or not masks work as it does not equate to people actually wearing masks.

    2. On 2021-07-06 02:38:59, user Nonyo Business wrote:

      This is consistent with the 2019 WHO mask guidelines, saying there is little evidence they work but that in an epidemic maybe give them a try.

    1. On 2021-09-01 20:00:22, user Peter Hanse wrote:

      This should be checked against "Experimental investigation of indoor aerosol dispersion and accumulation in the context of COVID-19: Effects of masks and ventilation" Physics of Fluids 33, 073315 (2021)

    1. On 2021-09-07 15:50:49, user Zach wrote:

      Im confused if it doesn't reduce death significantly statistically and increases serious adverse events by double. Why is this being pushed as effective or safe? This data proves both to be wrong. If your chance of death is unchanged and your hospitalization rate is nearly doubled it literally makes no sense to take this. What am I missing?

    2. On 2021-08-01 17:13:02, user Catriona wrote:

      No number needed to vaccinate for any outcome (death, severe COVID, hospital admission, ICU admission)? Of course, NNV to prevent death is infinite as no deaths were prevented. But what about the other outcomes?

      No number needed to harm?

      What were the severe adverse reactions?

      30 people in the placebo group were diagnosed with severe COVID and 1 in treatment group, so that’s a difference of 29.

      150 severe reactions in placebo group vs 262 in vaccine group is a difference of 112. I feel we need to know a lot more details on these adverse events to be able to make informed decisions. As it looks like the number needed to harm is greater than the number needed to prevent harm.

      I’m assuming that reducing PCR positivity if you have a mild illness isn’t what most people are really interesting. People are most worried about getting sick enough to be admitted to hospital, admitted to intensive care, dying or getting chronic fatigue syndrome and “long COVID” or some other autoimmune complication afterwards. Particularly if it affects their quality of life, employment, financial security, etc.

      This study wasn’t designed to detect how the vaccine affects contagiousness or herd immunity. So can’t assume that reduced mild COVID will improve those things as it could as easily cause atypical symptoms which didn’t warrant a PCR test and allowed participants to interact with others while infectious.

    3. On 2021-08-05 20:42:44, user Sven wrote:

      Where can I find Fig. S1A and Fig. S1B about adverse events? Can't find it in the supplementary appendix!

      As I just realized now, this was already asked by I. Bokonon<br /> It seems that these tables are indeed not provided. Please add the referenced tables about adverse events.

    1. On 2020-04-22 20:36:24, user Konstantin Momot wrote:

      The crucial issue here is sample selection. The participants essentially self-selected, but that is a potential source of huge bias. As a hypothetical scenario, if the people who chose to participate were predominantly people who’d had a cold and were curious to find out if it was COVID, then the cohort would a priori be hugely overweight with people who had a higher-than-average likelihood of COVID exposure. That would not be a good sample of the general population in that it would not represent the true percentage of CoV-exposed and CoV-naive people in the population as a whole. That would mean that the 2-4% figure is completely meaningless. Given how crucial this number is to any epidemiological modelling, I think it's important to remember that this is just one study with no guarantee of flawless methodology, and avoid making far-fetched conclusions based on limited evidence.

    2. On 2020-04-23 18:41:00, user Young-In Kim wrote:

      It’s not going to get peer reviewed well just like the studies on hydrochloroquine. That’s not a sample to apply to the whole population when we have data from all over the world. It’s picking and choosing data. I might well take my sample from worst hit parts of NYC n Italy to prove a point.

    3. On 2020-04-23 18:52:50, user Deplorable Kev wrote:

      So they tested for IgG antibodies in the population and found a high number of persons already had been exposed and likely immune. Interesting, but that depends on if their assay cross reacts with US Coronavirus and if so, then this is not correct. If their test is specific for COVID-19 antibodies then he is correct. PS: I believe it was here much earlier and we saw cases as early as mid November. Most of this early antibody testing has not been evaluated for cross reactivity yet.

    4. On 2020-05-16 23:36:43, user PANAGIOTIS AMPATZIS wrote:

      There is a hit piece from Buzz Feed news (I know lol) that says the Jet Blue founder (and anti lockdown proponent) gave 5000$ to the study to influence it. They make a huge deal out of it, so please release the full amount of the funding as an answer to show them that 5000 dollars is a small amount to make the study unreliable.

      Keep up the good work

    5. On 2020-04-20 20:43:45, user Kenneth Melendez wrote:

      So what these people are saying is that if you did the study LAST year, that you would get .5% false positives, about 10K people if you tested all of Santa Clara. You should declare them to be asymptomatic cases, and say that there are 9.5 times the number of asymptomatic cases as confirmed cases in Santa Clara. Time to call the media.

      And if you want to believe their results, then you surely want to believe the that the false positive rate (1-specificity) was accurately determined by 2 false positives out of 371 tests of old (assured non-infected) samples.

    6. On 2020-04-20 22:26:21, user Andy wrote:

      Preliminary results of USC-LA County are out. Unfortunately, that study also suffers from self-selection bias. We need to know how many people were contacted initially; should we count those who declined to participate as presumed negatives?

      In the Santa Clara study, people were contacted initially through Facebook. We need to know how many people saw the ad, and should we also count those who declined to participate as presumed negatives? (Surely the authors know how many people saw the Facebook ad.)

      Edit to add: Wait a minute, I just saw comment by Anon who participated with a link but without Facebook. So the study is even more flawed.

    7. On 2020-04-21 20:29:57, user rerutled wrote:

      The Santa Clara results calibrated the test using known gold-standard "negative for covid19" blood. They had 401 "pre-covid19" samples; of those, 2 resulted in "positives". Thus, the test produces a "false positive" rate of 2 out of 401.

      They gave the test to 3330 people, and 50 of those tests returned positive. How many of them can be produced by the observed "false positives"?

      On average, for a sample of size 3300, one expects approximately 3330*(2/401) = 16.6 false positive results. Is that statistically significantly different from the 50 detected? It depends on what the uncertainty in that "16.6" is. Playing fast and dirty, take the Gaussian uncertainty in that false positive rate to be fractionally, 1/sqrt(2) = 0.70. So the false positive rate is 16.6+/-11.7. Which means, the 50 detected is only (50-16.6)/11.7 = 2.9 sigma above the false positive rate.

      Nobody in science should claim a detection which is only 2.9 sigma different from the false positive number. This result is not significantly different from NO DETECTION of positive anti-bodies.

      It's even worse than that, though, because I used the Gaussian uncertainty; and as you know - since you have taken an undergraduate class in statistics and understand the binomial, poisson, and Gaussian distributions - the Gaussian approximation isn't accurate for this situation, and one should use the Poisson distribution, or better yet the binomial distribution; but, again, because of that undergraduate class, you remember that both the Poisson and binomial distributions have a broad wingy excess in the higher direction above the mode; which means, the significance of even the detection of ANY positive covid19 antibodies in the Santa Clara County results is even LESS significant than 2.9 sigma.

      The authors of this result should answer this criticism, stating what they believe to the the probability of detecting 50 positive results among 3330 tested, given their measured false positive rate.

    8. On 2020-04-22 16:02:39, user Texas Longhorns wrote:

      The research paper does not indicate how many of those that participated had already been tested for Covid and what those test results were.

      If they over sampled people that had already tested positive and recovered of course you will get a higher rate of positive antibodies. That would not be indicative of the general population.

      There is also the problem of false positives because the test can trigger for the common cold that is also a coronavirus.

      I don't think this research passes muster as any reliable indication of antibodies in the general population and should absolutely not be used as a basis to reopen businesses and large public gatherings.

      Having antibodies to one strain of the virus may not give you any immunity to the more than 8 strains of Covid we know are out there.

      Even if the test results are accurate at 2% that is nothing and you need at least 60% solid immunity to consider any large population to have herd immunity protection.

    9. On 2020-04-17 22:20:15, user Teddy Weverka wrote:

      To take this test, people had to violate the ORDER OF THE HEALTH OFFICER OF THE COUNTY OF SANTA CLARA DIRECTING ALL INDIVIDUALS LIVING IN THE COUNTY TO CONTINUE SHELTERING AT THEIR PLACE OF RESIDENCE EXCEPT FOR ESSENTIAL NEEDS. What do you bet that people willing to violate that order have a higher prevalence of covid-19 antibodies? The paper ought to mention this source of bias.

    10. On 2020-04-18 14:47:42, user Chris S wrote:

      Were participants asked if they had been tested? If so, could the rate of actual COVID-19 testing among participants relative to the general population (at the time of survey) be used to assess ascertainment bias?

    11. On 2020-04-19 03:40:52, user disqus_B1vk25qxNZ wrote:

      Possibly other coronaviruses that the population has been exposed to confer cross immunity to SARS-CoV-2. If so, the benign virus could be a vaccine analogous to cowpox and smallpox.

    1. On 2020-04-22 22:06:16, user Marie Benz wrote:

      Table 2 discrepancies which favored the non-treatment group, lack of randomization, lack of information on when treatment was begun as well as lack of number of doses completed make this paper unable to be interpreted since it is being heralded by news media as demonstrating that such treatment has now been proven ineffective. Clearly the jury is still out but the authors owe it to the country and the scientific community to point out in the media that this is not enough to conclude that HC or HC + AZ is ineffective or dangerous and that present therapies should not be altered one way or the other based on this report.

    2. On 2020-04-22 00:53:03, user Mike Cee wrote:

      This paper is flawed and should be withdrawn immediately.<br /> 1) This paper is flawed due to the limitation discussed on page 12 about the likelihood that the HCQ group, "However, hydroxychloroquine, with or without azithromycin, was more likely to be prescribed to patients with more severe disease, as assessed by baseline ventilatory status and metabolic and hematologic parameters."<br /> Doctors working the front lines have already noted that patients at SYMPTOMS ONSET + 14 days should not be prescribed HCQ<br /> 2) The important grouping by number of days since SYMPTOM ONSET was left out of this study. Previous studies, while anecdotal, suggest that patients should be prescribed HCQ early because it is believed to prevent the virus from infecting the type 2 lung cell through the ACE2 receptor and thus stops the progression of the disease.<br /> 3) This study did not document the dosage given to the patients. That would have been a helpful inclusion so we could understand if the patients who died were actually poisoned by excessive the treatment.<br /> 4) HCQ prescribed in patients in the first week after SYMPTOMS ONSET to include a zinc supplement which anecdotal evidence suggests a dual function of the combination: The HCQ provides an avenue for the zinc to enter the Type 2 lung cell where it interferes with the virus replication process.

    3. On 2020-04-22 03:32:32, user Cheri Trigg wrote:

      According to MedCram Zinc is what actually helps. Chloroquine is how the zinc gets into the cells. MedCram is a medical journal on Youtube. Update #32 and 34 go into depth on this. It is tragic to use a medication meant as a delivery system and not use it to deliver. Not sure to laugh or bawl. I hope the word gets out to use zinc before the zithromyacin at least.

    1. On 2020-04-23 08:03:58, user excivilservant wrote:

      I'm amazed nobody has commented on this, given the mass of media coverage about tracing and testing. Isn't a fairly obvious point that the model on social interaction is based on a pre-virus, no restrictions world. That's not the world we will be in for the coming months at least. Many/most/all hospitality businesses will be closed and these are the places where one would imagine a substantial number of social interactions occur, especially ones that widen the social circle beyond family and close friends. The survey data underlying the model will have the details. Even leaving aside formal restrictions, a lot of people will be acting in a much more conservative way about going out and about. Numbers travelling by train, tube and bus are likely to be greatly reduced. Again, the model could be adjusted, or a sensitivity test done. So the efficacy of tracing should be much greater than the article implies - or, put another way, the effort required to trace people should be a lot less, as there will be a (much) lower average number of contacts. No doubt the authors are actually on to this, but I thought I should point it out anyway.

    1. On 2020-04-24 01:46:41, user Mike wrote:

      Note: the original version has been previously discussed at length. medrxiv is redirecting that paper to v2 (this paper), making those comments are no longer available. This is a link to those comments: https://disqus.com/home/dis...

      Below is my original comment with updates for this version of the paper


      This was certainly an interesting paper. They’ve done a lot of work and the findings are notable. IMHO it warrants as much attention as the original pro-HCQ study via Dr. Raoult (~3/15). While it is entertaining, I will add that it is not conclusive, nor without fault. A double-blind study is still required.

      It is important to note that this is version 2 of a previously released paper and it is much the same, with no major differences in the conclusions reached compared to v1. Therefore, my previous comment still holds true. Below I’ve included them followed by a new list of observations.

      Observations/Questions (updated)

      1. "hydroxychloroquine, with or without azithromycin, was more likely to be prescribed to patients with more severe disease” (p.12)<br /> 2. “as expected, increased mortality was observed in patients treated with hydroxychloroquine, both with and without azithromycin” (p.12) — I assume it’s expected because the patients given drugs were in a more severe state (and more likely die regardless of treatment)<br /> 3. "we cannot rule out the possibility of selection bias or residual confounding” (p.13)<br /> 4. demographic: 100% male, 66% black, median age ~70 (59 youngest); (Table 2, p.17)<br /> 5. uses PSM, which despite a common practice, could be considered controversial (https://gking.harvard.edu/f... "https://gking.harvard.edu/files/gking/files/psnot.pdf)")<br /> 6. Unless I missed it, I didn't see any specifics about how the treatments were administered.<br /> - How long before death were patients treated?<br /> - What was the quantity/frequency/duration of the treatments?<br /> - Were the treatments consistent between hospitals?<br /> 7. The rate of ventilation was less in HC+AZ (half of the HC and no-HC rates). Why was that? What does that suggest?<br /> 8. Although they were statistically insignificant, what was the result of the 17 women not included in the study?<br /> 9. Why does the paper seem to address political points? It seems like the Abstract is editorialized, which I'm not used to. The result seems to address topical issues of the times, having awareness of other similar studies being conducted, rather than a standalone independent study of its own. I interpret this as potential for some analysis/deciphering bias. I don't mind in the Discussion sentence as it's normal, I'm just not as accustomed to seeing it in the Abstract.

      New List of Observations

      1. There seems to be some bias in the number of healthy people with no-HC treatment, but left in the study. Those people are going to be unlikely to die to begin with. This is not a comparison of apples to apples.??

      To clarify:<br /> ?- Dramatic difference in percentage of people people that had fever temperatures (38.1-39.0ºC / 100.58-102.2ºF); HC:11.3%, HC+AZ:11.5%, no-HC:7.6%. There’s ~4% difference between treated and untreated fever temps (more likely to die) in favor of untreated cohort. ?<br /> - Compare that with the percentage of people that had normal temperatures (35-37ºC/95-98.6ºF);HC:56.7%, HC+AZ:52.2%, no-HC:61.4%. There’s a 5-9% difference between treated and untreated normal temps (likely to not die) in favor of untreated cohort.

      ??So in this study, there was a larger proportion of people that did not have fevers, suggesting the data may be padded. In absolute numbers it's approx. a 40 person swing, which is a fairly large percentage in such a small study/survey. Similar observations are for systolic blood pressure and breaths per minute. There appears to be more healthy people? again.

      2. Creatinine is created when muscles breakdown creatine. It’s a waste product removed by the kidneys. Levels are elevated when the kidneys begin to fail. Notice, there’s a much larger presence in the HC-groups, which suggests there was a larger percentage of these patients experiencing kidney failure.?

      3. There are an awful lot of missing data in solely the no-HC group for statistically significant criteria. For instance Erythrocytes (red blood cells that transport oxygen), there 11.4% (!) missing in no-HC patients, yet that category has a P-value of 0.001 (<0.05 is statistically significant).??

      Hermatocrit is the same way (missing 11.4% for no-HC). It’s also related to red blood cells, it is the ratio of red blood cells to the blood volume. Same missing amount for Leukocytes (white blood cells) test. And also Lymphocytes —white blood cells in lymphatic system, which transports fatty acids from the digestive system and white blood cells from the lymph nodes into the bones— not only has a lot of missing data, but the disproportional low count (<800 per mm^3) may warrant further investigation.

      4. Even looking at the statistically significant Cerebrovascular disease, there are a much larger percentage of HC-only patients per its cohort.?

      5. Table 4 describes a greater percentage of people using HC+AZ being discharged (recovering) w/o ventilation; 5% more than no-HC patients. Keep in mind that 30% of the no-HC patients were given AZ.

    2. On 2020-04-28 00:30:59, user Mark Reeder wrote:

      I am advocating that the authors, in the interest of public health, fill in the blanks of the following statement:<br /> "It was found that ___ of the 7 patients reclassified from the 'No HC' to the HC group (after ventilation began) died. Likewise, ___ of the 12 patients reclassified form the 'No HC group' into the 'HC+AZ' group died."<br /> Based on a comparison of Tables 4 and 3, the first blank must be either 6 or 7 whereas the second blank must be between 7 and 12, inclusive.

      If the groups are compared based on whether they were given the drug(s) PRIOR to ventilation or prior to discharge, the HC+AZ is better by either a 20% margin over the 'No HC' group or by a FACTOR of 6. <br /> To wit, let's assume that all 12 of reclassified as HC+AZ died. That would mean that only 2 in the original HC+AZ group died. Since we have no idea when the HC+AZ drug was administered to those who died without ventilation, a fair comparison would show that HC+AZ, one might justifiably count only the 2/90 (2.2%) in that group (excluding deaths w/o vent.) as having had HC+AZ early enough in treatment. It would also mean that 3+(6 or 7) + 12 out of 162 (13.6%, also excluding deaths prior to ventilation) eventually died. <br /> This would be a huge difference with HC+AZ coming out as a terrific alternative (factor of 6 better) if given early enough. By the same logic (pre-vent treatment only, excluding non-vent deaths), the worst case for HC+AZ would still mean a 20% IMPROVEMENT over the control group! <br /> But we cannot know unless the authors (or others) are ethical and transparent enough complete the sentence above. Even if they disagree with the foregoing analysis, what is the downside in providing those numbers?<br /> I understand the difficulty of dealing with imperfect data. But for that very reason, good science demands that all information be placed on the table.

    1. On 2020-04-25 13:53:09, user Rosemary TATE wrote:

      Since this is an observational study, and you answered yes to "(I have) uploaded the relevant EQUATOR Network research reporting checklist(s), could you please explain why you have not uploaded the STROBE guidelines? So many people are doing the same and I'm wondering why. I'm trying to think of ways of improving the process of sorting the wheat from the chaff in this explosion of preprints. While this looks like wheat, it would be really helpful if you could<br /> 1. State the type of study in the title (some journals require this and it is very helpful)<br /> 2. Upload the checklist.<br /> Many thanks.

    1. On 2020-04-26 15:17:51, user peopletrees wrote:

      Can you post the r script that you wrote in order to process and analyze this data? I'm particularly interested in the rstanarm model specification.

    1. On 2020-04-27 03:28:31, user Sam Leumas wrote:

      How was the data even collected? Veracity of this information?How can you tell if the contracted virus is from within house, transport area,outdoors,malls etc? Common sense tells stagnation of air is a probability but still,seems like there is a major flaw in the data collection process.

    2. On 2020-04-10 16:34:41, user Brian wrote:

      this looks at "3 or more case outbreaks" if outdoor spread is happening it's probably on a smaller scale, as people are out less.

    1. On 2020-04-28 13:51:34, user alasdair hay wrote:

      I am not convinced this paper provides any evidence on the safety of HFNC in respiratory infections. My understanding is the study shows similar number of respiratory particles were measured in all conditions from breathing air to any 02 device or coughing.

      I would be more reassured if the articlce demonstrated an increased number of respiratory aerosol particles with some airway devices or proccedures but not HFNC. In fact there is no control demonstrating the SMPS aerosol measurement probe detects respiratory droplets. The only particles the probe appeared to detect were produced by a candle. If the probe is picking up a background number of particles in the air and those produced by a candle it is not picking up anything of medical interest

    1. On 2020-05-01 09:48:59, user Kasper Kepp wrote:

      This paper on the state-of-the-art Danish blood-donor data finds a IFR = 0.08% for people between 0-69 years of age. The study is very important because the sampling bias from case fatality ratios (the iceberg effect of knowing almost all deaths but only the most symptomatic cases, i.e. missing the dark number) is largely removed.

      By interpolation, the Danish population now has approximately 1.6% infection, corresponding to 100,000 people out of 6 million. The dark number stands at 12-fold the known cases (7-18).

      Some minor sampling biases remain (people who are blood donors need to be healthy and may be socioeconomically skewed) but considering the wide blood donor representativeness in Denmark, I think all Danish researchers will agree that sampling bias must be small.

      The IFR is also fully in line with the most representative data we have from Iceland (14% of population tested, 48000 tests), where the sampling bias is essentially eliminated, which stands at approximately 0.56% (10 deaths / 1799 cases as of May 1) and includes all the high-risk individuals >70 years. https://www.worldometers.in...

      Compared to the Santa Clara study, which caried potential major sampling bias, this issue seems to be now largely removed. Consensus in Denmark is now emerging that the overall whole-population crude mortality of covid-19 is of the order of 0.25-0.6%, in excellent agreement with the Iceland data.

      These two countries have not have their health care systems strained, making them the relevant data also for this reason for pinpointing the "real" mortality of covid-19 absent overmortality by capacity exhaustion as seen in some other countries.

      Obviously, the fact that the IFR is 0.08% for the 0-69 year old has enormous implications for political decision making in Scandinavia, as it evidences that most of the population can build immunity at much reduced mortality than previously assumed.

    1. On 2020-05-02 04:43:05, user Ilya Zakharevich wrote:

      Thanks! This is the first potentially meaningful text on such stats I have seen so far. However, there are some (possibly significant) apparent problems too… (Below, I only address points about false-positivity rates.)

      The presentation: for your Table 2 (and other related presentations), would not it be more clear if you add a row “Pre-Covid” before the row “1–5 day”?

      “Samples from UCSF and ZSFG were assigned a random well position in one of four 96-well plates.”

      I suspect this is not clear enough. The crucial question for estimating the seropositive population is the “degree of double-blindness” of mixing the 108 pre-CoVid samples among the other samples. Can you be more explicit about this? (Separately for ELISA and the rest, if possible.)

      “Binomial exact 95% confidence intervals were calculated for all estimates.”

      With 14 different schemes tested, this is a very questionable choice. Basically, for laymen in statistics (which, in my experience, probably covers >99.9% of potential readers), only the 99.74%-confidence intervals can be useful. (Here 5%/14?0.357%.)

      The people who understand the pitfalls of using 95%-confidence with 14 schemes could also be interested in 95%-confidence numbers — but these numbers would just create an unneeded confusion among the overwhelming majority. (As this XKCD shows.)

      “Four assays … maintaining >95% specificity.”

      Sigh… Do you understand the expected number of outliers with 14 groups? Especially when you, essentially, say “<=5 false-positive samples per group”?!

      “We based minimum sample size calculations on expected binomial exact 95% confidence limits.”

      I think 108 samples is ruefully small for any reliable conclusion. (As your numbers, and 99.64%-confidence intervals show.) I do not see any way than to start with >=1,000 samples (with one method), or >=3,000 (with 14 methods, as you use).

    1. On 2020-04-06 17:15:59, user Matheus Macedo-Lima wrote:

      It seems to me some of the countries used in this analysis are still under-testing (Brazil, USA etc). Would a more suitable dependent variable be deaths/confirmed COVID cases?

    1. On 2025-08-21 08:10:06, user Taukim Xu wrote:

      Comments from the authors-01:

      Since the preprint was made available, we have been honored to receive numerous publication invitations from various journals. We would like to express our sincere gratitude for your interest and recognition of our work.<br /> However, we feel it is important to clarify that this manuscript is currently under consideration at another journal. Therefore, we are unable to submit it to your publication at this time.

      Dr. Taukim Xu

    1. On 2025-10-18 04:34:17, user CDSL JHSPH wrote:

      Hello!

      Thank you for sharing this preprint. I found it very interesting, and I believe this is an excellent contribution to improving statistical approaches in antibiotic trial designs. I believe that the comparisons between MCP-Mod and Fractional Polynomials (FP) models provided a very strong case for adopting model-based frameworks to more accurately identify duration-response relationships and estimate the Minimum Effective Duration (MED). I also believe that the demonstration of these methods outperforming traditional pairwise comparisons, particularly in smaller sample sizes, helps highlight the potential to make trials more efficient without compromising reliability and ultimately the patient.

      I also appreciate how you clearly discussed the limitations of traditional methods and how there is a need to move towards more model-driven designs. The emphasis on the simulation as a proof-of-concept framework was well-structured and persuasive. I think it would be very interesting to see how these approaches would perform under real clinical data. This is where patient adherence, comorbidities, and variable response rates could truly test the robustness of these models.

      I think that in future works, expanding on various factors like dose-spacing, variability in treatment adherence, or non-monotonic response patterns that affect model performance could help further strengthen the clinical applicability of your framework. Overall, I found that this paper sets a solid foundation for refining antibiotic trial methodologies, and also bridges the gap between statistical modeling and real-world trial optimization.

    1. On 2025-11-17 18:55:33, user David Stapells wrote:

      Interesting.

      Small comment:<br /> Rather than head size explaining larger amplitudes in females, Don and colleagues suggest cochlear travelling wave differences:

      Don, M., Ponton, C.W., Eggermont, J.J., Masuda, A. (1994). Auditory brainstem response (ABR) peak amplitude variability reflects individual differences in cochlear response times. J Acoust Soc Am, 96, 3476–3491.<br /> Don, M., Ponton, C.W., Eggermont, J.J., Masuda, A. (1993). Gender differences in cochlear response time: An explanation for gender amplitude differences in the unmasked auditory brain-stem response. J. Acoust. Soc. Am., 94, 2135–2148.

    1. On 2025-11-30 14:34:14, user Jeff Lubell wrote:

      This article has now been published in a peer-reviewed journal. Please use that version. Thank you. Here is the citation: Torok RA, Lubell J, Rudy RM, Eccles J, Quadt L. Variant connective tissue as a risk factor for long COVID: a case-control study of data from a retrospective online survey of adults in the USA and UK. BMJ Public Health. 2025 Sep 17;3(2):e002949. doi: 10.1136/bmjph-2025-002949. PMID: 41001233; PMCID: PMC12458677.

    1. On 2020-04-20 18:14:27, user Wladimir J. Alonso wrote:

      Not being temperature-dependent does not imply that it might not be seasonal. For instance, in equatorial regions seasonality of influenza is driven by the rainy and dry seasons, not temperature (see references below).

      Tamerius J, Nelson M, Zhou S, Viboud C, Miller M, Alonso WJ. Global Influenza Seasonality: Reconciling Patterns Across Temperate and Tropical Regions. Environ Health Perspect. 2010

      Alonso WJ, Viboud C, Simonsen L, Hirano EW, Daufenbach LZ, Miller MA. Seasonality of influenza in Brazil: a traveling wave from the Amazon to the subtropics. Am J Epidemiol. 2007;165(12):1434–42.

      Alonso WJ, Yu C, Viboud C, Richard SA, Schuck-Paim C, Simonsen L, et al. A global map of hemispheric influenza vaccine recommendations based on local patterns of viral circulation. Sci Rep. 2015;5:17214.

    1. On 2020-04-01 18:46:10, user Miklos Kertesz wrote:

      Does this modeling shed any light on the approximate number of those who have been infected, and are now presumably immune?<br /> Your curves fall off quite fast. Is this realistic?

    2. On 2020-04-04 21:18:14, user Lucy Branum wrote:

      Does this data reflect the lack of PPE for healthcare providers, and other essential workers? I appreciate that is a difficult factor to quantify, but if your healthcare providers are becoming infected and therefore spreading the virus to others, and becoming ill themselves, I would have to assume that would have a massive impact on the trajectory of the curve.

    3. On 2020-04-05 15:41:10, user Art Mills wrote:

      This was supposed to be updated yesterday and had a notification it was to be, then after the daily press conference, suddenly was not updated and the notification it was to be updated was removed. Why?

      The ICU and Vent numbers in the model to support the 93K deaths by Aug. 4 are off by a factor of 10 (meaning they have modeled far too high). The update yesterday should have brought those lines down. Why would we not update to reflect a line closer to the actual numbers?

    4. On 2020-04-06 07:06:24, user Tom T Walker wrote:

      Many of the State projections radically changed when updated 4/5. For example Colorado's peak date went from 4/17 to 4/4, hence an ICU deficit of 762 to a surplus of 421. Other states had similar swings in the opposite direction. Wondering if there might have been some data entry issue. Is Colorado really suddenly 1-day past it's peak date?

    5. On 2020-04-07 07:38:54, user dreich wrote:

      Every place in the US is significantly under-testing.<br /> This means that the infection rate is underestimated.<br /> And because there aren't enough tests almost no post mortem tests on performed for any deaths.<br /> This has the "advantage" of under-counting the number of fatalities which makes the organized incompetence (intentional or otherwise) seem less bad. The bigger the number the worse it looks for the guilty politicians.

    6. On 2020-04-10 13:51:26, user steve rubin wrote:

      Does anyone know peak hospitalization during the 2017-2018 flu season? From the cdc summary there were 808,000 total hospitalizations with 61,000 deaths and I remember the flu season was pretty long. I wonder how the curves for new cases, deaths, hospitalizations and icu use looked. I remember stories that the hospitals were crowded but I don't remember stories about people dying because there weren't beds or ventilators.

      I know that it's unpopular to compare coronavirus to the flu. Underestimating a threat is dangerous and could (and maybe did) lead to delay in ramping up testing and beds and ventilators and other necessary medical resources.

      When people were predicting 2,000,000 deaths in the US and then 200,000 deaths I could understand the fears. But now they're predicting 60,000 deaths and it may end up half of that, so I think it's reasonable to make the comparisons.

      Comparing situations to past situations is usually our best way to understand how to react. In terms of how contagious and how lethal this epidemic/pandemic is, it now seems that it and the flu are similarly contagious and that covid is much less lethal. The big difference is that we have a pretty big number of people with significant immunity to the flu while it's likely there was little immunity in our population to covid-19. If we end up with 200,000,000 people becoming infected but with only 60,000 deaths, then covid was a fifth as lethal as the flu for 2017-2018 with 3 in 10,000 infections dying vs 14 in 10,000 infections dying from the flu.

      However the comparison to the flu can lead to some counter-arguments. For example, the cdc uses a multiplier of ~80 for estimated current flu infections vs confirmed flu infections. Applying that to covid-19 means that we have ~500,000 confirmed cases so we would have had 40,000,000 total infections leaving another 160,000,000 to go assuming 60% of the population for herd immunity. Projecting deaths would mean 17,000 + 68,000 for a total of 85,000. We'll soon know what that multiplier is for covid-19 because there are a number of antibody surveys going on in the US and internationally. You can bet that the same thing will happen for the flu next year and we'll have a more accurate estimate of infections and lethality for the flu rather than the current guesstimates.

      A big question is how social distancing will have affected the final number of infections and deaths. It seems so logical that social distancing will curb infections and deaths, but many suggest that it may end up only prolonging the length of the pandemic while not making a significant difference in total final infections and total final deaths. The antibody tests may give us the answer to that as well.

    1. On 2020-04-22 08:47:13, user Katri Jalava wrote:

      This is a really interesting study from the environmental epidemiology perspective. However, the authors fail to present some of the key findings they surely have. The clinical presentation of their cases is "not interesting", as they are anyway, typical Covid-19 cases, that could be shortened. You could also try to get some of your food control experts as authors, they might be able to give you good clues about the potential spread modes with employees and hygienic situation in the supermarket. Also if one of the epidemiologists from China-CDC would help you to clear the epidemiological messages, that would be useful. Your current conclusions are "not really interesting".

      If possible, it would be very important to know how many customers were screened, at which time points? Also the time line for the employees when they were at work and when did their symptoms begin, ie. how long they were working while infectious? What type of supermarket this was, how much selling, customer profile? Also the description of the supermarket employees job duties, were they at the cashier, providing meat or other over-the-counter food etc.? How many children were exposed either as case household contacts or in the supermarket?

      You should try to follow up the asymptomatics. How many of them went on to develop symptoms, even mild? If not, please state the follow up and initial examination date. Comparison of clinical symptoms between groups A and B is not really interesting (you could just state in your text that the groups did not differ, data not shown), but comparison of symptomatic vs. truly asymptomatic might be. However, there is plenty of studies on that and unless you detect something extraordinary, that is of interest. Were your cases really 48 year old, plus minus 2 years, ie. 46-50 year old. All of them? What were the isolation (cases) and quarantine (close contacts, yet not symptoms) measures, please describe and how this succeeded. Did you consider (or would you do a next study) to do serological survey among all cases, their close contacts and supermarket customers to see if they had SARS-CoV-2 antibodies?

      Thanks for submitting this, this is a really nice and interesting study. Looking forward for the final paper.

    1. On 2020-04-07 16:22:04, user Kevin Niall wrote:

      "All received standard treatment (oxygen therapy, antiviral agents, antibacterial agents, and immunoglobulin, with or without corticosteroids)"

      This could play a huge part in the difference considering teh very small sample size and the difference of 8 patients.

    1. On 2020-04-07 19:20:32, user Andrei Drabovich wrote:

      There was no experimental proof presented that these variants impact ACE2 structure or function in any way.<br /> It is also not possible that ~3% genotype frequency for women and 1.5% for men (1.5% allele frequency for p.Asn720Asp) will result in additional >6% lethality. Were would additional 3-4.5% come from?

    1. On 2020-04-08 00:15:34, user Sinai Immunol Review Project wrote:

      Clinical features and the maternal and neonatal outcomes of pregnant women with coronavirus disease 2019

      Keywords

      Pregnancy, SARS-CoV2, neonatal and maternal Covid-19 outcome

      Key findings

      33 pregnant woman and 28 newborns were included in this retrospective multi-center study, conducted at 5 hospitals in Wuhan and Hubei province, China, between January 1 and February 20, 2020. All women were diagnosed with Covid-19 by qPCR or viral gene sequencing based on the Chinese New Corona Pneumonia Prevention and Control Program, 6th edition, and were further subdivided into four groups based on clinical severity: (1) mild, presence of mild clinical symptoms without radiological abnormalities; (2) moderate, fever or upper respiratory symptoms as well as radiological signs of pneumonia; (3) severe, at least one of the following: shortness of breath/respiratory rate >30/min, resting oxygen saturation SaO2<93%, Horowitz index paO2/FiO2 < 300 mmHg (indicating moderate pulmonary damage); and (4) severe-acute, acute respiratory distress with need for mechanical ventilation; systemic shock; multi-organ failure and transfer to ICU. Maternal admission to ICU, mechanical ventilation or death were defined as primary outcomes; secondary study outcomes comprised clinical Covid-19 severity in both mothers and newborns, including development of ARDS, neonatal ICU admission as well as mortality.

      Maternal characteristics and outcome: 3 out of 33 women were in their second trimester of pregnancy (17, 20 and 26 weeks), and 15/33 (45.5%) had a previous history of underlying chronic health disorders including cardiovascular, cerebrovascular or nervous system disease. Common Covid-19 symptoms at presentation were fever (63.6%), dry cough (39.4%), fatigue (21.2%), and shortness of breath (21.2%). Less common symptoms included diarrhea, post-partum fever, muscle ache, sore throat and chest pain. 4 (12.1%) pregnant women had no apparent symptoms. The majority of cases were classified as mild (39.4%) or moderate (57.6%); however, one woman developed severe Covid-19. 40.6% of women were diagnosed with bilateral pneumonia, 43.8% presented with unilateral pneumonia, and 15.6% showed radiological ground-glass opacity. 87.9% of women required oxygen administration, and one (3%) woman had to be put on non-invasive mechanical ventilation (primary outcome). 81.5% of women had a C-section and only 5% had vaginal deliveries. Obstetrical complications were seen in 22.2% of women, including three cases of preterm rupture of membranes, two cases of hypertensive disorders of pregnancy, and one case of spontaneous preterm labor. Five pregnancies were ongoing at the end of the observation point of this study; one woman decided to have her pregnancy terminated. Neonatal outcome: Out of 28 newborns included in this study, 35.7% were born preterm at <37 weeks of gestation with Apgar scores ranging from 8-10/10 at 1 min and from 9-10/10 after 5 min, indicating normal heart and respiratory rates. 17.9% of newborns were of low birth weight (not specified) and 14.3% showed signs of fetal distress (also not specified). According to the authors of this study, none of the newborns presented with clinical Covid-19 symptoms. However, one newborn, delivered at 34 weeks of gestation, was diagnosed with (apparently Covid-19 unrelated?) ARDS and transferred to NICU (secondary outcome). Of 26 newborns tested for SARS-CoV2, only one was found positive and showed radiological signs of pneumonia, but no clinical symptoms of Covid-19. It remains unclear whether this was the same case as the newborn diagnosed with ARDS. The affected newborn did not require any treatment and was discharged at 16 days post birth. In summary, the primary outcome “mechanical ventilation” in pregnant women was rare (3%), no other primary outcomes were reached. Most Covid-19 cases in pregnant women were described as mild to moderate. Only one of 28 (3.57%) newborns was diagnosed with ARDS (secondary outcome).

      Potential limitations

      Major limitations of this study are its small size and the rudimentary and at times inadequate description of patient specifics. For example, underlying health conditions that might be affecting Covid-19 outcome in pregnant women should have been clearly specified (other than being of be listed (not just <37 weeks). Given that maternal infection status seemed mostly unknown at the time of birth and, more importantly, that the majority of cases in this study were clinically asymptomatic or mild to moderate, it remains unclear whether the C-sections performed were a medical necessity or elective procedures. This is of importance and should have been discussed. With regard to neonatal outcome, it is also not apparent whether the newborn found to be infected with SARS-CoV2 and the case diagnosed with ARDS were the same individual. If this was the case, it would be incorrect to refer to all newborns as asymptomatic. Additionally, it seems somewhat unlikely that a newborn with a near-perfect Apgar score would present with ARDS immediately after birth. Likewise, any individual diagnosed with ARDS would certainly be expected to receive supportive treatment including (invasive) mechanical ventilation. While it is highly relevant that overall clinical outcome in pregnant women diagnosed with Covid-19 seems better than in SARS or MERS (as discussed by the authors), it nevertheless needs to be stressed that more than 37% of newborns in this study were delivered preterm and that the obstetric complication rate of 22% seems higher than non-Covid-19 average.

      Overall relevance for the field

      Observations in this study confirm some of the findings published in a case series by Yu N et al. (Lancet Infect Dis 2020; https://doi.org/10.1016/ S1473-3099(20)30176-6). However, due to the relatively small study size of 33 pregnant women and 28 newborns, this study lacks statistical power and final conclusions on Covid-19 outcomes in pregnant women and newborns cannot be drawn. Yet, the data collected here are important and should be incorporated into larger data sets for more insight. Understanding the clinical course and effects of Covid19 in both pregnant women and newborns is essential, and while there are some recent publications on vertical SARS-CoV2 transmission between mothers and newborns (Dong L et al, JAMA March 26, 2020, doi:10.1001/jama.2020.4621; Zeng H et al, JAMA March 26, 2020, doi:10.1001/jama.2020.4861) as well as on neonatal infection at birth (Zeng L et al, JAMA March 26, 2020, doi:10.1001/jamapediatrics.2020.0878), our knowledge of how these patient subsets are affected is still very limited.

      This review was undertaken as part of a project by students, postdocs and faculty at the Immunology Institute of the Icahn school of medicine, Mount Sinai.

    1. On 2020-04-09 07:07:30, user Daniel Corcos wrote:

      This calculation is made by dividing the number of deaths by the number of days since the FIRST Covid death, and by comparing to the number of deaths per day by driving during the same period. This means that the death rate by day is diluted with the period of low mortality per day from Covid.

      But the highest mortality per day is yet to come. And, already, social distancing has been implemented and has some effect.<br /> This means that there is no evidence to support the hypothesis that there is overreaction to the pandemic.

    2. On 2020-04-17 14:53:53, user Petard Stamo wrote:

      "Absolute risk" is a very good measure to quantify the probability of dying from cancer and heart attack but it doesn't capture the dynamics of an infectious disease. It doesn't capture transmissivity. I am sure other people will find other features of an infectious disease that are missed with the "absolute risk" analysis. It has very complex dynamics. Ample Testing and exact classification of death as pointed out by dr. Ioannidis is the only way to APPROXIMATE "infection fatality rate". <br /> His analysis of the diamond princess cruise ship is very interesting IMO.

    1. On 2020-04-11 12:00:55, user Sherlock Holmes wrote:

      From an article in The Hill: "France has recorded 100 health incidents and four fatalities linked to experimental drugs for those with the coronavirus since late March..

      France has has some 13.500 deaths of coronavirus. Is someone doing the maths?

    2. On 2020-04-14 10:26:28, user Franko Ku wrote:

      New trials must include use of zinc sulfate in <br /> combination with HCQ as well as as in latest French study with over 1000 patients in early stages added to HCQ and azithromycin, using same regimen to prove the great reported results.

    1. On 2020-04-11 21:57:12, user H van Woerden wrote:

      I think that this is a really important paper as it starts to explore the cost effectiveness of different approaches.

      I am concerned about the effect of the fall in average incomes over the next decade on life expectancy and analysis of that issue by this team would be helpful. Particularly the fall in income for lower socio-economic groups.

    1. On 2020-04-12 04:43:17, user Bárbara Souto wrote:

      It seems that the authors want to reinforce an "almost-effect" of chloroquine. They wrote that matched 25 pairs of patients from Changsa, the only place using chloroquine, to compare the 25 individuals who have taken chloroquine, as a kind of case-control nested study. They said they found a difference of 12% and none patient in chlorquine group got worse (with a almost significance. p = 0.074). Well, two groups of 25 with zero outcome in one of them and 12% difference means that only three people in the comparison group got worse. A very small number to consider. But what about the p-value? 0.074 is a very promising p-value, suggesting that the non-significance was just a type II error. We should believe that this p-value is a probability derived from a Fisher exact test. But not. The p-value of a Fisher test from these numbers is 0.234. The 0.074 p-value came from a chi-square calculation. The authors have forgotten that the chi-squared test must not be used when there is a zero in one of the cells. I hope that the reviewers realize this shameful mistake and ask for correction.<br /> Francisco Souto<br /> Brazil

    1. On 2020-04-12 10:21:12, user Lev Yampolsky wrote:

      wait a minute - there is one more confounding co-variable that radically affects death per million for which no control has been done as far as I can tell - timing of epidemics onset in each county. Clearly everything started early in large cities which are both travel hubs and tend to have higher PM2.5 than some nice place in Montana. This analysis will only be possible after the outbreak is over everywhere.

    1. On 2020-04-13 08:45:15, user Alberto Villena wrote:

      Please, have a look to this preprint: Lianne Abrahams, 2020: Covid-19: acquired acute porphyria hypothesis. https://osf.io/4wkfy/<br /> "Macaques infected with SARS-CoV-2 also have<br /> decreased red blood cell numbers (Munster 2020) and susceptibility to<br /> SARS-CoV-2 appears to be determined by blood group; blood group A is most<br /> affected whereas blood group O seems to be protected (Yang 2020). This finding<br /> is concordant with previous studies showing that susceptibility to the 2003<br /> strain of SARS-CoV was determined by blood group (Guillon 2008)."

    1. On 2020-04-13 14:09:18, user Ian Sinclair wrote:

      This study seems to me potentially of enormous significance. I think it would gain greater acceptance if a) the authors explain why they chose to publish before they had reached the numbers specified in the protocol (100 for TAU and 100 for 4 mg group b) they say why they did not report the results for the 2 mg per day group c) they report the actual data on coughs temperature, numbers improved on radiology examination rather than just the significance levels d) they remedy a minor error in the summary (quotes 32 cases as against 31 e) they confirm that the measures were also made by staff who were blind to allocation f) they got themselves an editor who is a native English speaker. I absolutely do not think that the authors have anything to hide but they need to cope with a Western Audience that has been trained to be ultra critical, looking among other things for investigators who stop a trial the moment that it looks to be going their way. My guess is that this was not the case in this instance and that the study was running out of subjects or the authorities were asking for results or some other event that was out of the control of those running the trial. Given the potential world importance of this trial everyone should be trying to offer constructive suggestions for its greater acceptability rather than exercising their brains on ways in which mistakes might have been made.

    1. On 2020-04-13 17:55:08, user Izz Thatso wrote:

      The country benefiting the most from its BCG vaccination program, South Africa, probably has too few deaths in the defined 30 day period for it to be of statistical significance.

      Why do my comments need to be approved?

    1. On 2020-04-14 00:10:20, user Sinai Immunol Review Project wrote:

      Main findings<br /> This report is a retrospective analysis of clinical data collected from 47 patients with confirmed COVID-19 disease (median age 64.91 years; 26 males; 24 severe cases). Upon admission, patients were assessed with an APACHE II score (mortality risk) and SOFA score (organ system dysfunction). Among other clinical parameters, including lymphocyte count, CRP, and enzyme levels, serum lactate dehydrogenase (LDH) level was most positively correlated with APACHE II and SOFA. Additionally, strong positive correlations were found between LDH and observational assessment of severity of pneumonia and of underlying coronary disease. Notably, serum LDH was also positively correlated with other univariate parameters, including troponin and CRP. Meanwhile, LDH was negatively correlated with lymphocyte count across different subsets, including CD4+ and CD8+ T cells.

      Limitations<br /> Several studies have previously identified serum LDH as a predictive biomarker for COVID-19 progression (see references). Therefore, it is important to note that while LDH correlates with disease severity, it does not suffice as a reason for severe disease, so whether it serves as a robust predictive factor is not thoroughly supported; an explanation is still needed between the release of LDH from infected cells and disease severity.

      Significance<br /> The identification of serum LDH as a predictive biomarker does glean insight into possible links between its release into the circulation and immune response to SARS-CoV-2 pathogenesis. It is important to note that LDH release itself requires significant damage to and permeabilization of plasma membranes, leading to necrotic cell death. Therefore, this report, and others like it, should warrant investigation into the link between SARS-CoV-2 infection and potential necrosis or pyroptosis of infected host cells. It is important to note that previous studies concerning biomarkers for SARS-CoV-1 disease have also identified serum LDH as an important correlative factor with disease severity1. And subsequent investigations have studied the activation of inflammasome structures involved in pyroptosis in the presence of SARS-CoV2-3.

    1. On 2020-04-14 07:12:46, user Ole 500 wrote:

      So they dosed the subjects at x2 and x3 the normal dosage of 400 mg/day. If I'm reading this right, they overdosed the patients and rediscovered known overdose side effects?

    1. On 2020-04-14 08:38:01, user Quyen Vu wrote:

      I am not scientist or sort of , however, BCG vaccine has been already used in Vietnam in 70 or 80 years until now , now covid?-19 death rate in Vietnam today is zero (today April 14th) , yes 0? ?at current infected 265 and recovered 155 . BCG is total make sense to me .

    1. On 2020-04-15 13:58:04, user Buck Rogers wrote:

      I don't understand why there is nothing here about the fact that Hydroxychloroquine opens a Zync channel to the cell and prevents the virus from replicating. Hydroxychloroquine was given 200MG twice a day with Azithromycin 500MG once a day. And most important Zync sulfate 220MG once a day. If you don't test the correct protocol what is the value of the research?

    1. On 2020-04-15 14:38:14, user Lawrence Mayer wrote:

      I am very suspicious of articles that are data driven but do not allow us to see the data. Why aren't papers like this being submitted to peer-reviewed journals? They are guaranteeing two week turn around from submission to posted on line.

      I suggest people interested in learning about what has been reviewed by at least one expert join our discussion group on clinical epidemiology and science. We are over 2000 clinicians and epidemiologists discussing the latest empirical results on Covid19

      Lawrence S. Mayer, MD, MS, PhD<br /> Faculty Associate<br /> Harvard University<br /> Professor Emeritus of Public Health, Medicine and Biostatistics

      https/facebook.com/groups/covidnerds

    1. On 2020-04-16 06:58:38, user Kratoklastes wrote:

      It would have been useful to tabulate critical illness (and deaths) - both by age cohort - and to have given some indication of the statistical properties of the estimators (beyond p-values).

      The OR of 66× for the 75+ age cohort in the hospitalisation regression seems outlandish; the raw OR is 4× (i.e., the raw ratio of (Admitted|PosTest) for over-75s compared to the same quantity for 19-44 year olds).

      That looks (to me) like a collinearity issue in the regressor matrix - a really wide CI for one really-obviously-important variable is another clue. (Call me a Bayesian!).

      If your regressors were boolean (i.e., presence/absence) for comorbidities, VIF is not an appropriate test for collinearity: VIF performs poorly for categorical variables. Why not simply test the determinant of X´X, or its condition number, or its smallest eigenvalue?It's not a large matrix by modern standards - so it can't be a computational constraint. R's mctest package does a good job too (omcdiag includes the Farrar-Glauber test)

      I would suspect some rank deficiency caused by correlation between hypertension and variables that represent CVD-ish things; not necessarily pairwise - and this is the problem with booleans.

      Weak collinearity can happen because of weighted sums of columns - 3 boolean columns can give a run of '2' values, that correspond 'enough' with the '1's in the hypertension column. Add in other correlates with hypertension (age, obesity and maleness) and it would be suspicious if there wasn't collinearity.

      I don't think it would be viewed negatively if you dropped the 19 newborns (who are confounders in the 'hospitalisation' regression, since they are always hospitalised), so long as it was clearly disclosed: the presence of those 19 observations will also mess up the 'critical illness' regression as well (it would only require a handful of newborns to need critical care for things unrelated to COVID19, to bias up the OR for their age group, and their presence already biases up recovery rates).

      Lastly: it would be relatively straightforward to furnish the R script and the parameters (and residuals) without furnishing the data - that way interested people could generate their own pseudo-data and run some Monte Carlo experiments to get an idea of the asymptotic properties of the estimates. (To do it properly it would be good to have a covariance matrix so that the pseudodata could be generated by an appropriate cupola).

      .

      It's a pity that (as far as I am aware) Disqus does not permit any type of maths markup.

    1. On 2020-04-16 10:58:04, user Pete Quinn wrote:

      Hello, I expect that your results are far too conservative, by one or two orders of magnitude. I expect it's far more likely that by the end of this first wave (say mid-May, plus or minus), the total reported fatalities will lie in a range between about 1700 to 3500, given current restrictions (which are more lax than in neighboring Norway and Finland, which I estimate to report between about 200-450 and 100-250 respectively), and the new daily cases and fatalities will be small numbers by that point (handful or less fatalities daily). I hope I'm right and you are wrong, not for the sake of being right, but because, well, fatalities...

    1. On 2020-04-16 12:24:09, user Eric Zorrilla wrote:

      Hi,

      I tried to comment on the preprint, but I cant tell if my comment is going through. I noticed some sig/near-sig OR relations in your Suppl. Table 1 that I thought were potentially of interest to highlight in your paper in addition to those you highlighted?

      NON-HOSPITALIZATION also associated with<br /> Oseltamivir (of interest to field imo?)

      And sig or nearly sig with<br /> Atenolol<br /> Lisformin<br /> Metformin<br /> Rosuvastatin

      The joint presence of these anti-metabolic syndrome drugs (hypertension, diabetes, dyslipidemia) seems of interest as may support the emerging hypothesis/view that UNTREATED obesity/metabolic syndrome disorders may be worse than TREATED obesity/metabolic syndrome disorders as COVID-19 comorbidities.

      DECREASED VENTILATION was sig. associated with<br /> Cetirizine (Zyrtec)<br /> Diphenhydramine (Benadryl)<br /> Cholecalciferol (Vitamin D3)<br /> I’m not sure if the first two are an age confound or something interesting vis-à-vis immune reactivity/hypersensitivity, but may be of interest to deconvolute/deconfound?<br /> And the Vitamin D3 may be of interest given various other emerging preprints on it?

      Nice work and best wishes,<br /> Eric

    1. On 2020-04-16 17:51:35, user James Bell wrote:

      In the assumptions, was the ILI resulting from flu an average number or some lower number? This flu season peaked in November/December. A simple Google search reflects this through articles that were published about it at the time. If, in fact, an average ILI was used, doesn't that means the flu ILI are too high in this paper and that the overestimated value presumably belongs to COVID-19 instead? Also, since some in the media are arguing this paper "proves" we don't need the lockdown, is that really true (assuming all the assumptions in the paper are correct)? After all, we're talking about a virus for which none of us (except those who have recovered from it) have any immunity. If you have viruses with equal fatality rates, but we have herd immunity to one and not the other, doesn't common sense dictate there would be more fatalities with the latter (all else being equal)?

    1. On 2020-04-17 20:31:38, user Brian Byrd wrote:

      The authors have made a major error in asserting that

      In order to evaluate whether CCBs have therapeutic effect in COVID-19 patients, we<br /> 199 retrospectively analyzed the medical record of 487 adult COVID-19 patients with<br /> 200 hypertension...

      The effect of a drug is not discernible in this retrospective observational design.

      This major error is extended and compounded when the authors claim that have identified a mortality reduction effect:

      For the primary outcome of mortality, a beneficial effect in reducing the case fatality<br /> 234 rate (CFR) was observed in patients receiving amlodipine besylate, with the CFR<br /> 235 being significantly decreased from 26.1% (12/46) in non-amlodipine besylate treated<br /> 236 group to 6.8% (3/44) in amlodipine besylate treated group (P = 0.022).

      It is irresponsible of the authors to include these incorrect claims of public health significance in a manuscript that is likely to be widely read prior to peer review. It is not clear how the authors can fix any damage that may already have been done, but they should most certainly immediately remove these unsupported and inappropriate claims.

    1. On 2020-04-20 00:46:27, user deutsch wrote:

      A conclusion with only two days oh HCQ and such a smal sample is dishonest!<br /> The probability of missing real differences between the treatments is very high with such a small sample. For example with real death rate 2% vs 4% the probability of false conclusion is 90%....

    1. On 2020-04-21 08:34:15, user Maria wrote:

      Don't you think that lung capacity differences are instead the key to explain the lower incidence and severity of the disease in women and children? Thank you.

    1. On 2020-11-20 10:16:10, user Saar Wilf wrote:

      Thank you for this publication. Given that we already have two successful vitamin D RCTs and dozens of retrospective studies, I believe the null hypothesis is that vitamin D is effective in reducing COVID-19 severity. Therefore the failure of this study is not enough to accept a new hypothesis.

      Digging in to the details, the main problem is that the study was extremely underpowered. For example, the probability that it would have found a reduction of 50% in mortality with p<0.05 is only 25%. ICU admission and mechanical ventilation were a bit better with 61%, and 44%, and they indeed showed some effect (despite 1 treatment arm patient admitted to ICU before receiving treatment), although still not significant (as expected).

      Sample size was calculated assuming vitamin D would reduce hospital stay from a mean of 7 to 3.5 days. That is extremely unlikely, especially given that bolus vitamin D takes a day or two to be mostly converted to 25(OH)D (and I assume more for 1,25(OH)d. anyone knows how long?).<br /> In essence, the study had a low probability of achieving any of its endpoints..

      Patients were recruited 10 days after symptoms, and 90% were already on oxygen. It is possible that improving innate immunity is not very relevant at that point. Whatever effect it may have had, was further reduced due to 62% of patients receiving steroids.

      While the study does not show vitamin D is not effective, it does suggest that at a late stage, alongside steroids, it is not *immensely* effective, i.e. does not reduce odds of severe outcomes by 5x or more. Our analysis at Rootclaim suggests that a 5x reduction is a reasonable outcome when vitamin D is administered correctly.

      Thanks again.

    2. On 2020-11-18 13:37:42, user Dr Gareth Davies (Gruff) wrote:

      Thank you for this important study.

      I have some observations and suggestions regarding interpretation and reporting of this data.<br /> 1. A low or high P-value does not prove anything with regard to the effectiveness of an intervention. It only tells us whether we can confidently reject the null hypothesis or not. In this case, any treatment effect was obviously too small for a study of this power to measure. It does not mean there was no effect.<br /> 2. There appears to be a long time between onset of symptoms and randomisation and initiation of treatment (~10 days?). It takes up to a week for cholecalciferol to metabolise to 25(OH)D so it's not very surprising that by the time it did so, it was too late for these patients. The actions of 25(OH)D on renin gene supression and modulation of ACE2 expression need to occur much earlier to be beneficial in preventing an overactive RAS and cytokine storm.<br /> 3. The patient demographics show that the patient populations were overweight and obese, many with comorbidities (hypertension and diabetes) so again, it's unsurprising that this dose of D3 administered this late was not effective. This does not mean that D3 in general is innefective, merely that this protocol was not for patients of this type. I strongly suggest you make your conclusion statements more precise to reflect this, especially since Castillo et al have shown that administration of high dose calcidiol was very effective. Calcifediol is able to raise serum levels in hours compared to cholecalciferol.

      It would be a great shame to report this as a failure of vitamin D3 to treat COVID19 in general when it was simply this protocol for patients of this type and late disease progression which failed for enitrely comprehensible reasons.

      It was just too little too late.

    1. On 2020-11-21 02:14:12, user kdrl nakle wrote:

      Less than 5% acquired in schools. Except that you have no idea about asymptomatic transmissions because you do not test for that. And that is 50% or more among children. So when a parent gets infected you simply ascribe that to "community acquired." And we should really believe your tracing, right?

    1. On 2020-11-21 09:00:00, user Ton Soons wrote:

      the study didn’t address rhe intrinsic benefits of population-wide testing and it default assumes sensitivity and specifity will be a challenge. Suppose population-wide testing would be executed by using a pooled PCR-based testing strategy. Suppose the population-wide testing based on pooled PCR testing was addressed to a more realistic scenario by using it as a kind of front testing of HCW or at Nursing homes. More realistic scenario’s to study would als be:<br /> ->the added value of testing and monitoring of contacts during their quarantaine period (ttsi) in order to support backward contact tracing to find the super spreader. Keep in mind ‘mainly’ 20% is responsible for wide spreading!<br /> -> additional precaution taken to enter events, hospitals, restaurants, theaters, work envirionments.

    1. On 2020-09-14 19:25:30, user Vincent Fleury wrote:

      Can you provide the distribution by age of the deaths, I can't find it in the paper. What I read is that there are 8 times more people in the stratum age>65yo, while the mortality is only 3 to 4 times higher. If mortality occurs only in the >65yo, then this work shows 1-that HCQ is not given to elderlies and 2-potentially that HCQ is actually harmful.

    1. On 2020-09-16 13:44:38, user Ian wrote:

      Pretty bold to make a claim about the entirety of Milwaukee County with 100 observations at one grocery store (fewer than most counties on the list despite a million in population). The paper also mentioned their results "don't differ from Wisconsin at large using the KS test"-- but this is not cited. Based on what data? Your own convenience sample that overrepresents the white flight Twin Cities suburbs and appears to have looked at only one or two grocery stores per county, even in the large, diverse counties? This is much weaker data than you're making it out to be.

    1. On 2020-09-22 19:05:10, user fennudepidan wrote:

      If applicable, please cite the following peer-reviewed version of the manuscript: Wu, X., Nethery, R.C., Sabath, M.B., Braun, D. and Dominici, F., 2020. Air pollution and COVID-19 mortality in the United States: strengths and limitations of an ecological regression analysis. Science Advances (in press).

    1. On 2020-09-22 20:58:09, user Daniel Corcos wrote:

      1) The hazard ratio between twice weekly and once weekly HCQ have been inverted in the abstract as compared to the figure.<br /> 2) If there is no statistical difference between twice weekly and once weekly, why don't you pool the results to observe a statistical difference between HCQ and placebo?

    1. On 2020-09-22 22:58:23, user Stephen D wrote:

      It would be useful if you included "modelling" or "computer simulation" in your title. I've never been able to understand why so many advocates of draconian approaches to this virus have such a 'thing' against choirs and singing. Is there a lot of that going on where you live?

    1. On 2020-09-25 12:20:05, user Svet wrote:

      Please, can you analyse our samples from the surfaces, using this experimental, as a service or in collaboration? Can you measure kinetics of the virus on particular sample surfaces? What would be the price for 10 and 100 samples? Thanking you in advance, <br /> Dr. Katuscak

    1. On 2020-09-26 21:24:12, user Keith Robinson wrote:

      If one point could be made clear to me regarding Table 1, where a sensitivity of 84.7% and Positive Predictive Value of 89.7% are given: These are compared against the gold-standard qRT-PCR, right?

      How would these values change if we were to consider only individuals who carried a contagious level of the virus (viral load of 1 million/mL). As a contagiousness test, wouldn't the sensitivity and PPV get even better than they already are? Close to 99%?

    1. On 2020-10-12 15:02:39, user Tara Berger Gillam wrote:

      This article has been accepted for publication in the Journal of Public Health, published by Oxford University Press.

    1. On 2020-10-15 03:24:32, user correctnotright wrote:

      This is a flawed mathematical model that is contradicted by the facts on the ground. If 20% is the herd Immunity threshold, then why are there outbreaks in NYC, Italy, Spain and the UK that are all at or above 20%. there is ZERO evidence for pre-existing COVID-19 immunity. The existence of some cross-reactive T cell clones means nothing for immunity. There have been numerous examples of Covid-19 attack rates of over 85% (Choir in Mount Vernon Washington) - this disproves the notion of pre-existing immunity. We don't even know if people who have had the virus are immune from the exact same viral stain - but we already know the virus can mutate and people can get it again. It is not clear that neutralizing antibodies are effective and it is not clear that the cell mediated T cell response is protective for either transmission of disease or for prevention of serious disease and death. then there are the long term effects in people who did not die. So many problems with this poorly done paper. The infections decreased because people got scared, socially isolated, wore masks and changed their behaviors. How are infections starting up again if there is immunity? If there was some immunity it is either not sufficient or not long lasting.

    2. On 2020-10-26 12:37:26, user reader_DF wrote:

      Perhaps not at 20% but this research brings very good points that it should not be as high as 60 or 70%. Second wave is in part a terrible misunderstanding about the relation between confirmed cases and REAL number infected. Most likely, the peak of first wave was much higher as underreporting was back then much higher. Deaths numbers are a good indicator. What we are seeing now is a bit of improvement in treatment (or do you believe in some tremendous breakthrough?) and a pseudo second wave whcih is likely less pronounced if we could look at all cases not just confirmed ones.

    1. On 2020-10-15 09:13:31, user Ariel ISRAEL wrote:

      Our new large population study is out!<br /> We harnessed the power of Clalit unique database to identify drugs that significantly decrease severity of COVID. CoQ-10 with ezetimibe or rosuvastatin helps (a lot):<br /> Systematic analysis of electronic health records identifies drugs reducing risk of COVID-19 hospitalization and severity<br /> https://www.medrxiv.org/con...

    1. On 2020-10-16 10:49:03, user M&S wrote:

      "CONCLUSIONS ... Hydroxychloroquine ... <br /> appeared to have little or no effect on hospitalized COVID-19" HOSPITALISED!!! No one ever claimed it should help hospitalised patients; every proponent begs that the regime start after the first suspicion of infection!

    2. On 2020-10-17 03:11:04, user AB wrote:

      What constituted the standard of care arm in different countries and different centres? Did every centre use steroids? Or some did and some didn't for severe cases? In absence of specifying SOC, are we sure we are using a standard comparator for the intervention (experimental drug arms)?

    1. On 2020-10-16 13:31:39, user Benjamin Himes wrote:

      Encouraging results that should be tempered by further review. There are some large jumps to get from this test for viral particles to a clinical diagnostic. I've posted a full review on Zenodo

      https://doi.org/10.5281/zen...

      *edit new doi with proper file name extension. Previously .pdf.pdb which broke some ppls readers. Structural biologist faux pas : )

    1. On 2020-10-18 19:15:36, user Sam Wheeler wrote:

      How recent is recent enough? Should one repeat the influenza vaccine for example every 2 months until covid-19 vaccine is available?

      From the paper:<br /> "On average Covid-19 patients who received the inactivated trivalent influenza vaccine in 2020 – even if administered after the onset of SARS-CoV-2 infection-related symptoms – had significantly higher chances of surviving and less need for intensive hospital care than patients without recent influenza vaccination.<br /> If a long-lasting immunity had been the main mechanism of cross-protection, we should have also observed similarly protective effects for Covid-19 patients vaccinated in prior years, which was not the case in our analysis".

    1. On 2020-10-19 16:03:32, user Rogerblack wrote:

      This studies depression measure ASSUMES A HEALTHY PATIENT. 'little energy', 'trouble concentrating' 'moving slowly' = a minimum score of 3 due to physical symptoms of longcovid/fatigue.<br /> If very exhausted, this can easily rise into the 'severely depressed' range.

      It is not unreasonable to use the PHQ-9 or similar as a screening measure of disease severity.

      To use it in a patient population suffering from fatigue, concentration problems, ... is guaranteed to cross-read between those symptoms and anxiety - it is useless without a careful assessment of each question.

      It absolutely cannot justify sentances such as 'A significant proportion of COVID-19 patients discharged from hospital experience ongoing symptoms of breathlessness, fatigue, anxiety, <br /> depression and exercise limitation at 2-3 months ' without much more work, as it will lead to the conclusion that treating depression may benefit the patient when there is no depression, and it's a scale artifact.

    1. On 2020-10-20 16:12:43, user Martin Dugas wrote:

      The preprint by Westerhuis et al. looked <br /> only at severe COVID-19 cases (n=17). Our manuscript analyzes mild, moderate/severe and critical disease. We focus on antibody levels in the early phase of disease.

    1. On 2020-10-20 18:03:33, user Dinofelis wrote:

      One is again making the same fundamental error in this paper: not being able to reject significantly H0 is absolutely not a proof of H0 validity: it simply means the study wasn't accurate enough.

      For instance:

      "The overall mortality was not significantly different among patients who<br /> received hydroxychloroquine compared to the control group (OR: 0.94, <br /> 95% CI: 0.72 to 1.22; p = 0.63)"

      doesn't mean that one has shown that there is no effect on overall mortality. It means one doesn't have good enough data to conclude anything about it, and one could have an improvement of 25% in mortality (0.75 is within the CI) without being able to discriminate it from no effect or worsening.

    1. On 2020-10-22 03:50:45, user Shelley G wrote:

      When I look at Predicted Case Rate, I cannot find MO. It does show up on the death rate chart. Am I just missing it or is MO missing from the Case chart?

    1. On 2020-10-22 13:36:33, user Olaf Lange wrote:

      This is comprehensive study, which answers many topical questions. Still I am wondering about one part. If I start to observe an empty room and than at a time zero people start breathing I would assume, that the particles, which are able to transport a virus, increase. But they drop by 30% even without purifiers. Do we look at the right parameters, which are able represent the increasing production of respiration?

      Another explanation is that the humans acts as filters themselves, as stated in the text. Actually they would filter even more, than they produce. In this case, experiments with empty rooms, with and without purifiers might be a good supplement to determine absorption rate of the humans.

    1. On 2020-10-23 14:56:07, user Caetano Filho wrote:

      The study states that 38 participants of the<br /> nitazoxanide arm reported complete absence of symptoms after 1 week follow-up. How this number of participants could correspond to 78% of the 194 ones enrolled on that arm?

    1. On 2020-10-26 20:47:12, user Val wrote:

      I enjoyed your discussion on perinatal HIV and concerns with detectability. While this assay seems promising, I am a bit dubious about its accessibility in LMIC with high rates of pediatric HIV, which would theoretically need it the most. I do not work with ddPCR, but since many regions of LMIC countries lack the trained staff, lab space, and equipment to even perform RT PCR at the moment, and those that do a frequently only accessible by urban residents, do you think it is feasible for a ddPCR assay to be effective? Great paper, I enjoyed reading it.

    1. On 2020-10-27 03:23:50, user Anonymous wrote:

      Great analysis and synthesis of ideas pertaining to EV isolation. The utilization of EV-like liposomes are certainly supportive of the argument for NBI methods and back-up the validity of that isolation method. The only source of confusion I encountered were the error bars in the figures being quite variable in relation to the claimed increased or decreased biomarkers without much note of their presence. In a clinical setting, can one truly indicate a difference between ALS and SBMA, for example, if data pertaining to EVs/biomarkers tend to fall within those error bars? What is the implication here for misdiagnosis?

    1. On 2020-10-30 19:50:12, user Louis Rossouw wrote:

      There is a typo on page 8. Second last paragraph discusses excess mortality and refers to Appendix P. I think it should refer to Appendix Q.

    1. On 2020-11-05 08:26:01, user OMSK wrote:

      Nice work!<br /> I wonder whether brain DPP4 level is associated with BMI. Because GLP-1RA is known to exert its effect on BW via hypothalamus in the brain, it might be possible that higher brain DPP4 level leads to faster degradation of GLP-1, then leads to more weight. Have you done such analysis?

    1. On 2020-11-09 11:47:26, user Thomas Grünebaum wrote:

      I wonder if there is an influence on adults having childeren in kindergarden versus adults who do not have.

      The reason I am asking this is becaus I assume that the immune system of parents is more trained and has a better response to covid.

      Is there any study realated to this? I fully agree this might be too specific and quite dificult to determin due to lack of data.

    1. On 2020-11-17 01:51:32, user Chris Barker wrote:

      Maybe I missed in the article. where is the formal definition of MAPE? and did you have access to the actual computer programs to run the projections or was it based on evaluation of the literature?

    1. On 2021-08-11 15:21:36, user circleofmamas wrote:

      They need to include vaccination status of the cases, hospitalizations and deaths. Canada is one of the most widely vaccinated countries in the world, and we can see from Israel data and UK data that the vaccinated are particularly vulnerable to the Delta variant.

    1. On 2021-08-12 19:41:47, user Steven Ramirez wrote:

      I wonder if the study disentangled time since vaccination. If protection diminishes over time for both vaccines it needs to be controlled for.

    1. On 2021-08-15 12:39:25, user Jérémy LLL wrote:

      Thank you for the useful work. In your Figure 1, the green trend line does not seem to fit any data on the right half of the plot. Can you explain this? Did you exclude the negative detections (Ct 50) from the fit?

    1. On 2021-08-26 05:56:47, user William Brooks wrote:

      This is an interesting paper that fails to find an effect of early bar/restaurant closures during Japan's second state of emergency (SoE). However, I think it has several limitations.

      1) Were early closures actually justified?

      The authors fail to point out that the SoE started one week after the effective reproduction number (Rt) had peaked and 2/3 days after it had gone under 1 throughout Japan [1, slides 17-18], so the SOE was unnecessary for preventing the "collapse of the medical system", which was the government's justification. Also, the early closure of bars/restaurants in Tokyo/Osaka prior to the SoE didn't stop Rt increasing during the second half of December exactly the same as in the rest of Japan without early closures [1, slides 19-20]. This isn't surprising since even the extreme lockdowns in Peru and Argentina couldn't counteract seasonal rises in Covid infections [2].

      Furthermore, even if there were statistically significant reductions in self-reported coughs and sore throats, do the authors think these could justify the negative effects on employment, firm exit, and mental health mentioned in Introduction?

      2) Why suggest capacity limits but not ventilation improvements?

      In addition to early closures, the Japanese government also recommends mask-wearing while dining out (which is unlikely to be effective [3] [4]) and the use of plastic partitioning in restaurants/bars (which may actually increase infection risk [5]). The authors suggest capacity limits, but this doesn't solve the socioeconomic impacts mentioned in Introduction. Fortunately, even modest improvements in ventilation may be as effective as high-quality R95 masks [4], so investments in improved ventilation/air-purification could be a better solution.

      [1] https://www.mhlw.go.jp/cont...<br /> [2] https://www.scirp.org/pdf/o...<br /> [3] https://doi.org/10.7326/M20...<br /> [4]https://aip.scitation.org/d...<br /> [5] https://doi.org/10.1101/202...

    1. On 2021-07-03 04:25:21, user Sumedh Bhagwat wrote:

      COVAXIN preprint says: (Protocol, Supplementary appendix 2) but there is just a single supplementary file which is not the protocol .<br /> Please attach Supplementary appendix 2.

      Actual Vaccine Efficacy against symptomatic Covid when calculated using formula is 77.27% ~77.3%<br /> Paper reports 77.8% which is due to rounding off of fractions

    2. On 2021-07-05 05:35:47, user The Scrutinizer wrote:

      Sumedh Bhagwat you are really worried about the difference of 0.5% error? That is not the most imp thing to worry about right now. If you don't know this vaccine might be the best so far in comparison to all the genetic vax

    1. On 2021-07-04 13:13:09, user Sebastian Rosemann wrote:

      Dear authors,

      you write: "For each country we predicted the ‘baseline’ mortality in 2020 based on the 2015–2019 data (accounting for linear trend and seasonal variation; see Methods). We then obtained excess mortality as the difference between the actual 2020 and 2021 all-cause mortality and our baseline. For each country we computed the total excess mortality from the beginning of the COVID-19 pandemic (from March 2020) (Figure 2, Table 1)"

      For Germany Table 1 shows 36.000 excess deaths, which makes up 4% increase, suggesting a baseline of 900.000 for March 2020 - Mai 23th 2021.<br /> According to destatis<br /> https://www.destatis.de/DE/...<br /> the yearly number of deaths for 2015-2019 in Germany was always above 900.000.<br /> How can your baseline for a timespan of ~14 month be lower than the actual number of deaths for 12 months within the last years?

    1. On 2021-07-06 09:34:12, user David wrote:

      Just 188 participants were seropositive. So if long covid would occur in 1% of patients one would statistically expect 1,88 long covid sufferers in this group. Obviously, this sample size is too small to quantify the long covid risk. It could probably be anywhere between 0 and 3% according to this study. Assuming this survey covers long covid symptoms well which it obviously does not as already mentioned in the 2 previous comments.

    1. On 2021-07-18 17:48:18, user Olavo Amaral wrote:

      In their discussion, the authors attribute the imbalance between study arms within hospital sites in Table S1 to the fact that randomization was not stratified by study site and to batch delivery of drugs to remote sites. However, if the trial was truly randomized as described in the Supplementary Appendix, it is hard to understand how the extreme imbalance in the distribution of arms in each site could have happened by chance alone.<br /> When one considers the distribution of arms in the 3 centers located in Manaus, where patients are described to have been randomized independently, there is a total of 97 patients in the active group vs. 298 in the placebo group. Using the chisq.test function in R, the p value for obtaining this distribution if patients had an equal chance of being randomized to either group is 4.8e-24 (amounting to a chance of roughly 1 in 200 sextillion).<br /> For the remote sites, one cannot estimate this chance exactly, as the authors report that randomization was performed for batches of 5-50 patients. Nevertheless, the distribution is even more unbalanced (220 patients in the active group vs. 30 patients in the placebo group), and also seems unlikely to have occurred by chance (and in the opposite direction as that of Manaus, thus balancing the distribution in the whole trial almost exactly).<br /> Although this imbalance would not explain the difference in the main outcomes between groups by itself, as the stratified analyses in Tables S3 and S4 show effects favoring the active treatment at all sites, it does call into question whether randomization was really carried out as described. Can the authors please clarify on what happened?

    1. On 2021-07-25 17:46:53, user Ullrich Anton Schuler wrote:

      I recommend to reduce percentages to 3 significant digits. E.g. "33.76% of SARS-CoV-2 exposed children" means with 548 children: 1% is 5 children, 0,1% represents 0,5 children, 0,01% means 1/20th of a child. Therefore two digits after the decimal point in my opinion are nonsense (as it is the case in very many situations!). The unnecessary question arises as to whether the authors have a sensible feeling for numbers and sizes.

    1. On 2021-07-26 10:03:20, user Christoph Terasa wrote:

      In Table 2 on page 12, the heading on the rightmost column reads

      Control group of unvaccinated users testing negative

      Shouldn't that be

      Control group of unvaccinated users testing positive

      This is what I gathered from the text, and it makes more sense.

    1. On 2021-07-27 14:07:32, user VailShredBetty wrote:

      Additionally, the only conclusion I can see that this study and subsequent studies illustrate is the vaccine is stopping ore severe infections....but at what cost??

    1. On 2021-07-29 08:14:23, user Alenita Luz Mateo wrote:

      Thank you for this very informative research, my mom is currently on dialysis and suffering from comorbidities. As of today 7/29/31, we are on the MECQ status here due to a surge of covid19 cases, with delta variant on the loose. I think our health care system is breaking, from 5:18 nurse to patient ratio it now down to 3:18. Our health care providers are already contracting the virus, are now exhausted. I'm afraid they cannot function well anymore to do their duties for their patients. Yesterday another EKD patient died and contracted the covid19. We are a third world country and access to materials needed are very crucial. I hope we can survive this pandemic.

    1. On 2021-07-30 16:26:33, user Kirsten Elliott wrote:

      The search strategy for this review has not been adequately reported as there is not enough information in the appendix to replicate the search in any database, as it hasn't been made clear which fields were searched. It is therefore not compliant with PRISMA 2009 or PRISMA 2020 reporting standards. Furthermore, from what has been reported it appears that relevant search terms and subject headings have not been included, meaning that it is likely relevant papers have been missed. The other paper based on this search will also have the same problem (https://www.medrxiv.org/con... )

    1. On 2021-07-30 16:44:39, user Kirsten Elliott wrote:

      There seem to be at least three reviews based on the same literature search, so I'm posting similar comments on each of them.

      The search strategy for this review has not been adequately reported as there is not enough information in the appendix to replicate the search in any database, as it hasn't been made clear which fields were searched. It is therefore not compliant with PRISMA 2009 or PRISMA 2020 reporting standards. Furthermore, from what has been reported it appears that relevant search terms and subject headings have not been included, meaning that it is likely relevant papers have been missed.

      For this paper, there's also a mistake in the flow diagram in figure 1 - it appears as if papers have been excluded on the grounds of "non Spain" when that should be "non Africa".

    1. On 2021-08-05 06:56:54, user Piet Streicher wrote:

      How do you account for "survivorship bias" in the case of the vaccine. If prevalence was high during the vaccination phase, then infection prior to the 14d after 1st dose or prior to 7d after 2nd dose would remove people from the pool, creating a bias towards those that made it past this point. They then appear to have a higher level of protection (say 86%) when in fact this group has been through a selection process already.

    1. On 2021-08-08 03:47:33, user theasdgamer wrote:

      This is a very, very interesting paper. I'm not a molecular biologist nor medical, so please excuse any questions I have which may seem silly.

      First, are there any conditions which might lead to viral escape from lysozomes but not zinc escape? It seems to me likely that if zinc isn't escaping, then maybe the viruses aren't escaping the lysozomes either. And should there be conditions where the viruses escape, then I would expect zinc to also escape into the cytosol. Could you perhaps give a little background here? Am I totally mistaken, and if so, how?

      Second, are there bio-regulatory systems in human cells in vivo that might be triggered by high pH in lysozomes that might cause acidification of lysozomes to increase?

      Third, how much variability is there in CQ protonization over time from continual lysozome acidification, or am I mistaken in thinking that cells continually acidify lysozomes?

      If CQ is given twice daily, how much variation in CQ levels would there be in the lysozomes of cells in vivo?

      What differences in results might we expect if we were to use hypothetical human vascular endothelial cells as the cell medium for in vitro studies?

      Hydroxychloroquine is the more active form of chloroquine and I wonder if it might make some difference in results versus chloroquine.

      Thank you for a most interesting paper.

    1. On 2021-08-09 19:22:35, user Pedro Nomad wrote:

      As a 65 year-old participant in phase 2 of these Medicago clinical trials, I can say that after receiving my 2 doses of their candidate-vaccine in February 2021, I am still Covid free (tested twice, negative each time) and have been very active doing outdoor activities: skiing, skating, hiking, rollerblading, cycling. Couldn't feel better! Just hoping to get a 3rd (booster) shot in time to resume my teaching job this Fall. Hurry up and give Medicago the GO!

    1. On 2021-09-16 10:25:45, user kdrl nakle wrote:

      Yup, Alberto is right, the sample is really not much to speak of. I actually tend to believe conclusions but not based on this paper. Folks, you need to do a lot more work.

    1. On 2021-09-17 16:27:49, user Daniel Keyes wrote:

      I commented previously that the apparently long delay in peer review is understandable, and typical for many journals. However, given the extremely high profile of this article, it should have been considered for expedited review. Now even "regular" review seems to be taking rather long. This could be due to reviewer delay, reviewers or editors requiring a large number of changes, or it may even reflect a general reluctance to publish an article with this conclusion.

    1. On 2021-09-17 17:39:54, user Gabriel Lee wrote:

      I do not understand the numbers for vaccine doses administered during the June 1 to July 31 period. These seem far too low for a city with a population of almost 1 million and the two-month period during which some of the fastest vaccination rates in Ontario and Canada were recorded. I downloaded the CSV file from open.ottawa.ca and tallied the Moderna and Pfizer-BioNTech doses in this period, and found 352,687 and 481,264 doses, respectively. Could you please explain?

    2. On 2021-09-22 17:13:20, user dansari wrote:

      The authors do not describe their methodology for comparing incidence of myocarditis/pericarditis to background rates, and thus whether they have been disproportionately reported following vaccination, nor whether this comparison has indeed been performed. Reference (8) does include this detection.

      As mentioned in other comments, the "Vaccinated in Ottawa" data (is this the dataset that was used?) indicate that 838,442 vaccinations were given during the time frame of this study. This would yield approximately 4 cases per 100,000, a figure more in line with currently understood vaccine statistics.

    3. On 2021-09-25 19:07:25, user Peter Dimitrov wrote:

      Time frame : 60 days following vaccination; Location: who presented at a single academic institution/Ottawa Heart Centre. Patients were identified by admission and discharge records of the Ottawa Heart Centre. Sample size: 32 patients, 29 of which were male! Median time between vaccination dose and pain symptoms was 1.5 days. Startling findings by themselves, ought to raise curious questions not outright dismissal/withdrawal.

      However, obviously the incidence rate based on total MRna vaccines in Ottawa area is waay off. Am curious to know disaggregated sex/age etc data of Myo/peri cases recorded in the exact same time period in all the other hospital treatment centres in Ottawa area?

    1. On 2021-09-21 14:02:09, user Isatou Sarr wrote:

      If it works, why not!!!!! What is the genome similarity index between MEASLES-MUMPS-RUBELLA (MMR) and COVID-19? Is there any common clinical manifestation identifiers between the ailments?

    1. On 2021-09-25 09:47:21, user Jan Podhajsky wrote:

      Hi there. There is a question about personal and sensitive data protection during obtaining answers via questionnaire distributed through social web.

      1. Unsufficient introduction to the survey. No mention about sensitive data personal data being colellected via questionnaire in consent question
      2. Missing contact to authority responsible for Personal and sensitive data protection
      3. Doubts about processing of personal data especially electronic personal data like cookies, refferals, geolocation
      4. Questionnaire enabled continuation without previous login into the system which might lead to other person to access personal and sensitive data, e.g. on shared computers

      I raised these concerns to data protection authority of Faculty of Scince, CUNI.

      The questionnaire research did not met personal and senstive data protection standards. It is unethical research by my opinion.

    1. On 2021-09-27 18:07:42, user DiogenesNJ wrote:

      "early attention directed towards identifying SARS-CoV-2 in returning <br /> international travelers may have led to a failure to recognize locally<br /> circulating infections..."

      And why were we limited to that? A shortage of testing capability attributable to CDC's insistence on their own test, which turned out to be flawed, and FDA's refusal to allow any of several other tests developed by the private sector or universities to be used during the crucial first months of the pandemic. A classic NIH problem (Not Invented Here, not to be confused with the National Institutes of Health).

      Compare and contrast with South Korea, which called in every big pharma company in the country during their New Year holiday, approved a private-sector test in early February, and had massive testing up and running in less than a month (15k tests/day capacity by the end of February).

    1. On 2021-09-30 12:14:12, user Joren Buekers wrote:

      Exciting research! I hugely support focussing on continuous SpO2 measurements instead of discrete/spot check measurements only! I'm happy that this finally found its way to COVID research. <br /> A small remark/request: you indicated that "preprocessing of the raw SpO2 signal was performed using a block filter as in Levy et al.", but this block filter was actually developed in another study (https://doi.org/10.2196/12866) "https://doi.org/10.2196/12866)"). Would it be possible to acknowledge the original paper as well? (Sorry for the self-promotion, but I do think it's more correct to refer to the original paper).

    1. On 2021-10-01 13:49:12, user Jasper wrote:

      I'm wondering if there's any research on the long term effect of the single-shots. Most studies have antibodies, B and T cell checks at one week after the second dose. Doesn't that result interfere with a possible prolonged effect of the first dose? This study shows that there's a sound immune response, that apparently (according to the BNT162b2 clinical studies) provides ample protection 12 days after the first dose (especially if you don't take into account the positive cases during the first 12 days without protection). VE also seems really good in real-life - albeit the time period and groups are smaller - but still massive, considering the magnitude of this operation. Maybe we don't need to boost with 0,3ml, and so on.

    1. On 2021-10-03 18:28:32, user Joseph Psotka wrote:

      It would be much more useful to look ar discrepancies in the immune system of the twins. Even dizygotic twins can have the same immune system.

    1. On 2021-10-04 07:26:13, user Wolfgang Wagner wrote:

      Int J Mol Sci. 2021 Aug 27;22(17):9306.<br /> Epigenetic Clocks Are Not Accelerated in COVID-19 Patients

      PMID: 34502212 <br /> PMCID: PMC8431654 <br /> DOI: 10.3390/ijms22179306