On 2020-04-24 14:46:54, user Russel Future wrote:

THis is a really important article. Reports from New York hospitals also show evidence of positive outcomes using heparin to treat Covid-19. Reuters article, re. experiences at Mt. Sinai hospital, New York:<br /> https://www.reuters.com/art...

SARS-Cov-2 virus seems to cause small blood clots to form in lungs. Reports indicate Mt. Sinai using new protocol where dosages of heparin above typical prophylactic levels are now given to patients before lung blood clots are detected.

On 2020-04-25 14:05:52, user Rosemary TATE wrote:

Hi, I dont see the STROBE guidelines checklist uploaded, although you ticked yes to this<br /> "I have followed all appropriate research reporting guidelines and uploaded the relevant EQUATOR Network research reporting checklist(s) and other pertinent material as supplementary files, if applicable. "<br /> A lot of people seem to ignore these but they are important and any good journal will require them.<br /> Can you please upload? Many thanks.

On 2020-04-25 14:55:25, user Ivan Berlin wrote:

Fontanet et al. Cluster of COVID-19 in northern France: A retrospective closed cohort study<br /> medRxiv preprint doi: https://doi.org/10.1101/202...<br /> Commentary about the finding of lower prevalence of anti-SARS-CoV2 seropositivity among smokers compared to non-smokers.<br /> Ivan Berlin, Daniel Thomas, Anne-Laurence Le Faou, Paris, France<br /> This is a correctly run retrospective closed cohort study aimed to assess the prevalence of anti-SARS-CoV2 seropositivity among a group of 661 individuals in a region of France with high COVID-19 incidence rate. Seropositivity and clinical symptoms (questionnaires) were assessed. No RT-PCR data are provided about the presence or absence of SARS-CoV2. <br /> To note that among the 878 individuals first invited, only 326 (37%) agreed to participate. Further, the sample was completed by 345 individuals. This recruitment history cannot exclude selection bias e.g. smokers or former smokers were more likely to decline participation than nonsmokers.<br /> Of the 661 participants 452 (68,4%) reported respiratory symptoms and could be considered as having COVID-19. Among the 661 participants only 171 had anti-SARS-CoV2 seropositivity (25.9 %).The discriminative ability of the serological determination seems to be weak with respect of the clinical symptoms in particular between no symptoms and minor symptoms according to Suppl. Table 1: No symptoms: 13.9%; Minor symptoms: 16%, Major symptoms 37.7 % (but Minor symptoms: 26% according to Table 2. <br /> The reported smoking rate is 10.4 % (69/661). Out of the 661 individuals tested, among nonsmokers 167 (25.7%) and among the smokers 5/69 (7.2%) were tested serologically positive. No definition of “smokers”, no data about former smokers, no biochemical verification of smoking status are provided. It cannot be excluded that there were some occasional smokers, recent or long-term quitters in the nonsmoker group. <br /> Page 8 last paragraph: “Smoking was found to be associated with lower risk of infection ”after adjustment for age and occupation. No adjustment for gender was done, however the sample’s consists of largely more women than men (women: 62% men : 38%,) which is the opposite in the general population in France. The authors use “infection” for seropositivity all over the manuscript. However, infection can be implied if SARS-CoV2 RT-PCR is positive. The fact that 452 reported clinical symptoms of infection and only 25.9 % were seropositive leads to the question what are the factors contributing to seroconversion i.e. having a good immune response to SARS-CoV2. This is not explored in the paper (no SARS-CoV2 RT-PCR). A proxy answer to this question would be to provide detailed clinical and demographic information in Suppl Table 1, not only serological information. More specifically, what is the distribution of smokers and nonsmokers according to the clinical features and serological findings.<br /> Because of the low seropositive rates (25.9%) and lack of direct SARS-COV2 detection one cannot conclude about factors contributing to the seroconversion. One can hypothesize that the higher seropositive rate among nonsmokers can be due, among other factors, to the higher percent of women. <br /> In this sample the percent of women is higher than that of men contrary to other reports of individuals with COVID-19. In large samples, men have higher smoking rate than women and more men have COVID-19 than women. More than one third of the sample is <=17 years old; the smoking prevalence in this group is usually very low. It is likely that the nonsmokers with seroconversion are mainly women. One can hypothesize that smokers have a lower seroconversion rate that non-smokers explaining the lower percent of smokers with seropositivity.<br /> The observed lower anti-SARS-CoV2 seropositivity rate among smokers in this sample is an interesting unexpected but secondary finding. Before drawing any conclusion about this finding and generalize them, further studies are needed aiming to assess specifically the incidence of COVID-19, SARS-CoV2 infection and anti-SARS-CoV2 seropositivity among well documented smokers, correctly classified former smokers versus lifelong never smokers.

On 2020-04-25 15:03:10, user David Ian Walker wrote:

You detected virus at 100 fold lower levels in treated effluent than in untreated influent. However, you do not specify what type of treatment the wastewater was subject to. This is important to know. Even if we could just have an idea of whether it was secondary or tertiary etc, that would be useful, but ideally a little more detail such as the type of treatment processes that are used at the Parisian WWTPs would be very helpful.

On 2020-04-25 15:11:00, user beencensured wrote:

This study would have been much more helpful if they could give what type of patients with which types of comorbidities had died with Clinical Characteristics of the Patients and demographic data.

Of these types of patients this is the mortality rate for each type of patient including demographics. This data would have been then useful though limited. Right now there are much sicker patient in HCQ group, and so more dead. So to rationalize the study, we should know the death rate for each comorbidity per demographic group including clinical characteristics of patients. This would have given some direction to the study.

On 2020-04-25 16:58:34, user Mike wrote:

As a reminder, medrxiv.org displays this on their opening page:

Caution: Preprints are preliminary reports of work that have not been certified by peer review. They should not be relied on to guide clinical practice or health-related behavior and should not be reported in news media as established information.

(I added the emphasis)

On 2020-04-26 18:36:10, user Christopher Rentsch wrote:

We believe that Magagnoli et al failed to correctly identify intubation occurring in hospitalized patients testing positive for COVID-19. They used CPT codes 31500, 94002, 94003, and E0463 and ICD-10 procedure codes indicative of assistance with respiratory ventilation, or extracorporeal membrane oxygenation (ECMO). We identified 5,906 COVID-19 patients treated in the Veterans Health Administration between March 1 and April 21, 2020. In addition to the above CPT codes, we identified intubation according to ICD-10 procedure codes for insertion of endotracheal airway, and respiratory ventilation, which were usually concordant. We cross-validated with medications typically used during intubation, such as neuromuscular blocking agents (e.g., succinylcholine, rocuronium) and short acting sedatives (e.g., propofol, midazolam). We also found these intubation codes most frequently in the context of intensive care. We did not find similar evidence of face validity for ventilation assistance codes. No instances of ECMO were found as this procedure is unlikely to be used in the Veterans Health Administration.

We classified 307/5,906 = 5.2% patients as intubated. Using the Magagnoli algorithm, only 96/5,906 = 1.6% patients were said to be intubated. Of these, 37 were classified based on ventilation assistance codes, not indicative of intubation.

List of ICD-10 Procedure codes used to identify intubation

Codes in both Magagnoli and Tate lists<br /> - Respiratory Ventilation (5A1935Z 5A1945Z 5A1955Z)

Codes in Magagnoli list, but not Tate list<br /> - Assistance With Respiratory Ventilation (5A09357 5A09358 5A09359 5A0935B 5A0935Z 5A09457 5A09458 5A09459 5A0945B 5A0945Z 5A09557 5A09558 5A09559 5A0955B 5A0955Z)<br /> - Extracorporeal Oxygenation, Membrane (5A1522F 5A1522G 5A1522H)

Codes in Tate list, but not Magagnoli list<br /> - Insertion of Endotracheal Airway Into Trachea (0BH13EZ 0BH17EZ 0BH18EZ)

Janet P. Tate (Janet.Tate2@va.gov)<br /> Christopher T. Rentsch (@DarthCTR)<br /> Joseph T. King Jr.<br /> Amy C. Justice

VA Connecticut Healthcare System<br /> West Haven, CT

On 2020-04-23 16:10:24, user quillerm wrote:

Why are all the survey participants over 65 and such a large percentage diabetics?

On 2020-04-21 19:40:16, user Brandon B wrote:

Risk of ventilation was 6.9% in HQ + AZ group and 14.1% in no Tx group. That is double. It was stated that these numbers are similar in the article. Not significant?

On 2020-04-27 03:23:55, user buzzbree wrote:

Beyond the seroprevelance conclusions of the study which are generally consistent other reports, another very important issue that needs to be clarified by the authors is if the study fully adhered to Good Clinical Practice (GCP) standards.

To be fully compliant with GCP the Stanford IRB really needed to be informed of the the email Jay Bhattacharya's wife (https://www.buzzfeednews.co... "https://www.buzzfeednews.com/article/stephaniemlee/stanford-coronavirus-study-bhattacharya-email)") sent to potential subjects. The email had several erroneous statements- that the test was FDA approved (Its not) and they would know if they were now immune from COVID-19 and would know that they were free from getting sick and could no longer spread the virus. These statements could have impacted subject safety by encouraging riskier behavior (i.e. ignoring social distancing) from the study subjects if they believed that the test was FDA approved and a positive result was definitive proof of protective immunity.

In the Buzzfeed article Dr. Bhattacharya has stated that he did not know about the email or approve of it, but he still had an ethical duty to report it to the IRB when he found out. There is only one line in manuscript stating that IRB approved the study- how the IRB addressed this email should be expounded upon in final manuscript given these new issues that have come to light.

Relevant GCP sections:

"3.3.8 Specifying that the investigator should promptly report to the IRB/IEC:(b) Changes increasing the risk to subjects and/or affecting significantly the conduct of the trial (see 4.10.2).

4.10.2 The investigator should promptly provide written reports to the sponsor, the IRB/IEC (see 3.3.8) and, where applicable, the institution on any changes significantly affecting the conduct of the trial, and/or increasing the risk to subjects.

On 2020-05-01 15:58:00, user Diego Fleitas wrote:

First, thank you for your research.<br /> Second, I have some doubts about your weighting criteria. Because of: it is not known the relation between contagion and socio demographic aspects; weighted results are almost three times higher than raw prevalence what looks out of scale; and also the relation of 80 times fold between actual diagnosed and projected looks a bit out of scale.<br /> Best

On 2020-05-01 21:51:10, user John Schuna wrote:

Perhaps I am missing something, but the manuscript’s variance estimation method for test sensitivity and specificity remains unclear. It appears that sensitivity and specificity estimates were calculated from the summed raw counts across the described studies (Page 19) upon which exact 95% confidence intervals were constructed.

As sensitivity and specificity data were drawn from multiple studies, it would seem prudent to use something like a generalized linear mixed-effects model (with study as a random effect) for variance estimation and subsequent generation of confidence intervals around your sensitivity/specificity point estimates. This would seem more straightforward for specificity analyses; however, this could be problematic for sensitivity analyses due to the small number of studies.

On 2020-05-03 15:40:02, user Charles Rosa wrote:

The WSJ, referencing this study, has the following quote in one of their article: "It suggests that the large majority of people who contract Covid-19 recover without ever knowing they were infected, and that the U.S. infection fatality rate may be more than an order of magnitude lower than authorities had assumed. Based on this seroprevalence data, the authors estimate that in Santa Clara County the true infection fatality rate is somewhere in the range of 0.12% to 0.2%—far closer to seasonal influenza than to the original, case-based estimates."

On 2020-05-12 17:17:05, user Michael A. Kohn, MD, MPP wrote:

As I said in my comment on the first version of this pre-print, the authors did a great job of collecting this data and reporting their results and assumptions. From the 3439 people who showed up for testing, they were able to obtain 3330 valid specimens on which to perform the Premier Biotech antibody test. Of these, 50 were positive. That’s 50/3330 = 1.5% . They re-weighted their sample to reflect the county’s sex-race-zip code distribution and reported an estimated county-wide sero-prevalence of 2.8%. In the first version, they miscalculated their confidence intervals. In the original pre-print, they reported 2.81% (95CI 2.24-3.37%); in this one they are reporting 2.8% (95CI 1.3-4.7%). Their new confidence interval is 3 times as wide as that originally reported. This was not delta method versus bootstrapping; it was a simple matter of plugging the wrong numbers (variances) into the well-known Rogan-Gladen formula that adjusts apparent prevalence based on an imperfect test (which the authors apparently re-derived). We have posted an online calculator that calculates the confidence interval correctly: https://www.sample-size.net/prevalence-estimation/

On 2020-04-17 20:25:02, user Mortal Wombat wrote:

Hold on, they made no adjustment for self-selection of symptomatic people in their study?

Researchers, at least please tell us the number of people shown the Facebook ad so we can have some sense of the potential for self-selection -- i.e. how many saw it but chose not to participate.

This seems problematic given the population demographic adjustments that were necessary. The researchers say that white women were heavily oversampled while hispanics and Asians were heavily undersampled, and that population adjustments led them to adjust the observed prevalence of 1.5% up to a population-weighted prevalence of 2.81%.

It would appear highly likely that the relatively affluent population (white women) would have the interest and capacity to be a roughly random sampling -- people with no prior symptoms just interested in knowing. Whereas for lower-income populations, there may be less ability to simply participate out of interest, and may have been a higher self-selection drive of the previously symptomatic to get themselves tested.

Thus I find the upward adjustment in the numbers quite suspect. To come up with an overall result that's _higher_ than the raw outcome of the study when you know that people who've been sick will be the most motivated to get themselves tested just seems perverse.

Did the study not even ask people whether they had been sick over the past couple months? Why not? That at least could've given some sense of whether self-selection was biasing results in the samples.

On 2020-04-18 02:01:37, user mendel wrote:

First, he picked the county that had the earliest cases in California and had the outbreak the first, ensuring that the population would be undertested. This means that it's likely that every other county in California has fewer unregistered infections than Santa Clara.

Second, study participants were people who responded to a facebook ad. This is a self-selected sample, and this property completely kills the usefulness of the study all by itself. This is a beginner's error! People who think they had Covid-19 and didn't get tested or know someone who did are much more likely to respond to such an ad than people who did not. (By comparison, the Gangelt study contacted 600 carefully chosen households per mail, and 400 responded. Still somewhat self-selected, but not as badly.)

Third, age is the one most common predictor of mortality. He did not weigh the results by age, and old people are underrepresented in the study. Anything he says about mortality is completely useless if we don't know how prevalent the infection was in the older population. (In Germany, cases show that the prevalence among tested older people was low initially and took a few weeks to rise.)

Fourth, instead he weighs prevalence by zip code--why? This exacerbates statistical variations, since there were only 50 positive results, and Santa Clara has ~60 zip codes. If you have a positive result fall on a populous zip code by chance where only a few participants participated, then the numbers are skewed up. They must have seen this happen because their estimated prevalence is almost twice as high as the raw prevalence.

Fifth, the specificity of the test is "99.5% (95 CI 98.3-99.9%)". This means that theoretically, if the specificity was 98.5%, all of the 50 positive results could be false positives, and nobody in the sample would have had any Covid-19. This means the result is not statistically significant even if the sample had been well chosen (which it wasn't). (It's not even significant at the 90% level.)

Sixth, they used a notoriously inaccurate "lateral flow assay" instead of an ELISA test and did not validate their positive samples (only 50) with a more sensitive test -- why not?

Seventh, The Covid-19-antibody test can create false positives if it cross-reacts with other human coronavirus antibodies, i.e. if you test the samples of people who had a cold, your speficity will suffer. Therefore, a manufacturer could a) test blood donor samples, they not allowed to give blood if they have been sick shortly before; b) test samples taken in the summer when people are less likely to have colds than in March.

To state the previous three points this in another way, a large number of positive results (a third if the specificy is actually 99.5%, but probably more than that) are fake, and depending on which zip codes they randomly fall in, they could considerably skew the results.

On 2020-04-18 06:35:20, user DomesticEnemy wrote:

About 5 weeks ago based on a study of the Diamond Princess cruise ship I estimated the death rate to be around 0.23% or so. Welcome.

On 2020-04-18 18:39:18, user jj wrote:

Where is the discussion of selection bias? You invite folks to get tested by advertising on Facebook... I think there will be an over-representation of folks who fear they have COVID-19 based on their recent interactions in places with or around COVID-19 cases.

Without randomization to eliminate self-selection bias, the authors should not be making any far-reaching conclusions that are now being picked up and reported by the media without providing proper interpretation.

I think this publication should be rejected for not doing this study properly.. and then seeking publicity!

On 2020-04-23 05:57:15, user David Feist wrote:

It is always good to compare data within nations. But in fact preliminary, linear regression analysis, from a fellow maths major, now seems to indicate that the lockdowns had no statistically significant effect within the USA: https://www.spiked-online.c....

This Santa Clara study indicates why Sweden, Japan, South Korea and Australia have not had public health apocalypses, with no lockdowns; the mortality rate was miscalculated.

On 2020-05-01 12:52:41, user Scott Howell wrote:

Any chance that the data and code can be made available? Seems fitting given the topic and re-analysis.

On 2020-04-17 21:38:11, user Michael Stein wrote:

There could be a very large upward bias due to the participants in the study being people who responded to the Facebook ad. It stands to reason that people who suspected they might have been exposed to the virus would be more likely to respond to such an ad. The fact that randomization was used to select who got the ads and that corrections for demographics were made does not address this potentially serious source of bias. There is little doubt that many more people have been infected than the official numbers, but I find the factor of 50-85 rather hard to believe in a place like Santa Clara County that has not been overrun by cases.

On 2020-04-18 04:34:06, user Zev Waldman MD wrote:

I agree with prior commenters that people who suspected that had or were exposed to Covid would be more likely to seek antibody testing. I see that participants were asked about prior symptoms, but it would also have been nice to ask about prior possible exposure concerns, If both numbers are very low, that would provide some reassurance about this possible bias.

I really wanted to address another issue: the calculation of the infection fatality rate, i.e., estimated deaths/cases. It seems that a lot more thought went into trying to get an accurate case count than an accurate death count. They seem to take it as a given that 50 people died of Covid in the county as of April 10; however, like case counts, there are multiple reasons to suspect this number of deaths might be higher:

1. Reporting of deaths is well-known to be delayed - i.e., date of reporting does not equal date of death

2. People who actually died of Covid may never have been tested, and thus may not be included as cases or deaths

3. The doubling time of deaths that was used to project to April 22 is based also on reported deaths; if reporting of deaths is delays, the doubling time may appear slower.

I did appreciate the authors' efforts to validate the antibody testing. That's useful information.

I worry that, because these results support their prior beliefs, some readers may take these results at face value and push them for policymakers to use before they have been more widely vetted by the scientific community.

On 2020-04-18 05:16:02, user rodger bodoia wrote:

Deeply flawed methodology. Others have noted (as did the authors) the obvious inherent bias towards those seeking antibody testing (maybe they had symptoms, maybe they knew someone who had symptoms). Also note the bias that is inherent in the method of using Facebook as the messenger with a brief period between posting on FB and the actual testing. We would need significant information on the other behaviors of people who use FB this frequently and whether they are more or less likely to have engaged in practices that would have put them at risk of acquiring the virus.<br /> Back of the envelope "smell test": 48,000 infections and only 69 deaths (as of April 17) is an infection fatality rate of 0.14%. This is inconsistent with Diamond Princess data, even if we adjust for age differences. Also compare with https://www.nejm.org/doi/fu... in which they did UNIVERSAL screening of obstetric patients from March 22 to April 4 in NYC and found 15% positivity of SARS-CoV-2. Without lots of population-weighted adjustments we can interpret this as pretty good evidence of roughly 15% prevalence in NYC (say roughly 1.2 million infections) and roughly 9,000 deaths for infection fatality rate of 0.75%

On 2020-04-18 18:45:15, user S. MonDragon wrote:

Dear Dr. Jay et al.,

I am curious about a couple of other scenarios related to your study. Do those that have SARS-CoV-2 antibodies, also show any other antibodies that might be of particular research interest? And further, how many of these people actually had any symptoms? For example, how many of those who had COVID-19 antibodies also had antibodies for other types of coronaviruses, including SARS-CoV-1. Did the presence or absence of these other antibodies seem to have an effect on symptom severity? I guess what I am asking is, why do some people have such severe symptoms while others can walk around without even knowing that they may have the virus? And, can your samples help us to answer some of these questions?

On 2020-04-19 00:34:15, user SonoranSeeker wrote:

Considering that this virus is extremely contagious, two or three times that of the flu, and considering that this extremely contagious virus was circulating unabated for a relatively long time, this study is probably pretty accurate. It is also in line with the study in Germany and modelling of virus spread based on previous corona virus characteristics. <br /> This could be why there were so many deaths in such a short time. Let's hope it burned hot, but will flare out just as fast.

On 2020-04-19 02:02:27, user defragmentingthecode wrote:

THe CDC's guidelines for reporting Covid19 deaths is "where the disease caused or isassumed to have caused or contributed to death".

I really don't know how accurate any studies are when we don't know how many Covid deaths were assumed, and how many deaths were due to the patient's co-morbidities rather than the presence of virus? Surely, we could have got this bit right?

Here is the CDC link. https://www.cdc.gov/nchs/da...

On 2020-04-19 05:26:28, user chalkful wrote:

You can talk about selection bias, and that’s valid. But nit-picking every part of this study down to the assumptions made about manufacturer specifications seems ridiculous and very hypocritical when similar assumptions were made with RT-PCR tests that were rushed to market with questionable, if any, validation, and which are relied upon to make public policy decisions which dictate the lives of millions, even billions.

Did you apply precisely the same level of detailed, critical analysis and nitpick every minute inconsistency in all the other COVID-related “peer reviewed” studies which were rushed to print, and which were gloom-and-doom?

Why choose to write off all conclusions drawn by a study with such a substantial effect size because of minor statistical errors? As a non-academic, the study appears largely methodologically sound, despite a few flaws, and it is not logical to throw the baby out with the bath water. Even assuming the infection rate is half of what is concluded, that is still significant and the first study, and conclusion, of its kind.

I suspect that, with as with most things COVID-related, much of this is due more to politics than pure intellectual rigor, and that the conclusions drawn by the study shake the foundation of what many believe, which scares them.

On 2020-04-20 02:47:08, user Comfrey's Gone wrote:

Probably related to Dean Karlen's observations below - but in working through the statistical appendix, it seems like the calculation of the standard error is independent of the number of samples (371 or 401) used by the manufacturer/Stanford team to evaluate the number of false positives.

To determine the standard error, the authors first compute the cumulative variance by combining variances from each source of uncertainty (finite sample of respondents of 3,330, finite sample for false positives in the serology test and finite sample for false negatives). These separate variances are the variances of the binomial distribution (p(1-p)), not rescaled by the inverse of the sample size. The authors then take this cumulative variance and divide by the number of respondents (3,330), and apply the square root to arrive at the standard error. (.0039 = sqrt(.034/3330)).

Instead, when the cumulative variance is computed in the equation for Var(Pi) above, I believe that each of the contributing terms should be multiplied by its appropriate 1/N (where N is the relevant sample size, e.g. 3,330 for the Var(q) term, and 371 or 401 for the Var(s) term.)

One way to assess that the 'N' rescaling doesn't seem right is to think about the limit in which the number of respondents being tested is infinite, the sample size for determining the number of false negatives is also infinite, but there is a finite sample (e.g. 401) used to determine the number of false positives. If you trace through the appendix calculation, you'll then find (if I've done it correctly) that the standard error for 'Pi' (the infection rate) would then be zero, although some error certainly should exist, due to the uncertainty in false positives.

Other commenters have also raised concerns about the normality assumption in computing the standard error, but the way in which scaling by sqrt(N) has been applied here has a large impact on the calculation of the standard error and resulting confidence intervals.

On 2020-04-20 03:50:27, user Tomas Hull wrote:

Those who insist on the selection bias of this study: Would you rather see the ads targeting people working in hospitals and covid19 assessment centres, or those providing essential services to those institutions, like mailmen, delivery men, garbage men, cleaning and maintenance, and so on? <br /> How about people in self-isolation, COVID-19 observation and ICU wards? <br /> Would this kind selection bias satisfy anybody?

On 2020-04-20 10:50:57, user Douglas Gilliland wrote:

Peer review rejects this study... https://medium.com/@balajis...

On 2020-04-27 19:53:17, user eldhose poulose wrote:

https://data.gov.in/catalog...

This is the link that I received from your paper

On 2020-04-27 22:39:49, user pam garcia wrote:

Obesity, diabetes, and hypertension are clearly the major factors in hospitalizations and deaths from Covid-19.

Northwell Health just released a study of over 5000 Covid-19 patients that revealed 94 percent of the hospitalizations and deaths involved comorbidities, obesity being the most prevalent factor.

It appears crystal clear that the true pandemic is the comorbidities, most which are preventable by not overconsumption of sugary, salty, wheat and corn based processed foods and drinks.

We need to eliminate these fake foods from the world population diet, as this data is overwhelmingly similar across the world.

If these preventable conditions are eliminated, in 10 or 20 years we can afford health care for all.

On 2020-04-28 03:55:27, user nicky wrote:

CFR data as of 27th April<br /> https://uploads.disquscdn.c...

On 2020-04-28 16:26:32, user Philip Machanick wrote:

The description of what happened in Korea is grossly inaccurate. The Korean strategy included aggressive and comprehensive contact tracing and quarantining contacts.

NYC illustrates where relying on herd immunity takes you. The Bronx has about 0.22% of POPULATION dead, i.e., mortality rate, not case fatality rate. Extrapolate that to all of the US and you have over 700,00 fatalities.

This study is flawed because it does not take into account NPIs that have varied a lot - e.g., though Germany started late, they ramped up testing fast and adopted the S Korean strategy. Italy instead at the start focused on testing the most ill and hence miss mild and asymptomatic cases, resulting in a much higher case fatality rate.

On 2020-04-29 14:05:10, user Maxwell wrote:

Note the massive conflicts of interest for these "experts." So we are to trust a study whose authors receive money from an industry that will be profiting handsomely from the very thing they are studying? Unfortunately that is the current state of affairs in Academia particularly in the sciences.

For example:

DMW has received consulting fees from Pfizer, Merck, GSK, and Affinivax for topics unrelated to this manuscript and is Principal Investigator on a research grant from Pfizer on an unrelated topic.

Can we also get a full disclosure from Mr. Weinberger on any personal investments he may have in pharmaceutical companies or anything related to that industry.

Can we also get full disclosure on the funding that Yale School of Public Health receives from industry? That includes any and all foundations connected to the pharmaceutical industry.

On 2020-04-29 16:03:06, user Eugene Kutsin wrote:

1. comparing flu to covid death based on a hand-picked interval is misleading. The peak of flu distribution based on 2017-2018 CDC data comes in January and is not included in the statistics. based on 2017-2018 data, 34% of death fall into the interval, while covid peak is included, meaning that at least 50% of deaths is accounted;
2. excess death is calculated based on 2017-2019 (I guess average or mean). Based on country wide CDC data (https://www.cdc.gov/nchs/da... "https://www.cdc.gov/nchs/data/nvsr/nvsr68/nvsr68_09-508.pdf)") making adjustment for New York we see deaths count fluctuates in tens of thousands. So the 13K falls into marginal error. This makes measure "excess deaths" as computed not usable;
3. statement "number of COVID-related deaths [based on excess number] is in fact greater ... as trauma and even heart attacks, appears to have decreased during the shelter-in-place period of the pandemic" is a speculative one. while it's probably true that trauma related deaths would go down during lockdown, number of death related to cases when people are not getting adequate medical help by avoiding visits or lack of resources;

On 2020-04-29 19:47:21, user Frank Conijn wrote:

To the authors:

Thank you for this review, which is very useful.

However, I did find a small error. In the Flowchart you're listing the study by Molina et al, reference #24, as a non-randomized trial. That's incorrect, because in a non-randomized trial one is still supposed to compare one's experimental group with a comparable control group. Molina et al experimented with patients with severe covid, while the control group consisted of patients with light or mild covid from another study (mean symptom duration until treatment: 4 days). That's comparing apples and oranges. It should be listed as a short communication.

In table 3, you are describing it correctly.

On 2020-04-29 22:52:54, user Steven Markowskei wrote:

Could you kindly define what is included in "chronic cardiac disease". <br /> In particular does it include uncomplicated hypertension?

On 2020-04-30 00:28:52, user conceitedlawyers wrote:

The paper is very important and interesting. Almost all commentators didn't seem to read the full paper. The author points out that the dosage considered in the in vitro studies is unlikely to be high enough. However, unlike 9/10 commentators, the author of the paper does NOT say that Ivermectin should be ruled out. The author suggests alternatives (e.g combination therapy using a mix of antivirals). I respectfully concur with the author and dissent from the views expressed in almost all comments above.

On 2020-04-30 14:30:10, user Ewa Kirkor wrote:

One more difficulty is in establishing the start and end of the period of contagion after the infection takes hold. Some individuals still have positive rt-PCR test outcome 6 weeks after the COVID19 symptoms appear. How would such spread of parametrization of the model affect its predictions? Could you let me know at EKirkor@NewHaven.edu

On 2020-04-30 14:34:37, user Ivan Berlin wrote:

Rentsch CT et al. Covid-19 Testing, Hospital Admission, and Intensive Care Among 2,026,227 United States Veterans Aged 54-75 Years. <br /> medRxiv preprint doi: https://doi.org/10.1101/202... version posted April 14, 2020<br /> Review of the results concerning smoking related issues.<br /> Ivan Berlin<br /> The title is somewhat confusing. Only 3789 persons were tested for SARS-CoV-2.<br /> Data are extracted from the Veteran Administration (USA) Birth Cohort born between 1945 and 1965 electronic database. Between February 8 and March 30, 2020, 3789 persons were tested for SARS-CoV-2. Among them 585 were tested SARS-CoV-2 positive (15.4%) and 3204 SARS-CoV-2 negative. (Remark: the authors frequently confound testing for SARS-CoV-2 and having the disease: COVID-19.)<br /> Testing used nasopharyngeal swabs, 1% of the testing samples was from other unspecified sources. Testing was performed “in VA state public health and commercial reference laboratoires”, page 7. No further specification about the testing method is provided. Data are analyzed as if no between test-sources variability existed. However, it is unlikely that between test-source variability would influence the findings.<br /> Data extraction included diagnostics by diagnostic codes of comorbidities, non-steroid inflammatory drug (NSAID), angiotensin converting enzyme inhibitor (ACE) and angiotensin II receptor blocker (ARB) use, vital signs, laboratory results, hepatic fibrosis score, presence or absence of alcohol use disorder and smoking status.<br /> Smoking status data, never, former, current smokers were extracted using the algorithm described in McGinnis et al. Validating Smoking Data From the Veteran’s Affairs Health Factors Dataset, an Electronic Data Source. Nicotine & Tobacco Research, Volume 13, Issue 12, December 2011, Pages 1233–1239, https://doi.org/10.1093/ntr... used for HIV patients. According to this paper, the algorithm correctly classified 84% of never-smokers 95% of current smokers but only 43% of former smokers. The reported overall kappa statistic was 0.66. When categories were collapsed into ever/never, the kappa statistic was somewhat better: 0.72 (sensitivity = 91%; specificity = 84%), and for current/not current, 0.75 (sensitivity = 95%; specificity = 79%). Thus, classification error cannot be excluded in particular in classifying former smokers. <br /> In unadjusted analyses (Table 1) factors associated significantly with SARS-CoV-2 positivity were: male sex, black race, urban residence, chronic kidney disease, diabetes, hypertension, higher body mass index, vital signs but not NSAID or ACE/ARB exposure. It is to note, that among the laboratory findings, severity of hepatic fibrosis was associated with positive SARS-CoV-2 tests. <br /> Among those with alcohol use disorder, 48 (8.2%) tested positive versus 480 (15%) who tested negative (p<0.001).<br /> Among never smokers 216 (36.9%) tested positive vs 826 (25.8%) who tested negative. Among former smokers 179 (30.6%) tested positive vs 704 (22%) who tested negative. Among current smokers 159 (27.7%) tested positive vs 1444 (45.1%) who tested negative. Expressed otherwise, among SARS-CoV-2 negative individuals, there were less never smokers, less former smokers and more current smokers. To note: the reported OR for current smoking should be the inverse to that presented i.e. <1 and not >1. However, among individuals with SARS-CoV-2 positivity there were more persons with positive smoking history (former + current smokers): 57.8 % than with negative smoking history (never smokers): 36.9%. <br /> COPD, known to be strongly related to former or current smoking, was more frequent among SARS-CoV-2 negative (28.2%) than among SARS-CoV-2 positive (15.4%) individuals p<001).<br /> In multivariable analyses (Table 2), male sex, black ethnicity, urban residence, lower systolic blood pressure, prior use of NSAID but not ACE/ARB use and obesity were associated with SARS-CoV-2 positive test; current smoking (OR: 0.45, 91% CI: 0.35-057), alcohol use disorder (OR 0.58, 95%CI: 0.41-0.83) and COPD (OR: 0.67, 95%CI: 0.50-0.88) were associated with decreased likelihood of SARS-CoV-2 positive test. No association with age and SARS-CoV-2 positive test was observed. The association with hepatic fibrosis with SARS-CoV-2 positive tests remained significant in the multivariable analysis and the authors point out (page 15) that the “pronounced independent association with FIB-4 (fibrosis) and albumin suggest that virally induced haptic inflammation may be a harbinger of the cytokine storm.”, page 15. <br /> The main risk factors for hospitalization or ICU among SARS-CoV-2 positive persons are those that associated with a worse clinical signs (status). This is expected: clinical decision about severity is based on current clinical signs and not on previous history. <br /> Neither co-morbidities, nor smoking status or alcohol use disorder were associated with hospitalization/ICU. Surprisingly, age was inversely associated with hospitalization (Table 4) among SARS-CoV-2 positive individuals.<br /> Conclusion of the reviewer.<br /> This is the first report showing that there are less current smokers among SARS-CoV-2 positive persons. However, smoking history (former + current smoking) seems to be more frequent among SARS-CoV-2 positive individuals than never smoking. It is not known what is the percent of former smokers who were recent quitters; duration of previous abstinence from smoking is a crucial variable in assessing associations with smoking status. This raises the question of the validity of smoking status category classification. <br /> It is not known when smoking status is reported with respect of the SARS-CoV-2 testing. It is likely that individuals with clinical symptoms stopped smoking some days before testing and considered themselves as former smokers.

The fact that alcohol use disorder, which is frequently associated with tobacco use disorder, is also less frequent among SARS-COV-2 positive individuals raises the question of the specificity of the smoking finding and the raises the contribution of substance use disorders overall i.e. the finding about current smoking is part of a cluster of various previous or current substance use disorders e.g. cannabis use, potentially associated with SARS-CoV-2 negative test. <br /> COPD as well as current smoking are being reported to be more frequent among SARS-CoV-2 negative individuals raising the possibility that reduced respiratory function (entry of SARS-CoV-2 is by the respiratory tract) is associated with lower likelihood of SARS-CoV-2 positive tests. This hypothesis may suggest that reduced respiratory function and not smoking itself is associated with higher likelihood of SARS-CoV-2 negative tests. <br /> The paper does not report on analyses of smoking by clinical signs/co-morbidities interactions. It is likely that former smokers or those with alcohol use disoders are more frequent among individuals with comorbidities. Based on previous knowledge about smoking associated health disorders, one can assume that more severe clinical signs were associated with current smoking or among recent quitters; the smoking x clinical signs interaction is not tested. <br /> The authors conclude on page 14 “To wit, we found that current smoking, COPD, and alcohol use disorder, factors that generally increase risk of pneumonia, were associated with decreased probability of testing positive. While they were not associated with hospitalization or intensive care, it is too early to tell if these factors are associated with subsequent outcomes such as respiratory failure or mortality.”<br /> The reduced current smoking rate among SARS-CoV-2 positive individuals is an interesting but preliminary finding. It is likely that it is part of a more complex symptomatology and not specific to current smoking. Smoking status should have been assessed on a more detailed manner. The current findings, from a retrospective cross sectional analysis, certainly not support the hypothesis that current smoking protects against SARS-CoV-2 positivity.

On 2020-04-30 14:43:28, user Alan Beard wrote:

Repeating a question from a few minutes ago. Can you clarify whether the Smoking Status included in the data comes from <br /> 1) A current Smoking status only .......OR <br /> 2 Current and Former Smokers(irrespective of when smoking ceased)<br /> This is an important question that really should be answered

On 2020-04-30 23:52:08, user Dena Lester Arnold wrote:

I understand basic Biology. I obtained my BS 48 years ago. I recently taught HS Biology but this is beyond my understanding. Is there anyone who can explain this study to me in simple terms?

On 2020-05-08 17:47:49, user Markus Cornberg wrote:

Dear Leif,<br /> very important data.<br /> we are also looking at this in Hannover (CD8 though). Maybe we can talk about this if you like<br /> Please look at these studies in mouse models to understand cross-reactive T cell responses.

Anti-IFN-? and peptide-tolerization therapies inhibit acute lung injury induced by cross-reactive influenza A-specific memory T cells. Wlodarczyk MF, Kraft AR, Chen HD, Kenney LL, Selin LK. J Immunol. 2013 Mar 15;190(6):2736-46. doi: 10.4049/jimmunol.1201936. Epub 2013 Feb 13.

All the best<br /> Markus

On 2020-05-01 19:31:01, user Randy Jackson wrote:

Thank you for this contribution. I found it very useful for understanding the tradeoffs between deaths from the virus and duration of the epidemic with alternative social distancing scenarios. But, you don't discuss or address the immunity bit. With R0=2.75, ~100% immunity is achieved, but very low percentages of population with lower R0. So, my question: What happens when such a large part of those still alive aren't immune? Rebound?

On 2020-05-02 12:27:16, user Thomas Clarke wrote:

"Current social distancing measures may be argued to either increase or decrease variation in exposure, depending on the compliance of highly-susceptible or highly-connected individuals in relation to the average"

One aspect of this during lockdown is the separation of the population into essential workers (highly connected) and locked down population (minimally connected). The connectivity here for the highly connected group can be estimated, based on hygiene and PPE regimes in workplaces rather than compliance, and in many countries during the COVID epidemic has been shown to be high.

That therefore would drive one element of herd immunity quickly during lockdown: the immunity of the essential workers. How significant this is overall then depends on the compliance with lockdown measures as compared with the leakiness of lockdown through necessary contacts with essential workers, and whether the job of "essential worker" remains attached to individual identity, so that this type of analysis applies.

Modelling this bimodal behaviour explicitly might have some merit.

On 2020-05-03 13:39:59, user Cristine Carrier wrote:

Wow, I can't believe they made almost the exact same mistake as the doctors in Bakersfield. On what planet would people actively seeking out medical care a representative or random sample of the entire population of a city? If they had 33 cases of broken legs would they then use that number to calculate that the entire city of Kobe had the same rate of broken legs as the people going to the clinic?

On 2020-05-05 00:24:27, user Marc Imbert wrote:

The dosage is twice higher then the protocol raoult , without blood analysis to follow up toxicity apparently. The disparities between the groups are high with negative conclusion to be expected.

From figure 3 and 2, we can deduct that 56% (23/41) patient in the high dosage group and 35% (14/40) of the lower dosage group are in intensive care at enrolment. Not discussed in the study, although it has a clear impact on lethality, and CK.

A pertinent critera to look at, 51% hypertension profile in high dose CQ at enrolment, while the comparative group has less. To note, it is the exact figure ( and the highest) associated with the " poor clinical result outcome" in Roult's 1064 series ( with HCQ) and the "death outcome" the Chinese reference study without HCQ.

There are certainly more investigating data to explore out of this study,

On 2020-05-05 22:07:15, user Katri Jalava wrote:

I would be quite cautious about shielding. If possible to isolate to an island or similar, then ok, but within the community, I think that is a major challenge. I think it may only create highly vulnerable clusters within the population which then in turn increase the force of infection substantially. One of the main thing to be done in EU is to break these chains of transmission within care homes by e.g. excluding exchange of staff between units, removal of symptomatic from the care homes to isolation units and preventing staff from working even if a suspect case within their household. It may be that this outbreak is much driven by care home clusters.

On 2020-05-06 08:05:15, user Prof Pranab Kumar Bhattacharya wrote:

Dear Editor<br /> In the world, Corona virus cases jumped up till 3rd May 2020 from December 2019 is 3,51,743 with death 2,45,617 (18%) and 31.5 death per one million people of infected.Almost 212 countries worldwide and most affected countries are USA,( death rate 304, followed by Spain (540),Itali 475, UK 414, France 379 per million population when in India total cases of positive by RT PCR is 40,266 death 1300 per one million people and in West Bengal province of India total infected is 963 with death 48 cases as per ministry of health government of India records on covid 19. The question is why such a huge percentage of death from this dangerous virus ( no more should be considered simple like influenza virus) inspite of lockdown, social distancing ventilation guided treatment protocol for mild moderate and severe pneumonia from covid 19?<br /> Mortality from covid 19 is higher in groups at higher risks of thromboembolism including hypertension, types 2DM, obesity, coronary artery disease ,cardiomyopathy, pre existing renal pathology as co morbid condition known to all. It has been also seen world wide that the risk of thromboembolism ( both venous and arterial) are more likely to occur when patients are admitted at ICU or in PEP ventilation, ànd in aged over 60 yrs( approximately 63% of death in India from covid 19).<br /> What did the autopsy studies revealed of these death, though very limited autopsy were performed with covid 19 death as the virus is HG 3 category virus. Brane Hanely (1) eral published in journal of clinical pathology of BMJ group showed histopathology of lungs on HE stain oedema, Type Ii pneumocytes hyperplasia,large pneumocytes with ground glass viral inclusions bodies focal inflammation, multinucleated giant cells,when no hyaline membrane ( a histopathological features of ARDS) diffuse alveolar damage. The pulmonary vessels showed hyaline necrosis with thrombus formation and capillary congestion.inflamatory infiltrate composed of alveolar macrophages in alveolar lumen and lymphocytes in interstitium. Zhe Xu et Al (2) in journal Lancet reported also one 50 year old man died on day 14 of covid 19 after being treated with lopinovir+retinovir+moxiflixain and high nasal cannula oxygen therapy and niddle autopsy of lungs liver and heart tissue showed diffuse alveolar damage with cellular fibrimyxiod exudate,dissquamation of pneumocytes and hyaline membrane formation (sign of ARDS) , interstitial mononuclear inflammatory infiltrate dominated by lymphocytes ( CD8) multi nucleated syncitial giant cells, atypical pneumocytes and microvascular thrombosis in pulmonary vessels (2).Sufang Tian et Al (3) did post mortem needle core autopsy of four patients who died of severe covid 19 pneumonia and patients age range were 59-81 years and time of death 15-52 days were in ventilation. Histology of their finding in lungs were again diffuse injury to alveolar epithelial cells, hyaline membrane formation, hyperplasia of type II pneumocytes , diffuse alveolar damage and consolidation by fibroblasts proliferation with extra cellular fibrin forming clusters.All these tour cases had vascular congestion with intravascular thrombus suggesting an acute phase components reaction and fibrinoid necrosis of blood vessels.The autopsy finding of heart was that endocardia and myocardia didn't contain inflammatory cellular infiltrate, although focally myocardium appeared irregular in shape with darkened cytoplasm and fibrinoid necrosis of blood vessels in myocardia.There were various degrees of focal oedema interstitial fibrosis and myocardial hypertrophy which suggests patients had underlying hypertension associated with hypertrophy or past ischemic injury. A large series of 38 cases of autopsy of lung by Luca carsana etall (5) showed from death cases of covid 19 in northern Itali on H&E stain showed also diffuse alveolar damage, capillary congratulations, necrosis, necrosis of pneumocytes, hyaline membrane, interstitial oedema,type II pneumocytes hyperplasia, platelet fibrin rich thrombus(5) .Electron microscopy showed viral particles within cytoplasmic vaccoule of pneumocytes.<br /> So from above post mortem studies, besides ARDS like pictures in terminal event , platelet fibrin rich thrombosis of pulmonary vessels, myocardial vessels, hyaline necrosis of blood vessels of both lungs and of myocardium are prominent picture along with endothelial dysfunction according to this author.The severe cases of pneumonia from covid 19 also shows increased D Dimer value (4) prognostically bad , increased c reactive protein, increased pro calcitonin and increased FDP value<br /> All these suggest to me that pathogenesis behind so many death in ventilation or at ICU of covid 19 patients are not ARDS itself but some kinds of coagulopathy or DIC occurred before death in severe pneumonia cases<br /> Though lymphopenia, inflammatory cytokine stroms ( raised IL6,raised TNF are for cytokine stroms)are typical abnormalities described in almost all literature in highly pathogenic Covid 19 infection with disease severity ,only one rapid response in BMJ (4) suggest , based on post mortem finding use of low molecular weight heparin (LMWH) to be included in the treatment modules of covid 19, particularly those who have high D Dimer high FDP value in serum though TT,APTT,PT,INR may not show any significant difference.use of heparin therapy with constant monitoring for bleeding manifestation should be instituted in patients showing clinical signs of turning towards severe pneumonia,along with antiviral therapy with remdesvir (within 7 days onset of symptoms at scheduled disease)<br /> If the pathology behind the death of severe pneumonia in covid 19 patients is DIC ( according to Autopsy finding the pneumocytes are not killed or destroyed by the virus nor by cytotoxic T cells, rather proliferation occur with much viral replication and virus load) there will be DIC , vascular congestion, thrombosis there will be AMI stroke ) then treatment at ICU with ventilation become useless unless if thromboembolism is not resolved first with LMWH infusion <br /> Referencs <br /> 1) Brain Hanley, Sebastian B Lucas,Esther youd,Benjamin swift,Michael Asbron "Autopsy in suspected covid 19 cases " JCP 73,(5):2020 http://dx.doi.org.10.1136/jclinpath-2020-20652<br /> 2) Xu Z,Shi L,Wang y eral "Pathological finding of covid 19 associated with acute respiratory distress syndrome " The Lancet respiratory medicine 8 (4):420-22 :2020<br /> 3) Sudan Tian, young xiong,Shu yuan xiao,Liu H et all "Study of 2019 novel Corona virus disease ( covid 19) through post mortem core biopsy" Modern pathology (Nature.com ) 14 th April 2020 http://doi.org/10.1038/s 41379-020-0536-<br /> 4) William Atenio ,Nadu Okonkwo "should prognostic models for covid 19 not also incorporate markers of thrombosis" Rapid Response published BMJ on 14th April 2020 to article"Prediction model for diagnosis and prognosis of covid 19 infection: systematic Review and clinical analysis" The BMJ 2020:369:m1328 published on 7th April 2020 https://doi.org/10.1136/bmj...<br /> 5)Luca carsana, Aurelo sanzogoni ,Ahmed Nast, Roberta Rossi etall"pulmonary post mortem finding in a large series of covid 19 cases from northern Itali" MedRxiv https://doi.org/1101/2020.0...

On 2020-05-06 11:58:39, user Ran Israeli wrote:

Very interesting, good manuscript.

Is there a chance you expanded the data from Figure 1 (Especially 1A and 1D) to a more updated one?

On 2020-05-06 18:09:14, user Olujimi A-Williams wrote:

I was wandering if the writers can look for any Vitamin D levels in this cohort. Any level found in the EMR will do. It sure will be interesting to see

On 2020-05-07 19:08:36, user Cranmount wrote:

The author correctly notes that his model gives "surprising" conclusions, and it is not difficult to conjecture where the model goes wrong, with the result that it makes an incorrect extrapolation to future mortality (including the unwarranted prediction that imposing social distancing (after more than a minimal delay) increases cumulative mortality). Put crudely, the model incorrectly fits the logistic curve by conflating the effects of all control steps into herd immunity, giving the illusion that the epidemic is much farther along than it really is, and thus giving wildly inaccurate estimates of logistic parameters even though a close curve fitting. Unfortunately, the data that would be necessary to build a reasonable (though much more complicated) model is not available.

On 2020-05-08 06:17:03, user Yoshihiro Ishida wrote:

I read this manuscript with great interest. I see that K value may potentially be a good measure for assessing the efficacy of social interventions. However, I doubt how useful it is for making predictions.

My concerns are:<br /> 1. The authors seem to pick an arbitrary period for each country to fit the model. The model may fit nicely retrospectively, but it may merely be overfitting without much generalizability or predictive value.<br /> 2. In fact, the parameter does not seem to predict explosive increase in Russia past mid-April.

On 2020-05-09 15:50:32, user Everett wrote:

Hey featheryHen,<br /> here is the link you are asking for:<br /> https://www.medrxiv.org/con...

On 2020-05-15 23:52:54, user Milesy Mathis wrote:

At least they included Zinc. That's progress.

On 2020-07-06 18:41:52, user Fatnot wrote:

Unlikely that just zinc supplementation would work,,,a zinc ionophore is also required.. We also have the report from Dr Zelensky, in Rockland County, NY, who treated hundreds with<br /> a combo including zinc and HCQ, resulting that few required hospitalization. The report is anecdotal...but another term for a set of anecdotes is of course DATA And with data and analysis, one can draw conclusions and confidence intervals.

On 2020-05-17 07:28:24, user Robert Ray wrote:

Would this suggest that plasma transplants from recovered patients might supply the responsive T-cells that are missing in severe disease patients?

On 2020-04-05 01:22:40, user stephen zhang wrote:

See its associated journal publication in Psychiatry Research:

https://www.sciencedirect.c...

On 2020-04-05 13:17:57, user Robert Nachbar wrote:

According to the Australian Government Department of Health, more than half the cases of COVID-19 in Austalia have been imported (https://www.health.gov.au/n... "https://www.health.gov.au/news/health-alerts/novel-coronavirus-2019-ncov-health-alert/coronavirus-covid-19-current-situation-and-case-numbers)"). There are no details in the manuscript regarding this significant part of the model, other than they were a source term to the I_2 compartment, making it impossible to reproduce the results. Furthermore, the data in Figure 2 shows substantially fewer imported cases that reported by the Department of Health.

On 2020-04-06 11:13:41, user japhetk wrote:

I think this is another voodoo correlation study of BCG which keeps appearing one after another. BCG may be effective or not effective, but that cannot be revealed by country analyses due to many uncontrolled and complex factors.

The problems of this study's analyses, they are not controlling when the infection spread in the country properly.

Other analyses are controlling that (for example, number of patients (or deaths) 10 days after the 100th patients were detected, was used as a dependent measure). In this study, only the 3 categorical classification is used, which is apparently not appropriate nor objective.

Also, probably, the most accurate available BCG measure is "how long the country has advanced the BCG vaccination measure" (the year when the country stopped the BCG vaccination (or now, when the BCG vaccination is currently conducted in the country) - the year when the country started it). UK for example, advanced BCG for more than 50 years, so the majority of the have been experienced with BCG. So, the "current" "past" "never" classification is not appropriate.

I controlled these measures and have done the analyses.

The results were as follows.<br /> The partial correlation between "how long the country has advanced the BCG vaccination measure" and number of patients in the 10th day (when 1st day is 100th patients were detected in the country) after controlling the population of the country. P = 0.455, partial correlation coefficient = -0.116).

The partial correlation between "how long the country has advanced the BCG vaccination measure" and number of deaths in the 10th day (when 1st day is the 100th patients were detected in the country) after controlling the population of the country. P = 0.111, partial correlation coefficient = -0.243).

But the partial correlation between "how long the country has advanced the BCG vaccination measure" and when the 100th patients were detected in the country after controlling the population of the country was P = 0.078, partial correlation coefficient = 0.281.

Also "how long the country has advanced the BCG vaccination measure" is robustly and negatively correlated with GDP of the country after controlling the population of the country (p = 0.019, partial correlation coefficient:-0.292).

But the correlation between GDP of the country and when the 100th patients were detected in the country after controlling population of the country was more robust (p = 0.001, partial correlation coefficient: -0.438)(the greater the GDP, the faster the 100th patients appeared).

And the correlation between "how long the country has advanced the BCG vaccination measure" and when the 100th patients were detected in the country disappeared when the population and GDP is also controlled (p = 0.322).<br /> Most of correlations are tendencies levels and disappeared or substantially got weaker after GDP is controlled.

So, my guess is probably, there are number of spurious correlations going on authors' analyses due to lack of important control variables.

The higher GDP countries can do more tests, they are more popular from the <br /> tourists from Asia, but they were less inclined to use masks as far as I heard and less alert.

The list of potential confounding variables, as far as I can see is endless and diverse.<br /> Such as numerous nutritional components, the foods of western rich countries include.<br /> The temperature. <br /> How many Chinese (ratio) is included in the nation (apparently Taiwan, China (except Wuhan) and other countries where Chinese consist most of nations were as alert as possible). <br /> Preference of wearing masks.<br /> Popularity of religion. (Religious ceremonies are cited as the popular source of the cluster of infections).<br /> The strength of individual rights and freedoms.<br /> The factors related to Western rich countries.

In the case of Diamond princess, Japanese was apparently not invincible.<br /> About 300 patients were Japanese and 180 patients were from Western rich countries in the case of this ship, and among them, 7 Japanese died and 2 from Canada and UK died. Japanese were not biological invincible from this coronavirus, please stop spreading this rumor through country based correlation studies… Just please wait for RCT studies.

On 2020-04-06 18:06:44, user Ziv Gan-Or wrote:

I am sorry to say, but the title is somewhat misleading, the authors did not show that "ACE2 variants underlie interindividual variability and susceptibility to COVID-19 in Italian population". What the authors show is that there are ACE2 variants found in Italians and not in Asians, and that in-silico tools predict that they might have structural or functional effects. I am certain that there are variants in Asians that are not found in Italians and are with similar predictions. We need population-based genetic studies to determine what the authors suggest in the title. Nevertheless, this is interesting and worthy of additional studies.

On 2020-04-06 19:59:09, user Virginia Savova wrote:

Are you planning to release the count matrices to crowdsource analytics and speed up the impact of this data on drug discovery? The time is now.

On 2020-04-07 18:39:42, user Snap wrote:

Diffusion and mortality of COVID-19 in regions with poor air quality

The COVID-19 spread from China to the rest of the world in just over three months has turned into a pandemic that poses several humanitarian and scientific challenges. By comparing air quality levels with the spread and mortality of the virus, a significant correlation was found in China, Italy and the United States. People seem to become infected and die more often in those areas affected with poor air quality levels. Although the infection is still ongoing globally, these results are convincing because they are not influenced by different population densities. Similar to smoking, people living in polluted areas appear to have a respiratory system more vulnerable to this new form of coronavirus. This suggests the detrimental impact of climate change as a novel supplemental risk factor and prompts the need of its abatement.<br /> https://uploads.disquscdn.c...

On 2020-04-10 07:07:16, user Red Sage wrote:

One explanation re New York's higher stats is lack of health insurance leads to delays seeking medical assistance when illness gets serious

On 2020-04-11 00:34:41, user Tomas Hull wrote:

How do the deaths of almost 80 doctors age 57 to mid 60 in the hotbeds of COVID-19 in Italy fit into the computation with the probabilities of ...the COVID-19 death risk in people <65 years old during the period of fatalities from the epidemic was equivalent to the death risk from driving between 9 miles per day (Germany) and 415 miles per day (New York City)"?<br /> Have 80 physicians in Italy run the red light, statistically speaking?

On 2020-04-15 03:25:26, user Realist50 wrote:

On top of ideas for society - i.e., that most places should follow a path more like Sweden than full lockdowns - doesn't this strongly argue for Human Challenge trials among lower-risk groups to speed vaccine development?

As some background, my understanding is that Human Challenge Trials speed tests of the efficacy of vaccines by intentionally exposing vaccinated individuals to the virus. Otherwise, people are vaccinated and essentially told to go on with life, with vaccine efficacy only observed by comparing rates of infection to a control group over time.

The idea has already been proposed for COVID-19, but AFAIK it hasn't yet been implemented anywhere. Human Challenge Trials aren't a new idea: they've historically been used for diseases where the risk from infection is viewed as manageable. This analysis therefore supports the idea that Human Challenge Trials are ethical for those who are relatively young without obvious risk factors.

On 2020-04-11 00:52:38, user Moi wrote:

Reality:<br /> In Austria, recently an excess of 2.35 has been measured (8,500 officially infected vs. 28,500 immunized overall for A), see SORA-Prävalenzstudie.

vs.

Model:<br /> Here, we have to assume that 75,000 officially infected (JHU) are the tip of a 33 million iceberg of immunized ...That would be a factor of 440 instead of 2.35. Hard to believe, for me at least.

We need actual and reliable tests, and more tests, worldwide.

On 2020-04-11 05:44:42, user Serge wrote:

Please be aware and advised that the situation is still developing in many jurisdictions and <br /> the information in the paper represents only a snapshot in time. Only <br /> following the developments over certain period can provide more <br /> confident background for further statistical studies.

On 2020-04-11 06:37:47, user Marema wrote:

The paper discusses the money and medical education.

On 2020-04-12 07:46:37, user German Perez Vazquez wrote:

I don't see patients characteristics description, nor enough methodology explanation on how to ensure patients have enough follow-up to make sure outcome death/alive is reliable. I also miss comparisons with already known ICU scores. Interesting work anyway.

On 2020-04-12 15:01:25, user Xavier de Roquemaurel wrote:

Wouldn’t a multi factorial analysis put in perspective the question? Testing together these factors : confinement, social distancing, masks, bcg vaccinations (strain by strain), number of intensive care beds per 000, etc... It could avoid a frontal confrontation and at the same time open the discussion.

On 2020-04-13 09:54:09, user FabioMassimoFabrizio wrote:

The study of France was controverified by france INSERM: "No Evidence of Rapid Antiviral Clearance or Clinical Benefit with theCombination of Hydroxychloroquine and Azithromycin in Patients withSevere COVID-19 Infection": https://www.sciencedirect.c...

On 2020-04-14 21:47:58, user Dr Eric Grossi Neurocirurgia wrote:

I would like to highlight a serious methodological error in this study. What we want for a drug treatment of COVID-19, only two objectives, to avoid and / or treat SARS and reduce contagion, therefore pragmatism in the selection of patients must be as close as possible to the clinical reality, which did NOT occur, since only patients between the ages of years were analyzed. This alone invalidates any useful result, since the vast majority of human losses are over the age of 64.

https://uploads.disquscdn.c... <br /> (Registry in ChinaClinicalTrialRegistry site) - look the date of approved by ethic commitee 5 day after the begun of study, and look too the final date - isn't a preview release NO NO , the study is OVER

On 2020-04-15 11:13:02, user Dr Eric Grossi Neurocirurgia wrote:

Here I post the screenshot of Ch CTR data to this study, registry with number : ChiCTR2000029559, everyone can see ! https://uploads.disquscdn.c...

On 2020-04-16 11:58:45, user Stef Verlinden wrote:

When will this article be published in a peer-reviewed journal? I think this, so far, is the only RCT study that supports the hypothesis that HCQ can prevent exacerbation of disease in COVID-19 patients with CT confirmed mild pneumonia. If correct, its importance can not be underestimated. In that case, early treatment of high-risk patients with HCQ could lead us the way to a faster exit out of the corona crisis...

On 2020-04-13 13:39:56, user Rosemary TATE wrote:

Hi, I dont see the STROBE guidelines checklist (for observational studies) uploaded, although you ticked yes to this<br /> "I have followed all appropriate research reporting guidelines and uploaded the relevant EQUATOR Network research reporting checklist(s) and other pertinent material as supplementary files, if applicable. "<br /> A lot of people seem to ignore these but they are important and any good journal will require them.<br /> Can you please upload? Many thanks.

On 2020-04-14 15:45:03, user Euclides Hernández wrote:

In Costa Rica, BCG is one of the compulsory vaccines for the entire population and is applied by the social security system, it is universal. CR. As of Tuesday, April 14, it has 612 cases, 62 people recovered and 3 deaths. Costa Rica has a population size similar to that of Ireland. Costa Rica's social security system is one of the strongest in Latin America.

En Costa Rica la BCG es una de las vacunas obligatorias para toda la

población y es aplicada por el sistema de seguridad social, es

universal. CR. tiene, al martes 14 de abril, 612 casos, 62 personas

recuperadas y 3 fallecimientos. Costa Rica tiene una cantidad de

población similar a la de Irlanda. El sistema de seguridad social de

Costa Rica es uno de los más sólidos de América Latina.

On 2020-04-16 07:46:42, user Hamdi Torun wrote:

https://uploads.disquscdn.c... The dataset in the papers is now outdated. If the authors use the new dataset in terms of the metrics they chose to use, the correlations will be poor. This graph shows an updated version of their Fig. 1 as of April 16th. More importantly, the reliability of the released data by the governments should be questioned. Even if the datasets are reliable, correlation does not imply causality.

On 2020-04-07 21:56:51, user VesnaV wrote:

The idea of the article is very interesting. But I am afraid that the trends are changing. It would be very useful to update the data on covid-19 and replicate the analysis. Could you please do it? Thanks a lot!

On 2020-04-16 03:43:06, user Rhodes wrote:

Why 600 mg dose? (Compare Sermo international barometer showing most doctors prescribing 400 mg) Why no supplements? (zinc, vitamin c, d, copper).

On 2020-04-16 07:46:21, user Nuria Calvo wrote:

Where is the information about the equations of the model?

On 2020-04-16 20:15:01, user alg3257 wrote:

Will an N-95 mask be safe if it is used once and then left to sit for a week then used again?

On 2020-04-16 23:50:15, user Brian T wrote:

MISLEADING UV DOSE! DO NOT FOLLOW THIS 5 uW/cm2 dosing!

The UVA/B meter used in this study “General UVAB 137 digital light meter (General Tools and Instruments New York, NY)” (look up General Tools UV513AB Digital UVA/UVB Meter on Amazon) advertises measurement from 280-400nm, barely overlapping the wavelengths used in this study, 260 – 285 nm. Furthermore, this meter does not report on the energy at a given wavelength. Its possible that this study is grossly underreporting the dose of UV needed because their meter doesn’t read many of the wavelegths used.

Previous research on SARS and MERS used 254 nm and noted much, much higher energy needed to kill these corona viruses vs this study's reported 5 uW/cm2):

• MERS is inactivated at 90 uW/cm2 x 60 mins[1], dose of 0.324 J/cm2<br /> • SARS is inactivated at 4016 uW/cm2 x 6 min[2] , dose of 1.446 J/cm2

[1] http://www.diniesturkiye.co...

[2] https://www.sciencedirect.c...

On 2020-04-18 12:30:31, user Jack Prior wrote:

Is it possible that people are seeking treatment for flu-like symptoms at a higher rate than normal due to anxiety over consequences of severe Covid-19 infections?

On 2020-04-22 23:37:10, user Glenn Korbel wrote:

In the absence of tests for antibodies, which they don't have they are simply guessing/Yhere is NO way to predict deaths without knowing how many people have already been infected.<br /> None.

On 2020-07-16 21:40:47, user Marm Kilpatrick wrote:

Very nice study.<br /> Did you measure viral loads in patients? If so, would it be possible to include those to see if they might be implicated in 4 cases of infection? Sample size and power would be low, but it would be useful to at least take a look.<br /> It also wasn't clear if some HCWs had to engage in riskier activities (e.g. intubation) and this might have led to infection. Thank you!

On 2020-06-23 17:39:31, user Liam Golding wrote:

Hi great work by your team.

I'm curious whether you standardize the log inactivation to untreated masks or to viral stock added. You note that for bacterial contaminants that untreated coupons are compared to treated to obtain log reduction values. But, for example, you note that "For each decontamination method, each sample used for treatment had a corresponding no-treatment control. No-virus blank masks were also included to identify possible contamination." Was the control viral load extracted then compared to treatments to obtain log reduction values, or was a known quantity of viral load added to controls and used to determine the log reduction?

**Edit: you draw mention to this in the Material and Methods.

However, compared to other studies (Mills et al., 2018; Lore et al., 2011) your method of extracting viral load is minimal to say the least. Generally, coupons are cut; placed in a 15/50ml tube with ~ 15ml extraction solution then vortexed/mixed for 20 minutes. Can you comment on why you chose 1 minute vortex with 1.3mL solution over the common OP?

On 2020-06-24 18:16:30, user Gerard Cangelosi wrote:

Nice study, and a very valuable addition. I collaborated on one of the previous studies you cited (Tu YP et al, 2020). May I suggest an alternative explanation for the difference between your findings and ours? You used all-purpose flock swabs, and we used foam swabs. These differences aren't trivial (e.g. see https://www.medrxiv.org/con... "https://www.medrxiv.org/content/10.1101/2020.04.28.20083055v1)"). I would urge you to note this possibility in your manuscript. Thank you!<br /> Jerry

On 2020-06-24 18:56:17, user André GILLIBERT wrote:

Title : Proposal for improved reporting of the Recovery trial<br /> André GILLIBERT (M.D.)1, Florian NAUDET (M.D., P.H.D.)2<br /> 1 Department of Biostatistics, CHU Rouen, F 76000, Rouen, France<br /> 2 Univ Rennes, CHU Rennes, Inserm, CIC 1414 (Centre d’Investigation Clinique de Rennes), F- 35000 Rennes, France

**Introduction**

Dear authors,<br /> We read with interest the pre-print of the article entitled “Effect of Dexamethasone in Hospitalized Patients with COVID-19: Preliminary Report”. This reports the preliminary results of a large scale randomized clinical trial (RCT) conducted in 176 hospitals in the United Kingdom. To our knowledge it is the largest scale pragmatic RCT comparing treatments of the COVID-19 in curative intent. The 28-days survival endpoint is objective, clinically relevant and should not be influenced by the measurement bias that may be caused by the open-label design. While 2,315 study protocols have been registered on ClinicalTrials.gov about COVID-19, as of June 24th 2020, Recovery is, to our knowledge, the only randomized clinical trial on COVID-19 that succeeded to include more than ten thousands patients. The open-label design and simple electronic case report form (e-CRF) may have helped to include a non-negligible proportion of all COVID-19 patients hospitalized in the United Kingdom (UK). Indeed, as of June 24th 2020, approximatively 43,000 patients died of COVID-19 in hospital in the UK, of whom approximatively 0.24 × 11,500 = 2,760, that is more than 6% of all hospital deaths of COVID-19, where included in the Recovery study.<br /> Having read with interest version 6.0 of the publicly available study protocol (https://www.recoverytrial.n... "https://www.recoverytrial.net/files/recovery-protocol-v6-0-2020-05-14.pdf)") we had hoped for more details in the reporting of methods and results of this trial and take advantage of the open-peer review process offered by pre-prints servers to suggest improving some aspects of the reporting before the final peer-reviewed publication. Please, find below some easy to answer comments that may help to improve the article overall.

**Interim analyses and multiple treatment arms**

The first information would be about interim analyses. The protocol (version 6.0) specifies that it is adaptive and that randomization arms may be added removed or paused according to decisions of the Trial Steering Committee (TSC) basing its decision on interim analyses performed by the Data Monitoring Committee (DMC) and communicated when “the randomised comparisons in the study have provided evidence on mortality that is strong enough […] to affect national and global treatment strategies” (protocol, page 16, section 4.4, 2nd paragraph). The Supplementary Materials of the manuscript specifies that “the independent Data Monitoring Committee reviews unblinded analyses of the study data and any other information considered relevant at intervals of around 2 weeks”. This suggests that many interim analyses may have been performed from the start (March 9th) to the end (June 8th) of the study.<br /> Statistically, interim analyses not properly taken in account generate an inflation of the type I error rate which may be increased again by the multiple treatment arms. Methods such as triangular tests make it possible to control the type I error rate. Most methods of control of type I error rate in interim analyses require that the maximal sample size be defined a priori and that the timing and number of interim analyses be pre-planned. This protocol being adaptive, new arms were added, implying new statistical tests in interim analyses, and no pre-defined sample size as seen in page 2 of the protocol: “[...] it may be possible to randomise several thousand with mild disease [...], but realistic, appropriate sample sizes could not be estimated at the start of the trial.” This make control of the type I error rate difficult. The fact that the study has been stopped on the final analysis as we understand from the current draft rather than interim analysis does not remove the type I error rate inflation. The multiple treatment arms lead to another inflation of the type I error rate.<br /> The current manuscript does not specify any procedure to fix these problems. The Statistical Analysis Plans (SAP) V1.0 (in section 5.5) and V1.1 (in section 5.6) specify that “Evaluation of the primary trial (main randomisation) and secondary randomisation will be conducted independently and no adjustment be made for these. Formal adjustment will not be made for multiple treatment comparisons, the testing of secondary and subsidiary outcomes, or subgroup analyses.” and nothing is specified about interim analysis. Therefore, we conclude that no P-value adjustment for multiple testing has been performed, neither for multiple treatment arms nor for interim analysis. If an interim analysis assessing 4 to 6 treatment arms at the 5% significance level has been performed every 2 weeks from march to June, up to 50 tests may have been performed, leading to major inflation of type I error rate. In our opinion, the best way to assess and maybe fix the type I error rate inflation, is to report with maximal transparency every interim analysis that has been performed, with the following information:<br /> 1. Date of the interim analysis and number of patients included at that stage<br /> 2. Was the interim analysis planned (e.g. every 2 weeks as planned according to supplementary material) or unplanned (e.g. due to an external event, for instance the article of Mehra et al about hydroxychloroquine published in The Lancet, doi:10.1016/S0140-6736(20)31180-6), and if exceptional, why?<br /> 3. Which statistical analyzes, on which randomization arms, have been performed at each stage <br /> 4. If predefined, what criteria (statistical or not) would have conducted to early arrest of a randomization arm for inefficiency and what criteria would have conducted to arrest for proved efficacy?<br /> 5. If statistical criteria were not predefined, did the DMC provide a rationale for his choice to communicate or not the results to the TSC? If yes, could the rationale be provided?<br /> 6. The results of statistical analyzes performed at each step<br /> 7. The decision of the DMC to communicate or not the results to the TSC and which results have been reported as the case may be<br /> The information about interim analyses and multiple randomization arms will help to assess whether the inflation of type I error rate is severe or not. A post hoc multiple testing adjustment, taking in account the many randomized treatments and interim analyses, should be attempted, and discussed, even though there may be technical issues due to the adaptative nature of the protocol.

**Adjustment for age**

An adjustment for age (in three categories <70 years, 70-79, >= 80 years, see legend of table S2) in a Cox model was performed for the comparison of dexamethasone to standard of care in the article. This adjustment was not specified in the version 6.0 of the protocol but was, according to the manuscript “added once the imbalance in age (a key prognostic factor) became apparent”. This is confirmed by the addition of a words ““However, in the event that there are any important imbalances between the randomised groups in key baseline subgroups (see section 5.4), emphasis will be placed on analyses that are adjusted for the relevant baseline characteristic(s).” in section 5.5 page 16 of the SAP V1.1 of June 20th compared to the SAP V1.0 of June 9th which specified a log-rank test. The SAP V1.0 of the 9th June may have been written before the database has been analyzed (data cut June 10th) but the SAP of the 20th has probably been written after preliminary analysis have been performed. This is consistent with the words “became apparent” of the manuscript. Therefore, in our opinion, this adjustment must be considered as a post hoc analysis rather than as the main analysis. Moreover, even though the SAP V1.1 specifies that an “important imbalance” will lead to an “emphasis” on adjusted analyses, it does not change the primary analysis (see section 5.1.1 page 14). It is not clear what “important imbalance” means. To interpret that, we will perform statistical tests to assess balance of key baseline subgroups specified in SAP V1.1 (see section 5.4):<br /> 1. Risk group (three risk groups with approximately equal number of deaths based on factors recorded at randomisation). Its distribution is shown in figure S2. A chi-square tests on the distribution of risk groups in Dexamethasone 1255/500/349 and Usual care 2680/926/715 groups, lead to a P-value=0.092. A chi-square test for trend yields a P-value equal to 0.23.<br /> 2. Requirement for respiratory support at randomisation (None; Oxygen only; Ventilation or ECMO). P-value=0.89 for chi-square test and P-value=0.86 for chi-square for trend.<br /> 3. Time since illness onset (<=7 days; >7 days). P-value=0.17<br /> 4. Age (<70; 70-79; 80+ years). P-value=0.016 for chi-square test, p=0.019 for chi-square test for trend<br /> 5. Sex (Male; Female). P-value=0.97 for chi-square test<br /> 6. Ethnicity (White; Black, Asian or Minority Ethnic). No data found.<br /> The criteria to define “important imbalance” seems to be statistical significance at the 0.05 threshold, however that should have been stated and tests for all other variables should have been provided too.<br /> First, this adjustment, from a theoretical point-of-view, was not necessary since the study was randomized; if the exact condition of imbalance triggering the adjustment was pre-specified in the protocol or SAP before the imbalance was known, it could induce a very slight reduction of the type I error rate and power. However, as it was performed when the imbalance was known, there is a risk that the sign of the imbalance (i.e. higher age in the dexamethasone group) have influenced the choice of adjustment. Indeed, an adjustment conditional to a higher age in the dexamethasone group will increase the estimated effect of dexamethasone in these conditions, and so, provide an inflation of the type I error rate. If the same conditional adjustment were further considered for other prognostic variables, the inflation could even be higher. <br /> Unless there is strong evidence that the amendment to the SAP was performed without knowledge of the sign of the imbalance (higher age in the dexamethasone group), we suggest that the primary analysis be kept as originally planned, without adjustment, and that the age adjustment be performed in a sensitivity analysis only. The knowledge of the sign of the unbalance is unclear in the last version of the SAP (V1.1, June 20th) and in the manuscript. In addition, in an open label trial, it is always better to stick to the protocol.

**Results in other treatment arms**

The manuscript specifies that “the Steering Committee closed recruitment to the dexamethasone arm since enrolment exceeded 2000 patients.” It is not stated whether any other treatment arm has exceeded 2000 patients or not and whether the study is still ongoing. Results of treatment arms that have been stopped should be provided (all arms having enrolled more than 2000 patients?). If not, the number of patients randomized in other treatment arms should, at least, be reported. If the study is completely stopped, all treatments should be analyzed and reported, unless there is a specific reason not to do so; that reason should be stated as the case may be. This data would be useful to provide evidence on other molecules. It would also clarify the number of statistical tests that have been performed or not, providing more information about the overall inflation of alpha risk.

**Sample size**

The paragraph about the sample size suggests that inclusions were planned, at some time, to stop when 2000 patients were included in the dexamethasone arm. The amended protocol (May 14th), the SAP V1.0 (June 9th) and the SAP V1.1 (June 20th, 4 days after the results have been officially announced) all have a paragraph about the sample size but all specify that the sample size is not fixed and none specify any criteria of arrest of the research based on sample size. There are 2104 patients included in this arm, which is substantially larger than the target of 2000 patients. The exact chronology and methodology should be clarified: when was the sample size computed and what was the exact criteria to arrest the research? Could the document (internal report?) related to this sample size calculation and statistical or non-statistical decision of arrest of the research be published in supplementary material?<br /> Indeed, assessment of the type I error rate requires knowing exactly when and why the research has been arrested: arrest for low inclusion rate of new patients or for reaching target sample size cannot be interpreted the same as arrest for high efficacy observed on an interim analysis.

**Future of the protocol**

With the new evidence about dexamethasone, the protocol will probably be stopped or evolve. The future recruitment may slow as the peak of the epidemic curve in United Kingdom is passed. The past, present and future of the protocol needs also to be known to assess the actual type I error rate. Indeed, future analyses, that have not yet been performed influence the overall type I error rate. That is why we suggest that author’s provide the daily or weekly inclusion rate from March to June and discuss the future of the study.

**Loss to follow-up**

Table S1 shows that the follow-up forms have been received for 1940/2104 (92.2%) patients of the dexamethasone group and 3973/4321 patients of the usual care group (91.9%). The patients without follow-up forms (8.5% overall) may either be lost to follow-up or have been included in the 28 last days before June 10th 2020 (data cut). The manuscript mentions that 4.8% of patients “had not been followed for 28 days by the time of the data cut”, suggesting that 8.5%-4.8% = 3.7% of patients are lost to follow-up, but that is our own interpretation. We suggest that authors report the actual number of loss to follow-up and how their data have been imputed or analyzed. The number of loss to follow-up may differ for different outcomes. For instance, if the Office of National Statistics (ONS) data has been used for vital status assessment, there should be no loss to follow-up on that outcome.

**Vital status**

The current manuscript only specifies the data of the web-based case report (e-CRF) form, filled by hospital staff, as source of information, suggesting that it is the only source of information about the vital status. The document entitled “Definition and Derivation of Baseline Characteristics and Outcomes” provided at https://www.recoverytrial.n... specifies many other sources. For instance, the vital status had to be assessed from the Office of National Statistics (ONS). Other sources, including Secondary Use Service Admitted Patient Care (SUSAPC) and e-CRF could be used for interim analysis. The ONS was considered as the defining source (most reliable). Whether the ONS data has been used or not should be clarified. If the ONS data have been used, statistics of agreement of the two data sources (e-CRF and ONS) may be provided to help assessing the quality of data. If the ONS data have not been used, this deviation from the planned protocol should be documented.<br /> The manuscript as well as the recovery-outcomes-definitions-v1-0.pdf file specifies that the follow-up form of the e-CRF is completed at “the earliest of (i) discharge from acute care (ii) death, or (iii) 28 days after the main randomisation”. If the follow-up form is not updated further, patients discharged alive before day 28 (e.g. day 14) may have incomplete vital status information at day 28. The following information should be specified:<br /> 1. Whether the follow-up form of the e-CRF had to be updated by hospital staff at day 28 for these patients<br /> 2. If response to (1) is yes, whether there was a means to distinguish between a lost to follow-up at day 28 (form not updated) and a patient discharged and alive at day 28 (form updated to “alive at day 28”)<br /> 3. If response to (2) is yes, how many patients discharged before day 28 were lost to follow-up at day 28<br /> 4. If response to (2) is yes, how has their vital status at day 28 been imputed or managed in models with censorships (log-rank, Kaplan-Meier, Cox)<br /> Of course, this information is really needed if the ONS and SUSAPC data have not been used.<br /> The quality of the vital status information is critical in such a large scale open-label multi-centric trial, because there is a risk that one or more center selectively report death, biasing the primary analysis.

**Inclusion distribution by center**

A multicentric study provides stronger evidence than a single-center study but sometimes, few centers include most patients, with a risk of low-quality data or selection bias. The very high number of included patients in the Recovery trial suggests that many centers included many patients but the distribution of inclusions per center could be reported.

**Randomization**

The protocol specifies that “in some hospitals, not all treatment arms will be available (e.g. due to manufacturing and supply shortages); and at some times, not all treatment arms will be active (e.g. due to lack of relevant approvals and contractual agreements).” This is further clarified in the SAP V1 (section 2.4.2 Exclusion criteria, page 8) by the sentence “If one or more of the active drug treatments is not available at the hospital or is believed, by the attending clinician, to be contraindicated (or definitely indicated) for the specific patient, then this fact will be recorded via the web-based form prior to randomisation; random allocation will then be between the remaining (or indicated) arms.” Showing that randomization arms may be closed on an individual basis, when the patient is included, with the argument of contraindication or definitive indication. It seems that the “standard of care” group could not be removed and that at least another randomization arm had to be kept as suggested by the words “random allocation will then be between the remaining arms (in a 2:1:1:1, 2:1:1 or 2:1 ratio)” in section 2.9.1 page 11 of the SAP V1.0. Even exclusion of a single randomization arm can lead to imbalance between groups. For instance, if physicians believed that a treatment was contraindicated for the most severe patients, only non-severe patients could be randomized to the treatment’s arm, while most severe patients would be randomized to other arms. Several things can be done to assess and fix this bias. First, report how many times this feature has been used and which randomization arms have been most excluded. If it has been used many times, provide the pattern of use that help to assess whether this is a collective measure (e.g. 2-weeks period of shortage of a treatment in a center ? no major selection bias) or individual measure. If its use has been rare, a sensitivity analysis could simply exclude these patients. If it has been frequent, we suggest a statistical method to analyze this data without bias, based on the following principles: patients randomized between 3 randomization arms A, B and C (population X) are comparable for the comparisons of A to B. Patients randomized between A, B and D (population Y), are comparable for the comparisons of A to B. Population X and population Y may differ but, inside each population, A can be compared to B. Therefore, the within-X comparison of A to B and within-Y comparison of A to B are both valid and can be meta-analyzed to assess a global difference between A and B. This can be simply done with an adjustment on the population (X or Y) in a fixed effects multivariate model. Pooling of X and Y populations should not be performed without adjustment.<br /> A second problem with randomization exists although the dexamethasone arm is the least affected. Randomization arms have been added in this adaptative trial. When a new randomization arm is added, new patients may be randomized to this arm and fewer patients are randomized to other arms. Consequently, the distribution of dates of inclusion may differ between groups. This may have some impact on the mortality at two levels: (1) the medical prescription of hospitalization may have evolved as the epidemic evolved, with hospitalization reserved to most severe patients at the peak of epidemic and maybe wider hospitalization criteria at the start of epidemic and (2) evolution of patients included in the Recovery trial. Indeed, even if centers should have included as many patients as possible as soon as their inclusion criteria were met, it is possible that they have only included part of eligible patients and that this part evolved with time. This bias can be easily assessed and fixed: the curves of inclusions in the different arms and mortality rate in the Recovery trial can be drawn as a function of date (from March to June) and an adjustment on date of inclusion may be performed in a sensitivity analysis.

**Conclusion**

Recovery is the study with the best methodology that we have seen on COVID-19 treatments in curative intent and we salute the initiative of publishing transparently the protocol, its amendments, the statistical analysis plan and the first draft of the report. We hope that our reporting suggestions will be taken in account in the final version of the paper. We think that discussing these points will qualify the interpretation of results, further improve the transparent approach adopted by designers of the study and improve the reliability of the conclusions. We expect a high-quality reporting of these final results, with full transparency on interim analyses, statistical analysis plans and statistical analysis reports. We hope that these comments are helpful and again we acknowledge that this study is not solely outstanding in terms of importance of the results but is also a stellar example for the whole field of therapeutic research. We invite other researchers to provide comments to this article to engage in Open Science.

On 2020-06-24 05:32:27, user Gavin Donaldson wrote:

Were patients ineligible to the dexamethasone arm excluded completely from the recovery trial or allocated to the other arms including the standard care arm?<br /> Could the reasons for the exclusions be included in a consort diagram of the participant flow.<br /> There was an imbalance between the two arms in age. Is there any data on obesity or hypertension since these are important risk factors for covid-19 mortality.

On 2020-06-29 17:14:09, user Aiman Tulaimat wrote:

The study reports a mortality ~40% in patients on vent on the control arm. This is much lower than what is reported by the critical care audit from the UK, which reports mortality > 60% in such patients. The study reports ~25% mortality and ~60% discharge alive. Are we missing 15% of patients? If the analysis is a Cox hazard, why is the report using relative risk? How did 13% of patients with no oxygen therapy die? This is very high? Did their covid deteriorate or did they die from other reason? why were these patients hospitalized if they were not hypoxic? did they decline life support when it became needed? where they not on oxygen because they were in hospice like setting? how did the other patients die? Were patients on the ventilator made DNR early? Was prone position used? Was it used more in the dexa arm? Was there imbalance in the ramdomization by center?

On 2020-07-10 18:25:07, user Joanna Spencer-Segal wrote:

The authors speculate in the discussion that "It is also possible there is an effect via mineralocorticoid receptor binding in the context of SARS-CoV-2 induced dysregulation of the renin-angiotensin system." It is not clear what this means, but dexamethasone has minimal activity at the mineralocorticoid receptor, which distinguishes it from the other corticosteroids often used in critically ill patients (methylprednisolone, hydrocortisone). More clarification of what they mean regarding "mineralocorticoid effect" and rationale about why dexamethasone was chosen for this study would be welcome.

On 2020-06-25 04:04:22, user Greg WHITTEN wrote:

Thank you for your work. I am curious, however, about some parts of your article.

First, I read your paper and could not see where you tried to control for the introduction of other virus-containment measures such as school closures, lock-downs, and physical distancing. Did I miss something in your paper?

Second, I have a question about your model #4 on page 9. You wrote "All<br /> regression coefficients were statistically significant in this model." The coefficient for the non-mask wearing rate in late April and early May is significant but negative. I.e., not wearing a mask in late April and early may reduces deaths on May 13th. Do you have any thoughts about this?

Third, did you consider performing a panel regression using deaths on all days, say, starting from March 31st (about 2 weeks after the March mask non-wearing rate) instead of relying just on deaths from May 13? Although you did explain why you chose May 13th, it may be better to use all death dates after, say, the incubation period for the virus.

Fourth, your section "Prediction of mask non-wearing rates" suggests that your regression analysis suffers from multicollinearity. Do you have any concerns about this?

On 2020-06-25 11:29:20, user MAGB wrote:

Your basic reproductive number of 2.68 based on early Chinese data is at odds with the effective reproduction number of less than one in all Australian states by Easter, as tweeted by James McCaw. His data indicate that voluntary controls and border closures had the epidemic well under control before lock-downs had any effect.

On 2020-06-26 16:11:55, user Kevin McKernan wrote:

Very interesting work. There are many data points suggestion earlier introduction.<br /> It would strengthen the manuscript if you could provide the CQs for your NTC for the late PCRs.

I would also suggest sequencing the amplicons to see if these are ancestral or SARS-CoV-2.

https://youtu.be/guMqyrmni4I

https://youtu.be/2SfXZa7-JKM

On 2020-04-08 22:31:44, user Mansour Tobaiqy wrote:

I am glad to say that our manuscript Therapeutic Management of COVID-19 Patients: A systematic review has now accepted for publication at the Infection Prevention in Practice @IPIP_Open the Official Journal of the Healthcare Infection Society @HIS_infection

The last version will be available soon at their site. Thank you very much medRxiv for sharing our SR to a great and large audience .

On 2020-04-10 22:02:12, user Todd Johnson wrote:

Have any of the causal inference researchers at Harvard taken a look at this? Do we know enough to create a few candidate causal DAGs to know what to adjust for?

On 2020-07-01 14:02:04, user Dude Dujmovic wrote:

I don't believe this research has much in it. I think there is a richer social context for people vaccinated for Flu and that social context makes them less susceptible to COVID-19. For example if person lives in society where standards of care are higher then that person will have a longer lifetime and will also be more likely vaccinated against various diseases. You only accounted for education and that is not enough. But your research data does show connection between education and risk of death in COVID-19.

On 2020-08-15 14:01:48, user Dom_Pedulla wrote:

Joao not only had Dude made some very good points, but in observational trials like this, everything depends on the nitty gritty data. I notice the huge qualifier "recent" in the results sentence, noting that carefully since in many studies these kinds of adjectives disclose or hint at certain erroneous tendencies or conclusions in even in "meticulous peer-reviewed studies". I am requesting the paper to analyze for myself, and suspect strongly that what it may show is a "benefit" for only the very recently vaccinated, and that either long after it either ends up being a net liability as regards COVID death risk, or that the timing isn't possible to discern because the investigators avoided studying all but the recently vaccinated.

We'll see.

On 2020-07-02 15:42:32, user Kamran Kadkhoda wrote:

The entirety of covid serology remains questionable with lack of clinical usefulness; the specimen type therefore is irrelevant...

On 2020-07-02 18:00:55, user Wouter wrote:

Same model, new simulations: https://academic.oup.com/ci...

On 2020-07-05 20:02:42, user Rich Nunziante wrote:

There’s a word missing in the first paragraph of the abstract: “Of the 9 locations, 3 had one or employees infected with SARS-CoV-2,...” Should that be “one or two” since later you mention “both”?

On 2020-06-21 12:49:38, user OxImmuno Literature Initiative wrote:

https://www.immunology.ox.a...

On 2020-07-14 15:00:51, user Chyke Doubeni wrote:

The title should reflect the multicomponent nature of the intervention so that it is clear to readers that it used CHW to help people navigate the engagement

On 2020-04-17 14:41:47, user the.plummers@talktalk.net wrote:

Note that the link to the model on the front of the paper is out of data. It should be https://github.com/arp23/Ba... . The link on this medRxiv page is correct. Thanks.

On 2020-04-17 18:17:20, user LASD wrote:

So...uh...what about Sweden? Have yet to see any reasonable explanation for why the lack of lockdown there didn't lead to catastrophic consequences and bodies piled up in the streets?

Significantly lower number of confirmed cases/deaths than Switzerland and all the other major western European countries, Belgium, etc.

On 2020-04-19 15:53:35, user JGaltbna wrote:

Nothing happens “right now”. I suggest actually reading the WH plan to reopen and what has to happen before anything is “relaxed” per policy. 3 phases, each lasting at least 14 days? Ring a bell? The only restrictions being eased now are things that should never have been restricted like walking on a beach. The danger isn’t the policy but that people ignore the policy.

On 2020-07-16 10:16:25, user Hagai Perets wrote:

See https://arxiv.org/abs/2007.... for a possible explanation of the observed differences and origin of herd immunity, from a preceding low-virulence strain.

On 2020-07-18 18:45:51, user James Truscott wrote:

Hi, I think there is an error in the model as laid out in Supplementary Text File 1. The variable z represents all non-susceptibles, which includes the infectious-infected , y. The rate of loss of immunity term in equation 1 is gamma*z, but infectious individuals presumably don't have immunity to lose. They first recover (at rate sigma) and then can lose immunity. The term in equation 1 should therefore be gamma*(z-y). This change will affect the algebraic result, probably, and may change the dynamics significantly at some time points and/or parameter values.

On 2020-07-21 19:08:09, user Jeremy Rolls wrote:

Fascinating paper. Looking at the antibody data (such as there is any published here in the UK) about 18% of people in London have antibodies compared to about 8% nationally. On that basis alone 82% of Londoners may still get infected compared to 92% nationally - i.e. you would expect the mortality rate in London still to be pretty close to the national rate. Yet the hospital death stats for covid-19 in recent weeks shows London's rate consistently to be less than 40% of the national rate. Something else must, therefore, be going on - a) London is locking down better (unlikely), b) antibody immunity does not give the complete picture (possible given the data coming out of Sweden showing that for every person having antibodies two others have T-cell immunity) or c) there is a % of the population who have pre-existing resistance (from exposure to other corona-viruses) or are biologically incapable of getting infected. Ruling out a), a quick bit of maths shows about 75% of the population must fall into b) or c). So, on that basis, in London well over 90% have either been exposed to the virus or have pre-existing immunity and maybe 80-85% nationally. I suggest herd immunity has probably been achieved in London and is close in many other parts of the UK.

On 2020-07-16 13:02:24, user Dimy Fluyau wrote:

We humbly request your comments and constructive suggestions on this paper.

On 2020-07-16 18:06:03, user Marcos Woelz wrote:

What about recovered people´s blood serum? Any good news from that already? Untill that, let´s keep on helping people stay at home

On 2020-07-17 01:54:28, user Born in Akron wrote:

Is LD-RT a widely known specific therapy? This paper does not indicate the type of radiation. X-rays, Gamma rays, proton accelerator, sun lamp? The dose is 1.5 Grays = 1.5 Joules/kg = 150 rad. But the biological effect in rem or Sieverts depends on the type of radiation and duration of the exposure. Even if LD-RT is always, say, X-rays, shouldn't the effects depend on the energy of the X-rays? Unless LD-RT has a unique definition this preprint is deliberately irreproducible, perhaps to gain advantage for patent protection during a worldwide pandemic.

On 2020-07-17 12:36:05, user Hassaan wrote:

Very informative and well researched article.

On 2020-07-20 21:52:28, user Deborah Barr wrote:

It might be useful to correlate by medications taken. Depletion of magnesium and zinc affect clotting.

"drug-induced nutrient depletions are well known by pharmacists, many are underdiscussed and subsequently underdiagnosed and undertreated."<br /> 33 citations.<br /> https://www.uspharmacist.co...

Uwe Gröber's Magnesium and Drugs, https://www.ncbi.nlm.nih.go... with an excellent image of ways that drug interfere with nutrient levels in the body, and a table specific to Magnesium.

On 2020-07-21 15:50:10, user OxImmuno Literature Initiative wrote:

https://www.immunology.ox.a...

On 2020-07-21 16:14:03, user Kamran Kadkhoda wrote:

Baes on the current estimates, the sero-prevalence in Idaho is around 4% at most; such high percentages are most likely false positives; I refer authors to the study just posted here on medrxiv from China showing sero-prevalence of 2% or less in Wuhan! They used PRNT to confirm the results. That's the right way. <br /> Abbott is clear in their IFU by saying they did NOT use samples from cases with confirmed infection with common CoVs…<br /> Despite publications using "convenience samples" specificity shows its shortcoming while used large scale in the field...here's one example!

On 2020-07-22 17:00:56, user Robin Whittle wrote:

As Karl Pfleger suggested, I hope there will be more detailed information on 25OHD levels, symptoms at admission and as treatment progresses.

In light of a recent review (Charoenngam & Holick for a recent review https://doi.org/10.3390/nu1... "https://doi.org/10.3390/nu12072097)") which states that 40 to 60ng/ml 25OHD is required for proper immune system function, the 25OHD thresholds and D3 doses seem inadequate. This article also recommends an initial 12.5mg D3 (50,000IU) for all COVID-19 patients.

According to the present article, patients with 30ng/ml or more are given no D3 at all. Daily doses for those with lower levels are only 0.02mg (800IU) per day, which is a 20% or less of what most people would require to maintain 40ng/ml - assuming the supplement was taken with a fatty meal and well absorbed. https://journals.plos.org/p... indicates that average weight people need about 0.125mg (5000IU) a day to reach the middle of the 40 to 60ng/ml target range.

Surely all these low 25OHD levels (and the researchers report 21.6% of patients with initial levels below 6ng/ml and some below the 3.2ng/ml detection limit) warrant urgent action. What objection would there be to bringing all patients up to at least 40ng/ml with oral or IV 25OHD cholecalciferol (Rayaldee)? This would go into circulation immediately without relying on potentially hepatic conversion of D3 to 25OHD, which takes days or a week or so - even if the liver is functioning properly.<br /> The present article cites, as prior observations of low vitamin D levels correlating with COVID-19 symptom severity, an Indonesian article (26), an Indian article (27) and one from the Philippines (28). The first two have been withdrawn. Please see my page https://researchveracity.in... for the reasons which lead me believe that none of these three articles report actual research.

I think that the present article and a recent one An autocrine Vitamin D-driven Th1 shutdown program can be exploited for COVID-19 Reuben McGregor et al. 2020-07-19 https://www.biorxiv.org/con... are important steps in elucidating the role of vitamin D deficiency in COVID-19 severe symptoms. I have cites both articles at my page on vitamin D and COVID-19: http://aminotheory.com/cv19/ .

More research is urgently needed, but since vitamin D is a safe, inexpensive, nutrient which most people are deficient in (by the 40+ ng/ml standards we now know are important for immune system health) robust supplementation programs for all in need (most humans) need not await further research or clinical trials.

On 2020-12-19 17:11:21, user Gary Bayer wrote:

As an actuary whose required training includes construction of mortality tables, life tables and life expectancies, I attempted to verify the results. Unfortunately the details of the methods are too vague to be easily followed, so instead I attempted a standard approach to creating life expectancies. Starting with the 2017 US life tables, I explored modifying the "qx's" (probabilities of death in the next year for an individual aged x) but assuming a one time nature of Covid-19, only the specific current age (and perhaps the following age) should be adjusted for any age cohort. Therefore, for an individual age 10, only the qx for age 10, and perhaps age 11, should be adjusted to reflect the impact of Covid-19 on life expectancies. The age adjustment should be reflective of mortality risk at that age. At this point on time, based on the CDC's reporting of excess mortality, there is no evidence of increased mortality for idividuals under the age of 15. In other words, Covid-19 has not changed this cohort of individuals at all.<br /> The best guess that I can make as to what the authors were trying to express is that Covid-19 has, or is expected to reduce the average age at death this year by a year. I do not know if this is true or not but can see some merit in estimating that result.<br /> One final note, I visit the IHME Covid-19 website almost daily. It is a great tool for seeing the current state of Covid-19 in the United States, and a great tool for policy makers to get insights on what they may need to be planning for in the next couple of weeks. However, a simple look at it's various projections for daily deaths clearly shows the naivety of the estimates of what might happen in the beyond a couple of weeks. An adage that I always rely on as an actuary is the results can only be as good as the assumptions--even if the model being used is good.

On 2020-07-15 13:15:39, user E Y wrote:

Something is wrong here, the IHME projected 2020 total US death is about 250000, that's 0.08% of the US population, how can that cause 1% of reduction of life expectancy of US population?

On 2020-07-25 12:17:53, user John H Abeles wrote:

Hydroxychloroquine ( HCQ ) and Covid19

The negative observational and controlled clinical studies to date refer mainly to using hydroxychloroquine (HCQ) in serious, later stage, hospitalised Covid19 patients

In both the Solidarity/WHO study and the Recovery/UK study extremely high, even massive doses ( up to 6 times that recommended for early CoVid19 patients!) were used for unknown reasons - since the half-life of HCQ is around 21-30 days these daily massive doses could have caused very high blood levels and likely were fatal in some instances - so HCQ group deaths could have been caused by such high dose regimes, so probably skewed the results ..

Also this is likely the wrong group of patients to treat with maximum effect, in the first place — early Covid19 is the best arena for HCQ treatment in combination with zinc and either azithromycin or doxycycline...

It must be stated that no known oral antiviral for outpatients works maximally unless given quite early in disease eg oseltamivir/Tamiflu influenza; valacyclovir/Valtrex in herpes

Even iV remdesivir - a potent SARS-CoV-2 antiviral - didn’t achieve hoped for results in hospitalised patients

Later stage Covid19 patients are mostly suffering from the effects of hyperinflammation ( cytokine storm) and when viral titres are well beyond their peaks. Hyperinflammation can cause myocarditis which can certainly predispose to further cardiac toxicity.

[There are interesting thoughts that the hospitalised patients with cytokine storm / hyperinflammation in reality have a form of ADE ( antibody dependent enhancement of disease ) ie a hyperimmune reaction to a second SARS-CoV-2infection or as a result of a SARS-CoV-2 infection after a previous infection with a closely related virus]

HCQ was also used in the negative studies without added zinc which could be a design for failure, as one of the main, but certainly not only, antiviral actions of HCQ is as a zinc ionophore ie it gets zinc to enter cells much more easily where it can exert its added and established antiviral actions

HCQ is a known antiinflammatory and this action may be of some use in the hyperinflammation stage in hospitalised patients, but other more potent immunosuppressive ( and a few candidates that are nonimmusuppresive immunotherapies) could be more demonstrative in this regard.

Despite this there are some data to suggest benefit of HCQ even in hospitalised patients

For early Covid19 the usually prescribed course is for 5 to 7 days of around 400 mg daily HCQ with 100-200 mg zinc which would not invoke the long term side effects mentioned so often - and very few toxicities are reported even in long term therapy for autoimmune disorders. Any short-term arrhythmia concerns can be allayed by making sure of normal potassium blood levels

In the several thousands of outpatient Covid19 case reports published up to now , when used in early disease, there have been few if any major side effects noted.

(But in later stage, serious hospitalised patients many other drugs are also used, bringing into question the possibility of toxic interactions with HCQ. Also organ damage including myocarditis -heart inflammation-could be a particular predisposing factor in hospitalised patient toxicity predisposition to HCQ )

HCQ is a cheap, easily made generically available drug - and main manufacturers, like Novartis and Teva have donated billions of doses worldwide since the event of Covid19, so shortages, as some fear, for those taking it for malaria ( preventions or treatment) or for autoimmune diseases, like lupus or rheumatoid arthritis etc are highly unlikely

Here below are some pertinent positive references for further reading on the question of HCQ plus zinc plus either doxycycline ( my preferred choice because it isn’t associated with further small cardiac risk) or azithromycin

Note : Most of the successful reports of the use of HCQ plus zinc etc are in early stage, outpatients and not in late stage, hospitalised patients

The first link is a large data base (more than 50 studies ) on HCQ in Covid19 treatment

The second reference is an important review from a Yale University professor ...

The third and fourth are on a recent, large, well conducted observational study from Henry Ford Hospital ...

The fifth is an important outpatient study ...

https://c19study.com/

https://academic.oup.com/aj...

https://www.ijidonline.com/...

https://www.henryford.com/n...

https://www.preprints.org/m...

https://www.ijidonline.com/...

https://www.preprints.org/m...

https://aapsonline.org/hcq-...

https://www.medrxiv.org/con...

https://www.preprints.org/m...

https://www.evms.edu/media/...

https://link.springer.com/a...

https://pjmedia.com/news-an...

https://www.medrxiv.org/con...

https://www.medrxiv.org/con...

https://www.middleeasteye.n...

http://www.ijmr.org.in/prep...

https://aapsonline.org/hydr...<br /> decide/

https://www.indiatoday.in/i...

https://www.medrxiv.org/con...

On 2020-06-25 11:12:33, user Dude Dujmovic wrote:

Faulty study. The BCG cohort is older than non-BCG cohort, likely by 2-3 years on average. That is not a small difference when the samples are so big. Amazing how they did not notice that. There is always a good reason why randomized samples are used. Your samples are biased based on age. You need to have samples with about the same average age and about the same standard deviation. And more.

On 2020-07-25 19:31:23, user ???? ??? wrote:

It reflect the PK/PD pharmacological predication of efficacy <br /> The problem of LPV is complicated PK . Strong protein binding 98 % , extensive metabolism , long list of drug interaction. Therapeutic drug monitoring is mandatory to adjust dose in clinical setting. Moreover extrapolation of in EC50 to the current dose is not prefect. It was suggested to use PBA EC90 . the base line protein binding adjusted 90 % effective concentration. There is a debate about ability of current regimen to achieve Cmax > PBA EC 90 at lung tissues in severe cases

On 2020-07-26 14:06:30, user Gordon Erlebacher wrote:

I started to read the paper, but all the equations are missing. <br /> Here is an additional question. The contact matrix Mij measures to the average number of contacts between one person in group I and all members of group j. But are these different contacts or contacts with repetition? The different possible choices affects the spread of the virus. Any insight is appreciated.

On 2020-07-28 17:02:07, user Liam Golding wrote:

Nice research on an important matter. I really appreciate your work.

It would be nice to know the sample sizes you experimented on to obtain the statistical differences. Can you provide these on request?

Cheers,

Liam G

On 2020-07-28 19:39:18, user Dude Dujmovic wrote:

People that are vaccinated are usually more careful about their health so direct causality here is non-existent.

On 2020-06-30 21:10:32, user Stephen Cherniske wrote:

This is rather paradoxical, in that IL-13 is generally considered to be an anti-inflammatory cytokine, as in IBS. Even more surprising: the observed beneficial effect of DHEA treatment of murine IBS appears to result in part from increased IL-13 expression in colonic epithelial cells. REF: Immunobiology. 2016 Sep;221(9):934-43. doi: 10.1016/j.imbio.2016.05.013. <br /> Dehydroepiandrosterone (DHEA) Restrains Intestinal Inflammation by Rendering Leukocytes Hyporesponsive and Balancing Colitogenic Inflammatory Responses<br /> Vanessa Beatriz Freitas Alves , Paulo José Basso et al.

On 2020-08-03 13:54:21, user Charles R. Twardy wrote:

Forgive me if this is covered in the paper - today I am just skimming abstracts. But another preprint out today shows a mortality risk reduction of 0.7 per 100 kJ/m^2 of ultraviolet (UVA) exposure, in three countries measured at the county level. Is US altitude a proxy for UVA? Vice versa? Could you two combine models to look for residual effects?

On 2020-07-02 15:39:58, user Kamran Kadkhoda wrote:

The reported prevalence is very high suggesting high false positivity rate. The actual sero-prevalence for that county is estimated to be around 6% as of today (if we assume only 20% of cases are tested by for RNA). It would have definitely been much lower back in April. Another reason serology should not be used given it's high rate of false positivity mostly due to common CoVs like OC43 and HKU1.

On 2020-08-04 08:17:28, user OxImmuno Literature Initiative wrote:

https://www.immunology.ox.a...

On 2020-08-04 08:27:16, user OxImmuno Literature Initiative wrote:

https://www.immunology.ox.a...

On 2020-07-02 18:46:51, user Julio C. Spinelli wrote:

Having personally arquitected several clinical trials, the phase II results in young volunteers forces me to provide a word of caution to our collective desire to quickly develop a vaccine for COVID-19. <br /> The frequency and severity of many of the AE's described in this preprint for the young (18-55) and healthy population described here doesn't bode well for the results of a phase III clinical trial. Not until the phase II results of the older cohort are known. Furthermore, extrapolating these data to the Latin and Black populations would be pure hubris on our part. Further phase II data is required before we move into phase III trials<br /> Dr. Julio C. Spinelli

On 2020-07-02 20:46:36, user C'est la même wrote:

The claim of 99.3% specificity seems very high compared to other antibody tests when tested with large population samples.

But that aside, some readers seem to be inappropriately concluding that undersampling in specific regions during that period can be generalised to conclude that the true cumulative incidence is ten times the total number of confirmed cases.

This is unfounded for two reasons. The first is that regions with very high case numbers (Such as NYC) were temporarily overwhelmed in terms of testing capacity and correspondingly very high test-positivity rates. However over time, the testing caught up with demand and with expanded testing, the test-positivity rates dropped by the expected order of magnitude and likely "caught up" for at least some of the participants who were missed.<br /> The second reason is the sample in the study is not a true random population based sample, but a convenience sample which is also biased towards higher test-positivity rates.

Thus while I don't disagree with the conclusion of the authors, I urge strong caution among readers who are tempted to conclude that true case numbers are a magnitude of order higher than officially reported.

On 2020-07-03 20:21:43, user Marm Kilpatrick wrote:

Interesting paper. <br /> Could you clarify if the incidence values in Fig 1,2 and throughout are:<br /> Incidence = Cases in age group X/total population<br /> OR<br /> Incidence = Cases in age group X/population of age group X<br /> Since the age groups represent different fractions of the total populations this would change the intercept of the different incidence values/curves.<br /> Thanks!<br /> marm

On 2020-08-07 20:38:52, user Adam Garland wrote:

We are developing a test for SARS-CoV-2 in saliva. Is there any chance you have saliva samples leftover from this study that you'd be willing/able to share with us?

On 2020-08-10 08:17:44, user OxImmuno Literature Initiative wrote:

https://www.immunology.ox.a...

On 2020-07-06 16:17:27, user OxImmuno Literature Initiative wrote:

https://www.immunology.ox.a...

On 2020-07-06 16:20:42, user OxImmuno Literature Initiative wrote:

https://www.immunology.ox.a...

On 2020-08-12 08:06:03, user Marek Kochanczyk wrote:

The source code to analyze data and generate figures as well as updated figures are available at GitHub: https://github.com/kochancz...

On 2020-08-12 12:37:37, user Marc Imbert wrote:

It is worth to not that this study has more cormobity and symptomes for the group treated with HCQ and AZT. All patient not treated has a mild desease while about only 63% in the group treated. Further one should use a healthy scientific scepticims regarding hasting conlusions based on studies at the late stage of the desease. In particular with the description of the evolution of the disease which is now known,

.

On 2020-07-13 11:08:47, user Andrew D'Silva wrote:

In any infection IgM responses converting to IgG responses fall over time and rise when there is a secondary antigen exposure. Why do these findings suggest that there is loss of immunity with declining neutralising antibody levels? Surely the questions are: what happens after secondary antigen exposure? Do the neutralising antibody levels rise again? Do they protect from developing the same clinical disease again? Do they affect severity of disease after second exposure?

On 2020-08-17 22:15:16, user Roy Verdery wrote:

Neutralizing activity assessed using Vero cells may not reflect clinical<br /> activity. Nature (2020). https://doi.org/10.1038/s41....

On 2020-08-24 04:45:43, user Bill Pilacinski wrote:

Now it will be important to identify those in the population who are immune so that the early limited supply of vaccine can be used for those susceptible individuals of high priority as we attempt to reach herd immunity.

On 2020-08-20 08:37:09, user Mitchell X Kriegman wrote:

does second hand smoke carry covid shedding?

On 2020-08-21 19:06:14, user Heather Gray wrote:

I'm a layman but I want to express my appreciation for exploring the issue from this angle

On 2020-08-25 09:01:02, user Tjabbe wrote:

Evidence for what, that it doesn't work for late stage covid in hospitalised patients? Is that even news? How come at this stage in the pandemic we are still publishing reports that claim medication be ineffective "for treating covid19" when in fact it was only tested for patients with severe covid19 already in the hospital. We all know patients will not be sent to a hospital in the Netherlands for covid unless they have progressed pretty far. <br /> The report describes hcq being used on patients when deteriorating in several of the hospitals, affecting mortality, and media outlets conveniently leave out this part of the puzzle.

If you want to curb covid, or if you want to write off medication as being useless "for covid" , start doing trials on early outpatient treatment.

On 2020-08-30 11:33:10, user Martijn Weterings wrote:

One problem with those S(E)IR compartmental models is that they always assume/pretend that a virus is spread homogeneously among a well mixed population. According to such models, the chance that someone in a small village in the South infects somebody else, is the same chance for anyone. The same for somebody in the North as somebody in their immediate family or other people in close neighborhood.

Such compartments are obviously not realistic for modeling an entire country. More suitable are networked S(E)IR' or spatial S(E)IR models. In such models, the virus spreads more like an ink blot.

Due to the local saturation, growth rates are already decreasing early on. Models that do not incorporate local saturation will 'compensate' (in order to get the same early deflection) by either reducing R0, or the (effective) population, or the reporting factor (upscaling the number of infected). If you try to fit a simple compartment SIR model to real data, then you will get unrealistic epidemiological parameters.

What they are doing in this article, dividing the population into layers with different rates of infection, is effectively shrinking the population that is 'reached' by the virus.

So this effectively makes the population smaller, but the question is whether it is the right way to shrink the population? Instead of a parameter in a mechanistic model, it might better be regarded as a parameter in an empirical model. It is an extra variable to ensure that the unsuitable simple SEIR model corresponds somewhat better with the measurements.

In reality, there are several effects that cause the observed epidemiological curves to deviate from the simple models (Besides heterogeneity, the use of local distribution in spatial or networked S(E)IR models, instead of global homogeneous compartments, is another important one).

By only including only a single effect in fitting, you get that all other effects are absorbed by that one effect. The result is an unrealistic estimate of the epidemiological parameters, which will not be suitable for extrapolation (for example calculating the 'herd immunity' percentage).

It is to be expected that this model, with only the heterogeneity incorporated, will likely underestimate the percentage to reach herd immunity. This is because it is overestimating the effect to compensate for the lack of other non-incorporated effects (and spatial models will be able to model the same deflection of the curves, but with less reduction of the herd immunity).

The above is a severe systematical problem, which will result in a bias towards smaller herd immunity percentages.

In addition: The fit with the curve is strongly determined by an interaction of the population size and the factor between the reported infections and actual infections (in a simple S(E)IR model, the two have the same effect). Such correlation between the two parameters will cause great inaccuracy.

And these are considerations that do not yet mention the problems with measurements of the epidemiological curve. For instance, the inaccuracies in reporting are not easily solved with a single (constant) reporting fraction. In order to estimate epidemiological parameters we need more direct experimental data (e.g. detailed information about contact tracing). From those we can deduce more directly the variations in infection rates and estimate the potential impact on herd immunity. Just fitting a model to the curve is a bad idea.

On 2020-09-02 15:42:59, user Abhaya Indrayan wrote:

We are waiting for further doubling of cases to 4 million in India so that we can test our model.

On 2020-09-23 07:52:50, user Subhajit Biswas wrote:

Pleased to see other scientists are supporting with further evidences, the trend we had observed and reported as early as April 2020.

Based on non-overlap of dengue and COVID-19 global severity maps and evidences of SARS-CoV-2 serological cross-reactions with dengue, we proposed that immunization of susceptible populations in Europe, North America and Asia (China, Iran) with available live-attenuated dengue vaccines, may cue the anti-viral immune response to thwart COVID-19.

https://www.preprints.org/m...

Our publications in this area to support our proposition:<br /> 1) COVID-19 Virus Infection and Transmission are Observably Less in Highly Dengue-Endemic Countries: Is Pre-Exposure to Dengue Virus Protective Against COVID-19 Severity and Mortality? Will the Reverse Scenario Be True?

Clinical and Experimental Investigations, Volume 1(2): 2-5.<br /> https://www.sciencereposito...

1. Nath, H., Mallick, A., Roy, S., Sukla, S., & Biswas, S. (2020, June 19). Computational modelling predicts that Dengue virus antibodies can bind to SARS-CoV-2 receptor binding sites: Is pre-exposure to dengue virus protective against COVID-19 severity?. https://doi.org/10.31219/os...

2. This one in medRxiv!

Now, other scientists are observing the same trend in Brazil! Exciting!

See recent publication below and news coverage

1.https://www.medrxiv.org/content/10....

1. Study suggests dengue may provide some immunity against COVID-19.<br /> https://timesofindia.indiat...

Amazing! Nature has its own ways of controlling parasite aggression! Antigenic correlation between a flavivirus and a coronavirus was unprecedented.

Existing and licensed dengue vaccines could be tested in SARS CoV2 animal models and tried in dengue non-endemic countries.

Use in dengue-endemic countries may be problematic as such vaccination can elicit antibody-dependent enhancement of subsequent dengue infections.

On 2020-11-26 14:59:36, user Raul_Y wrote:

This paper is of ridiculously low quality. They even forgot to mention what dose of dutasteride was used.

On 2021-06-09 23:47:14, user Gnash wrote:

Is there any data on any group who received only AZM or only received HCQ?

On 2020-08-03 14:07:24, user Monica Sidén wrote:

I am a nurseryschool teacher in Sweden. My bloodgroup is AB+.<br /> As I can understand it is a rare bloodgroup and I can receive blood from any other bloodgroup( since I don´t have any antivirus against any other bloodgroup). Now I am very wooried that I am likely to be at a high risk. I would be very pleased if someone can explain.