On 2020-05-26 22:31:23, user guost wrote:

Unfortunately there is absolutely no info about what kind immunoassay this LIAISON XL platform is. The company Diasorin website for is totally useless in that regard.

On 2020-04-24 07:57:39, user Sinai Immunol Review Project wrote:

Repurposing Therapeutics for COVID-19: Rapid Prediction of Commercially available drugs through Machine Learning and Docking <br /> Mahapatra et al. MedRXiv [@doi:10.1101/2020.04.05.20054254v1]

Keywords<br /> • Drug prediction<br /> • Machine learning<br /> • Docking

Main Findings<br /> The COVID-19 pandemic has ravaged hospitals: the disease can present severe complications (acute respiratory failure in particular), and yet no specific drug exists to date. Time being of the essence, it is therefore essential to explore drugs already on the market for other indications. These drugs, however, must be tested in COVID-19 patients and thus selection of limited candidates is important. The authors argue that an important step in accelerating the selection of promising drugs can be done in silico.<br /> The authors use machine learning (ML), training their algorithm on a dataset obtained from in vitro targeting of SARS Coronavirus 3C-like Protease with existing drugs. The trained algorithm was then used to screen drugs available in the Food and Drug Administration’s Drug Bank. Using the Drug Bank dataset, the authors also performed a docking study -a process used to predict in silico the orientation and conformation of a molecule when bound to its receptor. Since SARS-CoV-2 spike protein is considered to play an important role in infection by binding ACE-2, docking was also applied to study the stability of drug-spike protein complexes. The results of the ML and docking were aligned, and antiretroviral Saquinavir was identified as a potentially promising therapy for COVID-19.

Limitations<br /> The authors train their algorithm on SARS Coronavirus 3C-like Protease, as inhibitors of this protein should prevent the virus from replicating in the host. However, the authors note that the most promising target seems to be SARS-CoV-2 spike protein. Moreover, the training dataset is the result of in vitro studies, and may have limited relevance in vivo.<br /> Overall, preclinical studies and then potential clinical trials would need to be performed before administering this drug to COVID-19 patients though, admittedly, clinical validation of an existing drug could happen faster than the development of new drugs entirely. Saquinavir has been studied in vitro by Yamamoto et al.[1], and shown little promise in SARS-CoV-2 treatment so far.

Significance<br /> Repurposing of existing drugs is can be advantageous to develop treatment strategies. An in silico approach could help identify potential therapies, although they must be confirmed in clinical trials before being administered on a large scale.

References<br /> Yamamoto et al. Nelfinavir inhibits replication of severe acute respiratory syndrome coronavirus 2 in vitro. BioRXiV preprint, 2020

Credit<br /> Reviewed by Maria Kuksin as part of a project by students, postdocs and faculty at the Immunology Institute of the Icahn School of Medicine, Mount Sinai.

On 2024-04-28 02:04:59, user Ian Myles wrote:

Small correction, the Clinical trial number is NCT04864886

On 2024-04-27 18:42:07, user Kim Brumfield wrote:

Thank you to Dr. Myles and his team for doing this research. My daughter has 20 years of prescribed steroid use for eczema without informed consent of the risk of topical steroid addiction and withdrawal. The dermatologists she saw used step therapy which eventually resulted in tachyphylaxis. The topical steroids stopped working after reaching hichest potency and now she is suffering from "eczema on steroids" which is really topical steroid withdrawal. Please help us find the cure for this horrible iatrogenic disease.

On 2024-04-27 19:07:49, user Toby wrote:

So glad to see that this problem is being taken seriously. I have suffered from eczema all my life and from TSW as described here for many years. I hope this research will result in new treatments for this awful condition.

On 2024-04-27 19:50:44, user Eve Benson wrote:

I am starting my 8th year of TSW. The first five years were torture, the last few have been manageable. I can identify with every symptom listed in this study. I can also identify with the stress of seeing several "well-meaning" doctors whose only choice of treatment was putting me back on steroids, which I refused thanks to the education I received from online communities of thousands of people who were suffering like myself. More studies like this one are needed. TSW is real. Sufferers of TSW deserve appropriate medical care and care from practitioners who understand the disease and thus provide appropriate treatment. It is time for medical institutions to step up and address TSW.

On 2024-05-02 15:45:09, user Kelly Barta wrote:

This is such an important and groundbreaking study in its showing a differentiation between Atopic Dermatitis and Topical Steroid Withdrawal Syndrome, which has been one of the big debates within the medical community. Patients need and deserve acknowledgement and support from their health care providers, but understandably, this is unable to happen without the science to back up what we are seeing anecdotally in the eczema patient population.

More research is needed to determine how and when topical steroids are creating these dysfunctions in patients in order to better understand their proper use and prescribing guidelines. This research is CRITICAL to the over 31 million Americans living with eczema and over 300 million worldwide (not to mention the countless other dermatology patients), who are prescribed topical steroids to manage a skin condition.

On 2024-11-15 06:07:16, user Nina T ang wrote:

May I have access to the FFA dataset please?

On 2020-02-13 21:52:14, user reuns wrote:

I deduce from "case 13: no viral RNAs were detected until the fourth upper respiratory samples" that in the graphic U means negative

On 2020-03-25 19:31:48, user Charles Haas wrote:

My concern with their disinfection experiments is that there is no indication that they neutralized the disinfectant prior to culturing. This is an absolute necessity.

On 2020-04-22 20:47:32, user John Stevens wrote:

Hope this is widely understood - each tiny bump in efficiency makes differences in daily KPIs

On 2022-12-24 03:16:24, user ArCH wrote:

Published version<br /> https://www.frontiersin.org...

On 2020-05-28 08:18:17, user Philippe Brouqui wrote:

The paper of KIM et all reports on the response to treatment of hydroxychloroquine or<br /> lopinavir-ritonavir with or without antibiotic in a retrospective cohort study<br /> with comparison with standard of care. The evaluation has been carried out on<br /> moderate case only and no death were reported. They assume that HCQ and ATB is superior to both SCO and LR plus ATB in time to viral clearance, length of hospital<br /> stay, duration of fever and cough. Adverse event been significantly different<br /> from SOC but not different between the two arms of treatment and only<br /> mild. <br /> Methodology: Is appropriate for the aim<br /> -The study is relevant to the aims: treat patient early (non-severe disease) at the time to diagnosis to avoid complication and death<br /> - The study is relevant toward bringing new data in time to outbreak by using repurposing of drugs HCQ and LR <br /> -The study is relevant toward bias related to heterogeneity of care as a single center study.<br /> - Classification of patients referred to NIH Guidelines (NEWS Score) as mild, moderate and severe COVID<br /> - Definition of negative PCR and viral clearance is in adequation to previous published literature.<br /> -Treatment was done within the range of recommended dosage ; HCQ 200mg/ twice a day as well as for lopinavir/ritonavir 200/50 mg, however higher doses are generally used for HCQ (600mg/d). For antibiotics there were given as recommended

Ethics

-Data were collected through the patient medical file of the hospital blinded and using patient dataprotection

Outcome measures<br /> -Correspond to aims delay between treatment initiation and viral clearance, discharge from hospital and symptoms resolution

Statistical analysis

This is classical analysis methods relevant to aim

Confounding bias/ Limitation:<br /> -Those relative to retrospective study but well a posteriori controlled <br /> -As a retrospective study of the 358 patients (270) 173 mild and 97 moderate covid-19 cases were analyzable for completion of treatment and data availability. <br /> - 3/270 patients were still ongoing treatment at time of release (1%)<br /> - 97 moderate Covid-19 patients were categorized HCQ ATB SOC (22) LR ATB Soc (35), SoC (40) and analyzed for treatment and outcomes<br /> -A posteriori comparative analysis shows that the two groups HCQ and LR were identical in terms of comorbidity and other known factors that may weigh on the outcome.<br /> -Comparison with the Soc group alone showed that this last was less severely ill<br /> (significantly less pneumonia), and factor associated with poor outcomes, such<br /> as low lymphocytes count, and elevated CRP were associated with the treated<br /> groups. Interestingly dyspnea was more prevalent in this SOC group but we know<br /> that absence of dyspnea (silent hypoxemia) is rather than dyspnea linked to<br /> outcome. This suggest that if a difference in treatment exist it will probably<br /> be under evaluated. <br /> - Retinopathy to HCQ as never been reported in such short treatment and HCQ in serum returns to negative in 15 days after end of treatment (unpublished) <br /> The limitation of the study as been well reported

Interpretation of results: adequate except comparison LR/LR and ATB<br /> Time to clearance are shorter particularly in the HCQ and ATB group for which time to PCR > 35CTis 12 days in what is published elsewhere<br /> -Cough resolve significantly better in the HCQ ATB and fever in the two treated groups compare to SOC.

Adverse events were more frequent but only mild

The subgroup analysis LR versus LR ATB should be interpreted carefully as we don’t know interval time to onset of the LR only arm which may interfere with viral clearance of this subgroup.

Conclusion<br /> This study appropriately show that HCQ & ATB is better than LR & ATB than to SOC<br /> to shorten viral clearance, resolution of symptoms, and to shorten hospitalization duration in moderate form of COVID-19. The role of ATB alone to shorten viral clearance is overestimated

Note : P BROUQUI , has no conflict of interest with the industry concerning this review<br /> but has already published study supporting the efficiency of HCQ and azithromycin in COVID-19.

On 2020-04-23 05:41:01, user FrezzaLab wrote:

Is it possible that some of these kids had already been infected, before the lock down, and passed it onto relatives?

On 2020-04-26 13:59:26, user Italian_in_london wrote:

There is an important element of this research widely mentioned in Italian TV interviews and on Italian generalist press: the isolation of all positive cases caused a sharp drop (60%) in intensive care cases, as if the high viral load of people exposed to multiple contacts with positive cases is the main cause as to why some people end up I intensive care. I am interested to understand why information allegedly coming from this research does not seem to appear here. It is obvious what the implications are for medical staff being asked to go back to work despite being still contagious.

On 2021-12-05 11:00:28, user Professor Ritual wrote:

I actually like the modelling idea, a noble effort - but as things move fast the data used in the study is now outdated. Please remodel for the current data: 71% senior breakthrus and 50% adult breakthrus.

On 2020-04-07 01:06:33, user Cristian Reyes P. wrote:

In Chile BCG vaccine has been mandatory since 1949. Everybody is vaccinated. Over 90% of the population. You can study us. We still have the lowest mortality in the region.

On 2020-03-31 00:14:04, user Alex wrote:

Have you analyzed West and East Germany?

On 2020-04-01 16:36:48, user japhetk wrote:

The study seems interesting.

However, the problems of this study's analyses, are as mentioned in the comments, <br /> they are not controlling when the infection spread in the country.

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).

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 <br /> year when the country started it). The current measure is not indicative as authors indicated.

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 <br /> 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 <br /> robustly and negatively correlated with GDP of the country after controlling the population of the country (p = 0.019, partial correlation coefficient: -0.292).

Also "how long the country has advanced the BCG vaccination measure" is robustly and negatively correlated with how fast the 100th patients were detected in the country (p = 0.078, partial correlation coefficient: 0.281).

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).

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).

The partial correlation between "how long the country has advanced the BCG vaccination measure" and number of deaths in the 10th day after controlling the population and GDP of the country was P = 0.178, partial correlation coefficient = -0.210.

So, my guess is probably, there are number of spurious correlations happening in authors' analyses due to lack of important control variables, even if there are real correlations, they apparently should not be that strong (studies of the BCG's universal effects have not <br /> indicated such things either).

In Korea, China, Japan (diamond princess), the virus infected a lot of people in some regions or situations, too.

The countries of higher GDP can do more tests, they are more popular to the tourists from Asia, but they were perhaps less inclined to use masks, they were confident of their medical system and less alert. And those are the countries where BCG was "no longer necessary".

Remember one month ago, the coronavirus is an infectious disease of Asian people. Now it is an infectious disease of Western countries, who knows if it is not the disease of developing countries in the next few months.

On 2020-04-01 22:04:32, user Mr. Andrew wrote:

Singapore, Taiwan, Hong kong are litterally next to China and have only double digit death rates, all added, in total. WHY? All vaccinate their kids for BCG versus Tuberculosis. It's not a coincidence, all other countries do not vaccinate for it. Other BCG Vaccinating countries: Romania, Malaysia, Thailand..

check this map bellow (in the link) of countries which never had BCG. In entire Europe, Italy is the only one which never had BCG vaccination. Thus, they have a huge deathrate.

https://www.researchgate.ne...

This is further proved by all countries with BCG at birth. Check out all countries and how bad they are doing with covid 19 by looking at their death rate and serious critical numbers, at the official WHO numbers: https://www.worldometers.in...

Till now all countries which have BCG at birth have extremely low death rate and people in serious critical condition, but huge infection rate (thus small percentage). Singapore, Taiwan Hong Kong were infected way back before Italy was infected.

=

Lets tell people about BCG and pressure more research on this, and if it actually is helpful give every other country which did not get it at birth: a shot. It might be a cure, I am predicting but the data does not lie.It provides viral immunity although its meant for bacteria. As your lungs are stronger. The data shows something, and look at all the countries death rate and serious critical numbers versus infected.

They are soo exceptionally good compared to all others like x30 times better. Would like your help to spread the word of BCG and more research to be done. So countries like Italy which never where vaccinated would get a shot. (the U.S is next, as they never vaccinated for BCG)

On 2020-04-02 12:47:23, user Anders Milton wrote:

The 70-plus Italians would have had BCG vaccinations when they were young, I believe. Still they die due to the covid-19 infection. How to explain that?

On 2023-05-30 13:27:21, user Nils Yang wrote:

This paper has been published in Journal of Child Psychology and Psychiatry, see http://doi.org/10.1111/jcpp...

On 2020-07-24 16:43:50, user Kamran Kadkhoda wrote:

The correlate of protection is not inferred this way it is typically inferred through prospective vaccine trials in SARS-CoV-2-native volunteers.

On 2020-04-08 01:57:26, user Eliot Abrams wrote:

This just fits a gaussian curve. Absurd. Among other reasons, there is a second wave as soon as the current shelter in place restrictions are lifted.

On 2020-04-03 16:31:40, user Alexander Siegenfeld wrote:

These projections likely severely underestimate the number of deaths and hospitalizations because they assume that any state that has implemented three out of four interventions they consider (school closures, non-essential business closures, travel restrictions including public transportation closures, stay-at-home recommendations) will see an epidemic trajectory similar to that reported in Wuhan, China.

The Imperial College report released on March 30 that quantifies the impact of nonpharmaceutical interventions in Europe predicts that even with the complete lockdowns implemented by 10 out of the 11 countries studied, the number of new infections may still increase. Given that the response in even the U.S. states implementing all four of the interventions considered by IHME may be less effective than the European lockdowns, there is a distinct possibility that without action beyond that assumed by the IHME study, the rate of new deaths and hospitalizations may not only not peak and decrease as quickly as IHME predicts but may also continue to exponentially increase (albeit at a slower rate).

See our full comment here: https://tinyurl.com/yx8xxqsv

On 2020-04-06 23:24:28, user Mastah Plannah wrote:

It is ridiculous that the model still says that Massachusetts has NOT implemented a Stay At Home order. Therefore the model is useless for Massachusetts.

Massachusetts did implement stay-at-home. They did it on the early side on March 23.

On 2020-07-27 03:48:45, user Hagai Perets wrote:

A potential explanation for such dynamics could be related to co-infectio, see <br /> https://arxiv.org/abs/2007....

On 2021-12-22 14:03:31, user Simone Davies wrote:

I had these symptoms (vibrations for months, muscle twitching for days) after my all 3 of my vaccine shots. I had not had covid before I was vaccinated. Would be interested to know if this is true in others?

On 2023-11-07 14:15:21, user Nils Yang wrote:

This preprint has been published at Sleep: https://doi.org/10.1093/sle...

On 2020-06-09 17:41:58, user Hamid Reza Marateb wrote:

This is a hospital-based cohort whose results could not be generalized to the population. Moreover, these patients usually have commorbidity, and thus avoid smoking. Here are justifications.

On 2020-08-08 06:57:06, user Dr-Beesan Maraqa wrote:

Thank you for this study. I am struggling to find studies assessed the associations between stress and demographic factors, job title, and relation to social life.

On 2022-01-01 05:17:59, user Ardiana wrote:

N501Y and E484K signals high spread ability and original vaccine antibody evasion. But there were many variants like this and that couldn't beat Delta.

On 2020-04-15 11:57:26, user Renato Prandina wrote:

Duration and extent of immune protection will be critical to the novel betacoronavirus SARS-CoV-2 and will unfold in coming years. Some speculative scenarios...

On 2020-04-16 06:06:53, user Hellbound Reaper wrote:

SARS-CoV-2 is the virus name..COVID-19 is the disease name..you can't catch a disease. :P

On 2020-04-16 12:20:10, user Marlowe Fox wrote:

The tests on the efficacy of HCQ are confounded by multiple variables, including comorbidities, symptom onset, prescription drugs (RAAS inhibitors appear to play a key role in viral intensity), and testosterone/estrogen level, to name only a few.

Geneticists, epidemiologists, and other scientists have long used casual diagrams to clearly show variables that may potentially confound their results (1). The Wuhan study at the very least would need to account for the following:

HCQ <— comorbidities —> recovery<br /> HCQ <— symptom onset —> recovery<br /> HCQ <— drug prescriptions —> recovery

Adjusting for the confounding variable would essentially smooth out the flow of information between the treatment (HCQ) and the outcome (recovery), allowing for the inference of causal effects.

Assuming observable data is not available to adjust for confounding variables, a casual mechanism (mediator) could smooth out the flow of information from the treatment to the outcome (so long as the mediator is not influenced by confounder).

Luckily, multiple in vitro studies have been performed. One study posits that HCQ lowers endosomal pH which ultimately inhibits COVID from binding to ACE 2 and decreasing viral intensity (3).

HCQ —> endosomal pH —>glycosylation of COVID cellular receptor —> ACE 2 binding —> viral intensity —> acute lung injury

Another in-silico study posits that HCQ blocks specific protein sites on the host ACE2 cell, thereby thwarting its attempt to infect it and preventing the cytokine storm (over-reaction of the lymphatic system) that some posit is responsible for Acute Lung Injury (3). So here we have an entirely different causal mechanism:

HCQ —> BRD-2 receptor sites —> cytokine storm —> acute lung injury

Despite these problems, some believe that the p-values obviate the need to control for potentially lurking variables. However, they are subject to myriad influences, known as p-hacking. Whether it is the number of tests performed or the number of comparisons made, it increases the chance of finding a statistically significant p-value (4). Three professional statisticians co-authored a paper reviewing the validity of the Wuhan study (5). There were several issues with the data upon which the two significant p-values were based.

I suppose there is also a pragmatic argument: The p-values, along with existing studies and reports, are sufficient enough evidence to offset any concern for lurking variables in these urgent times. In other words, how much evidence is sufficient to warrant large scale roll-out of a low-cost treatment that may have a beneficial effect, from saving individuals who would have otherwise died to curbing its spread?

The consequences of large roll-out: manufacturing, scaling, distribution chains, and so forth could result in a tremendous diversion of resources. How many pharmaceutical manufacturers even have the capacity to roll out production of this magnitude? What if they all start scaling their labor to produce this particular drug. You can’t just put this genie back into the bottle. Not to mention the scientific energy/intellectual capital that would go to proving or disproving this proposed treatment. And why? Because scientific evidence demanded it? No because a tortured p-value and unpublished/unsubstantiated anecdotal evidence caught the attention of some in the media, and it has been over-popularized as a panacea. What about the risk that HCQ is not an effective treatment despite large investments in cash and resources that have been invested? Do you think the wheels of capitalism turn so easily? Investors will want a return and if that means continually touting an ineffective drug through spurious science, they will continue to do so. What about individuals taking HCQ as a prophylactic, believing themselves to be protected against COVID? Or COVID+ individuals taking HCQ and believing themselves to be cured? Or individuals who think: Well, if I get it—I’ll just take HCQ and be fine. This would increase the spread of COVID. From my perspective, the ignorance to viral transmission and the required precautions is widespread. This is just one more reason not to acquiesce to the new social norms of wearing face masks, social distancing, and abiding by shelter-in-place rules. Here, I think an understanding of cognitive psychology is important to anticipate the future behavior of a society in which a cheap and easy-to-manufacture cure is published in the media.

To sum up, HCQ's efficacy is not sufficiently proven to warrant a widespread roll-out, because it could result in several downstream consequences, from the diversion of resources (both manufacturing capabilities and intellectual capital) to increasing the risk threshold of individuals--who spurious believe in an easy and cheap treatment--thereby increasing the infection rate. One of two things needs to happen. Clinical trials that properly adjust for all potential comorbidities. Or the discovery of a causal mechanism (in vivo), which would obviate the need to control/adjust for confounders. For me, this would tip the utilitarian scales in regard to the potential benefits versus the risks.

References

1. Judea Pearl and Dana Mackenzie. 2018. The Book of Why: The New Science of Cause and Effect (1st. ed.). Basic Books, Inc., USA.
2. https://www.ncbi.nlm.nih.go...
3. https://papers.ssrn.com/sol...
4. https://www.scientificameri....
5. https://zenodo.org/record/3....

On 2025-04-02 10:00:44, user Md Shahed Morshed wrote:

The published version can be found here: https://doi.org/10.3329/jacedb.v3i2.78642

On 2020-08-31 13:17:21, user Kamran Kadkhoda wrote:

Great work! This suggests the specificity of the Euroimmun assay is around 33%!

On 2019-10-30 08:25:55, user Marema wrote:

This paper is done to investigate how acute financial problems affect undergraduate students' clinical learning.The study used qualitative method to explore their experiences. I here for further information.

On 2020-05-06 06:53:53, user Hossein Mirzaei wrote:

Dear Georg<br /> i cant find table 1 in your article, Which you refereed to that for Main characteristics of patient.<br /> can you help me to find that?<br /> Thanks<br /> Hossein

On 2020-02-17 08:52:17, user Ellie_K wrote:

Once this paper has been peer-reviewed, could someone post here in the comments where (i.e. in which scholarly journal) it is published? Thank you!

On 2020-02-14 00:59:07, user acm_ian wrote:

Doesn’t the accuracy of the modelling depend on the input data. Identifying an unknown infectious agent in routine practice is not simple. It is feasible that the virus has been around longer without being recognized and the spread coincides with the usual winter influenza and other respiratory virus spread.

On 2020-02-24 21:00:31, user Alexei Vazquez wrote:

As predicted https://arxiv.org/abs/cond-...

On 2020-03-04 13:30:42, user Bìtao Qiu wrote:

I think there are some wrongly put numbers on Table 3 (page 35), e.g. n = 59 for Immunodeficiency patients. It should be 3 according to the abstract and n = 3 on page 36.

On 2020-09-19 14:10:12, user kdrl nakle wrote:

Given the low infection rates you most likely have many false positives and the adjustment is probably of uncertain quality. Basically, it is not very accurate to use serosurveys in low infection areas.

On 2020-03-12 14:56:17, user Reza Amini wrote:

This is great job! I've done something similar to this earlier but I have DATASET shortage.<br /> "imreza.ir" .<br /> This was really surprising

On 2020-03-17 03:45:06, user God Bennett wrote:

I foresaw this from February 9th, having started an ai based ct scan initiative:

https://github.com/JordanMi...

On 2020-03-24 19:18:13, user Luis Cabrera wrote:

In the Extended Data 2, there is another standard curve, instead of "Primer, reporter molecules, target gene fragments, and guide RNAs used in this study. "

On 2020-03-26 15:11:11, user Sinai Immunol Review Project wrote:

Study description: Plasma cytokine analysis (48 cytokines) was performed on COVID-19 patient plasma samples, who were sub-stratified as severe (N=34), moderate (N=19), and compared to healthy controls (N=8). Patients were monitored for up to 24 days after illness onset: viral load (qRT-PCR), cytokine (multiplex on subset of patients), lab tests, and epidemiological/clinical characteristics of patients were reported.

Key Findings:<br /> • Many elevated cytokines with COVID-19 onset compared to healthy controls <br /> (IFNy, IL-1Ra, IL-2Ra, IL-6, IL-10, IL-18, HGF, MCP-3, MIG, M-CSF, G-CSF, MIG-1a, and IP-10).<br /> • IP-10, IL-1Ra, and MCP-3 (esp. together) were associated with disease severity and fatal outcome. <br /> • IP-10 was correlated to patient viral load (r=0.3006, p=0.0075).<br /> • IP-10, IL-1Ra, and MCP-3 were correlated to loss of lung function (PaO2/FaO2 (arterial/atmospheric O2) and Murray Score (lung injury) with MCP-3 being the most correlated (r=0.4104 p<0.0001 and r=0.5107 p<0.0001 respectively).<br /> • Viral load (Lower Ct Value from qRT-PCR) was associated with upregulated IP-10 only (not IL-1Ra or MCP-3) and was mildly correlated with decreased lung function: PaO2/FaO2 (arterial/atmospheric O2) and Murray Score (lung injury).<br /> • Lymphopenia (decreased CD4 and CD8 T cells) and increased neutrophil correlated w/ severe patients.<br /> • Complications were associated with COVID severity (ARDS, hepatic insufficiency, renal insufficiency).

Importance: Outline of pathological time course (implicating innate immunity esp.) and identification key cytokines associated with disease severity and prognosis (+ comorbidities). Anti-IP-10 as a possible therapeutic intervention (ex: Eldelumab).

Critical Analysis: Collection time of clinical data and lab results not reported directly (likely 4 days (2,6) after illness onset), making it very difficult to determine if cytokines were predictive of patient outcome or reflective of patient compensatory immune response (likely the latter). Small N for cytokine analysis (N=2 fatal and N=5 severe/critical, and N=7 moderate or discharged). Viral treatment strategy not clearly outlined.

On 2020-05-18 12:43:08, user Sinai Immunol Review Project wrote:

Long period dynamics of viral load and antibodies for SARS-CoV-2 infection: an observational cohort study<br /> Huang et al. medRxiv [@doi.org/10.1101/2020.04.22.20071258]<br /> Main Findings<br /> The presence of serum IgM and IgG against SARS-CoV2 has been shown in several studies, however, a limited number of studies have shown the longitudinal relationship between viral RNA levels and antibody titers. This retrospective, observational study evaluated the dynamics of viral RNA, IgM and IgG specific for SARS-CoV2 proteins in patients with confirmed SARS-CoV-2 pneumonia over an 8-week period. <br /> Throat swabs, sputum, stool and blood samples from 33 patients with laboratory confirmed SARS-CoV-2 pneumonia were collected to analyze viral load and specific IgM and IgG against spike protein (S), spike protein receptor binding domain (RBD), and nucleocapsid (N). The demographics of the patients showed that 24 had respiratory symptom, two had symptoms in both the respiratory and the gastrointestinal tracts, one had gastrointestinal symptoms and six were asymptomatic. Chest CT revealed 27 patients had bilateral infiltrates and six had unilateral infiltrates. All the patients received antiviral treatment and atomized interferon during hospitalization. <br /> While viral load in throat swabs and sputum was higher at the symptom onset and undetectable by three weeks and five weeks respectively, viral load in stool started low but remained detectable for more than five weeks in many patients. The viral loads in sputum declined significantly slower compare to throat, so that the patients were divided into two groups based on load in sputum: short-persistence (viral RNA undetectable within 22 days, n=17) and long-persistence (viral RNA persists more than 22 days, n=16). The relationship between the persistence of sputum viral RNA and antibodies showed that short-persistence group had higher anti-S IgM, anti-RBD IgM and anti-RBD IgG levels compare to long-persistence group suggesting a potential protection by anti-RBD antibodies. The length of time from symptom onset to hospital admission was also associated with SARS-CoV-2 viral clearance. In addition, the higher seropositive rate for anti-S and anti-RBD IgM was seen in long-persistence patients. They suggest that delayed admission to the hospital resulted in higher seropositive and longer infection in patients with COVID-19.<br /> Limitations<br /> The separation of ‘low persistence’ vs ‘high persistence’ groups seems quite arbitrary, as only virus RNA levels for in sputum were considered, and viral loads in throat swabs and stool were not considered (these were no significant difference between the low vs high groups. The manuscript needs better and more detailed description of methods and figure legends. The same color codes for each patient could be used in figure 1, so that the readers could see if there was a trend in viral load between specimens in a same patient, and graphics for each patient will be useful to understand viral dinamics in the three types of samples for each person. The graphs on figure 5 seemed to correspond to one time point data, but there was no explanation which time point was used. In the results section, it was not clear which figures or tables were related to the text. The correlation between severity, viral load and persistency, and antibody titers could be analyzed.<br /> Significance<br /> It is important to understand the relationship between viral loads, disease progression and viral-specific antibodies in COVID-19 disease. More studies are necessary.

Reviewed by Miyo Ota as part of a project by students, postdocs and faculty at the Immunology Institute of the Icahn School of Medicine, Mount Sinai.

On 2020-10-30 23:23:26, user Leonidas Palaiodimos wrote:

This article has been reviewed by peers and published at Hormones-the International Journal of Endocrinology and Metabolism

On 2020-03-31 16:17:23, user Johannes Opsahl Ferstad wrote:

Link to tool: https://surf.stanford.edu/c...

On 2020-04-22 04:36:02, user Paul Hue wrote:

Has Covid19 been truly isolated? Have its purported surface proteins been linked to genetic sequences in recovered genetic material from a true isolation?

On 2020-04-30 16:32:41, user Tim McNaughton wrote:

I wrote a graphical model to help understand the "false positive problem"

in general: <br /> https://sites.google.com/vi...

and specifically for this paper:<br /> https://sites.google.com/vi...

On 2020-05-03 19:21:56, user ThetTruth wrote:

Can anyone use these findings to arrive at an estimate of how many US deaths from covid will occur 2019-2020?

On 2020-05-12 11:12:30, user Guest wrote:

I’m an advocate for ignoring cases & case fatality rates at this stage. Why? Because of variance in testing & “at risk” populations.

I only watch deaths & “excess deaths.” I don’t see a benefit from arguing CFR at this point if excess deaths are much higher in most countries than would be expected from any other cause. The severity of disease can be seen on the multiple excess death numbers by various countries (See Ecuador’s excess deaths.)

From memory: U.S. Flu season tests ~10,000,000, uses ~1,000,000 for statistical analysis, Flu season starts at the 10% positive rate & only ~6,000 confirmed seasonal deaths. There are ~8750 daily deaths from all causes & as the rate goes higher during December, January & February they try to calculate how many could be due to influenza. Ie. From all these #s they produce statistical modeling that shows ~35,000,000 infected & ~36,000 deaths.

ADDENDUM<br /> CFR numerator & denominator. <br /> To be clear, CFR should be based on all deaths out of all cases from a disease. The Flu example above shows it is not 6,000 / 1,000,000 but 36,000 / 35,000,000. The former is CFR of 0.6%, the latter near 0.1%. The numerator & denominator on the former are low based on lack of testing (especially both asymptomatic & deaths at home, nursing homes, hospice, etc.) The latter deaths are calculated out of analysis of all excess deaths. If we use the former numerator, with the latter denominator it greatly lowers the CFR. CFR of Flu is not 6,000 / 35,000,000.

Law of large numbers. <br /> Taking one example of infection rates doesn’t show the variance in the country nor the globe. Most believe both the case counts & deaths are a multiple higher than are posted.

Current estimates are CFR ~0.5-0.9 which is 4-9X more deadly than the Flu. This really doesn’t need to be true to assess the dangers of an infection; CFR is but one variable. If, CFR is 0.1 but reproduction numbers are large & the # of cases & deaths are drastically larger than Flu, it would still support “social distancing” measures.

On 2020-04-23 03:15:30, user Zachary Blair wrote:

Mortality projections that are based on a sample limited to one of the wealthiest counties in the country will likely be dangerously flawed. This methodology ignores wealth-based health disparities and is totally irresponsible from a public health perspective. A comparative study needs to be done if you seek to make conclusions that are valuable on a national scale. These results will ALWAYS be geographically specific unless you broaden your sample.

On 2020-04-23 12:38:51, user Tomas Hull wrote:

There is a significant number of populations tested of which there is a large number, mainly younger population, who have cleared the virus out, and yet, no detectable levels of antibodies where found by the antibody test in their blood plasma.

By what mechanism those groups of people, mainly younger population, were able to overcome SARS-CoV-2 infection, if their immune systems didn't produce the detectable levels of antibodies?

Also, what does this phenomenon imply how widespread really the virus is, if many more people, who have been infected with SARS-CoV-2, are among the many of false-negatives for antibodies?

On 2020-04-22 12:09:29, user Three-legged Crow wrote:

Where’s zinc mentioned in this study?

On 2020-04-23 17:39:53, user Patrick wrote:

"Rates of ventilation in the HC, HC+AZ, and no HC groups were 13.3%, 6.9%, 14.1%, respectively."

"The risk of ventilation was similar in the HC group [...] and in the HC+AZ group [...], compared to the no HC group."

Is there a mistake in one of these sentences, or am I reading this right?

On 2020-11-18 21:50:14, user Hamid Merchant wrote:

Very interesting findings. Can you post the Ct values of all individual patients at different time points in a table as a supplementary data file please?

On 2022-10-24 19:47:27, user Camille Sawosik wrote:

The main goal of this study was to create a new model in order to diagnose brain disorders as they are often complex and get misdiagnosed. It seems that the researchers here have created a base computer model in order to diagnose brain disorders. The main critique here, as even the researchers point out, is that the model needs greater development and study before it can be applied in a clinical setting. At this point in development of the model, I would say that this paper has moderate significance, but could provide a breakthrough in the field if further development occurs. For now, the brain samples looked at all came from a limited number at the NBB. In the future, perhaps applying this model in other populations would continue to develop its significance within the field. The large majority of this paper relied on methods based in computer design. Coming from someone with a smaller background in these methods, I would have liked to see greater description of what they did. For example, at one point the FuzzyWuzzy library is referenced, and it would have been helpful to include some explanation of what this is. At some points as well, I found that the methods were essentially repeated in the results. The concept was interesting, but it was often hard to follow exactly what was done as this study seemed much more programing and computing based. In some sections, as well, tables or figures were referenced, but then those tables did not exist within the paper. Moving forward, other researchers should definitely use this as a building block for the future in order to build off of and develop a more advanced model. The findings here are interesting and provide a good framework for future extrapolation of the model. I found this paper interesting and hope for future development to get this idea into a clinical setting!

On 2022-10-27 19:45:52, user Indi Trehan wrote:

This article has now been published in The Pediatric Infectious Disease Journal -- doi: 10.1097/INF.0000000000003740

On 2022-11-10 06:25:14, user David Hinds wrote:

This paper has been published: https://doi.org/10.1038/s41...

On 2022-11-17 03:39:53, user M. Cunningham wrote:

The FDA EUA specifies that Paxlovid's window of availability requires the patient to be within both 5 days of symptom onset and the first positive test, whichever comes first. Is this not the VA's protocol? I only saw the testing aspect mentioned. I would think that this would further narrow the margins of error (eg: confirming early treatment). Additionally, Paxlovid is nirmatrelvir and the co-drug ritonavir. Since the press is already reporting this the same as a peer-reviewed finding, IMHO it's important to correct these omissions so that the public is not confused about the use of this medication. But I am enthusiastic to re-read the study when it has been evaluated and reviewed! Thank you for your research.

On 2022-12-02 16:36:24, user Mark Czeisler wrote:

Note from the authors:

A revised version of this paper was published in Annals of Internal Medicine on 29 November 2022 following peer review. Below is a link to the article, along with the PubMed citation.

https://www.acpjournals.org...

Czeisler MÉ, Czeisler CA. Shifting Mortality Dynamics in the United States During the COVID-19 Pandemic as Measured by Years of Life Lost. Ann Intern Med. 2022 Nov 29. doi: 10.7326/M22-2226. Epub ahead of print. PMID: 36442062.

On 2022-12-13 18:27:20, user João Duarte wrote:

This article is now peer-reviewed and published in Frontiers in Neuroscience, here: https://www.frontiersin.org...

On 2022-12-26 13:26:47, user y wang wrote:

You did not indicate the method of calculating the chi-squre.<br /> Actually, your method does not seem correct. <br /> One can google "Comparing Two Independent Population Proportions" and find the formula and calculator.<br /> Using the calculator, I found z=6.00, i.e., chi-squre=36 (not your 35.67).

On 2023-02-07 14:43:14, user E Edlund wrote:

Interesting read. Have you seen this similar study, focused on stroke? https://pubmed.ncbi.nlm.nih...

On 2023-02-18 07:18:25, user Anikó Lovik wrote:

This preprint has now been published in an extended and revised version in PLOS One: https://doi.org/10.1371/jou...

On 2023-02-28 18:38:31, user Stéphanie C. Thébault wrote:

Link to published article:

PLoS One. 2023 Jan 12;18(1):e0278388. doi: 10.1371/journal.pone.0278388

On 2023-03-30 21:55:59, user Daniel wrote:

Can the authors please elaborate on the changes made, in particular the effect on TLR4 response?

On 2023-04-21 12:49:03, user antonia peros wrote:

I believe that the topic of your research is very important and current, however, I have several methodological objections. <br /> Although the authors pointed out the limitations, they made quite strong conclusions and recommendations despite too small, non-randomized sample and a cross-sectional design without a control group.<br /> The use of the used instruments is very questionable when it comes to recalling satisfaction, self-esteem, and reduction in depression from 6 months ago. I also think that the fact that the respondents were familiar with the purpose of the research could have contributed to the recall bias.<br /> An important factor in your research could be how long the subjects exercised, and you did not collect that data. What their target is in the research is also vaguely defined. I recommend including some more objective criteria for that.

On 2023-04-21 15:29:37, user Erin Jonaitis wrote:

This manuscript has now been published in Brain Communications. The final version is available at https://academic.oup.com/braincomms/article/5/2/fcad057/7070362.

On 2023-05-06 19:13:10, user Bret Nestor wrote:

Hello, If you are interested in a peer-reviewed subsequent work with a distilled message, please refer to our work in The Lancet Digital Health: https://doi.org/10.1016/S25...

On 2023-05-18 19:00:12, user Dave Fuller wrote:

Please add final peer-reviewed citation as:

Lin D, Wahid KA, Nelms BE, He R, Naser MA, Duke S, Sherer MV, Christodouleas JP, Mohamed ASR, Cislo M, Murphy JD, Fuller CD, Gillespie EF. E pluribus unum: prospective acceptability benchmarking from the Contouring Collaborative for Consensus in Radiation Oncology crowdsourced initiative for multiobserver segmentation. J Med Imaging (Bellingham). 2023 Feb;10(Suppl 1):S11903. doi: 10.1117/1.JMI.10.S1.S11903. Epub 2023 Feb 8. PMID: 36761036; PMCID: PMC9907021.

Thanks!!

CDF

On 2023-06-01 01:20:57, user Edmund Seto wrote:

This paper has been accepted for publication in the journal Science of the Total Environment under the title "Assessing the effectiveness of portable HEPA air cleaners for reducing particulate matter exposure in King County, Washington homeless shelters: Implications for community congregate settings"

On 2023-07-15 08:38:27, user readthecoran wrote:

Thanks for this paper. Your result may bne related to variance shrinkage of quantile averaging relative to equal weight combination of distributions. See https://onlinelibrary.wiley... For right skewed distributions medians tend to be lower than means

On 2023-08-08 10:45:14, user Robin Walters wrote:

Now published:

https://doi.org/10.1016/j.x...

On 2023-09-14 11:31:45, user Indi Trehan wrote:

This has now been published: J Glob Health 2023 Jun 9;13:04065. doi: 10.7189/jogh.13.04065.

On 2023-09-20 17:48:48, user ASH wrote:

Why did the authors investigate the associations between poultry fecal matters and E.coli, instead of other more poultry-specific zoonosis, like Salmonella? E. coli is commonly found in the lower intestine of warm-blooded organisms, of which most are harmless...<br /> Why didn't the authors check the DHS data? Similar data can be found in the DHS data which is publicly available.

On 2023-11-04 15:16:53, user Clive Bates wrote:

Two problems here.

First is scalability. This doesn't sound like an intervention that would engage many veterans, nor does it seem likely to be affordable or practical at the scale necessary to achieve a turnaround in the aggregate burdens arising from smoking.

Tobacco-related deaths exceed those resulting from homicides, suicides, motor vehicle accidence, alcohol consumption, illicit substance use, and acquired immunodeficiency syndrome (AIDS), combined.

Almost all of that excess mortality is attributable to smoking not nicotine. Tobacco harm reduction approaches may deliver more and sooner - e.g. encouraging migration to smoke-free alternative forms of nicotine use such as vaping.

Second, it is quite possible that veterans with forms of PTSD are benefiting in some way from the functional and therapeutic properties of nicotine. Again, an approach to smoking cessation that does not demand nicotine cessation may achieve nearly all the health benefits of quitting smoking without demanding withdrawal from nicotine use.

The trial could at least consider an additional arm to assess the utility of encouraging vaping for smoking cessation. It might achieve more for less.

On 2023-11-13 10:04:59, user Theo Peterbroers wrote:

"The duration from the day of index vaccination to the day of the survey completion was a median of 595 days (Interquartile Range<br /> (IQR): 417 to 661 days; range: 40 to 1058 days)."<br /> That is at least one participant vaccinated before the start of the pandemic.<br /> EDIT Make that one person from early in the vaccine trials. How time flies.

On 2023-11-15 16:42:12, user jhick059 wrote:

This article was published in the peer-reviewed journal PLoS One on October 30, 2023 (citation below), but the medRxiv page has not yet been updated to reflect the PLoS One publication.

Citation: Hickey J, Rancourt DG (2023) Predictions from standard epidemiological models of consequences of segregating and isolating vulnerable people into care facilities. PLoS ONE 18(10): e0293556. https://doi.org/10.1371/jou...

On 2023-11-27 21:08:47, user Judith Mowry wrote:

The recent paper on peripheral vasopressors by Yerke doi.org/10.1016/j.chest.202... is an important reference for your research. It is vital to note that they changed their protocol to add very specific protocols and rules regarding IV site inspection, defined who was responsible. Also note that an antecubital site (or any joint) is avoided to minimize movement and extravasation risk. I wish you success with your research.

On 2023-11-27 22:00:33, user Christos Proukakis wrote:

We note the results of the analysis, and look forward to submitting a detailed response soon.

On 2024-01-27 21:34:52, user kasia bera wrote:

now published here: https://pubmed.ncbi.nlm.nih...

On 2024-02-20 21:32:15, user Wally Wilson wrote:

It would be handy if the authors could get the abbreviations for Borderline Personality Disorder (BPD) and Bipolar Disorder (BD) corrected

On 2024-04-25 03:20:17, user Lena Palaniyappan wrote:

Very interesting work. We observed a similar 'amelioration' effect using a cross-sectional design a few years ago (Guo et al., 2016). Since then we made several cross-sectional and a few longitudinal observations supporting the possibility of compensation and reorganisation after first episode psychosis (Palaniyappan et al., 2019a; 2019b), including one with the largest untreated sample we could access at that time (Li et al., 2022). These observations compel us to spare more efforts to understand the compensatory processes in psychosis (Palaniyappan et al, 2017, Palaniyappan & Sukumar 2020, Palaniyappan, 2021; 2023).

Guo S, Palaniyappan L, Liddle PF, Feng J. Dynamic cerebral reorganization in the pathophysiology of schizophrenia: a MRI-derived cortical thickness study. Psychological medicine. 2016 Jul;46(10):2201-14.

Li M, Deng W, Li Y, Zhao L, Ma X, Yu H, Li X, Meng Y, Wang Q, Du X, Sham PC. Ameliorative patterns of grey matter in patients with first-episode and treatment-naïve schizophrenia. Psychological Medicine. 2023 Jun;53(8):3500-10.

Palaniyappan L. Progressive cortical reorganisation: a framework for investigating structural changes in schizophrenia. Neuroscience & Biobehavioral Reviews. 2017 Aug 1;79:1-3.

Palaniyappan L, Das TK, Winmill L, Hough M, James A, Palaniyappan L. Progressive post-onset reorganisation of MRI-derived cortical thickness in adolescents with schizophrenia. Schizophr Res. 2019a Jun 1;208:477-8.

Palaniyappan L, Hodgson O, Balain V, Iwabuchi S, Gowland P, Liddle P. Structural covariance and cortical reorganisation in schizophrenia: a MRI-based morphometric study. Psychological Medicine. 2019b Feb;49(3):412-20.

Palaniyappan L, Sukumar N. Reconsidering brain tissue changes as a mechanistic focus for early intervention in psychiatry. Journal of psychiatry & neuroscience: JPN. 2020 Nov;45(6):373.

Palaniyappan L. The neuroscience of early intervention: Moving beyond our appeals to fear. Australian & New Zealand Journal of Psychiatry. 2021;55(10):942-943.

On 2024-04-26 17:02:43, user Gary Goldman wrote:

We broadened our analyses (of IMRs) to explore potential relationships between childhood vaccine doses and NMRs (neonatal mortality rates) and U5MRs (under age 5-year mortality rates). Using 2019 and 2021 data, 17 of 18 analyses (12 linear regressions and six ANOVA and Tukey-Kramer tests) achieved statistical significance and corroborated the trend reported in our original study, demonstrating that as developed nations require more vaccine doses for their young children, mortality rates worsen. Please see https://pubmed.ncbi.nlm.nih...

On 2024-05-02 18:11:05, user Keith Robison wrote:

There is a great degree of interest in this preprint due to it being the first extended description of using the iCLR technology.

It would be very valuable to have details on how much Illumina short read data was generated from iCLR libraries and how much of that data contributed to the iCLR reads vs. what could not be used

It would also be valuable to report the read length distribution of the iCLR reads in greater detail - particularly since many interested parties cannot perform that analysis themselves on the clinical data sets

On 2024-05-28 13:49:19, user Jenn Ferris wrote:

This preprint is now published: https://www.neurology.org/d...

On 2024-11-08 19:59:59, user Andre Boca Ribas Freitas wrote:

Important Observations on Underreported Chikungunya Mortality in Light of Global Burden Analysis

Dear Authors,

I thoroughly appreciated your recent preprint on the global burden of chikungunya and the potential benefits of vaccination. Your work provides critical insights into the widespread impact of this disease and emphasizes the significant potential of vaccine interventions.

However, I wanted to highlight a critical issue that our research and that of others in the field have identified: the substantial underreporting of chikungunya-related mortality across many regions. While chikungunya is often categorized as a non-fatal disease, a growing body of evidence reveals severe and sometimes fatal cases that frequently go unrecorded by epidemiological systems. Our recent studies in Brazil documented excess mortality rates from chikungunya far surpassing those officially reported, with mortality rates up to 60 times higher than recorded by standard surveillance systems?Freitas et al., 2024?. Additionally, studies like those by Mavalankar et al. (2008) in India and Beesoon et al. (2008) in Mauritius underscore the elevated mortality associated with chikungunya during epidemic outbreaks, further reinforcing this critical gap in mortality surveillance.<br /> This growing evidence highlights the critical need for increased investment in molecular diagnostics, integrated surveillance, and more comprehensive mortality tracking for chikungunya. These measures are essential for aligning public health responses with the true impact of the disease and ensuring the full scope of chikungunya’s burden is addressed.

Thank you for advancing this essential conversation. Through improved surveillance and research collaboration, we can work toward effective strategies to mitigate the severe impact of chikungunya globally.

Best regards,

Dr. André Ricardo Ribas Freitas<br /> Faculty of Medicine, São Leopoldo Mandic, Campinas-SP, Brasil

Freitas ARR, et al. Excess Mortality Associated with the 2023 Chikungunya Epidemic in Minas Gerais, Brazil. Front Trop Dis. 2024. doi: 10.3389/fitd.2024.1466207.

Mavalankar D, Shastri P, Bandyopadhyay T, Parmar J, Ramani KV. Increased mortality rate associated with chikungunya epidemic, Ahmedabad, India. Emerg Infect Dis. 2008 Mar;14(3):412-5. doi: 10.3201/eid1403.070720. PMID: 18325255; PMCID: PMC2570824.

Beesoon S, Funkhouser E, Kotea N, Spielman A, Robich RM. Chikungunya fever, Mauritius, 2006. Emerg Infect Dis. 2008 Feb;14(2):337-8. doi: 10.3201/eid1402.071024. PMID: 18258136; PMCID: PMC2630048.

Manimunda SP, Mavalankar D, Bandyopadhyay T, Sugunan AP. Chikungunya epidemic-related mortality. Epidemiol Infect. 2011 Sep;139(9):1410-2. doi: 10.1017/S0950268810002542. Epub 2010 Nov 15. PMID: 21073766.

Freitas ARR, Donalisio MR, Alarcón-Elbal PM. Excess Mortality and Causes Associated with Chikungunya, Puerto Rico, 2014-2015. Emerg Infect Dis. 2018 Dec;24(12):2352-2355. doi: 10.3201/eid2412.170639. Epub 2018 Dec 17. PMID: 30277456; PMCID: PMC6256393.

Freitas ARR, Gérardin P, Kassar L, Donalisio MR. Excess deaths associated with the 2014 chikungunya epidemic in Jamaica. Pathog Glob Health. 2019 Feb;113(1):27-31. doi: 10.1080/20477724.2019.1574111. Epub 2019 Feb 4. PMID: 30714498; PMCID: PMC6427614.

On 2024-12-03 21:03:36, user xPeer wrote:

Courtesy review from xPeerd.com

This manuscript introduces DeepEnsembleEncodeNet (DEEN), an innovative polygenic risk score (PRS) model integrating autoencoders and fully connected neural networks (FCNNs) to address limitations of existing PRS methods. By disentangling dimensionality reduction and predictive modeling, DEEN enables the capture of both linear and non-linear SNP effects, improving prediction accuracy and risk stratification for binary (e.g., hypertension, type 2 diabetes) and continuous traits (e.g., BMI, cholesterol). Evaluation using UK Biobank and All of Us datasets highlights superior performance over established methods. While conceptually and methodologically compelling, areas such as interpretability, generalizability across diverse populations, and computational efficiency warrant further refinement.

Major Revisions<br /> 1. Interpretability and Practicality<br /> Black-Box Concerns: The complexity of the DEEN model limits its interpretability compared to simpler PRS methods like Lasso or PRSice. While the manuscript acknowledges this limitation, incorporating efforts to visualize model predictions (e.g., feature importance maps or SNP clustering analysis) would enhance its usability (Section: Discussion, p.16).<br /> Clinical Translation: The manuscript emphasizes the potential of DEEN for clinical utility but lacks discussion on the challenges of implementing deep learning models in healthcare. Addressing regulatory barriers and clinician engagement would add value (Section: Discussion, p.17).<br /> 2. Population Generalizability<br /> Demographic Bias: Both datasets used (UK Biobank, All of Us) consist predominantly of European-ancestry individuals. This limits the model's applicability to global populations. Expanding the discussion on efforts to improve DEEN’s cross-ancestry generalizability is essential (Section: Results, p.11).<br /> Validation Across Diverse Cohorts: While DEEN is validated on two datasets, additional external validations across non-European populations would strengthen claims of generalizability and reliability.<br /> 3. Comparative Analyses<br /> Missing Baseline Methods: Although DEEN is compared with multiple PRS methods, inclusion of additional machine learning benchmarks (e.g., gradient boosting models, convolutional neural networks for SNP effects) would better contextualize DEEN’s advantages (Section: Results, p.8).<br /> Risk Stratification Assessment: The risk stratification results are promising but need more rigorous evaluation metrics beyond odds ratios, such as net reclassification improvement (NRI) or integrated discrimination improvement (IDI).<br /> 4. Computational Efficiency<br /> Resource Requirements: DEEN’s reliance on high-performance computing resources (e.g., GPU usage) is noted but not sufficiently quantified. Providing benchmarks of computational costs and runtime against alternative methods is crucial for practical implementation (Section: Methods, p.19).<br /> Optimization: While grid search was used for hyperparameter tuning, exploring automated optimization frameworks (e.g., Bayesian optimization) could reduce computational overhead.<br /> 5. Data Filtering and Variant Selection<br /> Potential Bias from Variant Filtering: The preselection of SNPs based on p-values may exclude rare variants or those with small effects. A sensitivity analysis on SNP filtering thresholds would clarify the robustness of DEEN’s predictive power (Section: Methods, p.20).<br /> Minor Revisions<br /> 1. Typos and Formatting<br /> Figure Legends: Some figures (e.g., Figure 5) lack clear explanations of axes and statistical methods.<br /> Grammar: Line 124: Replace "similarly drive CRC progression" with "similarly drive progression."<br /> 2. AI Content Analysis<br /> Estimated AI-Generated Content: ~20-25%.<br /> Implications: Repetitive phrasing in methodological descriptions and literature summaries suggests potential AI assistance. While the technical content appears valid, manual rephrasing can enhance originality and scientific depth.<br /> 3. Statistical Reporting<br /> Insufficient Confidence Intervals: Odds ratio enrichment results lack 95% confidence intervals in several places, undermining statistical rigor (Section: Results, p.9).<br /> Inconsistent Metric Definitions: Terms like “improved R²” and “higher AUC” are used loosely. Precise numerical values and effect size comparisons would improve clarity.<br /> 4. Terminology Consistency<br /> Key terms like "dimensionality reduction" and "risk stratification" should be consistently defined and applied across sections to avoid ambiguity.<br /> Recommendations<br /> Enhance Model Interpretability:

Integrate explainability tools (e.g., SHAP values, visualization of autoencoder layers) to clarify how SNPs influence predictions.<br /> Discuss the potential for hybrid models balancing interpretability and performance.<br /> Address Demographic Bias:

Validate DEEN using datasets from underrepresented populations (e.g., African, Asian ancestries).<br /> Incorporate transfer learning techniques to enhance generalizability.<br /> Benchmarking and Evaluation:

Compare DEEN against additional advanced machine learning models for PRS.<br /> Introduce advanced evaluation metrics like NRI and IDI to strengthen claims.<br /> Refine Computational Analysis:

Provide detailed resource utilization benchmarks.<br /> Explore alternative hyperparameter optimization methods to improve training efficiency.<br /> Expand Data Analysis:

Perform a sensitivity analysis on variant filtering thresholds.<br /> Investigate the inclusion of rare variants to improve model robustness.

On 2024-12-20 20:46:20, user Jakub wrote:

You have stated: "We performed targeted metabolomics to quantify the absolute abundance of known uremic toxins, including (...) 4-ethylphenyl sulfate (4-EPS) (...) in plasma of this cohort. As expected, CKD and PAD+CKD groups had significantly higher levels of all these uremic toxins (Figure 3A)." Unfortunately, Figure 3A does not provide data on 4-ethylphenyl sulfate. May you add data on this solute?

On 2025-01-10 21:50:46, user Harold Bien wrote:

Fascinating article. Given that each individual VOC in Fig 1 appears to have significant overlap between each group and wide distributions, it would be interesting to learn how the various machine learning algorithms used each VOC and the resulting model. Could the authors provide more information on the ML algorithms used, how it was trained, and how the ROCs were constructed?

On 2025-03-17 20:22:07, user Adrian Barnett wrote:

A well planned and executed paper that highlights a serious and likely common issue in research practice.

On 2025-04-11 20:29:22, user Scott Olesen wrote:

Now published: https://doi.org/10.1016/j.vaccine.2025.127109

On 2025-10-09 02:52:59, user sid moose wrote:

I can’t tell, not a scientists here.. but did they test for whether or not the participants had the flue before the start of the study?

On 2025-10-15 20:49:15, user jpirruccello wrote:

This has been published; please see https://pubmed.ncbi.nlm.nih.gov/37019578/

On 2022-05-24 20:25:23, user Carol Taccetta, MD, FCAP wrote:

If a subject was still on immunotherapy at time of "recovery," the outcome of the adverse event cannot be considered as "resolved." It will also be important to follow these subject for relapse after therapy discontinuation, as immune-mediated conditions can sometimes relapse months, even years, after immunotherapy discontinuation.

On 2022-06-14 12:41:29, user Robert Clark wrote:

I was puzzled in Fig. 3 that the numbers for the severe cases was 39 for placebo and 51 for IVM. I thought this was measuring the comparative effect of ivermectin for the severe cases. But I see in the Supplementary appendix in eTable 1 that this is just giving the numbers in this category on entrance to the study.

But this raises another problem. For a randomized trial the number of severe cases assigned to the placebo and treatment groups should be close. Yet the IVM group got 24% greater number of severe cases. That’s a discomfortingly large difference for a randomized trial. Clearly this could create a bias against the treatment regimen.

Reviewing the further eTable 1of baseline symptoms, we see several categories of symptoms that would be key indicators of severe disease such as dypsnea, difficulty breathing, were assigned significantly more severe cases to the IVM group compared to the placebo group.

That shouldn’t happen in a randomized trial. I suspect something went wrong with the randomization. This could create such a serious bias against the treatment that a disclaimer should be placed on this study that its randomization procedure is being reviewed.

Robert Clark

On 2022-07-10 23:39:20, user Charles Warden wrote:

Hi,

Thank you very much for posting this preprint.

I consider the topics raised by this study to be important and interesting.

However, I have some comments and questions:

1) I agree that confirmation bias can be a contributing factor. However, I think true limitations in utility are also important. So, I am not sure if I completely agree with the statement "When results were not consistent with participant’s personal or family history, many participants found reasons to dismiss or discredit these results. This indicates a role for confirmation bias in responses to [self-initiated] PRS." For example, I might really want to understand the genetic basis for a disease, but the percent heritability explained by the PRS may be low and I could therefore be disappointed with the usefulness of a PRS due to a discordant result.

I have a blog post where I share my impute.me scores (along with others):

https://cdwscience.blogspot.com/2019/12/prs-results-from-my-genomics-data.html

I don't know if I would exactly say my response was "negative," but I certainly got the impression the PRS that I saw may have limited utility. In that sense, my view of the method was not positive, even if it did not evoke a strong emotional "negative" response.

Within that blog post, “ulcerative colitis” would be an example where there were different PRS for the same disease but very different percentiles (for the same SNP chip). So, I would consider that an example of the reaction that is described being due to something other than confirmation bias.

2) Did the interviewers respond when there were possible points of misunderstanding during the interview process?

It was acknowledged as a limitation in the discussion: "the researchers did not have access to participant’s PRS results and were unable to evaluate people’s understanding of their results".

However, it seems like that could be important. For example, there is a quote "Unfortunately, I do regret getting a PRS… I would have rather not known. I like uncertainty". Assuming that there were appropriate limitations to communicate, I believe a response from the interviewer might cause that quote to no longer reflect the subject’s opinion.

In general, there appears to be a noticeable emphasis on mental health in the article. My opinion is that this is an area where limitations are particularly important. If it helps, I think there are some additional details in this blog post for the book Blueprint.

In terms of my own impute.me results, I thought the "anxiety" PRS seemed reasonable (to the best of my ability to assess that). However, I also thought changes in conditions over time were important, and I thought there was potential for misuse.

3a) I think it is a minor point, but I don't remember receiving an invite to join a Zoom meeting for a discussion about my impute.me results.

I hope that I was one of the 209 candidates, but I was not sure if I could confirm that. I also noticed mention of categories like “medium” or “low” for one quote referencing a z-score of 2.5, but I only saw the continuous score distribution in the screenshots from my blog post.

3b) Perhaps more importantly, I tried to go back to sign in to check if I missed something.

In the Folkersen et al. 2020 paper, the link provided is for https://www.impute.me/. However, that link currently re-directs to a Nucleus website (https://mynucleus.com/).

Can you please provide some more information about the re-direction of the impute.me link?

For example, I submitted an e-mail to register on the new website, but I don't think I can see my earlier results anymore?

Additionally, I was confused when I couldn’t find the GitHub code provided with that paper: https://github.com/lassefolkersen/impute-me

4) Finally, but I don't think either of the 2 models that I see ("dismissed medical concerns" and "medical distrust") are a great description for myself. I think something like "curiosity" and "critical assessment" would be more appropriate for myself.

For example, I wouldn't say I distrust the healthcare system or medical research broadly, but I do think feedback and engagement is important. Thus, when I encounter problems, I submit reports to FDA MedWatch. Likewise, I contribute data/experience to projects like PatientsLikeMe.

Thanks Again,<br /> Charles

On 2022-07-11 04:26:35, user E Hansen wrote:

It would be useful if the authors could clarify if "unvaccinated" means "never injected", or if this group also includes subjects not yet defined as vaccinated, but who has received a vaccine within the last week/14 days. <br /> The same goes for the vaccinated group; does it include everyone who received a shot from the day injected, or only those who have passed the first 14 days after injection and then being consideres "vaccinated"? This was not entirely clear to me, anyway. <br /> Thank you

On 2022-07-29 09:25:01, user Dr. D. Miyazawa MD wrote:

Please also refer to previous studies.

Hypothesis that hepatitis of unknown cause in children is caused by adeno-associated virus type 2 (08 May 2022)<br /> https://www.bmj.com/content...

Daisuke Miyazawa. Possible mechanisms for the hypothesis that acute hepatitis of unknown origin in children is caused by adeno-associated virus type 2. Authorea. May 16, 2022.<br /> DOI: 10.22541/au.165271065.53550386/v2

On 2022-08-06 10:16:33, user Jef Baelen wrote:

No inclusion/exclusion criteria are mentioned? It includes a study from 2004, long before SARS-CoV-2 emerged. The Caruhel study was not performed on COVID-19 patients. The çelebi study used a cycle treshold cut-off of 38 and evaluated 20 parameters of which masks was only 1. The results of this study were wrongly extrapolated in table 1. Very dubious studies included in this meta-analysis!

On 2022-08-09 02:04:53, user Roel Ceballos wrote:

This is already published. The link to the publication is this https://mjst.ustp.edu.ph/in...

On 2022-08-09 12:40:13, user PhillyPharmaBoy wrote:

The authors conducted a thorough evaluation of the impact of ivermectin on SARS-CoV-2 clearance. On the surface their results differ from those of Krolewiecki, et al. (below). In a post hoc analysis these investigators found that ivermectin accelerated viral decay when drug concentration (4 hr) exceeded 160 ng/ml. It would be useful for the PLATCOV Group to mention this study and discuss potential reason(s) for the discrepancy.

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

On 2022-08-14 18:37:59, user Jason wrote:

Interesting manuscript, can see how it can easily be used in practice looking at the predictive variables chosen

On 2022-08-17 07:08:46, user Partho Sen wrote:

This paper is now published in iScience, Cell Press. Here is the link of the journal pre-proof. https://www.sciencedirect.c...

On 2022-08-29 13:47:58, user Titos wrote:

This was published in NEJM as https://www.nejm.org/doi/fu...

On 2022-09-14 16:05:23, user Roy Miller wrote:

There may be more evidence for this immunity than previously thought, Since Queen Elizabeth's death on September 8, 2022 the half-dozen British Royals have mingled with countless people, shaking hands, touching random surfaces, and breathing air all over the Scotland, England, and northern Ireland. I assume the Royals all have had their COVID injections and I have to assume that they have come in contact with the COVID virus on numerous occasions. Current thinking would lead me to assume the crowds gathered to see them would make transmission of the virus more likely.

So why hasn't COVID spread like wildfire over the grieving population? Perhaps it is too soon to tell. But since COVID symptoms appear 2 to 14 days after infection, at least a large uptick in the COVID infections should be noticeable by now But no such event has occurred according to the British Press. So I have to conclude that there are additional factors to consider such as the one postulated by this article.

PS: I have no medical training although my job involves helping the medical community take advantage of high technology.

On 2022-09-22 09:16:27, user Wera Pustlauk wrote:

Dear authors,

thanks for the valuable effort to set up a new assay for the determination of PPi.

Regarding table III samples remained somewhat vague to me. Clarification in the table header including the unit of the determined PPi might be helpful. Calculation of the standard deviation in addition to the average would make the data more roboust and would establish a more substantial link to the variability discussed in the paragraph before. Moreover, time differences in the addition of EDTA to the CTAD tubes as discussed in the text should be clearly stated for each sample in the table III as well.

In addition, a hands on protocol (stored in a repository or as supplement) allowing the direct usage of the assay based on the optimized procedure would make the usage more accessible.

Best regards,<br /> Wera Pustlauk

On 2020-04-24 13:43:14, user Anand wrote:

Congratulations for leading the systematic review team. Great effort highlighting long term outcomes and rehabilitation needs.

On 2020-04-24 15:03:58, user Iba net wrote:

I am Masum Billa from Bangladesh. Now situation going to volcano speed, we don't know what will be happen in the next but our government cannot control all of us meanwhile how is it possible to serve them. We are trying our best to serve them in camp. If anybody wants research or visit Rohinga Camp. Please contact me billaasia@gmail.com

On 2020-04-24 16:08:53, user Rajendra Kings Rayudoo wrote:

To<br /> Prof.Dankmar Boehning<br /> I'm.from india and IAM f glad to hear a realistic approach to realtime infectious cases including asymptomatic and presymptomatic.and interested to know how it works<br /> Please explain how can I calculate the cases in india through your model .<br /> Regards <br /> Rajendra

On 2020-04-24 20:55:11, user wangkon936 wrote:

The issue with a "retrospective" study is that populations cannot be randomized. It is clear just looking at the data that the vitals and the biochemistry of the hydroxlchloroquine ("HC") or hydroxlchloroquine and azithromycin group ("HC + AZ") was inferior vs. the non HC or HC + AZ group. In other words the HC and HC + AZ groups were significantly unhealthier vs. the the non HC or HC + AZ groups. A randomized true clinical trial would have filtered this bias out, but a retrospective study structurally solidified this bias in. Thus, this study has a structurally solid bias that render's its ultimate conclusions suspect and of limited use.

On 2020-04-25 08:48:51, user Huan Mo wrote:

Of course the patients who were treated with hydroxychloroquine were sicker at the baseline. The no-hydroxychlroquie patients have tons of missing values in basic labs and apparently they are mostly mild and outpatients.

This study is like saying pneumonia treated in tertiary medical center has worse outcome than treated in an outpatient clinic.

On 2020-04-25 22:54:50, user wbgrant wrote:

This open access article supports the present modeling study and should be cited in the final version:<br /> Low Temperature and Low UV Indexes Correlated with Peaks of Influenza Virus Activity in Northern Europe during 2010?2018.<br /> Ianevski A, Zusinaite E, Shtaida N, Kallio-Kokko H, Valkonen M, Kantele A, Telling K, Lutsar I, Letjuka P, Metelitsa N, Oksenych V, Dumpis U, Vitkauskiene A, Stašaitis K, Öhrmalm C, Bondeson K, Bergqvist A, Cox RJ, Tenson T, Merits A, Kainov DE.<br /> Viruses. 2019 Mar 1;11(3). pii: E207. doi: 10.3390/v11030207.<br /> http://www.mdpi.com/resolve...

On 2020-04-25 23:09:56, user wbgrant wrote:

One more publication in support<br /> Environmental predictors of seasonal influenza epidemics across temperate and tropical climates.<br /> Tamerius JD, Shaman J, Alonso WJ, Bloom-Feshbach K, Uejio CK, Comrie A, Viboud C.<br /> PLoS Pathog. 2013 Mar;9(3):e1003194. doi: 10.1371/journal.ppat.1003194. Epub 2013 Mar 7. Erratum in: PLoS Pathog. 2013 Nov;9(11). doi:10.1371/annotation/df689228-603f-4a40-bfbf-a38b13f88147.

On 2020-04-27 04:52:01, user Krishna Undela wrote:

It is the first information on knowledge and beliefs of general public of India on COVID-19. In this article we can understand the false beliefs / myths circulating among the general public of India about transmission of novel coronavirus and prevention and treatment of COVID-19.

On 2020-04-27 10:34:04, user Elena Sharova wrote:

They don't randomize health care workers to find an effect in comparable conditions (hospital, patients management, patient burden). Jan21 to Feb23 - incubation period before pneumonia detection is approximately 1-2 weeks. So how to compare pneumonias in the 1st week in control group with abcence in test group starting interferon - regurding the 1st week test group DO NOT infect a 1-2 week ago, not during the research. Are these groups equal?

On 2020-04-27 16:44:44, user Dr. Amy wrote:

Obesity increases the density and upregulates ACE2 receptors. I couldn't figure out how women were protected given the higher incidence of obesity, but a Japanese researcher in this area shared this with me (which I do not fully understand.) "ACE2 is not only an entry receptor for the SARS-CoV-2 virus, but also protects against the pathogenic effects of RAS and the ACE/AngII/AT1R axis. I believe the balance between ACE/AngII/AT1R axis and ACE2/Ang1-7/MasR axis is important. Higher expression of ACE2 in old female rats than male represents protective effects. That's why women would have lower morbidity and mortality. This is my speculation."

On 2020-04-27 19:52:19, user eldhose poulose wrote:

You mentioned that you colllected the airline data, I see only airline data for the year 2017, February. can you provide the latest data? Also can you give more details on the data that you used for the study?

On 2020-04-27 21:36:26, user Marc Bevand wrote:

Page 2: NYDOC should be NYDOH

On 2020-04-27 23:05:18, user Jink wrote:

Hope the higher Ct's are due to low yield from filters. Having a synthetic standard/control would have justified the Ct numbers. Think about it if expanding the sample size.

On 2020-04-30 02:42:13, user Tyler Chen wrote:

I appreciate the authors’ urgency in addressing SARS-CoV-2 decontamination for reuse of N95 filtering facepiece respirators (FFRs). In the spirit of that urgency and health impacts, I note two concerns with the current preprint that could accidentally cause confusion: (1) The paper claims N95 filtration is preserved after microwave-generated steam, whereas the test listed in the methods was a TSI quantitative fit test, which is primarily designed to test fit, not necessarily filtration. (2) The paper’s claim of a universally accessible N95 decontamination protocol may accidentally overstate the N95 models for which this protocol is verified. N95 models vary widely in their construction and resistance to steam heat, so any models other than the one used in this experiment will likely require thorough testing before this method is applied.

I would suggest that the authors make the following changes:<br /> (1) Clarify whether or not filtration is verified at larger particle sizes and charge (e.g. 0.26 microns, uncharged). If filtration is not yet verified at larger particle sizes, this test may be important to verify N95 performance following microwave steam treatment.<br /> To provide some background: The TSI 8026 Particle Generator generates 0.04 micron particles [1] that are intended to be readily filtered by the N95, and are assumed to only enter the N95 through gaps in the face seal and not through the mask material itself [2]. It is possible for N95 FFRs to pass quantitative fit tests while still failing filtration tests at different particle sizes--one study “observed [protection factors] <100 even for subjects who passed fit testing (fit factor > 100)” [3]. Therefore, fit testing using the 8026 particle generator does not imply that N95 filtration is necessarily preserved at larger particle sizes which are most relevant for filtration effectiveness in the SARS-CoV-2 pandemic, especially given the fact that the decontamination treatment may shift what particle size is most penetrating for the N95. Recent non-peer-reviewed research shows N95 material suffering a decrease in filtration of 0.26 micron particles after 4-5 cycles of 10min stovetop steam treatment [4], though it is unclear from this manuscript if the MGS treatment did not reach this limit, or that the limit was not observed due to the different particle size used. Therefore, testing for quantitative fit should perhaps be supplemented by filtration tests at larger particle diameters (whether from this study or by citing others), especially when a decontamination process is involved such as steam heat that has an unknown effect on the most penetrating particle size for the N95. Given the potential for widespread implementation of this protocol it seems important that this point be clarified.

(2) Secondly, it may be important to notify readers that the performance of the N95 model in this paper likely cannot be generalized to all N95 models without further testing. N95 models vary widely in construction and resistance to steam, and each model should be individually verified to maintain both fit and filtration by this protocol before use. This is supported by the fact that N95 models can vary widely with respect to the most penetrating particle size, and each country may only have access to certain N95 models [3,4,5]. Furthermore, there is literature evidence that the mask performance in response to steam and heat also varies across N95 models (see the table of results in Appendix B https://www.n95decon.org/fi... "https://www.n95decon.org/files/heat-humidity-technical-report)"). Therefore, it is important to clarify in the text that this method is not yet universally-validated -- N95 fit has only been verified for the 3M 1860 molded N95 in particular, and other models are likely to have significantly different behavior. Independent verification of both fit and filtration may be needed for other N95 models.

On 2020-04-30 04:49:44, user Aaron wrote:

It would be a useful addition to show the % positive of tests performed in the same range as Fig 1C. If a greater proportion of tests are coming back positive following April 7, that could suggest that testing is occurring more and more frequently among those most likely to be positive. This is problematic as it can obscure mild cases that may not appear to be positives. In general, this information would help contextualize what types of cases are being captured by these data in Wisconsin, mostly severe cases, or mild and severe cases alike.

On 2020-04-30 21:06:27, user Tim Lawes wrote:

A great paper by very respected researchers and clinicians, ruined by a bizarre press-release saying COVID-19 'as deadly as Ebola'. Let's take a fact-check on that one:<br /> 1. Case fatality rates not equivalent: Ebola CFR 50-75% in recent outbreaks, not 33% as in COVID-19. If referring to high-income country (HIC) stats only, talking handful cases treated in HIC with Ebola, all working age occupationally fit health workers, CFR ~18% (n=5/27 quoted).<br /> 2. Totally different age of death: Ebola median 30-35 yrs, COVID 80yrs. Ebola 95% deaths <60 yrs, Covid ~10% < 60yrs.<br /> 3. Translate age at death to Years of Life Lost (YLL): comparing age at death profile provided by Doherty et al to Ebola papers, each Ebola death 'costs' 45 YLL, vs. each COVID death costs 1.4 YLL.<br /> 4. Translated to population impact. To exceed an equivalent burden of YLL per capita in West Africa in 2013-16 would require >1 million deaths in UK from COVID-19.

To compare the mortality profiles of Ebola and COVID-19 is at best bad science, at worst an example of misinformation that perpetuates global health inequities. I don't imagine authors intended this, but I'm afraid its this sort of comment that creates hysteria, rather than appropriate responses. As a paediatrician we are seeing children coming to harm from avoiding hospital and not being seen in community due to misjudged risk assessments. A genuine thank you for your contribution to science, but please ensure reporting is responsible.

On 2020-05-14 10:44:33, user Dom Mcelhinney wrote:

In the at risk group for covid19 almost 50% of population are being treated for Hypertension and Hypercholesterolaemia. Can you explain why you have been unable to list this as a possible comorbidity.

On 2020-05-01 15:18:00, user Philip Machanick wrote:

Seems prior comments have been removed? Your S Korea story does not mention extensive contact tracing, a key part of the strategy.

On 2020-05-02 01:27:06, user fennudepidan wrote:

Notice: In the revision on April 24, 2020, we have updated our analysis using data up to April 22, and importantly in which we have adjusted for additional confounding factors that also reflect the timing of the epidemic's spread, the timing of the social distancing policies and the population age distribution. Consequently, we revised our finding as that an increase of 1 ug/m3 in PM2.5 is associated with an 8% increase in the COVID-19 death rate (95% confidence interval [CI]: 2%, 15%).

On 2020-05-02 11:10:41, user Enrique Povo wrote:

I don't think the use of linear regression is warranted to see the trend of your gamma function.

On 2020-05-02 11:52:48, user Malcolm Lightbody wrote:

Hi, just wanted to say that this is an interesting piece of work. I think there is a typo in equation 1 - there should be a -ve sign in the exponent - as in c(t) = exp( - (t-T)^2 / tau^2).

On 2020-05-03 00:49:59, user chess wrote:

Pr Raoult claimed that HCQ had an anti viral effect at the onset of the infection.<br /> So HCQ would also have an effect on cytokin storm later ?<br /> It's a fact that mortality rate 0.4% is very low at IHU_Maseille and someday we will discover why.

On 2020-05-05 21:07:04, user Bozo wrote:

Please review this preprint...figure 2 in the appendix says it all.

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

On 2020-05-05 22:35:44, user Alan Bell wrote:

"If worse respiratory health and aggravated symptoms in polluted areas are the main channels of action, higher COVID-19 case hospitalization rates should also be expected in these locations."

Not sure that follows:<br /> If two locations each have 150 cases, some of which may remain asymptomatic and undetected.<br /> Location A with low pollution detects 80 cases, 4 of which are hospitalized.<br /> Location B with higher pm2.5 levels detects 100 cases, 5 of which are hospitalized.<br /> Location A and location B have the same hospitalization rate of 5% of detected cases but due to aggravated symptoms location B has a higher detection rate.

The analysis in 6.2.5 is interesting but I am not sure it is conclusive. Maybe random antibody sampling across the whole population will eventually reveal more information.

On 2020-05-06 15:17:26, user Sinai Immunol Review Project wrote:

Summary: Based on peripheral blood samples from 117 COVID-19 inpatients and convalescent patients, the authors demonstrate that all patients sampled became seropositive with neutralizing antibodies within 20 days since onset of symptoms and stayed seropositive until day 41-53. Seropositivity of neutralizing antibodies was defined as a geometric mean titer (GMT) of 1:4 higher and the titer was calculated using a modified in vitro cytopathogenic assay where the dilution number of 50% protective condition from cytopathic effect of the virus represented the titer. The GMT of neutralizing antibodies (average: 1:271.2) was the highest at 31-40 days since onset and multivariate Generalized Estimating Equations (GEE) model controlling for clinical variables (i.e. gender, age, clinical severity, etc.) showed that antibody titer at 31-40 days was significantly higher than 10-20 days past onset. In addition, their multivariate GEE analysis showed that age- and clinical severity-dependent rise in antibody titers with the youngest (age 16-30) and patients with mild or asymptomatic conditions having a lower antibody tier than its elderly and moderately-sick counterparts.

Limitations: Several shortcomings limit the impact of this study. While it has been the intent of the authors to sample PBMCs from patients at various time points in order to establish a robust profile of antibody response against SARS-CoV2, in reality, sampling has been limited and inconsistent across different time points. For instance, PBMCs of only 12 out of 117 patients have been collected three or more times and it is not clear from the data whether samples from patients whose blood has been collected only once (n = 37) are evenly distributed across the time frames under analysis. Furthermore, the authors have tried to show differences in kinetics of antibody response between patients with mild and moderate conditions by sampling their blood at four different time points. However, not only do two of eight patients sampled in this study have only two data points, but also the authors have found that the antibody response varies considerably across individuals—further underscoring the need to have PBMCs sampled from each patient at multiple time points and normalizing their response before comparing the titers across individuals. In addition, due to the fact that patients were enrolled using convenient sampling instead of random sampling methods, it’s evident that the authors could not control for disease severity as they only had four patients in severe condition. Beyond the sampling issues, the modified cytopathic assay used to calculate the neutralizing antibody titers may be less sensitive and specific than ELISA-based assays that use purified antigens from the virus.

Significance of the finding: Limited. While it is informative to have descriptive studies like this one showing the dynamics of the antibody response against COVID-19, the failure of the study to collect samples in controlled manner prevents the reader from using the data to answer key questions regarding the humoral immune response against COVID-19: do differences in clinical severity manifest in different kinetics of antibody response? When controlled for age, is higher antibody titer predictive of their clinical severity and prognosis? Future studies may address those questions with more controlled experimental setup.

Review by Chang Moon as part of a project by students, postdocs and faculty at the<br /> Immunology Institute of the Icahn school of medicine, Mount Sinai.

On 2020-05-06 20:06:21, user MS wrote:

Need to be clear on "false-negative RT-PCR". This study is looking at the presence of virus and the viral load in the upper respiratory tract through-out infection as much as it looks at sensitivity and is dependent on a deep sample taken correctly using appropriate sampling kits as well as the sensitivity of the test.

On 2020-05-06 20:25:49, user Frank Conijn wrote:

A furthermore well-written paper, in which in particular the section about the used dosage in vitro and in vivo is interesting.

But two crucial questions are not answered:

1. What determined that the HCQ group got it and the other group not? It's a single-hospital study, so that shouldn't be difficult to answer.

2. What other drugs, if any, did the groups get? HCQ has a strong antiviral effect in vitro, but in vivo seemingly also an immunomodulatory one, since it's effective in rheumatoid arthritis. So, several interactions may be possible.

On 2020-05-07 21:23:21, user Dan T.A. Eisenberg wrote:

How much spit did you collect and how much of it + the PBS was needed for sufficient sample for extraction. Thanks!

On 2020-05-08 14:37:27, user Merilee Brockway, PhD RN IBCLC wrote:

I think that you need to consider the possibility of retrograde flow contaminating the breastmilk from the infant's saliva. Sicker infants would likely have a higher viral load in their saliva/respiratory secretions. A study in the Lancet found that "The mean viral load of severe cases was around 60 times higher than that of mild cases, suggesting that higher viral loads might be associated with severe clinical outcomes." https://www.thelancet.com/j.... This may help to explain why the virus was present in the milk of mother 2, but not mother 1. Infant 2 was much sicker than infant 1 and as such the viral load was likely much higher.