April 12th, 2020

IDSA’s COVID-19 Treatment Guidelines Highlight Difficulty of “Don’t Just Do Something, Stand There”

Winner, 1923 White House Easter Egg Roll

The Infectious Diseases Society of America (IDSA) gathered a series of experts for what were undoubtedly many late-night calls, reviews of published and pre-print literature, and revisions (of revisions), and admirably generated a set of treatment guidelines for COVID-19.

The problem — there is no proven effective treatment for COVID-19.

That is, there’s no proven treatment based on our usual highest standard metric for efficacy, the randomized clinical trial — nor the next-best thing, a carefully done observational study that meticulously accounts for potential confounders.

Which means these guidelines have a Groundhog Day-like quality. In a series of clear and comprehensive sections, they review the available evidence, then repeatedly conclude the same thing:

The IDSA guideline panel recommends the use of [insert putative COVID-19 treatment here] in the context of a clinical trial. (Knowledge gap.)

Well, not exactly the same thing — for some of these treatments they insert the word “only”, yielding “… only in the context of a clinical trial.”

Here’s the difference between the two, according to the lead author:

For interventions with certainty regarding risks and benefits, the expert panel recommended their use “in the context of a clinical trial”. The guideline panel used “only in the context of a clinical trial” for interventions with higher uncertainty and/or more potential for harm.

But the message is clear. We don’t have sufficient evidence now to recommend any specific treatment.

That’s right — for chloroquine/hydroxychloroquine (HCQ), hydroxychloroquine with azithromycin, tocilizumab, corticosteroids for acute respiratory distress syndrome (ARDS), lopinavir/ritonavir — all are readily prescribable by clinicians (each is FDA-approved for other indications), yet none is proven to work for COVID-19.

That might be hard to believe given the publicity surrounding some of the approaches, in particular hydroxychloroquine. But those are the facts as of today.

So where does that put clinicians on the front lines managing this new disease?

Highly conflicted.

Those favoring the use of hydroxychloroquine for COVID-19 say that there is at least some evidence that hydroxychloroquine helps — enough so that controlled studies are ongoing. We certainly don’t have anything else to offer, and people are sick!

Plus, there’s a plausible mechanism of action, with in vitro antiviral activity. Maybe even two mechanisms if we consider the anti-inflammatory effect.

In addition, there’s this comment, posted by a critical care specialist in response to my poll:

Yes. No idea if it works, but it’s plausible, and it’s part of the YNHH [Yale New Haven Hospital] treatment algorithm for now. Wonder if those on the only-clinical-trials high horse have ever prescribed Haldol for agitated delirium?

High horse, ivory tower, unconnected to “real practice” — these are common charges levied at academic medicine, with some justification. Certainly not everyone has access to clinical trials.

And even when clinicians do have access to these studies, not all patients meet inclusion criteria, and some others might choose not to participate.

Indeed, at our hospital — which, like many academic medical centers, both has clinical trials for COVID-19 and prides itself on following evidence-based medicine — approximately a third of our COVID-19 cases have received hydroxychloroquine.

(Thanks to our crack ID PharmD Jeff Pearson for the quick data review.)

But why just a third? Why not all of them?

Let’s take up the nay-sayers view. They cite the weakness of the data. One study was, on further scrutiny, so flawed the journal publishing it raised concerns about the low quality of the study.  How often do we see that?

Another trial has not yet been published in a peer-reviewed journal, was quite small, and showed improvement in some minor endpoints only — with tremendous heterogeneity in other treatment approaches.

Furthermore, people already receiving hydroxychloroquine for rheumatologic indications have already acquired COVID-19 — how effective can it be? Plus, there’s an abstract of a study (inadvertently circulated before publication) that not only shows no benefit, but also suggests harm.

If we have questions about clinical benefit, all must acknowledge that any treatment can cause harm. Of particular concern with hydroxychloroquine for elderly patients — those at greatest risk of severe COVID-19 disease — is QT prolongation, a problem worsened with concomitant azithromycin, many other medications, and underlying heart disease.

Is it any wonder the poll results are so split? This is a real tough one.

Often in such circumstances, it’s helpful to ask what one would do for a loved one — or yourself — if having to make the decision.

Personally, I would not take hydroxychloroquine for COVID-19, concerned about side effects and not so sanguine about its potential antiviral activity. The road is littered with drugs that have in vitro activity for respiratory tract infections, and yet do nothing when given to people.

In fact, our list of effective antiviral treatments for these infections is very short! For common respiratory tract infections, we have only the influenza drugs — and even with the flu, some argue the benefits are marginal.

I do understand the opposite view. I would listen carefully to a patient who strongly wanted treatment and go forward with prescribing it, provided they understood the risks and there were no contraindications.

But a well-designed clinical trial? Sign me up. We’ve got to learn more about this disease, and fast.

So take it away, Mabel and Olive. You two are giving me great pleasure at a time when we all really need it.

74 Responses to “IDSA’s COVID-19 Treatment Guidelines Highlight Difficulty of “Don’t Just Do Something, Stand There””

  1. Unk says:


  2. Matt Vail says:

    Yes, this is all crazy. Prescribing hrdoxychloroquine etc breaks what I have prided myself on for 20 years: evidence-based medicine.

    The problem is a significant number of these patients deteriorate so quickly from a respiratory standpoint. You fell like you have to “try” something.

    Based on experience with severe influenza hard to imagine finding benefit to any of these “meds.”

  3. Dr U N Thakur says:

    In spite of lot of deliberations, no consensus is emerging either in favor or against use of Hydroxychloroquine, reason being the lack of large number of randomized clinical trials. Mode of action of HCQ favors its use.

  4. Dr U N Thakur says:

    In spite of lot of deliberations, no consensus is emerging either in favor or against use of Hydroxychloroquine, reason being the lack of large number of randomized clinical trials. Mode of action of HCQ favors its use and I am in favor of its use particularly in early stage of Covid-19 infection

  5. Dr U N Thakur says:

    I am in favor of just do something.

  6. Liz Jenny-Avital says:

    1. Some of the academicians who preach the gospel of RCTs have themselves used HCQ/AZI in desperation.
    2. The preference for RCTs by academics who do not venture to the bedside is misguided by a lack of understanding of the natural history of the disease.
    3.Raoult’s observational cohort data would be more persuasive if we knew the natural hx better—we don’t know how often mild disease progresses to life-threatening disease.
    4. My clinical colleagues from across disciplines have discovered the importance of thrombosis to the pathogenesis of the disease. AC is being used more.
    5. Penicillin was not tested in any RCT against syphilis. Experienced syphiloogists knew it worked.
    6. Clinicians deserve respect.

  7. John Clark MD says:

    Would we even be having this discussion if hydroxychloroquine etc hadn’t been pushed by the president and Dr. Oz? If that is the case, then we might as well add zinc and a “My Pillow” because they’re coming next.

    • colleen says:


    • Alfredo Blanco says:

      It was first “pushed” by one of the most important virologist of France.
      I do not understand the resistence of people that just oppose whatever Trump says.
      If I am a patient, and dying, I’ll take anything that gives me a chance.

  8. Bernhard Lämmle says:

    From what I read here: Obviously no clearb benefit of any anti-viral treatment or antiinflammatory strategy. I wonder whether you do not mention antithrombotic Prophylaxis/treatment as a probably highly important measure to be applied in severe Covid-19 cases.

  9. Paul Cook says:

    Don’t just do something; stand there!

    Sometimes doing nothing is the right thing. The problem is that we don’t know when it is best to let nature take its course.

  10. Lily says:

    I am a geriatrician practicing in Thailand. Over the last month my colleagues and I are hearing over and over anecdotally from our ID and pulmonology colleagues that early use of Favipiravir seems to be making a difference. (They are using it added on to HCQ +/- Lopinavir/Ritonavir). I don’t see it being mentioned in the US or Europe at all.

  11. Frank Domino says:


    Thank you for this review and adding your perspective. Would i take hydroxychloroquine if I had mild symptoms: No. Would I want my daughter treated with it while on a vent with ARDS? Most likely. for the latter question, should steroids be used for ARDS seems to have its controversy as well. I hope my peers will be responsible with what they prescribe and take for patients in the community. You may believe the “benefit outweighs the risk” but you would be wrong. And I leave those who practice critical care medicine to decide what to do with the critical patient in front of them.

  12. Yehuda Lebowicz MD says:

    It was part of most institutional guidelines in China, Italy, and the “initial hotspots” in the US.
    Almost everyone of them at minimum had it as a consideration for those that were deteriorating.
    It seems Only when The President started pushing it, did it become a political hit button!

  13. MT says:

    Wondering if this drug is being shoed away because it’s cheap and effective, at early stage..For late stage, no drug might turn out to be beneficial even in RCT given the mortality in sick patients. .

  14. Charles Lawson says:

    The data I have seen shows good results on mild to moderate disease. It shows poor results on late stage crashing disease.
    That is not unusual as most treatments that work well with early infections , but work poorly with bacteremia and sepsis.
    I have decades of experience with plaquinil. I believe the concern about retinal disease to be inappropriate as it is generally a problem after years of treatment .

    • David K says:

      “The data you have seen” is the same data the rest of us have seen. In my opinion it doesn’t show “good results.” It’s simply uninterpretable.

  15. Peter H Proctor says:

    It is true that the only antivirals for treatment of respiratory illness are the antiflu drugs and those don’t work all that well. However, we do have a myriad of agents that affect the symptoms of viral respiratory illness. Assuming they work, I strongly suspect that hydroxychloroquine and azithromycin belong to this group and not to the antivirals per se. Just like, e.g., low molecular weight heparins, which are generally conceded to be helpful in Covid-19.

  16. Arkan Shammari says:

    The question is not only “ do we cause harm by using hydroxychloroquin?” We should question its benefit on deteriorating patients when myriad of factors play role.

  17. Alex Chen says:

    From what I’ve seen so far, hydroxychloroquine works just about as well as Xigris in ICU patients.

  18. Dr. MFM says:

    Perfectly said. Thank you.

  19. Michael P Dailey says:

    You have missed the risk benefit ratio issue. In my unscientific survey, as many as 80% do not make it off the ventilator!!! That has to be factored into any equation when you discuss this with a patient.

    • Fran Ferder, Ph.D. says:

      It seems critical to know why so many (80%) do not make it off the ventilator. What are their similarities? Is it underlying disease factors, age, etc.?

  20. Anibal Maldonado says:

    Very well said

  21. S d seltzer md says:

    I bet every single one of the No votes would take it themselves or prescribe to a family member who tested positive. Those who still deny it are simply not being truthful

    • Alex Chen says:

      I wouldn’t have taken it last week outside a clinical trial, and am even less inclined today.
      Would I give it to my grandmother? Definitely not.

  22. Richard L Gruenewald says:

    Great article on the physician’s dilemma. However, in Kentucky, the Kentucky Board of Pharmacy has been rather dictatorial; the board will not release any of the supplies of hydroxychloroquine for the physician to prescribe unless it is for previously prescribed rheumatoid arthritis or SLE. I do not know what the pharmacy board is waiting for or why they would interfere with the physician-patient options. The decision should rest with the prescribing physician and the patient. I have to wonder if the pharmacy board members would consider options for themselves or family members if push came to shove.

  23. Keith Klein MD says:

    RCT help but are not necessary if the evidence is overwhelming. We do not need one to justify using penicillin in pneumococcal pneumonia as its effects are dramatic. We don’t see that response with .HC, it is far more ambiguous and there may be greater deaths from prolonged QT intervals. I suspect if RCT did show efficacy, it would not be so dramatic as to alter the course of the disease. If you need statistics to prove something, it is probably not that efficacious or the effect would be more striking

  24. Jason Wilson says:

    There are many people advocating hydroxychloroquine on the basis that it might help (and downplaying risks of harm) and others who avoid corticosteroids on the basis it might harm (and ignoring the chance of benefit). Often they are the same person.

    If you list the solid evidence and biologically plausible mechanisms for/against there is no good reason for the different approaches.

    People find it hardest to do nothing but much easier to join the “we must do something” or the “precautionary principle” groups.

  25. ID PharmD says:

    The nomenclature within the IDSAs guidance seems heavily edited by their legal team.

    The actual risk of QTc prolongation is considered on a daily basis as we prescribe it empirically for our CAP pts. We know this drug , the risk of QTc prolongation is overinflated. The agenda for an expensive solution seems evident.

  26. Atul bhatia says:

    HCQ is time tested for malaria prophylaxis and well tolerated . Has demonstrable in vitro activity against Covid-19
    I don’t see any rationale not to prescribe a well tolerated medicine for prophylaxis for 7 weeks for those deemed as high risk I.e health care workers and close family contacts of patients . Based upon the ICMR recommendations in India , multinational trails for HCQ Prophylaxis have already started with Oxford Univ taking the lead .
    Dosing schedule is not different than what CDC recommends for malaria prophylaxis. Except for day one dose is 800 mg rather than 400 mg and then 400 mg once a week for 7 days
    I would support such an approach for heath care individuals on the front lines of the covid epidemic . There is practically no downside to this approach pending results of the studies .

  27. Courtney Thomason says:

    I guess inflammatory panels are being done to look at markers such as IL-6, CRP, Did you see high inflammatory markers and if you did, do something to reduce the inflammation? What does the Hgb look lik?. Are they anemic? Would an anti-inflammatory increase the hgb enough so that there would be more RBC’s to carry O2? What about lying in a prone position to give more room for lungs to inflate. I guess if my patient looks like they are dying that automatically means the risk has nothing to do with any of this. The benefit is that you may live.

  28. Ousama Moammar, MD says:

    Excellent blog, very happy to see a lot of reaction.
    Few quick points:
    – In general HCQ/AZTH works in early stages , within 72 hs of symptoms with estimated success rate 69%, but pt needs EKG, Electrolyte panel , Dont give with Digitalis, follow closely as OP.
    -It does not work in late stages for different reasons but mainly because the virus causes myocarditis
    -Some reports says micro-embolism is one of the causes of hypoxia in late stages
    -use on anticoagulation seems to be beneficial
    -Proning is very important early before intubation, more reports says they are avoiding intubation due to improved oxygenation with proning
    -Proning still important for patients on the Vent
    -Remdesivir (early news) seems to be more effective in advanced stages but only available for studies
    -Ivermectin seems to be very effective in Vitro (blocks viral replication), awaiting clinical trials
    -Thank you

  29. Barry Saxe says:

    I recommend looking at Leronlimab for intubated as well as moderately ill patients. The results from UCLA use were quite amazing.

  30. Dr Srinivas Chilukuri says:

    People say that Hydrxychloroquine has been popularised by President Trump as though nobody thought about it before. Several countries especially in Asia have incorporated into their respective protocols. There was some sketchy rationale albeit unproven one.

    1.Is there a need for clinical trials? Yes.
    2. Has the safety been proven? No, but there is a lot of experience of mass scale use of Chloroquine and Hydroxychloroquine in Malaria endemic countries. As a medical student in India, during rural medical camps we used to routine prescribe Chloroquine routinely for people of all ages.
    3. Is there a concern with QTc prolongation? Absolutely. We need to be cautious. In a hospital setting, it should be possible.
    4. Should it be given till the results of clinical trials are know? I think so. Nobody is talking ill of the antivirals and anti-inflammatory agents, but people are making a big fuss over the use of HCQ just because it’s cheap and unglamorous.

  31. MaryAnn Banerji says:

    I would use it. It has a long history in clinical use. The oxford prevention srudy is using a large dose. There is evidence that HCQ has a long tissue half life.
    I suspect it works in some people if given at the right stage, likely earlyin the disease .

  32. Mauro Schechter says:

    Excellent, as usual. I was quite sick with Covid–19, spent 5 horrible days in a hospital, and did not accept any of the drugs that were offered to me, from chloroquine to nitazoxanide. I suppose I’m
    Old fashioned: I still practice what I preach, evidence-based medicine.

    • Robert Stephens says:

      You may have avoided those 5 horrible days in hospital had you been taking HCQ. There is no EBM that it won’t help.

  33. isaac says:

    Maybe the use of hydroxichloroquine at the starts of disease has better outcome in some patients. The antiviral mechanism action of HQ reinforces the idea to give more time to inmune system to recognize the virus before a massive cells infection. HQ in critical patients dont helps because the virus had a great replication at the past days. Thank you for this blog.

  34. Mark Scapini says:

    Because President Trump is for the drug, many physicians and media people are actually rooting against it.
    Absolutely preposterous!

  35. Stan Deresinski says:

    We all suffer from the “therapeutic imperative” – the feeling of a need to so something, even when the appropriate thing is to do nothing. Besides being driven by empathy (and perhaps, in some cases, less noble causes), it is further driven by the “illusion of control” or more specifically in medicine, the “therapeutic illusion” which has been described as “the unjustified enthusiasm for treatment on the part of both patients and doctors”. This is then further enforced by confirmation bias – the tendency to only pay attention to evidence that supports our beliefs.

    We can’t totally avoid this, but it is imperative that we acknowledge the limitations of data, even as we choose to ignore those limitations.

  36. Barry Saxe says:

    The latest Chinese and French studies showed no benefit. Swedish hospitals have been issued guidance not to use it. The Pentagon did a study in March which showed no benefit.
    The Brazilian study
    “we have 16 deaths out of 41 in the high dosage, and 6 out of 40 in the low dosage.
    Hard to understand the interest in this drug

    • Robert Stephens says:

      In the Brazilian study, they used chloroquine (not HCQ) and it was trialled only for severe disease (RDS).

      The authors stated: ” Unfortunately, this study’s randomization, probably due to the low sample size, assigned older patients with heart disease to the high dosage arm”

      and they concluded: ‘More studies initiating CQ prior to the onset of the severe phase of the disease are urgently needed’.

  37. Tommy L Thompson Jr says:

    Most of the negative replies here speak of large scale clinical and the lack of them. It’s not only about size. It takes time, knowledge and a well trained/experienced staff to design and implement a study that is worthy of inclusion in the knowledge base to inform practice. The size of the study is determined by how powerful you want your inferences to be, the number of endpoints and the number of variables you want to control. Multi-site, randomized, double-blinded controlled clinical trials tend to be the “gold standard” because they include as much control of each defined variable as is possible given the knowledge base available at the time the study was designed. “Just do something” confounds efforts to design gold standard studies with inferior results and incitement of public opinion.

  38. Robert Shreck says:

    Amazed that the blogger did not mention serotherapy. Convalescent plasma has more than a century-long history of success against both viral and bacterial infections (never forget Balto!) largely obviated with the emergence of anti-microbials. Potential adverse effects include usual blood product transfusion risk, etc. but these are routine in clinical medicine. Availability is currently an issue but in communities with active and involved blood centers that problem should disappear in weeks. Under current nationwide IND protocols anyone sick enough to be admitted qualifies for administration of one dose (200cc). Up to 800 cc can be obtained from a donor with routine pheresis and they can donate weekly.

    When admitted I will ask for a dose on first day. And enoxaparin, 40mg daily if not therapeutic.

  39. R Chandler MD says:

    As an over 70 vulnerable I plan to take HCQ/AZTH if I develop fever, sore throat, cough or some variation being at high risk. If it works well enough to avoid ARDS then it is worthwhile. People like me rarely make it off the ventilator alive.

  40. Denzel says:

    I believe in evidence based care but right now there isn’t time so might as well use it

  41. Blaine says:

    Have you done any research on Leronlimab?

  42. Jeff Harries says:

    Ironically, Trump advocating for it has made it more difficult to prescribe because of “The Establishment’s” insistence on RTC’s being the only evidence we should be considering for this often fatal disorder. That is misguided and mis-charcterizing the real chance of side-effects speaks to the FUD being spread.
    Am I eager to find out what the RCT on HQ say? Absolutely. But if I contract Covid-19 or my family does, I would not hesitate in taking HQ. I live in the real world where I consider all factors in making clinical decisions.

  43. Mary Schwenkler says:

    I am a lay person wondering if the is any data showing that people already on Hydroxychloroquine have less severe symptoms or diminished progression of Covid19 than those not on the drug?

  44. Barry Saver says:

    We learn nothing from cowboy medicine. Since there are no proven, beneficial treatments, we need national/international default opt-in RCTs. No time wasted on recruitment/signed consent. If a patient declines, they get usual care. if they do not decline, their data are entered in the study coordinating center database and then they are randomly assigned to a treatment arm. Ideally, treatment med. would be blinded, but that could even depend on site capabilities, with adjustment in data analyses. That would allow for fastest accumulation of data on treatments, including fastest identification of treatment harms – and we could just as easily be harming as helping critically ill patients with “What have they got to lose?”, evidence-free treatment. Waiver of traditional informed consent is justifiable given the absence of any proven treatment and the fact that untested treatments are being given widely, also without traditional informed consent, but with the certainty that information about both benefits and harms will not be gained. Heck, I could put up a REDCap site in a week to enroll, assign, and track patients. Participating sites would just have to safely track study IDs and commit to responding to emailed surveys with outcome data. Why are we NOT doing this??? NIH and/or WHO should have gotten an effort like this up and running and cleared by their IRB a couple of months ago. We don’t need multiple, separate trials right now, nor bigger, uncontrolled case series (looking at you, Gilead!) we need one, big, rapid, adaptive RCT.

  45. james threadgill says:

    its about money. heck yes, you should use it. we dont time for all this .fine tune it later.

  46. Carl DiPerna says:

    I have more questions than answers. Dr. Sax stated, “people already receiving hydroxychloroquine for rheumatologic indications have already acquired COVID-19”. What is the severity of their symptoms? Also, just because those with rheumatologic indications on hydroxychloroquine do acquire COVID19, it does not necessarily indicate that the drug is ineffective in a non- rheumatologic population. Do inflammatory diseases put individuals in a higher risk group for COVID19?

  47. Gene Schulze says:

    Dr. Alan Green reports on zinc (up to 75 mg per day) if given immediately upon the first sign of symptoms, along with green tea or CoQ10 to transport it inside the affected cells, is a benign potential treatment. I have read of another doctor treating patients with Hydroxychloroquine and zinc, who stated that without the zinc the treatment is not effective.

  48. Matthew Wack says:

    Typically there are no good answers when the phrase “I believe” is stated more often than “the data shows”.

  49. Jason Stout says:

    The IDSA guidelines have been very helpful in emphasizing the point that we should not be throwing random drugs at people; rather, we should be treating every patient possible in a study so we can learn and move forward. At Duke we responded by designing and opening a pragmatic, partial factorial, Bayesian trial of hydroxychloroquine +/- azithromycin. The inclusion criteria are very broad; we’re taking pregnant women, children down to 12, pretty much everyone without a hard contraindication to one of the drugs. We are working with VA and other partners to bring this trial to sites across the country–it is relatively easy to get up and running, and provides clinicians with something to offer patients besides “doing nothing.”

  50. Beth says:

    Examples of positive results: Dr. Robin Armstrong, Texas City – 56 elderly, no side effects, all able to go outside for first time after starting treatment. Dr. Daniel Wallis, LA Cedar Sinai 75 under regiment — 100% recovery. Dr. Oz confirms legitimate sources of study and medical results from France where it is considered a game changer with positive results to recovery within 6 days! Dr Daniel Varga 3/4 NY patients– nursing homes and front line workers. 1400 patients…300 on ventilators — revisit to find results. Dr. Tsamasfyros treated 40 patients — 100% recovered within a day or two. Dr. Stephen Smith — 75 patients or so — 100% recovery. Dr. Zev Zelevsky currently 350 patients currently under regiment. wait to see results.. There is a democrat who took it and struggled for months and recovered within a day! What I read was that Hydroxychloroquine should be used with Azithromycin and the blogger who mentioned that ZINC is ESSENTIAL is right. WORKS WITH THE ZINC regiment because it allows the Hydroxychloroquine to be more effective in entering the cells because zinc is a powerful agent. In very latent and sepsis stages, there is too big of a battle and is not a cure. It is a considered preventative and reduces symptoms early for early recovery so the body wins before the virus does. Follow up and do research. Maybe there is some good info here.

HIV Information: Author Paul Sax, M.D.

Paul E. Sax, MD

Contributing Editor

NEJM Journal Watch
Infectious Diseases

Biography | Disclosures | Summaries

Learn more about HIV and ID Observations.