COVID-19 Treatment Update: Can We Just Stop Wasting Time on Hydroxychloroquine

Background: Hydroxychloroquine (HCQ) is an antimalarial and immunomodulatory drug that is postulated to exert an antiviral effect by increasing intracellular pH resulting in decreased viral binding at the ACE2 receptor.

Azithromycin is a macrolide antibiotic that also has anti-inflammatory and immunomodulatory properties which could help decrease viral replication and viral binding.

Both of these medications have been used to treat COVID patients based on in vitro findings. However, in vitro studies often do not extrapolate to patient oriented outcomes.

In June 2020 the US FDA revoked the prior emergency use authorization to HCQ and chloroquine (CQ) in patients with COVID-19. We now have yet another retrospective observational trial of HCQ, azithromycin, and the drugs in combination. 

Paper: Arshad S et al. Treatment with Hydroxychloroquine, Azithromycin, and Combination in Patients Hospitalized with COVID-19. International Journal of Infectious Diseases 2020. [Epub Ahead of Print]

Clinical Question: Is hydroxychloroquine (HCQ) alone and in combination with azithromycin (AZM) in hospitalized patients positive for COVID-19 associated with a reduction in in-hospital mortality?

What They Did:

  • Multicenter, comparative retrospective observational study performed at 6 hospitals in southeast Michigan
  • HCQ 400mg BID for 2 doses on day 1, followed by 200mg BID on days 2 – 5
  • AZM 500mg qD on day 1 followed by 250mg qD for the next 4 days
  • Combination of two drugs was reserved for selected patients with severe COVID-19 and minimal cardiac risk factors
    • QTc >500ms was considered an elevated cardiac risk and HCQ use was reserved for patients with severe disease and telemetry monitoring for serial QTc checks
  • Hospital protocols included adjunctive immunomodulatory therapy with corticosteroids and tocilizumab


  • Primary: In-hospital mortality


  • ≥18 years of age
  • Treated as inpatient for at least 48 hours unless expired within 24 hours
  • First admission only for patients with multiple admissions 


  • Not discharged from hospital
  • Left against medical advice
  • Transferred to another healthcare facility
  • Readmission


  • 2,541 patients
    • Overall in-hospital mortality was 18.1% (45% in ICU patients)
    • 24% of patients required ICU care
    • 18% of patients required mechanical ventilation
    • Median time to follow up was 28.5d
    • 9% of patients had not been discharged at the time of analysis
    • 68% of patients received methylprednisolone
    • 5% of patients received tocilizumab
  • In-Hospital Mortality
    • HCQ + AZM: 20.1% (95% CI 17.3 to 23.0%)
    • HCQ Alone: 13.5% (95% CI 11.6 to 15.5%)
    • AZM Alone: 22.4% (95% CI 16.0 to 30.1%)
    • Neither Drug: 26.4% (95% CI 22.2 to 31.0%)
    • P<0.001
    • No patient had documented torsades de pointes
  • Hazard Ratios of Medication vs Neither Medication:
    • HCQ Alone fs Neither Medication: HR 0.34; 95% CI 0.254 to 0.455
    • AZM Alone vs Neither Medication: HR 1.050; 95% CI 0.682 to 1.616
    • HCQ + AZM vs Neither Mecication: HR 0.294; 95% CI 0.218 to 0.396
  • Cox regression model using neither drug as reference:
    • Treatment with HCQ alone decreased mortality by a HR of 66% (p<0.001)
    • Treatment with HCQ + AZM decreased mortality by a HR of 71% (p<0.001)
  • Matched propensity scoring of HCQ vs no HCQ (190 patients in each group):
    • Treatment with HCQ vs no HCQ decreased mortality by a HR of 51% (p<0.009)


  • Consecutive patients hospitalized with COVID-19 from March 2020 until May 2020
  • Treatments were protocol driven and uniform in all hospitals
  • Reviewed all deaths to mitigate any potential limitations associated with missing or inaccurate documentation in EMRs


  • All patients were kept on cardiac monitors throughout the course of treatment to monitor for prolonged QTc. This may not be feasible in some hospitals that are currently overwhelmed by a surge of COVID-19 patients (i.e. not enough monitors). Additionally, these results could lead to indication creep with clinicians wanting to treat outpatients (i.e. without QTc monitoring)
  • Unclear why some patients received certain treatments and others did not
  • Retrospective, non-randomized, non-blinded study design
  • Duration of symptoms prior to hospitalization are not available
  • This was a select group of patients without cardiac risk factors
  • Groups were not well balanced at baseline which can bias results
    • No medication group was older which an independent risk factor for death
    • More steroids given in the HCQ group (78.9%) and HCQ+AZM group (74.3%) compared to AZM alone (38.8%) and neither med group (35.7%). See discussion section on the RECOVERY trial and dexamethasone
  • Unblinded study which could lead patients getting medications to have unmeasured improved care
  • This study an only establish association, not causation


  • This retrospective observational study with all of its methodological issues does not replace what we know from the well-done RECOVERY trial randomized clinical trial. In the pyramid of evidence-based medicine, an RCT is methodological superior trial compared to this trial. The RECOVERY trial looking at HCQ, out of the UK, which is a large randomized clinical controlled trial showed no beneficial effect of HCQ in patients hospitalized with COVID-19. In this trial 1542 patients were randomized to HCQ and 3132 patients were randomized to usual care alone.  There was no significant difference in the primary endpoint of 28d mortality (25.7% vs 23.5% respectively).
  • Another arm of the RECOVERY trial out of the UK randomizing patients to dexamethasone vs no dexamethasone showed profound 28d mortality benefit favoring steroids:
    • Patient requiring invasive mechanical ventilation: NNT = 9
    • Patients requiring oxygen support without invasive mechanical ventilation = 29
    • As most patients in this trial receiving HCQ or HCQ + AZM received steroids and the patients receiving AZM alone or neither therapy had far fewer patients receiving steroids, the likely mortality benefit of this trial is due to the steroids and not the HCQ or HCQ + AZM
  • There was minimal cardiovascular disease in this trial with only 8.7% of patients having reported cardiovascular disease
  • There were more patients requiring ICU level of care and mechanically ventilated in the HCQ and HCQ+AZM groups. As this group has the highest mortality, we would expect to see more of a mortality benefit as their overall mortality was simply higher compared to the AZM alone and no drug groups
  • More patients received steroids in the HCQ and HCQ+AZM groups compared to the other two groups, which can confound the results

Author Conclusion: “In this multi-hospital assessment, when controlling for COVID-19 risk factors, treatment with hydroxychloroquine alone and in combination with azithromycin was associated with reduction in COVID-19 associated mortality.  Prospective trials are needed to examine this impact.”

Clinical Take Home Point: This methodologically flawed study, with unbalanced use of steroids simply bias the results to favor patients receiving HCQ and HCQ + AZM.  This study should not change clinical practice of not prescribing these medications.


  1. Arshad S et al. Treatment with Hydroxychloroquine, Azithromycin, and Combination in Patients Hospitalized with COVID-19. International Journal of Infectious Diseases 2020. [Epub Ahead of Print]

Post Peer Reviewed By: Anand Swaminathan, MD (Twitter: @EMSwami)

Cite this article as: Salim Rezaie, "COVID-19 Treatment Update: Can We Just Stop Wasting Time on Hydroxychloroquine", REBEL EM blog, July 6, 2020. Available at:
The following two tabs change content below.

Salim Rezaie

Emergency Physician at Greater San Antonio Emergency Physicians (GSEP)
Creator & Founder of REBEL EM

Like this article?

Share on facebook
Share on Facebook
Share on twitter
Share on Twitter
Share on linkedin
Share on Linkdin
Share on email
Share via Email

Want to support rebelem?

10 thoughts on “COVID-19 Treatment Update: Can We Just Stop Wasting Time on Hydroxychloroquine”

  1. To balance the groups, they used propensity-score matching, which means they tried to find a patient in the treatment group (the group receiving HCQ) matching every patient in the control group (the group not receiving HCQ). But why did they discard so many patients from the control group (which was even smaller in the beginning)? From 556 patients in the control group (and 1985 in the HCQ-group), only 190 made it to the analysis after matching. Why discard so many patients? It seems their matching algo was very strict, which could have resulted in an extremely select population.

    • Hello Andreas,
      That was a big issue that when they tried to balance they were only able to do it 190 of the total cohort. It seems they had difficulty matching the groups which validates the point of more steroids in the HCQ groups. Glad you read the whole paper and agree with your thoughts. There were lots of issues methodologically unfortunately, which makes it difficult to hang our hats on the results of the trial.


  2. I agree that the methodology of this study is not able to give any kind of confident conclusions. However, the propensity matching process used to attempt to equalize the HCQ and No-HCQ groups also attempted to match the steroid use between these two groups.

    Table 3 shows the propensity score matched patients, which included those matched for steroid use; Upon propensity score matching, 44,2 % of patients matched for steroid use received HCQ and 44.2% did not.

    Though I am not a statistician (!) it appears to me that they DID attempt to match the steroid use differences and that this factor should not be called upon to explain the differences in out come between HCQ and no-HCQ groups.

    • Hello Mark,
      You are correct that they attempted to match with propensity matching, however they were only able to match 190 patients per group. In other words, so many were unmatched that they couldn’t even match the groups that well…huge portion of the population therefore excluded. Numbers too small to make a confident statement about HCQ. Had they been able to match most of the population, I would be willing to believe that outcome.


    • Hey Matt,
      Please see my comment to Mark…in order to match patients they had to exclude a large portion of the registered patients to achieve this…190pts per group…this is too small to draw any conclusion from…could sincerely be statistical noise or random chance. Would need much more patients matched to make the claim.


  3. Hi, just found your blog. Why would anyone waste time using antivirals in the hospital for an acute ILI? By that time antivirals are way too late. Focus should have been on early, outpatient treatment from the start, not “compassionate use” in the hospital. The CDC’s recommendation made no sense from a medical perspective.

    In fact, in Jan. 2020, the CDC said to give antivirals early to high risk outpatients. Why would the CDC say to reserve antivirals for use in the hospital?

    • All great questions and statements…all would be something you would have to ask the CDC. I have said since early in the pandemic that we are treating many patients too late into illness to make a difference if one exists…I will say however several medications have been used early on in illness and not shown a benefit either.



Leave a Comment

Time limit is exhausted. Please reload CAPTCHA.