Our large tertiary care ED in Queens, NY, usually sees more than 100,000 adults a year. As we write this, we have 753 (last week, 458) COVID-19-likely or -positive patients, more than 100 of them on ventilators. We have had a few successful extubations!

We have learned so much from treating COVID-19 patients in the emergency department, on the floor, and in the intensive care unit and by examining the available literature. We hope this summary will help with acute ED management. We will update it here as often as we can.

Initial Presentation

Symptoms of COVID-19 may vary, and may include fever, chills, headache, myalgia, rhinorrhea, throat pain, dyspnea, chest pain, cough, and sputum. Less frequent symptoms are loss of smell and taste, nausea, vomiting, and diarrhea. (EM:RAP. March 31, 2020; https://bit.ly/2Rftgo7.)

The clinical course trends toward two different stages. The first typically includes these symptoms, and is primarily experienced during the first week after onset of symptoms. For some, the second stage starts between days five and seven when sudden rapid clinical deterioration may occur.

We have not yet found any predictive symptoms of subsequent deterioration.

Workup

The workup for patients suspected to have COVID-19 is variable and depends on their clinical presentation at the time of evaluation. For those who have normal vital signs and appear clinically well with an uncomplicated medical history, it would be reasonable to treat and release them with quarantine directions, instructions on supportive care, and strict, detailed return precautions.

These steps below should be taken to evaluate and treat patients who will likely require hospitalization or intensive care.

Lab Evaluation

Complete the following:

• CBC with differential: The white blood count is usually normal, lymphopenia is common, and mild thrombocytopenia is common. (Clin Infect Dis. 2020. doi: 10.1093/cid/ciaa248. [Epub ahead of print]; https://bit.ly/2UQO8CT.)
• CMP with magnesium and phosphorus: Liver function tests (ALT, AST) commonly elevated
• Coagulation studies with D-dimer: PT/PTT/INR is usually normal on initial presentation. Some develop DIC. The D-dimer is commonly elevated.
• COVID PCR: (RVP if you suspect alternate viral etiology, though coinfection is possible.) We have found false-negative COVID testing in up to 10 percent of cases.
• Procalcitonin: This is usually not increased with COVID-19. If elevated, it may indicate an alternate diagnosis or superimposed bacterial infection. Procalcitonin is not routinely elevated higher than 0.5 ng/mL in these patients. (N Engl J Med. 2020. doi: 10.1056/NEJMoa2002032. [Epub ahead of print]; https://bit.ly/3bDobxu.) It does seem to go up as disease progresses, however. An elevated procalcitonin in the ED should lead you to consider an alternative or additional diagnosis more strongly.
• CRP (sometimes ESR too, but difficult for some because it is often performed manually): This is elevated in COVID-19 patients, and it seems to trend upward with the progression of the disease. It may have some prognostic correlation.
• LDH, ferritin, urine legionella, blood cultures, lactate, troponin, CK, CKMB, ABG, and G6PD: Chloroquine causes hemolytic anemia in G6PD; these are helpful for inpatient teams as well.
  
  

Poor Prognostic Factors:

• Neutrophil to lymphocyte ratio >3.13 (medRxiv. 2020. doi: 10.1101/2020.02.10.20021584; https://bit.ly/3dIp8GG.)
• Absolute lymphocyte count <0.8
• LDH >245 U/L
• Ferritin >300 ug/L
• CRP >100 mg/L
• D-dimer >1000 ng/mL. (Lancet. 2020;395[10229]:1054; https://bit.ly/3bIGQYT.)

Imaging

A chest x-ray is typically the most accessible means of imaging. Findings include:

• Patchy ground glass opacities: The literature suggests this is present bilaterally and predominantly peripherally around 70 percent of the time. (N Engl J Med. 2020. doi: 10.1056/NEJMoa2004500. [Epub ahead of print]; https://bit.ly/2WT5LES; Lancet Infect Dis. 2020;20[4]:425; https://bit.ly/2WWLyOv.) This would fit what we are seeing as well.
• Clear lungs: Early in the course of the disease, we are seeing many mildly symptomatic patients with clear chest x-rays that quickly progress over days to patchy ground glass opacities.
• Uncommon: Pleural effusions are uncommon, but Wuhan reported pleural effusions in 13 percent of cases. (Lancet Infect Dis. 2020;20[4]:425; https://bit.ly/2WWLyOv.) Washington state reported no pleural effusions in its case series of 24 critically ill patients. (N Engl J Med. 2020. doi: 10.1056/NEJMoa2004500. [Epub ahead of print]; https://bit.ly/2WT5LES.) Twenty-eight percent of patients had pleural effusions in another study from Washington state. (JAMA. 2020. doi: 10.1001/jama.2020.4326. [Epub ahead of print]; https://bit.ly/2UO0EDm.) This would indeed fit with what we are seeing: Pleural effusions are only present about 15 percent of the time.
• We initially had been progressing toward CT scans because the sensitivity is higher for picking up subtle opacities, but we have moved away from these unless we are concerned about an alternative diagnosis because a CT scan prompts no real change in management.
• Ultrasound may be used as an alternative modality as well. There is emerging evidence of findings with COVID-19, though it is not yet well documented. (Intensive Care Med. 2020. doi: 10.1007/s00134-020-05996-6. [Epub ahead of print]; https://bit.ly/3aC5zOg.)
  

Treatment: Fluids

Patients with COVID-19 appear to be very sensitive to fluid overload. We have found success at preventing the need for intubation by keeping patients net negative despite tachycardia and acute kidney injury.

Hypotension is not the most common presentation we have seen but is certainly a feature for some. We have been leaning toward starting patients on low-dose pressors to support mean arterial pressure rather than multiple fluid boluses if they are on the verge of intubation because of hypoxia.

We have been utilizing the following approach to fluid resuscitation for ED patients with suspected COVID-19:

• If the patient is normotensive, do NOT give a fluid bolus.
• If the patient is hypotensive or appears hypovolemic on clinical exam, carefully consider a very small fluid bolus v. pressor for hypotension, especially if the patient's volume is clinically overloaded.
• In the patient who is hypoxic and normotensive and requires oxygen, consider giving a dose of Lasix. Anecdotally, this seems to delay intubation hours to days. A side note about acute kidney injury (AKI): There appears to be a disproportionate number of patients with AKI from COVID-19 compared with typical viral pneumonia. This may suggest another contributing factor, possibly a microangiopathic component. If this is the case, AKI should not trigger us to give fluids or avoid Lasix. This needs further investigation.
• Do not fluid-resuscitate to clear the lactate. The elevated lactate in a nonhypotensive patient is likely not from hypovolemia but possibly from a catecholamine surge, severe hypoxia, and/or respiratory distress. DO NOT bolus for lactate, please.
• The WHO recommendations suggest a fluid bolus only in patients with septic shock, but this recommendation likely carries the same caveat that was used in the 2016 Surviving Sepsis campaign. (WHO. March 13, 2020; https://bit.ly/2UOPjCW.) The idea was that some patients would need to be intubated for the sole purpose of giving more fluids. This is particularly troublesome during a pandemic with a huge uptick in ventilator need. The pathophysiology of shock caused by hypovolemia does not seem to be playing out in this disease.
• Shock at presentation is rare. (Lancet Respir Med. 2020. doi: 10.1016/S2213-2600[20]30079-5. [Epub ahead of print]; https://bit.ly/3dOx4Gz.) If a patient is in a shock state, consider late-stage COVID-19 complications such as myocarditis or cytokine storm, both of which will not respond to fluids. Obviously, consider other causes of shock as well.
• Not that any of us normally do, but DO NOT start maintenance fluids.

Antibiotics

Generally, our EM training would tell us to initiate CAP coverage upfront in an undifferentiated hypoxic febrile patient, and we would not recommend discontinuing this practice in the era of COVID-19. Our job in the ED is to cast a wide lifesaving net while minimizing harm from our interventions. Our hospital system has updated its recommendations AGAINST using azithromycin in COVID-19. The reasoning for this change is that strong evidence for efficacy is lacking, and it can contribute to QT prolongation." We would like to address this treatment update in two parts:

• First, the risk of QT prolongation from azithromycin is present, however minimal, and likely does not contribute to increased risk of cardiac events. A paper from Korea compared the ECGs of 402,607 patients before and after taking azithromycin with those prescribed amoxicillin. (BioMed Research International. 2018; Article ID 1574806; https://bit.ly/2JKU8YG.) They found that a risk of mild QT prolongation (>450 male, >460 female) was increased by an OR 1.4, and a risk of severe QT prolongation (>500) carried an OR of 1.43 for azithromycin. This risk was not seen in the subgroup of those under age 29. The highest risk group was patients in the 60-80 year range. Even though this paper showed a mild increased risk of QT prolongation, not all QT prolongation is created equal. Rather than prolonging repolarization, which can lead to Torsades, azithromycin prolongs the action potential itself. A meta-analysis from 2014 found that azithromycin had no increased risk of death and no increased risk of cardiovascular events in 15,588 patients compared with placebo. (Am Health Drug Benefits. 2014;7[6]:318; https://bit.ly/3bZO8rj.) Indeed, several papers have now published data that argue against the original New England Journal of Medicine paper from 2012 that started the controversy of azithromycin and increased risk of cardiac death.
  
  
• Second, severe CAP is still a concern in the era of COVID-19. Legionella is not an insignificant cause of severe CAP, possibly up to 15 percent of cases. Even extremely broad-spectrum beta-lactams will not cover for legionella. Azithromycin should still be the first-line treatment for severe CAP while the legionella antigen is pending. Azithromycin has also shown a mortality benefit in pneumococcal pneumonia, likely from an anti-inflammatory effect. Azithromycin should likely be continued if pneumococcal pneumonia is suspected even if the legionella antigen is negative. If there is significant concern over the QT prolonging properties of azithromycin at your institution, then doxycycline should replace azithromycin in your initial CAP coverage.
  
  
• From an ED perspective, therefore, we would still recommend utilizing early ceftriaxone and azithromycin for severely ill patients who you are going to admit. Remember PO azithromycin is just as good as IV. This also allows you to limit volume of IVF by using PO instead. Don't forget to send blood cultures as well as procalcitonin and urine legionella antigen. As these results come back, the inpatient team can quickly de-escalate your antibiotics. Early antibiotic administration is still one of the few interventions that has shown to save lives in sepsis. We would discourage assuming that every severe respiratory sepsis patient you see from now on has COVID-19 and nothing else that we used to worry about.

We would STILL consider sending home patients who are not critically ill on azithromycin because the French study showed hydroxychloroquine plus azithromycin resulted in lower viral loads. (Int J Antimicrob Agents. 2020:105949. doi: 10.1016/j.ijantimicag.2020.105949. [Epub ahead of print]; https://bit.ly/2WZbyJc.) Of note, a Bayesian reanalysis of that paper addressed some of the controversies surrounding the original data that showed no likely benefit. (BioMed Research International. 2018; Article ID 1574806; https://bit.ly/2JKU8YG; (April 3, 2020; https://bit.ly/2Vcr4Po.) We are left to draw our own conclusions until better data are available. There are some data that hospitalized patients show no benefit from this drug combination when it comes to viral clearance. (Médecine et Maladies Infectieuses. March 30, 2020; https://bit.ly/34pTp97.) Indeed, the email from our institution would suggest the same, but we think that the best evidence available on nonsevere patients still suggests that azithromycin may be helpful.

AVOID vancomycin. We historically tend to give this to our undifferentiated infectious patients, but COVID-19 patients are developing renal failure upstairs. If there is significant concern for MRSA, consider linezolid instead.

Also consider calculating the patient's Shorr score to see how worried to be about MRSA pneumonia. (Calculate that here: https://bit.ly/MDCalc-ShorrScore.)

Send a procalcitonin if you are admitting the patient. This will help determine if CAP coverage needs to be continued by the floor teams.

Other Treatments

• Vitamin C: Two reasons to consider using vitamin C in COVID patients are the CITRIS-ALI trial and the data that suggest high serum levels of vitamin C seem to protect the lung epithelium in mice. (JAMA. 2019;322[13]:1261; https://bit.ly/3bCBaiU; YouTube. https://bit.ly/3bIEb1l.)  There is little downside to giving it because of the strong safety profile of IV vitamin C (other than resource utilization). The dose is IV vitamin C 1500 mg Q6 x 6 days. Despite the trial safety data, as mentioned by Paul Marik, MD, vitamin C does pose the theoretical threat of working as a prooxidant when there are high levels of serum iron. He suggests monitoring ferritin and CRP and then possibly reducing the dose of vitamin C if this is a concern. (Paul E. Marik. "COVID Protocol," March 24, 2020; https://bit.ly/2R1pyhx.) From an ED perspective, a single dose of 1500 mg is a good idea, especially when reviewing the animal data, which suggests that you need to reach a steady serum state before you see an effect. (YouTube. https://bit.ly/3bIEb1l.) Of note, effective concentrations were not reached through PO dosing in those studies.
• NSAIDs: There has not been compelling evidence of harm, but we try to avoid NSAIDs in patients who do not absolutely need them given the anecdotal reports. We know that they are nephrotoxic, and downstream renal failure in these patients is occurring. There is also an increase in ACE2 receptor expression theoretically leading to higher availability of viral binding sites.

• Hydroxychloroquine: The evidence of benefit seems to be lacking now, especially with the first randomized, controlled trial from China. (Medrxiv. April 14, 2020; https://bit.ly/3bpVGnb.) Prior data were based on a case series in France showing possible earlier viral clearance (Int J Antimicrob Agents. 2020:105949. doi: 10.1016/j.ijantimicag.2020.105949. [Epub ahead of print]; https://bit.ly/2WZbyJc) and in vitro data suggesting the drug inhibits viral entry into cells. The Chinese RCT showed no benefit to viral clearance and, among the multiple lab trends, only a possibly lower CRP in the treatment group but no clinically significant differences.

           Secondary outcomes showed possible reduction in duration of cough and fever. Adverse events were found in 30 percent of the treatment group and nine percent of the control group with the most common adverse event being diarrhea. It still stands that anti-viral therapy may be effective primarily early in the disease course because treatment was initiated on average after two weeks of illness in this RCT. At this point it seems that the best available evidence suggests hydroxychloroquine is not helpful, and therefore we feel that it should not be started in the ED. If you are still going to start hydroxychloroquine, please make sure that you consider baseline QTc as well as G6PD for patients getting chloroquine; that is why it should be sent with labs at admission. See below for QTc considerations.

• Remdesivir is emerging as a frontrunner for antiviral therapy, but it is being given on a clinical trial basis and is not widely available. Data from the New England Journal of Medicine article published on April 10 gives no evidence of benefit given the lack of a control group and no strong inclusion or exclusion criteria. (https://bit.ly/2KkbYSK.)

• Thiamine is of unclear benefit. It is currently being used in our ICUs. The dose is 200 mg IV Q12 for six days.
  

• Azithromycin: We are not starting everyone on this. We are using this (or doxy) on those patients who meet SIRS criteria with a presume respiratory infection. Keep in mind that both azithromycin (although mild) and {hydroxy}chloroquine are going to prolong the QT, so be careful and aggressively replete Mg and K as needed. The efficacy has yet to be proven.

• Steroids: The evidence is definitely mixed here. At this time, the limited evidence indicates we SHOULD NOT be giving steroids in the ED for the majority of nonsevere cases unless there is clear adrenal suppression or the patient is in the late stage of disease (severely elevated D-dimer, CRP, ferritin, or ARDS requiring advanced supplemental oxygenation via nonrebreather and/or  intubation). Even then, consider collaborating with the ICU team on this decision. The reasoning against giving steroids in early disease is that early steroids increased viral load in SARS patients. (J Clin Virol. 2004;31[4]:304; https://bit.ly/3bCBgqM.) Steroids would only serve to suppress adaptive immunity theoretically and not help modulate inflammatory response during the early stage of illness. The case study from Wuhan suggested that steroids might help those with ARDS. (Lancet Respir Med. 2020. doi: 10.1016/S2213-2600[20]30079-5. [Epub ahead of print]; https://bit.ly/3dOx4Gz.) They used methylprednisolone 40-80 mg IV daily. At our institution, our ICU is currently using methylprednisolone 40 mg IV BID.

  As Josh Farkas, MD, noted, dexamethasone could be preferable (7-15 mg daily) because it causes less fluid retention due to lower mineralocorticoid effect, and fluid is a huge issue in these patients, as we have already discussed. (IBCC. March 2, 2020; https://bit.ly/3bGL9nu.)
  
  QTc: Typically:

  If the QTc is < 470ms, you can give hydroxychloroquine and then repeat an ECG two hours after the second dose. The dose may be adjusted based on changes in the QTc on repeat ECG.

  If the QTc is >470ms and <500, then it's a risk/benefit discussion, but it still might be used with telemonitoring. If the QTc is >500, we do not start it, with very few exceptions.

  We are also trending QTc daily. I would monitor anyone at risk for arrhythmia (i.e., LV dysfunction), no matter what their QTc is if starting these medications. There are other caveats to this, but this is the basic overview. Please check your own hospital protocols for restrictions with the QTc.

  There are other caveats to this, but this is the basic overview. Please check your own hospital protocols for restrictions with the QTc.

  Anticoagulation: There is still no strong evidence on the importance of anticoagulation in these patients. Indeed, we are seeing more and more cases of arterial and venous thrombi in the ED and ICU. It is unclear when the best time to initiate anticoagulation is and if we should be sending patients home on anticoagulation who have elevated D-dimers and no evidence of thrombi on imaging. At this time, we would consider anticoagulation for patients with a D-dimer >1,500 ng/mL. Calculate the patient's HAS-BLED score prior to initiating anticoagulation. If admitting, speak with the hospitalist prior to starting it in the ED to make sure that it will be continued on the floor.

Intubation and Oxygen Therapy

Vents are a resource, and many COVID-19 patients need 10 days or more of ventilation. (Lancet Respir Med. 2020. doi: 10.1016/S2213-2600[20]30079-5. [Epub ahead of print]; https://bit.ly/3dOx4Gz; JAMA. 2020. doi: 10.1001/jama.2020.4783. [Epub ahead of print]; https://bit.ly/2wQiXQd; N Engl J Med. 2020. doi: 10.1056/NEJMoa2004500. [Epub ahead of print]; https://bit.ly/2WT5LES.

Saving a vent for several days is meaningful. Try to utilize every resource available to prevent intubation of these patients. At our hospital, there is no need to call the MICU for severe hypoxia at this point. Unfortunately, every patient on the floor is developing severe hypoxia. We are currently recommending an NRB at 15L with an NC at 10L underneath with persistent O2 saturation <90% before we even consider intubating for hypoxia. Please send an ABG, and look for other signs of low oxygen delivery to organs such as tachycardia and altered mental status before intubating patients who have low SpO2s and appear well.

We are currently not allowed to use high-flow nasal cannula (HFNC) at our facilities because of concern that this will cause aerosolization of viral particles. The Surviving Sepsis Campaign and ANZICS guidelines currently recommend the use of HFNC prior to intubation. (JAMA. 2020. doi: 10.1001/jama.2020.4914. [Epub ahead of print]; https://bit.ly/2RfuShF; ANZICS. March 16, 2020; https://bit.ly/3bz6z5M.)  Michelle Ng Gong, MD, the director of critical care research at Montefiore Medical Center and the chief of critical care and pulmonary medicine at Albert Einstein College of Medicine, both in the Bronx, also recommends HFNC for COVID patients. (AMA EdHub JN Learning. March 23, 2020; https://bit.ly/3dGVcdW.)

The WHO guidelines currently recommend HFNC in select patients because there is an unknown amount of aerosolization. (WHO. March 13, 2020; https://bit.ly/2UOPjCW.) Some literature suggests that HFNC does not increase the risk of infection to health care professionals in other disease processes. (PLoS One. 2012;7[4]:e35797; https://bit.ly/2QXGeGA; J Hosp Infect. 2019;101[1]:84; https://bit.ly/3dJz42E; J Crit Care. 2012;27[5]:434; https://bit.ly/2WWBmFD.)

Given the increasing degree of hypoxia in these patients, we believe that we should begin allowing HFNC. More adjuncts will save us more ventilators and more ICU beds. HFNC may increase aerosolization but likely not by much, definitely less than intubation, and it likely does not increase the risk of infecting health care workers if airborne precautions are maintained. At a minimum, we should be using HFNC on floors where there are only COVID patients. The practicality of utilizing this in the ED with undifferentiated patients is unclear at this time.

• If patients are able, have them lie prone while on supplemental oxygen. We have seen significant symptomatic improvement as well as improvement in oxygen saturation on patients who lie prone even if they are not on ventilators.
  In the ED, if a patient requires a significant amount of supplemental oxygen, please send an ABG. We are considering intubating these patients if they have a PaO2 <60 despite the NRB and NC.
  We are not intubating anyone just for the work of breathing. These patients generally look terrible on the floor. Please get an ABG and intubate them if they develop respiratory acidosis. They should also be intubated if they are obtunded on arrival. If the ABG shows low CO2 or normal CO2, please try to hold off intubating. See how they do first with all the CO2 flushed out of their system by the NRB and NC

• Goals-of-care conversations with patients and families should be done as early as possible when appropriate. This should be the priority in patients with a very poor expected clinical course (age >80, COPD, diabetes mellitus [DM], coronary artery disease [CAD]). (JAMA Intern Med. 2020. doi: 10.1001/jamainternmed.2020.0994. [Epub ahead of print]; https://bit.ly/2V3vFTZ; N Engl J Med. 2020. doi: 10.1056/NEJMoa2002032. [Epub ahead of print]; https://bit.ly/3bDobxu;  China CDC Weekly. 2020;2[8]:113; https://bit.ly/2wHYAVv;  Lancet. 2020;395[10229]:1054; https://bit.ly/3bIGQYT.)

• A meta-analysis of Chinese data generated the following odd ratios for severe v. nonsevere COVID-19: DM OR=3.04, HTN OR=2.31, CAD OR=2.76, and COPD OR=3.56. Of note, smoking showed no increase in risk of severe pneumonia (OR=1.4), though elevated BMI increased risk of severe disease. (ePub ahead of print; https://bit.ly/2xIB3DQ.) .) Our own institutional data appear to be consistent with these findings so far.
  

• Let the patient and family know if this is COVID-19, then we are likely talking about 10 or more days on the vent. (Lancet Respir Med. 2020. doi: 10.1016/S2213-2600[20]30079-5. [Epub ahead of print]; https://bit.ly/3dOx4Gz; JAMA. 2020. doi: 10.1001/jama.2020.4783. [Epub ahead of print]; https://bit.ly/2wQiXQd; (N Engl J Med. 2020. doi: 10.1056/NEJMoa2002032. [Epub ahead of print]; https://bit.ly/3bDobxu.)

• And the likely mortality is 50 to 80 percent for patients needing intubation, especially those with pre-existing lung or cardiac comorbidities. (JAMA Intern Med. 2020. doi: 10.1001/jamainternmed.2020.0994. [Epub ahead of print]; https://bit.ly/2V3vFTZ; N Engl J Med. 2020. doi: 10.1056/NEJMoa2002032. [Epub ahead of print]; https://bit.ly/3bDobxu; China CDC Weekly. 2020;2[8]:113; https://bit.ly/2wHYAVv; N Engl J Med. 2020. doi: 10.1056/NEJMoa2004500. [Epub ahead of print]; https://bit.ly/2WT5LES; JAMA. 2020. doi: 10.1001/jama.2020.4326. [Epub ahead of print]; https://bit.ly/2UO0EDm.)

• If you have to intubate in the ED, consider using rocuronium over succinylcholine. This is not the time to be redosing paralytics. (EMCrit RACC. March 27, 2020; https://bit.ly/2QU4ppB.) Make sure to target deep sedation post-intubation, and assume paralysis for at least 60 minutes after using roc.

• After paralytics have worn off, please pay close attention to vent synchrony. Target RASS -5 with your sedation package. Discuss starting a Nimbex drip if the patient is still asynchronous with the vent. Start patients on higher PEEPs early. "Changing Ventilation Parameters," a good summary of how to adjust the vent settings, is available in the Brigham and Women's Hospital critical care guidelines at https://bit.ly/2QXxXT2.

Operational Changes

COVID-19 hit us hard and fast. We have found that the key to addressing the challenges successfully that came with its onslaught was the ability to adapt to a rapidly changing environment and to reevaluate available resources and layout daily. Mobilizing resources early and quickly ensured that we had the best chance at safely and quickly evaluating and treating patients.

Some examples of the changes we made within our hospital and hospital system included:

• Immediate cancellation of all elective surgeries to open up beds within the hospital.
  
  

• Redeployment of all newly available health care providers to assist with COVID care.
  
  

• Mobilizing as many ventilators as possible.
  
  

• Ensuring video airway equipment is functional and acquiring additional units if needed. We are intubating COVID-19 patients frequently, and this is one of the highest risk procedures for its transmission. Current recommendations are to use video laryngoscopy for the best chance of first-pass success and to keep the intubator as far from the patient as possible.

• Accumulating PPE, already in short supply, to use for every potential COVID-19 patient.

• Reformatting spaces near the ED, including our lobby, to be utilized for screening lower-acuity, treat-and-release patients to keep them separate from the general populace.

• Designating specific spaces in the ED for non-COVID patients, though these are now in the minority. We have found that even those who are evaluated for unrelated complaints have concurrent symptoms of COVID-19 on top of their original complaint and/or are found to be testing positive for COVID-19 on admission. We therefore now use full PPE in these areas as well.

  The findings we describe are dynamic and subject to rapid change based on new evidence as it emerges. We will update these here as often as we can.

The authors thank Michael Pallaci, DO, an emergency physician in Chillicothe, OH, for promoting knowledge sharing.

Dr. Johnson is the chief resident in emergency medicine at North Shore University Hospital/Long Island Jewish Medical Center. Dr. Gottlieb is an assistant professor of emergency medicine at the Donald and Barbara Zucker School of Medicine at Hofstra/Northwell Health.