Last updated 09/09/20, new additions in pink. Full reference citations here.

Compiled by Kathryn Bolles, MD with contributions from Shawn Cohen, MD; Justin Granstein, MD MPH; Christopher Kovach, MD MSc; Patrick Marcus, MD; Kristen Rogers, MD MPH; Eric Tanenbaum, MD

A compilation and synthesis of COVID information created by and for frontline MDs, representing our interpretations of available data and personal experiences; any data should be personally reviewed and considered alongside guidelines and expert advice. Please email to notify of new studies, omissions, or errors. Any views here are our own and do not reflect the views of our employers.

Sharing link:


COVID ICU One Pager by Nick Mark, MD (in slightly longer publication form in Jamil et al)


IDSA COVID19 Resource Center (with info on testing protocols)

UpToDate COVID19 (no paywall)

BMJ Living Systematic Review/Infographic for COVID-19 Drug Treatments 

Cochrane Reviews Coronavirus Collection

UW IDEA Program COVID-19 Information & Teaching Resources

EmCrit COVID Review

EB Medicine COVID19 Guide

Evergreen Hospital Lessons Learned (guidance from the first US hospital to handle cases)

Essential Evidence Plus COVID-19 Chapter (no paywall)

ACP Guide to COVID19

Harvard Medical School Med Student Created COVID Curriculum

EM:RAP COVID-19 chapter


WHO Pandemic Situation Reports

Johns Hopkins Pandemic Map

The COVID Tracking Project

IHME COVID19 Projections

COVID Tracking by City

HealthMap COVID-19


Guidelines: CDC, WHO, NIH

Protocols: University of Washington, MGH, UCSF, Mount Sinai, University of Nebraska, Department of Defense, Walter Reed Medical Center, AliBaba/Global MediXchange Handbook

Critical Care-specific: SCCM, ACLS/BLS, Brigham & Women’s, U Michigan airway protocols


Emory Infectious Disease Weekly COVID-19 Literature Round Up


WHO COVID Research

NEJM, Lancet, Wiley, JAMA COVID articles (no paywall)

COVID Clinical Trials Registry

CORD-19 COVID19 Open Research Dataset

N3C National COVID Cohort Collaborative


IN SHORT: Prevalence is high and a lot of transmission appears to be asymptomatic/ presymptomatic. Transmission most likely droplet, contact +/- airborne; R0 seems to be 2-3 but can be lowered with public health measures. Infections likely convey immunity, but duration unclear.

Incubation period: most estimates between 4-7 days with 98% within 14 days (Lauer et al, Tian et al, Zhang et al, Park et al, Sanche et al, Weirsinga et al)

  • Serial interval (time between symptom development in one person and symptom development in a person they infect) appears to be between 3-8 days (Du et al, Zhang et al, Park et al)

Prevalence: unclear and will require extensive serologic testing, but early studies indicate prevalence may be much higher (50-85x) than number of confirmed cases from initial testing (Bendavid et al)

Transmission: droplet, contact (fomites up to 7hrs on some surfaces), likely airborne (1hr half life of aerosolized virus), likely fecal-oral (RNA and infectious virus found), and possibly maternal-fetal (van Doremalen et al, Xiao et al, Santarpia et al, Chen et al, Zeng et al, Leung et al, Wölfel et al, Guo et al, Liu et al, Xiao et al [different paper], Klompas et al, Parasa et al)

R0: ~2.0-3.0 (WHO, Li et al, Park et al, Kucharski et al); some estimates much higher (5.7) & 11 aboard cruise ships (Sanche et al, Mizumoto & Chowell); can be lowered significantly through early public health controls (Wang et al, Zhang et al, Pan et al, Salje et al, Lyu et al, Pei et al, Chu et al, Yehya et al, Islam et al, Auger et al)

Viral shedding: can be present 1-3 days prior to onset of symptoms, median duration 17-20 days with durations up to 11 weeks in recovered patients (Zhou et al, Kai-Wang To et al, Zou et al, Lan et al, Xiao et al, Cevik et al) but long durations may represent more viral debris than live virus capable of causing infection (Ranawaka et al); appears longer in male patients and those with more severe disease (Xu et al)

High risk populations: elderly; residents of nursing homes, assisted living facilities, shelters, and psychiatric facilities (McMichael et al, Roxby et al, Tobolowsky et al, Mosites et al, Baggett et al, Yao et al, Meyerowitz et al); healthcare workers (Wu & McGoogan, CDC, Mani et al, Baker et al); immunosuppressed patients (Haberman et al, Pereira et al, Akalin et al); refugees (Kluge et al); prisoners (Kinner et al); minority & poor populations (Johnson & Buford, Grantz et al, Lancet, ICNARC, Martinez et al, Emeruwa et al, Alsan et al, Tai et al, Meyerowitz et al, Vahidy et al); family members of infected patients (Rosenberg et al, Li et al)

Immunity: seroconversion appears to occur between day 7-14 but levels appear to vary significantly person-to-person with some patients without detectable levels; appear to neutralize SARS-CoV-2 in a small human study and in animal studies (Zhao et al, Bao et al, Wölfel et al, Long et al, Qu et al, Wajnberg et al, Ni et al, Wu et al, Wu et al [different paper]))

  • Level and persistence of antibodies may be related to severity of illness (Long et al)
  • There have been reports of patients recovering from clinical illness, testing negative, then subsequently testing positive again (An et al, Lan et al, Ye et al) but no documented cases of reinfection and discordance likely due to known issues with PCR testing sensitivity (Petherick, Xiao et al, Kirkaldy et al, Yang et al, KCDC)
  • If immunity develops, it is unclear how long it will last -- immunity to other coronaviruses appears to wane after 1-2 years (Memish et al, Li & Xu, Kirkaldy et al)
  • There may also be a coronavirus-specific T-cell response, which could be more durable than the antibody response alone (Ni et al, Braun et al), may be cross-reactive with other coronaviruses (Grifoni et al)
  • Some have called for continued opening in pursuit of herd immunity, but low antibody seroprevalence in areas with rampant infections argues this approach may be unethical due to the high likely death toll (Pollan et al, Silveira et al, Stadlbauer et al, Moscola et al)


IN SHORT: Universal community and healthcare masking recommended. Contact + droplet isolation with airborne if aerosolizing procedures, cohort patients as possible, limit room entry, and use caution in donning/doffing.


  • Contact and droplet for COVID+ and rule out; airborne if aerosolizing procedure (CDC, WHO, SCCM, IDSA); avoid aerosolizing procedures if possible
  • Aerosolizing procedures include intubation, extubation, NIPPV, HFNC, suctioning, nebs, sputum induction, bronchoscopy, endoscopy, TEE, autopsy; possibly laparoscopy
  • Limit staff to minimum during aerosolizing procedures and use the most experienced operator available (SCCM, Cook et al)
  • Data is conflicting on whether HFNC or NIPPV are aerosolizing (Ne-Hooi et al, Cook, Iwashnya et al, Miller et al) though many institutions view it as such, consider using surgical mask over HFNC to reduce spread (Vapotherm, UW)
  • Laparoscopy should be considered a potentially aerosolizing procedure and viral filters should be used, avoid ALL elective procedures (SAGES)
  • Work to cohort patients in COVID-specific wards (or even hospitals) if possible (Anesi et al)
  • Negative pressure rooms if possible, mandatory for all aerosolizing procedures (SCCM)
  • Limit room entry: no visitors, increase telemedicine, no non-essential personnel, minimize lab draws, bring IV and vent controls into the hallway (Ferioli et al, CNN on UW ICUs, Islam)
  • CDC recommends discontinuing isolation based on testing (2 negative tests 24hrs apart) and/or symptoms (10-20 days from symptom onset depending on severity and 3 days from recovery), no longer recommend extended isolation based on viral PCR outside this window

PPE: universal masking recommended + at least gown, gloves, mask (surgical vs N95) and goggles or face shield for general care; for aerosolizing procedures add N95 or PAPR/CAPR, shroud/neck cover, and consider “aerosol box” to contain secretions (see Critical Care section for further details) (CDC, WHO, SCCM, IDSA, Luo et al, Wang et al, Liu et al, Bhaskar & Arun)

  • All PPE use should be combined with proper hand hygiene and regular cleaning of commonly used & contaminated surfaces (printers, door knobs, and keyboards) (Ran et al, Guangming et al)
  • Use extreme caution in donning/doffing PPE, especially if re-using, as virus has been detected on surgical masks 7 days after inoculation (Chin et al)
  • N95s  surgical masks >>>>> homemade cloth masks > nothing against droplet transmission (MacIntyre et al, Davies et al, van der Sande et al, Chu et al)
  • N95 vs surgical masks: meta-analyses indicate likely pragmatic equivalence (Offeddu et al, Bartozsko et al, IDSA) while some individual influenza studies show superiority of N95 over surgical face masks this appears to be mostly present in vitro with likely real-world equivalence (Smith et al, MacIntyre et al, Radonovich et al, Loeb et al, Bae et al) possibly due to improper use or lack of fit testing
  • Cloth masks are clearly inferior protection, with HCW wearing cloth masks vs standard clinical masks having RR 13 for contracting ILI (MacIntyre et al) though CDC & Joint Commission have released statements allowing for cloth masks as a last resort
  • Reuse is associated with escalating risk of mask failure (Degesys et al)

Coping with PPE Shortages

  • The CDC has created a PPE burn rate calculator to help institutions calculate needs
  • PPE distribution sites working to match HCWs needing PPE with people looking to donate:
  • Some institutions are having providers use single-use surgical mask over multi-use N95 to try to decrease contamination (IDSA), can also use rubber bands to improve surgical mask seal
  • Mask decontamination being attempted as supplies run low, but little/no testing or indication of safety and efficacy
  • Can consider homemade/3D printed face shields (here, here, & here) and face masks for use with changeable filters (here, here, & here; consider trying with vacuum bags/HEPA filters inside); none of these designs are tested or validated but they're likely better than running out


IN SHORT: Symptoms are highly variable, including not only fever, cough, and silent hypoxemia but also cardiac, GI, neurologic, and hypercoagulable symptoms. Asymptomatic infection exists and may be common, have a high index of suspicion for isolation and testing. Consider universal testing.


  • Show “silent hypoxemia” -- profound hypoxemia without respiratory distress, due to ARDS and likely shunt physiology (Xie et al, Gattinoni et al)
  • Unlikely hemoglobinopathy -- unable to find reputable evidence (Hardin) and patients exhibit low PaO2, which cannot be attributed to hemoglobinopathy
  • Severe respiratory failure due to ARDS, present in all intubated patients (Arentz et al, Hardin) & on pathology from patients who died prior to intubation (Barton et al)
  • High altitude medicine experts view COVID as a clearly different phenomenon with pathophysiology most consistent with ARDS (Luks & Swenson, Luks) despite some arguments pathophysiology could be similar to HAPE (Solaimanzadeh)
  • Cardiac symptoms may present in absence of respiratory symptoms (Zheng et al)
  • Presence of shock highly variable (1%-35, 54-95% of ventilated patients in NYC required prossors) (SCCM, Cummings et al, Goyal et al)
  • Some patients have non-obstructive disease, even when presenting with ECG findings of apparent STEMI (Bangalore et al)
  • Echo findings variable including hyperdynamic function, RV failure, stress cardiomyopathy, global hypokinesis, pericardial effusion (Peng et al, Ammirati & Wang)
  • Pathology series show atypical scattered cardiac myocyte degeneration, lymphocyte infiltration; some but not all with lymphocytic myocarditis (Fox et al, Tavazzi et al, Barton et al, Schaller et al)
  • Likely at least some degree of direct viral effect with case report of postmortem virus isolation from frontal lobe in patient with AMS (Paniz-Mondolfi et al), though other pathology series with predominantly hypoxic damage (Solomon et al)

Special populations:

Co-infection with other respiratory pathogens seems uncommon at presentation but may be more common in children and later in hospital/intubation course (Ding et al, Sommer et al, Xia et al, Blasco et al, Kim et al, Seattle Chest Grand Rounds, Mitja et al), secondary HAP, aspergillosis, and TB have been identified (Torrego et al, Arastehfar et al, Bartoletti et al, Tham et al)

Spectrum of disease: 81% mild/moderate, 14% severe, 5% critical (Wu & McGoogan) in adults; 4% asymptomatic, 51% mild, 39% moderate, 6% severe in children (Yuanyuan et al)


  • Generally begins with mild infection, progressing to pulmonary symptoms with or without hypoxia and hospitalization; unclear if further progression is due to ongoing viral damage or to hyperinflammation (MGH FLARE); treatment options may benefit from targeting to a particular disease phase (Siddiqi & Mehra) but there is no definitive data to support this
  • Average 4-7 days from symptom onset to hospitalization (Tian et al, Korean CDC)
  • 26-29% of hospitalized patients require ICU admission (indications: ARDS in 60-70% of ICU patients, shock in 30%; multi-organ dysfunction less common); most patients admitted or transferred to ICU within 24 hours (Wang et al, Cao et al, Phua et al, Arentz et al)
  • Long course of severe disease--average 25 days from symptom onset to discharge or 10-18 days from symptom onset to death (Arentz et al, Wang et al, Grasselli et al, Verity et al, Korean CDC)
  • Symptoms, esp. fatigue and dyspnea, may persist long after acute infection is over (Carfi et al)


IN SHORT: Viral studies may lead to false negatives but accuracy improves with repeat testing. Labs are variable but lymphopenia, elevated immune markers (CRP, LDH, D-dimer), and hypercoagulability appear common and end-organ damage (elevated troponin, Cr, LFTs) may occur.

Viral studies:

  • RT-PCR: sensitivity is variable and appears low (51-67%) for single nasopharyngeal PCR, increases with repetition but may depend on timing since exposure (Fang et al, Wang et al, Tao et al, Guo, Kucirka et al); false negatives may be less common than initially thought (Long et al) though 21% of patients in one study with positive result after two prior negatives (Xiao et al)
  • Guides and Bayesian calculators may help you guess post-test probability of COVID-19 more accurately based on your patient’s symptoms & local prevalence (Watson et al)
  • Increased sensitivity with BAL (93%), sputum (72%) (Wang et al) but unclear BAL or sputum induction advisable given aerosolization risk (Bouadma et al, NIH); endotracheal aspirate may be safer (NIH); studies also suggest use of saliva, nasal swab, and tongue swab self-collection (Guo et al, Williams et al, Tu et al)
  • Quantitative monitoring of viral load may correlate with disease progression (Yu et al)
  • Lab specimen preparation methods such as thermal inactivation may impact PCR yield and lead to false negative results, especially in lower viral loads (Pan et al)

Estimated variation over time of viral tests with nasal PCR in blue, IgM in purple, & IgG in green (Sethuraman et al)

  • Serological tests: early studies appear to show antibody titers inversely correlated with viral load and seroconversion ~10 days from symptom onset (Amanat et al, Kai-Wang To et al) but may be delayed in immunosuppressed patients (Zhao et al)
  • Antibody testing may have a high false-positive rate, esp in populations with low prevalence, and should not be used alone for disease confirmation (IDSA)
  • The COVID-19 Testing Project is performing head-to-head tests of commercially available serologic assays to assess performance
  • Combining PCR with IgM testing may significantly increase detection rate (Guo, Xie et al)
  • Antigen testing: much faster than PCR or antibody testing but less accurate → may allow for rapid mass screening with follow up PCR confirmation, FDA has issued an EUA for one version

Blood counts

Inflammatory markers

  • Fewer studies available in lab findings in children but may not show the same increase in inflammatory markers (Zimmerman & Curtis)


  • Seeing some elevation in LFTs (usually transaminases) but typically just above upper limited of normal & no liver failure reported; pathogenesis is unclear (Zhang et al, Hansheng et al, Musa)

Cardiac markers: elevated trop, CK-MB, BNP, myoglobin and ECG changes (incl. myocarditis w/ ST elevation) especially in severe disease (Zhou et al, Ruan et al, Han et al)

Coagulation: coagulopathy with elevated PT, PTT, thrombocytopenia appears common; DIC and even positive APLS serologies and lupus anticoagulant in some patients (Zhang et al, Bowles et al)


IN SHORT: Imaging findings are highly variable, especially early in disease, but may precede symptoms or lab abnormalities. Most studies report predominantly multilobar, bilateral, diffuse peripheral and subpleural ground glass opacities with lower lobe predominance. CT and US are much more sensitive than CXR.

Pulm CCM Guide to Lung Imaging in COVID19

Italian Radiology COVID19 Image Database

  • CXR with hazy, bilateral reticular opacities or GGOs; sensitivity appears variable, but increased after 72hrs (Arentz et al, Rubin et al)
  • Initial CXR may be normal, even in patients who later progress to severe disease requiring intubation (Goyal et al)
  • CT is highly sensitive (98%) but less specific than viral testing (Fang et al, Tao et al), shows progression over disease course (Shi et al)
  • However, screening with CT not generally recommended due to nonspecific nature of findings, radiation exposure, and resource utilization (esp. given disinfection times) (Hope et al, Rubin et al)
  • May see unilateral findings early on (including before symptom onset) (Shi et al)
  • Consolidation with halo sign may be seen in pediatric patients (Xia et al)
  • Progression to multifocal consolidation, air bronchograms, traction bronchiectasis, crazy paving appears correlated with more severe disease (Zhao et al, Kanne et al)
  • Vascular thickening and PEs may be associated with severe disease (Quanadli et al)
  • Echocardiography shows severe RV or LV dysfunction in half of patients and changes management in 33%, including those w/o known pre-existing cardiac disease (Dweck et al)


IN SHORT: Mortality rates vary but are mostly 1-3%, rising with age, comorbidities, male sex, and minority status. Labs indicating an appropriate immune response to a viral infection seem to be associated with a good prognosis while those indicating an inappropriate or dysregulated response, excessive inflammation, or end-organ damage seem to be associated with a poor prognosis.

Case fatality rates:

  • Rates may be underestimated as there are also many excess deaths during the pandemic not yet attributed to COVID (Weinberger et al)
  • Cause of death: 53% resp failure, 33% resp failure + heart failure, 5% heart failure (Ruan et al)

Patient/clinical indicators of severity/mortality:

  • Poorer & working-class communities more likely to be exposed as they are unlikely to be able to use telecommuting/working from home to socially distance (Yancy, Wise)
  • Health department data shows huge disparities by race/ethnicity, with disproportionately high minority mortality (esp. black, Hispanic, Native American patients); the Navajo Nation has more confirmed cases per capita than almost every U.S. state (NYC, Louisiana, Michigan, Yancy, Miller, Wortham et al)
  • Unclear if delivery changes disease course or outcomes or if delivery method has an impact on course or transmission (Breslin et al, Sutton et al)
  • ROX index (SpO2/FiO2/RR) < 3 may help predict need for intubation (Roca et al)
  • High mortality with cardiac involvement though pathway (stress CM vs direct viral injury vs hyperinflammation) currently unclear, may reflect multiple pathways including stress, demand, and direct injury (Chen et al, Zheng et al, Bonow et al)
  • Type of immunocompromise may be important -- organ transplant, malignancy appear to have increased mortality while HIV does not appear to impact outcomes (Sigel et al, Fung & Babik, Williamson et al)
  • Important contributor to severity in children -- few severe cases but occur much more in children with pre-existing conditions (Shekerdemian et al)
  • Profession: HCWs appear to get sicker than age would otherwise predict with 14.8% severe or critical cases in one study and 79% in another, with higher rates of infection in overwhelmed areas (Wu & McGoogan, Chu et al, Li et al, Zhan et al); this is theorized to be related to intensity of exposure and high viral inocula
  • Environment: mortality appears to increase with long-term air pollution exposure (Wu et al)
  • Surgical procedures: surgery on asymptomatic/presymptomatic patients associated with high rates of ICU admission (44%) and mortality (21%) (Lei et al)

Laboratory indicators of severity/mortality:

  • Inflammatory markers:
  • Blood counts:
  • Lymphocyte count: lymphocyte percentage of total WBC may be predictive of outcome based on a small study, with percentages >20% generally associated with good outcomes and percentages <5% were associated with poor outcomes and death, esp. later in disease course (Li et al, Han et al, Du et al, Li et al [different paper])
  • Platelets: platelet count may track severity of disease, higher platelet to lymphocyte ratio (i.e., platelet count/absolute lymphocyte count) may predict severe disease (Qu et al), and thrombocytopenia associated with severe disease (Lippi et al, ICNARC)
  • Suspect these may be indicative of two competing deleterious processes -- reactive thrombocytosis/hyperinflammation and thrombocytopenia/DIC
  • Viral load: higher viral loads may be associated with more severe disease and longer viral shedding (Liu et al, Lescure et al), though other studies show no difference in VL between symptomatic and asymptomatic patients (Cereda et al)
  • Troponin: elevated troponin appears to be a predictor of mortality regardless of presence of known CAD (though higher in presence of CAD) (Guo et al, Ruan et al)
  • Hemostasis/DIC: presence of DIC a marker of poor outcome (Tang et al); it is unclear if this is reflective of severity of disease or causing worsening disease through micro-/macro-thrombi
  • Liver enzymes: some studies have indicated elevated liver enzymes may be an independent predictor of mortality (Jiang et al, Chen et al, Laing et al) but others have found LFT elevation correlated with overall inflammation and was not an independent predictor of outcomes (Zhang et al, Hansheng et al); most elevations are mild and no liver failure has been reported
  • Renal function: AKI is an independent risk factor for mortality (Naicker et al, NephJC, ICNARC)
  • Blood type: early reports (Zeitz & Tatonetti) concerning for increased infections and mortality with blood type A & B and Rh positivity but later studies have not borne this out (Latz et al)

Calculators: early drafts show promise in predicting critical illness but will require further testing and validation to ensure accuracy (Liang et al, Li et al)


IN SHORT: Use guideline-based typical respiratory failure, ARDS, and shock treatment; there are many resources to brush up if needed. Closely monitor for respiratory decompensation and consider proning. Ensure you and your team have proper PPE for all care and procedures (“no emergencies in a pandemic”).

SCCM Surviving Sepsis COVID-19 Guidelines

AHA Interim Guidance for BLS/ACLS in patients with confirmed or suspected COVID-19

Critical Care protocols: University of Washington, Brigham & Women’s

MGH FLARE Fast Critical Care Literature Updates

JAMA COVID Crit Care podcast

Resources for critical care for non-intensivists:

Respiratory Failure

  • Clear plastic drapes or “aerosol boxes” have been suggested, but make intubation more technically challenging and may actually worsen aerosol exposure (Matava et al, Canelli et al, Rosenblatt et al, Begley et al, Simpson et al)
  • HEPA filters on all PPV devices (Brown et al)
  • Optimize preoxygenation to give as much time as possible for and increase chance of success of intubation attempt (Brown et al)
  • Use rapid sequence intubation and video laryngoscopy; work to preserve continuous vent circuit and consider connecting directly to vent as opposed to bagging first to minimize aerosol generation (Mark, Phua et al, Brown et al)
  • Mechanical ventilation: all patients with severe respiratory failure appear to have ARDS and societies recommend following typical ARDS protocols (SCCM, ATS)
  • Lung-protective ventilation with low tidal volumes (4-8cc/kg), PEEP titration with plateau pressure 30 as per traditional ARDS protocols (ARDSNet on LPV & PEEP, Berlin  et al)
  • Sedation: goal is to avoid vent dyssynchrony with analgesics (opiates), add amnestics as needed (propofol, benzos, ketamine); may need to consider alternative agents given shortages (MGH, Seattle Chest Grand Rounds)
  • Lighten sedation & mobilize early as able (including while intubated & proned) to combat profound weakness that appears to be a prominent feature (Seattle Chest Grand Rounds)
  • If having difficulty with tachypnea in febrile patients, before increasing sedation consider aggressively treating fever with antipyretics, cooled IVF, and external cooling (Seattle Chest Grand Rounds)

From SCCM guidelines

  • Paralytics: consider use to help ameliorate vent dyssynchrony per ARDS protocols but ensure patients are appropriately sedated and on an amnestic (SCCM, PETAL)
  • Dexmedetomidine is not amnestic, need to use an additional agent to achieve amnesia (including for paralysis) (Seattle Chest Grand Rounds)
  • Prone mechanically ventilated with severe hypoxemia early to promote lung recruitment, appears to show recruitment benefit in COVID-related ARDS (SCCM, ATS, PROSEVA, Pan et al, Ziehr et al)
  • Recruitment maneuvers: guidelines recommend use of traditional but not staircase recruitment maneuvers to help open up compressed alveoli (SCCM)
  • Pulmonary vasodilators can be considered as rescue therapy but guidelines recommend against inhaled NO (SCCM), there are discussions of potential similarities to HAPE but altitude medicine experts argue that the process is clearly ARDS and treatment with vasodilators could worsen V/Q mismatch by releasing appropriate, compensatory hypoxic vasoconstriction and perfusing areas with poor ventilation (Luks et al)
  • Tracheostomy: patients appear to have a prolonged need for mechanical ventilation, not much data on tracheostomy yet but recommendations to avoid until stable, 2-3 weeks from intubation, and ideally PCR negative given high aerosolization risk (AAO)

From SCCM guidelines

  • If capacity exists, consider VV use in young patients with few/no comorbidities, single organ dysfunction, and P:F ratio < 100 and poor trajectory despite aggressive ARDS management (Badulak, ELSO guidelines); not using VA or E-CPR at this time
  • Limited vent resources: may be able to repurpose other types of vent supports (transport vents, OR vents, home ventilators) to increase ventilatory resources (Seattle Chest Grand Rounds); ASA has created a guide to repurposing anesthesia machines as ICU vents
  • Many of the largest critical care organizations (SCCM, CHEST, ASA, AARC, and others) are recommending against splitting vents due to risk of increasing mortality by providing inadequate care to multiple patients
  • Vent liberation/weaning: ensure each patient has regular spontaneous awakening and spontaneous breathing (minimal pressure support 0-5mmH20) trials; use rapid shallow breathing index (RSBI = RR (breaths/min)/TV (L)) 105, work of breathing, mental status, cough, cuff leak, level of secretions, and clinical judgment to guide decisions (Ely et al, Meade et al, ATS)
  • Consider extended SBT in patients with prolonged time on ventilator (Wolf)
  • Consider extubating to NIPPV/HFNC in high-risk patients (ATS) but remember that extubation, NIPPV, and HFNC are all aerosolizing and ensure appropriate PPE (Wolf)


  • Fluid resuscitation: recommend conservative over liberal fluid resuscitation with typical preference for buffered crystalloids > NS > colloids (SCCM, NIH)
  • Pressors: norepi first line, vaso second line to target MAP 60-65 (SCCM)
  • If cardiac dysfunction or rapidly escalating pressor requirements despite fluids and norepi, recommend dobutamine and assessment for cardiomyopathy (SCCM, NIH)
  • Trial of low-dose corticosteroids for refractory shock (SCCM, NIH)

Cardiac arrest:

  • Outcomes are very poor (0.7% survival with good neuro outcome at 30 days) (Shao et al)
  • Don airborne + contact PPE prior to entering room without exception (AHA, Kramer et al)
  • Use PAPR/CAPR as N95 may not function as well during compressions (Hwang et al)
  • Limit personnel to only absolutely necessary and consider use of mechanical CPR devices to further limit personnel exposure (AHA)
  • Close doors of rooms in which ACLS/BLS is being conducted to reduce spread (AHA)
  • Prioritize intubation, pre-oxygenate with passive O2 or bag masks with HEPA filter (AHA)
  • Most experienced available provider should intubate using RSI and VL as above, ideally with cuffed ETT (AHA, SCCM, Cook et al, UW Intubation Guidelines)
  • In intubated patients, try to maintain a closed circuit and limit aerosolization by leaving on vent in asynchronous mode (AHA)
  • Attempt to leave proned patients in position unless able to turn without risk of equipment disconnection and aerosolization (AHA)
  • Out-of-hospital arrest has been significantly more common in the COVID era and had worse outcomes (CFR 90%), with most patients presenting with non-shockable rhythms (Lai et al)


IN SHORT: Supportive care is clearly beneficial and remdesivir and dexamethasone may show some benefit in a small subset of patients, but overall there remains no definitive targeted therapy. Use caution outside of clinical trials and ensure patients understand questionable benefit when consenting. Please remember to check for drug interactions and use caution in prescribing as treatment can cause proven harm.

Antiviral Therapies

Investigational, some clinical data available

Investigational, no clinical data available

  • Nitazoxanide: broad antiviral, good in-vitro activity, no clinical data yet (McCreary and Pogue)
  • Ivermectin: antiviral effect in vitro, no clinical data (Caly et al)
  • Favipiravir: approved in China per news reports, not available in US; an early study seemed to show improved clinical recovery vs another antiviral but no control group (Chen et al), another study is widely cited but was subsequently withdrawn
  • Metabolized in liver and may have significant drug-drug interactions (Du & Chen)


  • Lopinavir-Ritonavir, ribavirin, interferon: RCTs without clear efficacy of lopinavir-ritonavir alone (Young et al, Cao et al, Hung et al) but an open-label randomized trial in mild to moderate illness showed improvement in viral clearance in combination with one or more of ribavirin and interferon (Hung et al) which is more consistent with prior SARS-CoV-1 studies (Chu et al, Stockman et al, EmCrit); trend towards benefit in early disease
  • IDSA recommends use clinical trial use only
  • Contraindications: QTC > 500, caution in liver & cardiac disease; many drug-drug interactions; avoid ribavirin in pregnancy, asthma, and COPD

Ineffective/Likely ineffective

  • Use for COVID-19 is endangering stock available for proven beneficial use in rheumatologic disease (Yazdany & Kim)
  • Concomitant use of hydroxychloroquine + azithromycin is associated with increased risk of mortality, heart failure, and angina in large non-COVID cohorts (Lane et al)
  • Data from patients on hydroxychloroquine for other indications does not show any benefit in disease prevention or hospitalization rates (Konig et al, Macias et al) and has not shown efficacy in post-exposure prophylaxis (Boulware et al)
  • Contraindications: QTc > 500 (Giudicessi et al), history of ventricular arrhythmias, retinal disease; may cause ↑QTc, cardiomyopathy, arrhythmias (Bessière et al, Mercuro et al)
  • Overdose/Poisoning: hypotension, prolonged QRS/QTc, hypokalemia, apnea, seizures, ventricular arrhythmias (Poison Control); toxicity being seen across the world
  • Management: no specific reversal agents; early intubation, epinephrine superior to norepinephrine for hypotension, Na bicarb for conduction abnormalities, benzodiazepines for seizures (Clemessy et al)
  • Darunavir/Cobicistat: no evidence of impact on negative conversion (Chen et al)
  • Neuraminidase inhibitors (e.g., oseltamivir): no in vitro benefit in SARS-CoV-2 (Tan et al) but should still use if influenza coinfection

Immune therapies & immunomodulators

Investigational, some clinical data available

  • Corticosteroids: appear to have reduced mortality, improved SpO2 (though FiO2 not reported), improved imaging resolution, and reduced need for care escalation in COVID-related ARDS in small uncontrolled trials (Wu et al, Wang et al, Fadel et al) and large RCT (Horby et al) though benefit limited to patients on O2, intubated, or highly inflamed (CRP > 20) and trend towards harm in less ill patients and those with lower CRP (Keller et al)
  • SARS-CoV-2 vaccination: some early human and animal data is promising (Jackson et al, Mulligan et al, Zhu et al, Kim et al, Folegatti et al, Corbett et al, Xia et al), clinical trials enrolling; development likely to take time (>12 months) and efficacy required will depend on prevalence but likely need to be >70-80% (Bartsch et al)
  • Siltuximab: could theoretically help hyperinflammation, small case series in severe disease (no control group) with 33% clinical improvement and 24% clinical worsening (Gritti et al)
  • Anakinra: could theoretically help hyperinflammation, early cohort study with possible improvements in mortality (though negated if censorship effect excluded), resp symptoms, and inflammatory markers (Cavalli et al), clinical trials enrolling
  • Mesenchymal stem cell infusion: may help hyperinflammation based on prior studies in influenza pneumonias and GVHD, promising small open-label trials (Chen et al, Leng et al) and anecdotal experience (Atluri et al)
  • Prior ARDS study showed single mesenchymal stem cell infusion was likely safe (though non-statistically significant increase in mortality in treatment arm), not powered to assess efficacy but no clear signal of improvement (Matthay et al)
  • IVIG: may decrease hyperinflammation, very small series (three patients) with rapid improvement in fevers, respiratory status but many confounders and would need further testing to argue for use (Cao et al); note that IVIG is already on critical shortage in much of the US

Investigational, no clinical data available

  • Sarilumab: could theoretically help hyperinflammation, clinical trials enrolling; early press release indicates trend toward negative results in less severe illness but possible trend toward benefit in critical illness → trials now continuing with only critical illness arm
  • Llama antibodies: single-domain camelid antibodies appear to bind and neutralize SARS-CoV-1, MERS, and SARS-CoV-2 in vitro; no human testing yet (Wrapp et al)
  • Convalescent sera: May have clinical benefit if administered early, particularly to mildly ill older patients. Early small series of critically ill patients with recovery from ARDS, extubation, O2 wean (Shen et al, Casadevall & Pirofski, Duan et al, Salazar et al, Krammer et al). A double-blinded, randomized, placebo-controlled clinical trial of 160 patients age >65 in Buenos Aires showed benefit of HIGH titer sera, with relative risk reduction of progression to respiratory failure of 48% and a number needed to treat of 7 and a mortality risk difference of -0.46% (10.9 (Libster et al). The sera was administered early, within 72 hours, to patients with mild COVID (no hypoxemia, no tachypnea). Trial was stopped early due to poor enrollment as cases in the region declined. .  Early RCT and other studies without significant benefit (Li et al had non-significant trend but ended early, Gharbharan et al); studies v low quality overall (Cochrane); efficacy may depend on neutralizing antibody titer in sera (Bradfute et al),  SCCM currently recommends against use of convalescent plasma or IVIG and IDSA recommends use clinical trial use only
  • Tocilizumab: may help hyperinflammation in severe disease; small series with clinical and laboratory improvement (Luo et al, Xu et al, Zhang et al, Guaraldi et al) but RCT without improvement in clinical status or mortality per press release; IDSA recommends use clinical trial use only
  • Adverse reactions: serious infections, HBV reactivation, anaphylaxis, liver damage and hepatic failure, intestinal perforation, acute hypertriglyceridemia (IDSA, Morrison et al)

Ineffective/Likely ineffective

  • BCG vaccination: could theoretically improve “trained,” non-antigen-specific immune response (Netea et al) but current analyses touting this approach (Miller et al, Dayal & Gupta) are deeply methodologically flawed and appear to cherry-pick data (Vinarsky); subsequent analysis showed no difference in infection rates in BCG vaccinated vs unvaccinated groups (Hamiel et al)

Other therapies

Investigational, some clinical data available

  • Anticoagulation: hypercoagulability, thrombi observed likely due to generalized inflammatory state, ARDS, immobility though most studies without control group so relative rate unclear
  • Prophylactic dose: chemoprophylaxis (LMWH preferred) recommended by multiple societies and guidelines (ASH, ISTH, MGH); likely mortality benefit to prophylactic dose UFH/LMWH (Tang et al, Yin et al, Tang et al [different paper])
  • Consider use of LWMH daily instead of BID/TID heparin SQ for chemoprophylaxis to minimize staff exposure (JACC)
  • Ensure that dosing is sufficient, esp. in obese patients (Poissy et al)
  • Therapeutic dose: uncontrolled observational trial points to possible mortality benefit but unclear how groups selected, immortal time bias, and control group mortality extremely high (higher than other reported ICU cohorts in any study) raising the possibility of an additional confounders) (Paranjpe et al); recommended by Lin et al but appears to be based on hypothesis of pathogenesis, not clinical data; no consensus exists and most recommendations suggest avoiding use unless other clinical evidence suggestive of thrombosis (and ideally confirmatory US/CT imaging) (ASH, JACC, MGH, Seattle Chest Grand Rounds); clinical trials enrolling
  • Prior studies in hospitalizations for infection indicated decreased VTE but increased major bleeding with NNT=NNH (Cohoon et al)
  • If heparin gtt is required, recommend monitoring with anti-Xa levels instead of PTT due to PTT lability in COVID (MGH, Beun et al)
  • If on prior therapeutic AC, continue but monitor closely for drug interactions (esp. between VKA/anything and DOAC/immunomodulating medications)

Investigational, no clinical data available

  • Recombinant ACE2: dose-dependent efficacy in vitro at preventing viral entry (Monteil et al) but no clinical data yet, undergoing small observational trial in China, no results yet
  • Angiotensin II: could theoretically help in shock (Chow et al) but no trials yet
  • tPA: in consideration given high rate of thromboses, initial series of three patients without bleeding events but with unclear (and if present, transient) benefit (Wang et al)

Ineffective/Likely ineffective

  • High dose IV vitamin C can cause spuriously elevated POC glucose readings with certain glucometer brands (Kahn & Lentz), may also contribute to hypernatremia (Chang)

Concomitant medications

  • Continue ACE/ARB for patients already taking, new initiation for lung protection in patients without other indications not currently recommended (Vaduganathan et al, Sparks et al, ACC)


IN SHORT: Ethics and standards of care may change in a crisis to emphasize community benefit, but there are specific frameworks to familiarize yourself with and follow. Unless specifically told you are entering contingency or crisis standards, assume normal standards of care apply. Centers should create triage experts to help with decision making.

AMA Journal of Ethics COVID19 Resource Center

NEJM on Fair Allocation of Scarce Medical Resources in the Time of Covid-19

JAMA Framework for Rationing Ventilators and Critical Care Beds During the COVID-19 Pandemic

SCCM Resource Availability for COVID19

SCCM on Acute Surge Planning

  • Standards of care may shift during emergency situations, with three prominent levels (IOM):
  • Conventional: normal levels of staff, supplies; provide conventional care consistent with usual practices & standards
  • Contingency: abnormal staffing, supplies but can be adapted to provide care functionally equivalent to usual patient care; typical standards of care predominantly apply
  • Crisis: staffing, supplies, space cannot be adapted sufficiently to provide typical care due to a catastrophic disaster or emergency (including pandemic), work to provide best possible care to patients given the circumstances and resources available but activation constitutes a significant adjustment to standards of care with focus on community instead of simply individual patients
  • Many US states are starting to come up with crisis standards of care (WA State here) and plans for triage; look up your regional standards and keep in touch with your hospital to ensure appropriate response to changing conditions
  • Do not apply crisis standards of care until level reached and declared by authorities; many communities are currently in contingency but not crisis standards and applying crisis too early may needlessly sacrifice lives
  • Many experts recommend creation of triage panels or officers separate from patient care to ensure any crisis decisions are objective and take pressure off of frontline providers (Emanuel et al, Truog et al, Anesi et al)
  • Protections for caregivers participating in triage from criminal or civil suits are currently piecemeal and leave frontline workers at risk, reform is urgently needed (Cohen et al)
  • Ethicists argue that providers should not be required to perform care in absence of adequate PPE, but also that they cannot abdicate the responsibility to care if PPE is available (Kramer et al)


IN SHORT: Goals of care discussions should occur early and often with clear, objective predictions of prognosis and definitive care recommendations offered. End-of-life care should focus on promoting family connections as able and managing dyspnea.

VitalTalk site & app for discussions with families, including a targeted guide for discussions with patients and families about COVID (Back et al) and COVID-specific family discussion videos

San Francisco VA Advanced Care Planning Communication Guidelines

NHS Scotland poster on Difficult Conversations

Goals of care:

  • Early & repeated goals of care discussion, trying to focus first on patient’s goals and then their code status (Curtis et al)
  • Consider early Palliative Care consults (even in ER) to help with discussions
  • Be clear about poor prognosis in pre-existing terminal conditions such as dementia as families may not have heard this information before (as one Palliative Care provider put it, “to be clear is to be kind” even if it feels blunt or cruel to deliver bad news)
  • Consider working on informed assent instead of informed consent-- making a recommendation regarding code status based on goals and likely outcomes to which a patient/family may agree or disagree rather than asking them what they would like (Curtis et al)
  • Provide specific, objective outcome predictions -- thus far, COVID-19 CPR data indicates extremely poor outcomes with only 13.2% ROSC, 2.9% 30-day survival, and 0.7% 30-day survival with good neurologic outcome (Shao et al)
  • The goals of care discussion process may differ depending on level of emergency/standards of care (see Ethics above) -- in general, we should adhere to traditional norms of patient autonomy in determining code status and allow for disagreement with our recommendations
  • Medical futility exists in some patients regardless of care conditions, iIf medically appropriate and legal in your region, may still declare a treatment medically futile
  • If entering crisis standards of care, families may face triage team decisions they do not understand and may hate; triage re-consideration should be based on objective medical information as per crisis guidelines and not on family discomfort with decision

End-of-life care:

  • Help patients and families connect via visitation (if permitted) or video or audio calls (Etkind et al); if unable to arrange a call consider helping a patient to record a last message to loved ones
  • Dyspnea/air hunger control appears to be one of the greatest challenges with this disease and may require admission for symptom management (Etkind et al)
  • O2 supplementation should still be via non-aerosolizing means (NC, NRB) if possible and should focus on dyspnea, not hypoxemia (Fausto, Ting et al)
  • Consider early use of opiate & benzodiazepine infusions for end-of-life symptom management as may be able to control symptoms while minimizing PPE use esp. if IV poles are placed outside rooms (Fausto, Ting et al)


Counseling & therapy:

Guides to self-care, coping:

Meditation/mindfulness apps:

Moments of zen: