Practice Rapid Recommendations A living WHO guideline on drugs for covid-19 BMJ 2020; 370 doi: https://doi.org/10.1136/bmj.m3379 (Published 04 September 2020) Cite this as: BMJ 2020;370:m3379 Visual summary of recommendation Last updated 10 Nov 2023 Interventions Population Disease severity Non-severe Severe Critical Requires life sustaining treatment Acute respiratory distress syndrome Sepsis Septic shock Absence of signs Risk of admissionto hospital: of severe or critical disease This recommendation applies only to people with these characteristics: Patients with confirmed covid-19 Oxygen saturation <90% on room air Signs of pneumonia Strong recommendations in favour Weak or conditional recommendations in favour Weak or conditional recommendations against Strong recommendations against Lopinavir-ritonavir Casirivimab and imdevimab Sotrovimab Hydroxychloroquine Colchicine Ivermectin Convalescentplasma Corticosteroids Ruxolitinib and tofacitinib Should be considered only if neither baricitinib nor IL-6 receptor blockers are available Remdesivir Remdesivir Molnupiravir Mitigation strategies to reduce potential harms should be implemented Remdesivir Molnupiravir Mitigation strategies to reduce potential harms should be implemented Molnupiravir Nirmatrelvirand ritonavir Nirmatrelvirand ritonavir Remdesivir Remdesivir All three may be combined IL-6 receptor blockers Corticosteroids Baricitinib M M M L L L Use the interactive multiple comparison tool to compare and choose treatments for patients at moderate or high risk of hospital admission Fluvoxamine Fluvoxamine Only in research settings The panel inferred that most patients would want to receive fluvoxamine only in the context of a randomised trial, given the uncertainty around potential benefits and the possibility of harms Convalescentplasma Only in research settings The panel inferred that most patients would want to receive convalescent plasma only in the context of a randomised trial, given the uncertainty around potential benefits and the possibility of harms There are also several practical issues related to the use of convalescent plasma, including but not limited to: Collection of plasma Storage and distribution of plasma Infusion of convalescent plasma into recipients Identification and recruitment of potential donors Only in research settings The panel inferred that most patients would want to receive ivermectin only in the context of a randomised trial, given the uncertainty around potential benefits and the possibility of harms Ivermectin Only in research settings The panel inferred that most patients would want to receive VV116 only in the context of a randomised trial, given the uncertainty around potential benefits and the possibility of harms VV116 UPDATE New recommendation UPDATE Ivermectin is no longer recommended for people with non-severe disease, even in research settings Nirmatrelvirand ritonavir H H H UPDATE The following recommendations for people with non-severe disease are now stratified by how likely it is for someone to be admitted to hospital Nirmatrelvir and ritonavir Remdesivir Molnupiravir UPDATE Low L Patients at low risk of hospital admission (0.5%)Includes people who are neither moderate nor high risk. Most patients are at low risk Moderate M Patients at moderate risk of hospital admission (3%)Includes people: over 65 years old with obesity with diabetes with active cancer with disabilities with comorbidities of chronic disease with chronic kidney or liver disease with chronic cardiopulmonary disease High H Patients at high risk of hospital admission (6%)Includes people who have: been diagnosed with immunodeficiency syndromes been diagnosed with immunodeficiency syndromes autoimmune illness, and are receiving immunosuppressants undergone sold organ transplant and are receiving immunosuppressants Signs of severe respiratory distress In adults: Accessory muscle use Inability to complete full sentences Respiratory rate > 30 breaths per minute In children: Very severe chest wall indrawing Grunting Central cyanosis Inability to breastfeed or drink Reduced level of consciousness Lethargy Convulsions Corticosteroids Corticosteroids Suggested regimen Acceptable alternative regimens Dexamethasone 6 mg Oral or intravenous Hydrocortisone 50 mg Intravenous Every 8 hours for 7-10 days Daily for7-10 days Every 6 hours for 7-10 days Methylprednisolone 10 mg Intravenous Daily for7-10 days Prednisone 40 mg Oral Recommendation 1Supportive careCorticosteroidsorPatients withnon-severe covid-19We suggest no corticosteroidsStrongAll or nearly all informed people would likelywant the intervention to the left. Benefitswould outweigh harms for almost everyoneWeakMost people would likely want the interventionto the left. Benefits would outweigh harms forthe majority, but not for everyoneWeakMost people would likely want the interventionto the right. Benefits would outweigh harmsfor the majority, but not for everyoneStrongAll or nearly all informed people would likelywant the intervention to the right. Benefitswould outweigh harms for almost everyoneEvidence profileFavours supportive careFavours corticosteroidsNo important differenceThe panel found that this difference was not important for most patients,because the intervention effects were negligible and/or very imprecise, forexample confidence intervals that include both important benefit and harmEvidence qualityEvents per 1000 peopleMortality5 fewer2328LowMoreBased on data from 1535 patients in 1 studyCorticosteroids may increase the risk of28-day mortality in patients withnon-severe covid-19LowGRADE rating,because of:Risk of biasSeriousImprecisionSeriousIndirectnessNo serious concernsInconsistencyNo serious concernsPublication biasNo serious concernsGRADE certainty ratingsHighThe authors have a lot ofconfidence that the trueeffect is similar to theestimated effectModerateThe authors believe thatthe true effect is probablyclose to theestimated effectLowThe true effect might bemarkedly different fromthe estimated effectVery lowThe true effect isprobably markedlydifferent from theestimated effectMechanical ventilation30 fewer11686ModerateMoreBased on data from 5481 patients in 2 studiesCorticosteroids probably reduce the needfor mechanical ventilationModerateGRADE rating,because of:Risk of biasSeriousImprecisionNo serious concernsIndirectnessNo serious concernsInconsistencyNo serious concernsPublication biasNo serious concernsGRADE certainty ratingsHighThe authors have a lot ofconfidence that the trueeffect is similar to theestimated effectModerateThe authors believe thatthe true effect is probablyclose to theestimated effectLowThe true effect might bemarkedly different fromthe estimated effectVery lowThe true effect isprobably markedlydifferent from theestimated effectGastrointestinal bleedingNo important difference4851LowMoreBased on data from 5403 patients in 30 studiesCorticosteroids may not increase therisk of gastrointestinal bleedingLowGRADE rating,because of:Risk of biasNo serious concernsImprecisionSeriousIndirectnessSeriousInconsistencyNo serious concernsPublication biasNo serious concernsGRADE certainty ratingsHighThe authors have a lot ofconfidence that the trueeffect is similar to theestimated effectModerateThe authors believe thatthe true effect is probablyclose to theestimated effectLowThe true effect might bemarkedly different fromthe estimated effectVery lowThe true effect isprobably markedlydifferent from theestimated effectSuperinfectionsNo important difference186188LowMoreBased on data from 6027 patients in 32 studiesCorticosteroids may not increase therisk of superinfectionsLowGRADE rating,because of:Risk of biasNo serious concernsImprecisionSeriousIndirectnessSeriousInconsistencyNo serious concernsPublication biasNo serious concernsGRADE certainty ratingsHighThe authors have a lot ofconfidence that the trueeffect is similar to theestimated effectModerateThe authors believe thatthe true effect is probablyclose to theestimated effectLowThe true effect might bemarkedly different fromthe estimated effectVery lowThe true effect isprobably markedlydifferent from theestimated effectHyperglycaemia46 fewer286332ModerateMoreBased on data from 8938 patients in 24 studiesCorticosteroids probably increase therisk of hyperglycaemiaModerateGRADE rating,because of:Risk of biasNo serious concernsImprecisionNo serious concernsIndirectnessSeriousInconsistencyNo serious concernsPublication biasNo serious concernsGRADE certainty ratingsHighThe authors have a lot ofconfidence that the trueeffect is similar to theestimated effectModerateThe authors believe thatthe true effect is probablyclose to theestimated effectLowThe true effect might bemarkedly different fromthe estimated effectVery lowThe true effect isprobably markedlydifferent from theestimated effectHypernatremia26 fewer4066ModerateMoreBased on data from 5015 patients in 6 studiesCorticosteroids probably increase therisk of hypernatremiaModerateGRADE rating,because of:Risk of biasNo serious concernsImprecisionNo serious concernsIndirectnessSeriousInconsistencyNo serious concernsPublication biasNo serious concernsGRADE certainty ratingsHighThe authors have a lot ofconfidence that the trueeffect is similar to theestimated effectModerateThe authors believe thatthe true effect is probablyclose to theestimated effectLowThe true effect might bemarkedly different fromthe estimated effectVery lowThe true effect isprobably markedlydifferent from theestimated effectNeuromuscular weaknessNo important difference6975LowMoreBased on data from 6358 patients in 8 studiesCorticosteroids may not increase therisk of neuromuscular weaknessLowGRADE rating,because of:Risk of biasNo serious concernsImprecisionSeriousIndirectnessSeriousInconsistencyNo serious concernsPublication biasNo serious concernsGRADE certainty ratingsHighThe authors have a lot ofconfidence that the trueeffect is similar to theestimated effectModerateThe authors believe thatthe true effect is probablyclose to theestimated effectLowThe true effect might bemarkedly different fromthe estimated effectVery lowThe true effect isprobably markedlydifferent from theestimated effectNeuropsychiatric effectsNo important difference3528LowMoreBased on data from 1813 patients in 7 studiesCorticosteroids may not increase therisk of neuropsychiatric effectsLowGRADE rating,because of:Risk of biasNo serious concernsImprecisionSeriousIndirectnessSeriousInconsistencyNo serious concernsPublication biasNo serious concernsGRADE certainty ratingsHighThe authors have a lot ofconfidence that the trueeffect is similar to theestimated effectModerateThe authors believe thatthe true effect is probablyclose to theestimated effectLowThe true effect might bemarkedly different fromthe estimated effectVery lowThe true effect isprobably markedlydifferent from theestimated effectEvidence qualityMean daysDuration of hospital stay1 fewer1312LowMoreBased on data from 6425 patients in 1 studyCorticosteroids may result in animportant reduction in the duration ofhospital stayLowGRADE rating,because of:Risk of biasSeriousImprecisionSeriousIndirectnessNo serious concernsInconsistencyNo serious concernsPublication biasNo serious concernsGRADE certainty ratingsHighThe authors have a lot ofconfidence that the trueeffect is similar to theestimated effectModerateThe authors believe thatthe true effect is probablyclose to theestimated effectLowThe true effect might bemarkedly different fromthe estimated effectVery lowThe true effect isprobably markedlydifferent from theestimated effectSee patient decision aidsSee all outcomesNon-severely illIndividual considerations Key practical issues The panel inferred that most fully informed individuals with non-severe illness would not want to receive corticosteroids, but many could want to consider this intervention through shared decision-making with their treating physician. When treating patients with non-severe disease, even after 7 days of symptoms, the panel concluded that it was preferable to err on the side of no corticosteroids Values and preferences Corticosteroids Usual supportive care In order to help guarantee access to therapy for severe or critical covid-19 patients, it is reasonable to avoid administering corticosteroids to patients who are less likely to derive benefit No additional practical issues Recommendation 2Supportive careCorticosteroidsorPatients with severe orcritical covid-19We recommend corticosteroidsStrongAll or nearly all informed people would likelywant the intervention to the left. Benefitswould outweigh harms for almost everyoneWeakMost people would likely want the interventionto the left. Benefits would outweigh harms forthe majority, but not for everyoneWeakMost people would likely want the interventionto the right. Benefits would outweigh harmsfor the majority, but not for everyoneStrongAll or nearly all informed people would likelywant the intervention to the right. Benefitswould outweigh harms for almost everyoneConditions for use of treatment IL-6 receptor blockers and baricitinib are also recommended, and may be administered in combination with corticosteroids. The incremental survival benefit afforded by baricitinib exists even among patients also treated with IL-6 receptor blockers and baricitinib Caution is advised when considering combinations in patients with an increased risk of opportunistic infections Consider combining medications in a stepwise fashion in patients who are deteriorating Evidence profileFavours supportive careFavours corticosteroidsNo important differenceThe panel found that this difference was not important for most patients,because the intervention effects were negligible and/or very imprecise, forexample confidence intervals that include both important benefit and harmEvidence qualityEvents per 1000 peopleMortality34 fewer160126ModerateMoreBased on data from 1703 patients in 7 studiesCorticosteroids probably reduce the riskof 28-day mortality in patients withcritical illness due to covid-19ModerateGRADE rating,because of:Risk of biasSeriousImprecisionNo serious concernsIndirectnessNo serious concernsInconsistencyNo serious concernsPublication biasNo serious concernsGRADE certainty ratingsHighThe authors have a lot ofconfidence that the trueeffect is similar to theestimated effectModerateThe authors believe thatthe true effect is probablyclose to theestimated effectLowThe true effect might bemarkedly different fromthe estimated effectVery lowThe true effect isprobably markedlydifferent from theestimated effectMechanical ventilation30 fewer11686ModerateMoreBased on data from 5481 patients in 2 studiesCorticosteroids probably reduce the needfor mechanical ventilation at 28 daysModerateGRADE rating,because of:Risk of biasSeriousImprecisionNo serious concernsIndirectnessNo serious concernsInconsistencyNo serious concernsPublication biasNo serious concernsGRADE certainty ratingsHighThe authors have a lot ofconfidence that the trueeffect is similar to theestimated effectModerateThe authors believe thatthe true effect is probablyclose to theestimated effectLowThe true effect might bemarkedly different fromthe estimated effectVery lowThe true effect isprobably markedlydifferent from theestimated effectGastrointestinal bleedingNo important difference4851LowMoreBased on data from 5403 patients in 30 studiesCorticosteroids may not increase therisk of gastrointestinal bleedingLowGRADE rating,because of:Risk of biasNo serious concernsImprecisionSeriousIndirectnessSeriousInconsistencyNo serious concernsPublication biasNo serious concernsGRADE certainty ratingsHighThe authors have a lot ofconfidence that the trueeffect is similar to theestimated effectModerateThe authors believe thatthe true effect is probablyclose to theestimated effectLowThe true effect might bemarkedly different fromthe estimated effectVery lowThe true effect isprobably markedlydifferent from theestimated effectSuperinfectionsNo important difference186188LowMoreBased on data from 6027 patients in 32 studiesCorticosteroids may not increase therisk of superinfectionsLowGRADE rating,because of:Risk of biasNo serious concernsImprecisionSeriousIndirectnessSeriousInconsistencyNo serious concernsPublication biasNo serious concernsGRADE certainty ratingsHighThe authors have a lot ofconfidence that the trueeffect is similar to theestimated effectModerateThe authors believe thatthe true effect is probablyclose to theestimated effectLowThe true effect might bemarkedly different fromthe estimated effectVery lowThe true effect isprobably markedlydifferent from theestimated effectHyperglycaemia46 fewer286332ModerateMoreBased on data from 8938 patients in 24 studiesCorticosteroids probably increase therisk of hyperglycaemiaModerateGRADE rating,because of:Risk of biasNo serious concernsImprecisionNo serious concernsIndirectnessSeriousInconsistencyNo serious concernsPublication biasNo serious concernsGRADE certainty ratingsHighThe authors have a lot ofconfidence that the trueeffect is similar to theestimated effectModerateThe authors believe thatthe true effect is probablyclose to theestimated effectLowThe true effect might bemarkedly different fromthe estimated effectVery lowThe true effect isprobably markedlydifferent from theestimated effectNeuromuscular weaknessNo important difference6975LowMoreBased on data from 6358 patients in 8 studiesCorticosteroids may not increase therisk of neuromuscular weaknessLowGRADE rating,because of:Risk of biasNo serious concernsImprecisionSeriousIndirectnessSeriousInconsistencyNo serious concernsPublication biasNo serious concernsGRADE certainty ratingsHighThe authors have a lot ofconfidence that the trueeffect is similar to theestimated effectModerateThe authors believe thatthe true effect is probablyclose to theestimated effectLowThe true effect might bemarkedly different fromthe estimated effectVery lowThe true effect isprobably markedlydifferent from theestimated effectNeuropsychiatric effectsNo important difference3528LowMoreBased on data from 1813 patients in 7 studiesCorticosteroids may not increase therisk of neuropsychiatric effectsLowGRADE rating,because of:Risk of biasNo serious concernsImprecisionSeriousIndirectnessSeriousInconsistencyNo serious concernsPublication biasNo serious concernsGRADE certainty ratingsHighThe authors have a lot ofconfidence that the trueeffect is similar to theestimated effectModerateThe authors believe thatthe true effect is probablyclose to theestimated effectLowThe true effect might bemarkedly different fromthe estimated effectVery lowThe true effect isprobably markedlydifferent from theestimated effectSee all outcomesIndividual considerations Key practical issues The panel inferred that all or almost all fully informed patients with severe or critical covid-19 would choose to take corticosteroids. In considering potential contraindications, clinicians must determine if they warrant depriving a patient from a potentially lifesaving treatment Values and preferences Corticosteroids Usual supportive care Absolute contraindications for 7 to 10 day courses of corticosteroid treatment are rare No clear differences in efficacy or adverse events between different corticosteroids Recommended regimens are available globally, and relatively inexpensive No additional practical issues Interleukin-6 (IL-6) receptor blockers Suggested regimen Tocilizumab Max 800 mg 8 mg per kg Intravenous Initial dose over 1 hour or Sarilumab 400 mg Intravenous Initial dose over 1 hour A second dose may be administered after 12 to 48 hours Recommendation 1Supportive careIL-6 receptor blockersorPatients with severe orcritical covid-19We recommend treatment with IL-6 receptor blockers(tocilizumab or sarilumab)StrongAll or nearly all informed people would likelywant the intervention to the left. Benefitswould outweigh harms for almost everyoneWeakMost people would likely want the interventionto the left. Benefits would outweigh harms forthe majority, but not for everyoneWeakMost people would likely want the interventionto the right. Benefits would outweigh harmsfor the majority, but not for everyoneStrongAll or nearly all informed people would likelywant the intervention to the right. Benefitswould outweigh harms for almost everyoneConditions for use of treatment Corticosteroids and baricitinib are also recommended, and may be administered in combination with baricitinib. The incremental survival benefit afforded by baricitinib exists even among patients also treated with corticosteroids and baricitinib Caution is advised when considering combinations in patients with an increased risk of opportunistic infections Consider combining medications in a stepwise fashion in patients who are deteriorating Evidence profileFavours supportive careFavours IL-6 receptor blockersNo important differenceThe panel found that this difference was not important for most patients,because the intervention effects were negligible and/or very imprecise, forexample confidence intervals that include both important benefit and harmEvidence qualityEvents per 1000 peopleMortality15 fewer130115HighMoreBased on data from 6526 patients in 17 studiesIL-6 receptor blockers reduce mortalityHighGRADE rating,because of:Risk of biasNo serious concernsImprecisionNo serious concernsIndirectnessNo serious concernsInconsistencyNo serious concernsPublication biasNo serious concernsGRADE certainty ratingsHighThe authors have a lot ofconfidence that the trueeffect is similar to theestimated effectModerateThe authors believe thatthe true effect is probablyclose to theestimated effectLowThe true effect might bemarkedly different fromthe estimated effectVery lowThe true effect isprobably markedlydifferent from theestimated effectMechanical ventilation23 fewer8663HighMoreBased on data from 5686 patients in 9 studiesIL-6 receptor blockers reduce the needfor mechanical ventilationHighGRADE rating,because of:Risk of biasNo serious concernsImprecisionNo serious concernsIndirectnessNo serious concernsInconsistencyNo serious concernsPublication biasNo serious concernsGRADE certainty ratingsHighThe authors have a lot ofconfidence that the trueeffect is similar to theestimated effectModerateThe authors believe thatthe true effect is probablyclose to theestimated effectLowThe true effect might bemarkedly different fromthe estimated effectVery lowThe true effect isprobably markedlydifferent from theestimated effectSerious adverse eventsNo important difference95Very lowMoreBased on data from 815 patients in 2 studiesThe effect of IL-6 receptor blockers onserious adverse events is uncertainVery lowGRADE rating,because of:Risk of biasSeriousImprecisionVery seriousIndirectnessNo serious concernsInconsistencyNo serious concernsPublication biasNo serious concernsGRADE certainty ratingsHighThe authors have a lot ofconfidence that the trueeffect is similar to theestimated effectModerateThe authors believe thatthe true effect is probablyclose to theestimated effectLowThe true effect might bemarkedly different fromthe estimated effectVery lowThe true effect isprobably markedlydifferent from theestimated effectBacterial infectionsNo important difference10196LowMoreBased on data from 3548 patients in 18 studiesIL-6 receptor blockers may not increasesecondary bacterial infectionsLowGRADE rating,because of:Risk of biasSeriousImprecisionSeriousIndirectnessNo serious concernsInconsistencyNo serious concernsPublication biasNo serious concernsGRADE certainty ratingsHighThe authors have a lot ofconfidence that the trueeffect is similar to theestimated effectModerateThe authors believe thatthe true effect is probablyclose to theestimated effectLowThe true effect might bemarkedly different fromthe estimated effectVery lowThe true effect isprobably markedlydifferent from theestimated effectEvidence qualityMean daysDuration of ventilation1.2 fewer14.713.5LowMoreBased on data from 1189 patients in 10 studiesIL-6 receptor blockers may reduceduration of mechanical ventilationLowGRADE rating,because of:Risk of biasSeriousImprecisionSeriousIndirectnessNo serious concernsInconsistencyNo serious concernsPublication biasNo serious concernsGRADE certainty ratingsHighThe authors have a lot ofconfidence that the trueeffect is similar to theestimated effectModerateThe authors believe thatthe true effect is probablyclose to theestimated effectLowThe true effect might bemarkedly different fromthe estimated effectVery lowThe true effect isprobably markedlydifferent from theestimated effectLength of hospital stay4.5 fewer12.88.3LowMoreBased on data from 6665 patients in 9 studiesIL-6 receptor blockers may reduce lengthof hospital stayLowGRADE rating,because of:Risk of biasSeriousImprecisionNo serious concernsIndirectnessNo serious concernsInconsistencySeriousPublication biasNo serious concernsGRADE certainty ratingsHighThe authors have a lot ofconfidence that the trueeffect is similar to theestimated effectModerateThe authors believe thatthe true effect is probablyclose to theestimated effectLowThe true effect might bemarkedly different fromthe estimated effectVery lowThe true effect isprobably markedlydifferent from theestimated effectSee all outcomesIndividual considerations Key practical issues The benefit of IL-6 receptor blockers on mortality was deemed of critical importance to patients, despite the small magnitude of effect and the very low certainty around serious adverse events. A majority of the panel inferred that almost all well informed patients would want to receive IL-6 receptor blockers Values and preferences IL-6 receptor blockers Usual supportive care Caution is advised when considering the use of tocilizumab in patients with a history of recurring or chronic infections All patients should be monitored for signs and symptoms of infection No additional practical issues Janus kinase (JAK) inhibitors Suggested regimen Baricitinib 4 mg Oral Daily Ruxolitinib 5 mg Oral Twice daily Tofacitinib 10 mg Oral Twice daily For 14 days or until hospital discharge Recommendation 1Supportive careBaricitiniborPatients with severe orcritical covid-19We recommend treatment with baricitinibStrongAll or nearly all informed people would likelywant the intervention to the left. Benefitswould outweigh harms for almost everyoneWeakMost people would likely want the interventionto the left. Benefits would outweigh harms forthe majority, but not for everyoneWeakMost people would likely want the interventionto the right. Benefits would outweigh harmsfor the majority, but not for everyoneStrongAll or nearly all informed people would likelywant the intervention to the right. Benefitswould outweigh harms for almost everyoneConditions for use of treatment Corticosteroids and IL-6 receptor blockers are also recommended, and may be administered in combination with baricitinib. The incremental survival benefit afforded by baricitinib exists even among patients also treated with corticosteroids and IL-6 receptor blockers Caution is advised when considering combinations in patients with an increased risk of opportunistic infections Consider combining medications in a stepwise fashion in patients who are deteriorating Evidence profileFavours supportive careFavours baricitinibNo important differenceThe panel found that this difference was not important for most patients,because the intervention effects were negligible and/or very imprecise, forexample confidence intervals that include both important benefit and harmEvidence qualityEvents per 1000 peopleMortality20 fewer130110HighMoreBased on data from 10815 patients in 4 studiesBaricitinib reduces mortalityHighGRADE rating,because of:Risk of biasNo serious concernsImprecisionNo serious concernsIndirectnessNo serious concernsInconsistencyNo serious concernsPublication biasNo serious concernsGRADE certainty ratingsHighThe authors have a lot ofconfidence that the trueeffect is similar to theestimated effectModerateThe authors believe thatthe true effect is probablyclose to theestimated effectLowThe true effect might bemarkedly different fromthe estimated effectVery lowThe true effect isprobably markedlydifferent from theestimated effectMechanical ventilation11 fewer116105ModerateMoreBased on data from 8412 patients in 3 studiesBaricitinib probably reduces mechanicalventilationModerateGRADE rating,because of:Risk of biasNo serious concernsImprecisionSeriousIndirectnessNo serious concernsInconsistencyNo serious concernsPublication biasNo serious concernsGRADE certainty ratingsHighThe authors have a lot ofconfidence that the trueeffect is similar to theestimated effectModerateThe authors believe thatthe true effect is probablyclose to theestimated effectLowThe true effect might bemarkedly different fromthe estimated effectVery lowThe true effect isprobably markedlydifferent from theestimated effectSerious adverse eventsNo important difference05ModerateMoreBased on data from 1611 patients in 2 studiesBaricitinib probably results in littleor no increase in adverse effectsleading to discontinuationModerateGRADE rating,because of:Risk of biasNo serious concernsImprecisionSeriousIndirectnessNo serious concernsInconsistencyNo serious concernsPublication biasNo serious concernsGRADE certainty ratingsHighThe authors have a lot ofconfidence that the trueeffect is similar to theestimated effectModerateThe authors believe thatthe true effect is probablyclose to theestimated effectLowThe true effect might bemarkedly different fromthe estimated effectVery lowThe true effect isprobably markedlydifferent from theestimated effectEvidence qualityDays (median/mean)Hospital length of stay1.4 fewer12.811.4ModerateMoreBased on data from 2652 patients in 3 studiesBaricitinib probably reduces length ofhospital stayModerateGRADE rating,because of:Risk of biasNo serious concernsImprecisionSeriousIndirectnessNo serious concernsInconsistencyNo serious concernsPublication biasNo serious concernsGRADE certainty ratingsHighThe authors have a lot ofconfidence that the trueeffect is similar to theestimated effectModerateThe authors believe thatthe true effect is probablyclose to theestimated effectLowThe true effect might bemarkedly different fromthe estimated effectVery lowThe true effect isprobably markedlydifferent from theestimated effectDuration of mechanical ventil...3.2 fewer14.711.5ModerateMoreBased on data from 328 patients in 2 studiesBaricitinib probably reduces duration ofmechanical ventilationModerateGRADE rating,because of:Risk of biasNo serious concernsImprecisionSeriousIndirectnessNo serious concernsInconsistencyNo serious concernsPublication biasNo serious concernsGRADE certainty ratingsHighThe authors have a lot ofconfidence that the trueeffect is similar to theestimated effectModerateThe authors believe thatthe true effect is probablyclose to theestimated effectLowThe true effect might bemarkedly different fromthe estimated effectVery lowThe true effect isprobably markedlydifferent from theestimated effectTime to clinical stability1 fewer9.98.9LowMoreBased on data from 2558 patients in 2 studiesBaricitinib may reduce time to clinicalstabilityLowGRADE rating,because of:Risk of biasNo serious concernsImprecisionVery seriousIndirectnessNo serious concernsInconsistencyNo serious concernsPublication biasNo serious concernsGRADE certainty ratingsHighThe authors have a lot ofconfidence that the trueeffect is similar to theestimated effectModerateThe authors believe thatthe true effect is probablyclose to theestimated effectLowThe true effect might bemarkedly different fromthe estimated effectVery lowThe true effect isprobably markedlydifferent from theestimated effectSee all outcomesIndividual considerations Key practical issues The panel inferred that most patients would want to receive baricitinib. The benefit on mortality was deemed of critical importance to patients and the panel was reassured by the moderate certainty evidence of little or no increase in serious adverse events Values and preferences Baricitinib Usual supportive care Administered as a pill, but can be crushed and administered by a nasogastric or gastrostomy tube Baricitinib must be adjusted for patients who have renal impairment Baricitinib has not been studied in patients with severe hepatic impairment Baricitinib may cause leucopenia, lymphopenia, thrombocytosis, anaemia, clotting abnormalities, hepatic impairment, and secondary infection No additional practical issues Recommendation 2Supportive careRuxolitiniborPatients with severe orcritical covid-19We suggest not using ruxolitinibStrongAll or nearly all informed people would likelywant the intervention to the left. Benefitswould outweigh harms for almost everyoneWeakMost people would likely want the interventionto the left. Benefits would outweigh harms forthe majority, but not for everyoneWeakMost people would likely want the interventionto the right. Benefits would outweigh harmsfor the majority, but not for everyoneStrongAll or nearly all informed people would likelywant the intervention to the right. Benefitswould outweigh harms for almost everyoneConditions for use of treatment Clinicians should only consider using ruxolitinib if neither baricitinib nor IL-6 receptor blockers (tocilizumab or sarilumab) are available Evidence profileFavours supportive careFavours ruxolitinibNo important differenceThe panel found that this difference was not important for most patients,because the intervention effects were negligible and/or very imprecise, forexample confidence intervals that include both important benefit and harmEvidence qualityEvents per 1000 peopleMortalityNo important difference130115Very lowMoreBased on data from 472 patients in 2 studiesThe effect of ruxolitinib is veryuncertainVery lowGRADE rating,because of:Risk of biasNo serious concernsImprecisionVery seriousIndirectnessSeriousInconsistencyNo serious concernsPublication biasNo serious concernsGRADE certainty ratingsHighThe authors have a lot ofconfidence that the trueeffect is similar to theestimated effectModerateThe authors believe thatthe true effect is probablyclose to theestimated effectLowThe true effect might bemarkedly different fromthe estimated effectVery lowThe true effect isprobably markedlydifferent from theestimated effectMechanical ventilationNo important difference116108Very lowMoreBased on data from 472 patients in 2 studiesThe effect of ruxolitinib is veryuncertainVery lowGRADE rating,because of:Risk of biasNo serious concernsImprecisionVery seriousIndirectnessSeriousInconsistencyNo serious concernsPublication biasNo serious concernsGRADE certainty ratingsHighThe authors have a lot ofconfidence that the trueeffect is similar to theestimated effectModerateThe authors believe thatthe true effect is probablyclose to theestimated effectLowThe true effect might bemarkedly different fromthe estimated effectVery lowThe true effect isprobably markedlydifferent from theestimated effectSerious adverse eventsNo important difference05LowMoreBased on data from 484 patients in 1 studyRuxolitinib may not result in anincrease in serious adverse eventsLowGRADE rating,because of:Risk of biasNo serious concernsImprecisionVery seriousIndirectnessNo serious concernsInconsistencyNo serious concernsPublication biasNo serious concernsGRADE certainty ratingsHighThe authors have a lot ofconfidence that the trueeffect is similar to theestimated effectModerateThe authors believe thatthe true effect is probablyclose to theestimated effectLowThe true effect might bemarkedly different fromthe estimated effectVery lowThe true effect isprobably markedlydifferent from theestimated effectEvidence qualityDays (median/mean)Length of hospital stayNo important difference12.811.4Very lowMoreBased on data from 472 patients in 2 studiesThe impact of ruxolitinib on length ofhospital stay is very uncertainVery lowGRADE rating,because of:Risk of biasNo serious concernsImprecisionVery seriousIndirectnessSeriousInconsistencyNo serious concernsPublication biasNo serious concernsGRADE certainty ratingsHighThe authors have a lot ofconfidence that the trueeffect is similar to theestimated effectModerateThe authors believe thatthe true effect is probablyclose to theestimated effectLowThe true effect might bemarkedly different fromthe estimated effectVery lowThe true effect isprobably markedlydifferent from theestimated effectTime to clinical stabilityNo important difference9.99.8Very lowMoreBased on data from 472 patients in 2 studiesThe impact of ruxolitinib on time toclinical stability is very uncertainVery lowGRADE rating,because of:Risk of biasNo serious concernsImprecisionVery seriousIndirectnessSeriousInconsistencyNo serious concernsPublication biasNo serious concernsGRADE certainty ratingsHighThe authors have a lot ofconfidence that the trueeffect is similar to theestimated effectModerateThe authors believe thatthe true effect is probablyclose to theestimated effectLowThe true effect might bemarkedly different fromthe estimated effectVery lowThe true effect isprobably markedlydifferent from theestimated effectSee all outcomesIndividual considerations The panel inferred that, given the uncertain benefits and the possibility of serious adverse events, the majority of patients would not want ruxolitinib. However, a minority of patients might choose to receive it if neither baricitinib nor IL-6 receptor blockers were available. Values and preferences Recommendation 3Supportive careTofacitiniborPatients with severe orcritical covid-19We suggest not using tofacitinibStrongAll or nearly all informed people would likelywant the intervention to the left. Benefitswould outweigh harms for almost everyoneWeakMost people would likely want the interventionto the left. Benefits would outweigh harms forthe majority, but not for everyoneWeakMost people would likely want the interventionto the right. Benefits would outweigh harmsfor the majority, but not for everyoneStrongAll or nearly all informed people would likelywant the intervention to the right. Benefitswould outweigh harms for almost everyoneConditions for use of treatment Clinicians should only consider using tofacitinib if neither baricitinib nor IL-6 receptor blockers (tocilizumab or sarilumab) are available Evidence profileFavours supportive careFavours tofacitinibNo important differenceThe panel found that this difference was not important for most patients,because the intervention effects were negligible and/or very imprecise, forexample confidence intervals that include both important benefit and harmEvidence qualityEvents per 1000 peopleMortalityNo important difference13078Very lowMoreBased on data from 289 patients in 1 studyThe effect of tofacitinib is uncertainVery lowGRADE rating,because of:Risk of biasNo serious concernsImprecisionExtremely seriousIndirectnessNo serious concernsInconsistencyNo serious concernsPublication biasNo serious concernsGRADE certainty ratingsHighThe authors have a lot ofconfidence that the trueeffect is similar to theestimated effectModerateThe authors believe thatthe true effect is probablyclose to theestimated effectLowThe true effect might bemarkedly different fromthe estimated effectVery lowThe true effect isprobably markedlydifferent from theestimated effectMechanical ventilationNo important difference11668Very lowMoreBased on data from 289 patients in 1 studyThe effect of tofacitinib is uncertainVery lowGRADE rating,because of:Risk of biasNo serious concernsImprecisionExtremely seriousIndirectnessNo serious concernsInconsistencyNo serious concernsPublication biasNo serious concernsGRADE certainty ratingsHighThe authors have a lot ofconfidence that the trueeffect is similar to theestimated effectModerateThe authors believe thatthe true effect is probablyclose to theestimated effectLowThe true effect might bemarkedly different fromthe estimated effectVery lowThe true effect isprobably markedlydifferent from theestimated effectSerious adverse events77 fewer077LowMoreBased on data from 284 patients in 1 studyTofacitinib may increase serious adverseeventsLowGRADE rating,because of:Risk of biasNo serious concernsImprecisionVery seriousIndirectnessNo serious concernsInconsistencyNo serious concernsPublication biasNo serious concernsGRADE certainty ratingsHighThe authors have a lot ofconfidence that the trueeffect is similar to theestimated effectModerateThe authors believe thatthe true effect is probablyclose to theestimated effectLowThe true effect might bemarkedly different fromthe estimated effectVery lowThe true effect isprobably markedlydifferent from theestimated effectEvidence qualityDays (median/mean)Length of hospital stay1.1 fewer12.811.7LowMoreBased on data from 289 patients in 1 studyTofacitinib may reduce length ofhospital stayLowGRADE rating,because of:Risk of biasNo serious concernsImprecisionVery seriousIndirectnessNo serious concernsInconsistencyNo serious concernsPublication biasNo serious concerns