Alteplase in COVID-19 severe hypoxemic respiratory failure: the TRISTARDS multicenter randomized trial

Abstract

Background: Pulmonary intravascular thrombus formation has been widely observed in patients with respiratory failure, for example, in patients with SARS-CoV-2 infection (COVID-19). The aim of this study was to evaluate the efficacy/safety of alteplase thrombolysis in COVID-19 severe hypoxemic respiratory failure. In this multicenter, open-label study, patients were randomized to receive alteplase (low- or high-dose) over 5 days plus standard of care (SOC), or SOC alone. The primary endpoint was time to clinical improvement (≥ 2-point decrease on WHO Clinical Progression Scale, or hospital discharge) up to Day 28. Secondary endpoints included all-cause mortality at Day 28, treatment failure at Day 28 and change in arterial oxygen partial pressure/fractional inspired oxygen (PaO₂/FiO₂) ratio at Day 6 versus baseline.

Results: Sixty-nine patients were randomized to alteplase (low- or high-dose) and 35 to SOC; 65% were on high-flow oxygen or non-invasive ventilation at baseline. Median time to clinical improvement was 25 days in the alteplase group and > 28 days (median not reached) in the SOC group. All-cause mortality was 8/69 (12%) versus 10/35 (29%) in the alteplase versus SOC groups, respectively (unadjusted risk difference [RD], - 17% [95% confidence interval (CI) - 34 to 0], p = 0.047; adjusted RD, - 16% [95% CI - 31 to 1], p = 0.058). The PaO₂/FiO₂ ratio (mean [standard deviation]) increased by + 30 (84) mmHg in the alteplase group and decreased by - 12 (59) mmHg in the SOC group (adjusted mean difference vs. SOC, p = 0.052). Differences were greater in patients receiving high-dose alteplase, and in those not receiving invasive ventilation. Eighteen patients (26.1%) in the alteplase group discontinued treatment due to adverse events. Major bleeding was more frequent with alteplase than with SOC (9 vs. 0 patients); no bleeding was fatal. The study closed early due to insufficient patient recruitment.

Conclusion: Alteplase was not associated with faster clinical recovery from COVID-19 severe hypoxemic respiratory failure. A numerical difference in survival and PaO₂/FiO₂ ratio was observed, particularly in patients not receiving invasive ventilation. These exploratory findings merit further investigation in larger patient cohorts that are adequately powered to confirm the hypotheses generated in this study regarding the impact of alteplase on treatment outcomes. Trial registration ClinicalTrials.gov: NCT04640194 (November 23, 2020); https://clinicaltrials.gov/study/NCT04640194 (early discontinuation due to insufficient patient recruitment).

Keywords: ARDS; Alteplase; COVID-19; Severe hypoxemic respiratory failure; Thrombolysis.

Fig. 1

Patient flow diagram (pooled data). All randomized patients were treated and included in the full analysis set. *A full list of adverse events leading to discontinuation is provided in supplementary material (Table S12). ^†Other patients discontinued due to loss of follow-up due to transfer out of ICU (n = 1), improvement in PaO₂/FiO₂ ratio of > 300 (n = 1), low fibrinogen level (n = 1), and mild bleeding (n = 1). FiO₂ fractional inspired oxygen, ICU intensive care unit, PaO₂ arterial oxygen partial pressure), SOC standard of care

Fig. 2

Time to clinical improvement up to Day 28. Alteplase pooled: 0.3 or 0.6 mg/kg over 2 h, followed by daily long-term (12-h) infusion of 0.02 or 0.04 mg/kg/h over 5 days (added to SOC). The symbols on the graphs represent censoring. *HR adjusted for treatment, baseline D-dimer status, age, baseline ventilation, and Part 1 or 2 of the study. ^†RD corresponds to Day 28. HR hazard ratio, RD risk difference, SOC standard of care

Fig. 3

Time to all-cause mortality up to Day 28*. Alteplase pooled: 0.3 or 0.6 mg/kg over 2 h, followed by daily long-term (12-h) infusion of 0.02 or 0.04 mg/kg/h over 5 days (added to SOC). *Time to all-cause mortality was not a predefined endpoint. Kaplan–Meier analysis was carried out retrospectively, based upon the results of the key secondary endpoint (all-cause mortality). ^†Cox proportional hazard model adjusted for fixed effects for treatment, baseline D-dimer status, age, baseline ventilation, and Part 1 or 2 of the study. ^‡RD corresponds to Day 28. ^§Unadjusted RD. ^¶Adjusted RD based upon the binary endpoint, adjusted for treatment, baseline D-dimer status, age, baseline ventilation, and Part 1 or 2 of the study. CI confidence interval, HR hazard ratio, RD risk difference, SOC standard of care