Brian Locke

RECOVERY TRIAL

Tocilizumab

Article Title

“Tocilizumab in patients admitted to hospital with COVID-19 (RECOVERY): preliminary results of a randomised, controlled, open-label, platform trial”

Study question and design

IL-6 and standard of care vs standard of care alone in COVID-19.

Open label (not blinded) Randomized 1:1, intention to treat analysis Platform trial (adding and finishing experimental treatments while the trial is ongoing) Adaptive – using pre-specified rules and interim analysis to modify enrollment/randomization

(https://ajkdblog.org/2016/03/10/nephmadness-2016-statistics-in-nephrology-region/)

Rationale: Prior evidence from REMAP-CAP (mortality benefit when added to patients newly admitted to ICU on NIV or HFNC) and EMPACTA (reduced progression to needing intubation) support. However: “Taken together, these previous trials did not show a significant 300 mortality benefit for treatment with tocilizumab (death rate ratio 0·91, 95% CI 0·72-1·14)”

Patients Included

Inclusion criteria: hospitalized with COVID-19 in UK (94% lab confirmed), <92% on RA, CRP over 75 mg/l Exclusion criteria: clear evidence of bacterial, fungal, or other viral infection. No toci available at the hospital.

“28-day mortality in the usual care group was above 25% then recruitment of around 4000 patients to this comparison would provide 90% power at two-sided P=0.01 to detect a proportional reduction in 28-day mortality of one-fifth”

N=4116. At randomization: Age 63.6 +/- 13.7

  • 14% ventilated, 41% HFNC or NIV, 45% NC
  • Mean 9-10d from symptom onset
  • Half have major comorbidity
  • Median CRP 143 ng/dl
  • 82% receiving corticosteroids at randomization

CRP correlated with IL-6 levels (= why that was chosen)

Note: CRP = hsCRP * 10 (roughly)

Intervention

28 day mortality, compared by log-rank test

Prespecified secondary: death or receipt of mechanical ventilation Subgroups: age, sex, ethnicity, level of resp support, days since symptom onset, use of steroids

6 month follow-up planned

Results

Primary outcome: 29% 28d mortality in Toci, 33% in standard of care (rate ratio 0.86, 95 CI 0.77 – 0.96).

Subgroup: Benefit still apparent in subgroup receiving corticosteroids (the majority of the cohort), and similar benefit across all subgroups

Prespecified Secondary: -Improved rate of alive and discharge from hospital in 28d among patients in Toci arm (54% vs 47%) -receiving invasive mechcanical ventilation or death prior to 28d (33% toci vs 38% usual)

Exploratory secondary: similar effect if randomized more or less than 2 days from hospital admission

Safety: Does not appear to an increase risk of infection or arrhythmia

Critque

  • Open label – provider behavior influence? mostly direct application to discharge outcome (which is directly measuring behavior), less so for mortality. However, other care may is likely affected, and this cannot be assess by the trial (they didn't collect all that information, as this is a pragmatic trial)
  • High mortality rate – approaches the mortality rate of intubated patients in the USA. UK has different norms around offering organ support for critically ill paients – might more of these patients go on to be intubated in USA? Alternatively, might this just be a reflection of the inclusion criteria? (Needing to have a high CRP for inclusion)
  • At time of preprint, primary outcome is available for 92% of patients – with sicker patients/longer hospital courses, might still be outcome differences, particularly in the sickest subgroups. Hence, 6 month follow-up
  • 17% randomized to toci didn’t get - are these patients different than the ones that did receive it? (ie, if you were randomized to it, but didn't receive it on open label, it suggests your clinicians might have had other concerns e.g. possible infection?). ~20% loss to follow-up