Read on Twitter
With the Jan 7th announcement of the @remap_cap IL-6ra results: https://www.medrxiv.org/content/10.1101/2021.01.07.21249390v2
and last nights @BMJ_Latest RCT drop: https://www.bmj.com/content/372/bmj.n84
let’s talk about tocilizumab [thread, 1/n]
and last nights @BMJ_Latest RCT drop: https://www.bmj.com/content/372/bmj.n84
let’s talk about tocilizumab [thread, 1/n]
COI: I am the pharmacist lead for @remap_cap @UPMC, however, we didn’t enroll to the tocilizumab domain due to supply issues in the US—drug was reserved for RA and CAR T patients. I saw the data with the rest of the world the day the pre-print was posted. [2/n]
In @BMJ_latest, Veiga et al enrolled 129 patients to tocilizumab or standard care. Enrollment was stopped early since tocilizumab was associated with increased death at day 15 (16.9% vs 3.1%, OR 6.42; 95% CI 1.59 to 43.2). Woof. That is… suboptimal. We’ll come back to this.[3/n]
REMAP-CAP enrolled 353 patients to toci, 48 sarilumab, or 402 standard care in an adaptive, open-label platform ( https://pubmed.ncbi.nlm.nih.gov/32267771/ ). Primary outcome here was ordinal scale combining inhospital mortality (-1) and days free of organ support to day 21 [4/n]
Patients were critically ill—they had to be enrolled *within 24h* of new requirement of 30L+ HFNC or MV or vasopressor (28.8% HFNC, 41.5% NIV, 29.4% MV). Due to pandemic nature of repurposed facilities, physical location didn’t matter-- ICU meant requiring organ support. [5/n]
Median organ support-free days were 10 (IQR -1, 16), 11 (IQR 0, 16), 0 (IQR -1, 15) for toci, sari, and control, respectively. Hospital mortality was 27.3% for patients receiving IL-6ra, 35.8% control. Progression to intubation or death: 41.3% toci vs 52.7% control. [6/n]
![Median organ support-free days were 10 (IQR -1, 16), 11 (IQR 0, 16), 0 (IQR -1, 15) for toci, sari, and control, respectively. Hospital mortality was 27.3% for patients receiving IL-6ra, 35.8% control. Progression to intubation or death: 41.3% toci vs 52.7% control. [6/n]](https://pbs.twimg.com/media/EsRMwnWXMAEaTnR.png "Median organ support-free days were 10 (IQR -1, 16), 11 (IQR 0, 16), 0 (IQR -1, 15) for toci, sari, and control, respectively. Hospital mortality was 27.3% for patients receiving IL-6ra, 35.8% control. Progression to intubation or death: 41.3% toci vs 52.7% control. [6/n]")
![Median organ support-free days were 10 (IQR -1, 16), 11 (IQR 0, 16), 0 (IQR -1, 15) for toci, sari, and control, respectively. Hospital mortality was 27.3% for patients receiving IL-6ra, 35.8% control. Progression to intubation or death: 41.3% toci vs 52.7% control. [6/n]](https://pbs.twimg.com/media/EsRMxMiXAAIloV6.png "Median organ support-free days were 10 (IQR -1, 16), 11 (IQR 0, 16), 0 (IQR -1, 15) for toci, sari, and control, respectively. Hospital mortality was 27.3% for patients receiving IL-6ra, 35.8% control. Progression to intubation or death: 41.3% toci vs 52.7% control. [6/n]")
But what about live-saving steroids, you ask? Good question. >80% received steroids w/in 48h of enrollment (exact N unclear). 30% received RDV, fwiw. So this is benefit of tocilizumab in addition to steroids, which is hugely important. Also, ADRs similar between groups. [7/n]
Would love to know more on what drug/dose/duration of steroid received as well as info about therapeutic anticoagulation. Reached out to authors—this is unknown at this time, BUT, toci effect is independent of other treatment domains per the model [8/n] https://www.nytimes.com/2020/12/22/health/coronavirus-blood-thinner-trial.html?referringSource=articleShare
15.6% didnt have confirmed COVID, ~10% didnt get IL-6ra in rx group. Regardless—these data provide insight on 1)incremental benefit of toci in a world of steroids (toci+steroids > than either drug alone) 2)pop that may optimally benefit, which was debated in previous studies[9/n]
So, how to reconcile? Quite honestly—not sure. In Veiga, MV d28 death was 72.7%(8/11) toci v 30% (3/10) SC; HFNC/NIV 20% (3/15) v 0% (0/26). The extraordinarily high MV toci death rate vs nonexistent event rate in HFNC/NIV standard care seem more to chance than drug effect [10/n]
“But all the other trials were negative!” details extremely important here before generalizing that conclusion. previous RCTs limited by sample size, low event rate (good for patients, bad for finding differences), and spectrum of severity of disease included [11/n]
Somers et al gave toci only to pts requiring MV, 74% w/in 48h of MV (same REMAP sweet spot!!) Toci assoc. with 45% reduction in death (HR 0.55, 95% CI 0.33-0.90), being discharged alive (56% v 40%), but more infections (54% vs 26%) [12/n] https://pubmed.ncbi.nlm.nih.gov/32651997/
![Somers et al gave toci only to pts requiring MV, 74% w/in 48h of MV (same REMAP sweet spot!!) Toci assoc. with 45% reduction in death (HR 0.55, 95% CI 0.33-0.90), being discharged alive (56% v 40%), but more infections (54% vs 26%) [12/n] https://pubmed.ncbi.nlm.nih.gov/32651997/](https://pbs.twimg.com/media/EsRNI70XEAosSJ7.png "Somers et al gave toci only to pts requiring MV, 74% w/in 48h of MV (same REMAP sweet spot!!) Toci assoc. with 45% reduction in death (HR 0.55, 95% CI 0.33-0.90), being discharged alive (56% v 40%), but more infections (54% vs 26%) [12/n] https://pubmed.ncbi.nlm.nih.gov/32651997/")
![Somers et al gave toci only to pts requiring MV, 74% w/in 48h of MV (same REMAP sweet spot!!) Toci assoc. with 45% reduction in death (HR 0.55, 95% CI 0.33-0.90), being discharged alive (56% v 40%), but more infections (54% vs 26%) [12/n] https://pubmed.ncbi.nlm.nih.gov/32651997/](https://pbs.twimg.com/media/EsRNJubXEAIsl7w.png "Somers et al gave toci only to pts requiring MV, 74% w/in 48h of MV (same REMAP sweet spot!!) Toci assoc. with 45% reduction in death (HR 0.55, 95% CI 0.33-0.90), being discharged alive (56% v 40%), but more infections (54% vs 26%) [12/n] https://pubmed.ncbi.nlm.nih.gov/32651997/")
These data were always encouraging, but we were hesitant to adopt early in the pandemic because of observational design, cost of drug, and by the time it was published we had steroids which we assumed served the same purpose, so we moved on despite these impressive findings[13/n]
![These data were always encouraging, but we were hesitant to adopt early in the pandemic because of observational design, cost of drug, and by the time it was published we had steroids which we assumed served the same purpose, so we moved on despite these impressive findings[13/n]](https://pbs.twimg.com/media/EsRNNbfXIAIr1-t.png "These data were always encouraging, but we were hesitant to adopt early in the pandemic because of observational design, cost of drug, and by the time it was published we had steroids which we assumed served the same purpose, so we moved on despite these impressive findings[13/n]")
COVACTA preprint = no diff. in 28d clinical status or mortality for toci v placebo. Caveats: ~30% on low-flow, MV pts were enrolled mean 5 days (range 0-28!!) after intubation so perhaps too late/too progressed [14/n] https://www.medrxiv.org/content/10.1101/2020.08.27.20183442v2
In non-MV pts 29% toci v 42% plcbo progressed to failure (weird endpoint including ICU transfer made this sig. v MV/death alone). Toci=shorter time to discharge(20 v 28 days), ICU LOS(9.8 v 15.5 days), less MV(27.9% vs 36.7%), more vent free days, less salvage steroid use [15/n]
![In non-MV pts 29% toci v 42% plcbo progressed to failure (weird endpoint including ICU transfer made this sig. v MV/death alone). Toci=shorter time to discharge(20 v 28 days), ICU LOS(9.8 v 15.5 days), less MV(27.9% vs 36.7%), more vent free days, less salvage steroid use [15/n]](https://pbs.twimg.com/media/EsRNVjEW4AEWDps.png "In non-MV pts 29% toci v 42% plcbo progressed to failure (weird endpoint including ICU transfer made this sig. v MV/death alone). Toci=shorter time to discharge(20 v 28 days), ICU LOS(9.8 v 15.5 days), less MV(27.9% vs 36.7%), more vent free days, less salvage steroid use [15/n]")
Also—looking at improvement in clinical status of patients at day 14 based on ordinal scale category in COVACTA compared to ACTT-1, they are… similar. And we certainly thought ACTT-1 demonstrated remdesivir benefit. Table credit: @jpogue1 [16/n]
![Also—looking at improvement in clinical status of patients at day 14 based on ordinal scale category in COVACTA compared to ACTT-1, they are… similar. And we certainly thought ACTT-1 demonstrated remdesivir benefit. Table credit: @jpogue1 [16/n]](https://pbs.twimg.com/media/EsRNYDFW8AACoQu.png "Also—looking at improvement in clinical status of patients at day 14 based on ordinal scale category in COVACTA compared to ACTT-1, they are… similar. And we certainly thought ACTT-1 demonstrated remdesivir benefit. Table credit: @jpogue1 [16/n]")
Finally, I know all of COVID has been subgroup analysis paralysis, but COVACTA pts requiring HFNC at baseline had significantly improved clinical status at day 14 (OR 2.1; 95% CI 1.07-4.10) & less death (13.8% vs 20.5%) with toci—this is intriguing, since it’s the REMAP pop[17/n]
Next, EMPACTA. Progression to MV/death by d28: 12% toci v 19.3% placebo (HR 0.56, 95% CI 0.33-0.97, P = 0.04). Note this is VERY SIMILAR to baricitinib RR for progression in ACTT-2 (0.69 (0.50 to 0.95)) [18/n] https://pubmed.ncbi.nlm.nih.gov/33332779/
![Next, EMPACTA. Progression to MV/death by d28: 12% toci v 19.3% placebo (HR 0.56, 95% CI 0.33-0.97, P = 0.04). Note this is VERY SIMILAR to baricitinib RR for progression in ACTT-2 (0.69 (0.50 to 0.95)) [18/n] https://pubmed.ncbi.nlm.nih.gov/33332779/](https://pbs.twimg.com/media/EsRNkZNW4AIb5yb.png "Next, EMPACTA. Progression to MV/death by d28: 12% toci v 19.3% placebo (HR 0.56, 95% CI 0.33-0.97, P = 0.04). Note this is VERY SIMILAR to baricitinib RR for progression in ACTT-2 (0.69 (0.50 to 0.95)) [18/n] https://pubmed.ncbi.nlm.nih.gov/33332779/")
Also, decent amount of dex and rdv in EMPACTA, so again showing toci effect in light of current standard care [19/n]
![Also, decent amount of dex and rdv in EMPACTA, so again showing toci effect in light of current standard care [19/n]](https://pbs.twimg.com/media/EsRNnUiXMA0GSKn.png "Also, decent amount of dex and rdv in EMPACTA, so again showing toci effect in light of current standard care [19/n]")
BACC Bay. 96% no O2/low-flow, no dex use. Toci not effective for preventing MV/death(aHR 0.66 (95% CI, 0.28-1.52). This is most “negative” study but least sick pts, least steroid use, and adj point estimate is similar to above data [20/n] https://pubmed.ncbi.nlm.nih.gov/33085857/
![BACC Bay. 96% no O2/low-flow, no dex use. Toci not effective for preventing MV/death(aHR 0.66 (95% CI, 0.28-1.52). This is most “negative” study but least sick pts, least steroid use, and adj point estimate is similar to above data [20/n] https://pubmed.ncbi.nlm.nih.gov/33085857/](https://pbs.twimg.com/media/EsRNqXqXcAUtPGf.png "BACC Bay. 96% no O2/low-flow, no dex use. Toci not effective for preventing MV/death(aHR 0.66 (95% CI, 0.28-1.52). This is most “negative” study but least sick pts, least steroid use, and adj point estimate is similar to above data [20/n] https://pubmed.ncbi.nlm.nih.gov/33085857/")
Hermine etal enrolled low-flow pts, 19% toci v28% usual care worsened by day 4. This was an 89% probability of negative absolute risk reduction so didn’t meet Bayesian platform trigger for efficacy of 95%, hence “negative”— but only 131 pts in trial [21/n] https://pubmed.ncbi.nlm.nih.gov/33080017/
17% toci v 27% placebo progressed to MV or death at day 14. No difference in 28d mortality: 11% v 12%. Steroid use in a pre-steroid world was 33% toci vs 61% placebo—so less salvage needed. The effect of toci greater with steroids (HR 0.38, 90% CI 0.13-1.11) [22/n]
![17% toci v 27% placebo progressed to MV or death at day 14. No difference in 28d mortality: 11% v 12%. Steroid use in a pre-steroid world was 33% toci vs 61% placebo—so less salvage needed. The effect of toci greater with steroids (HR 0.38, 90% CI 0.13-1.11) [22/n]](https://pbs.twimg.com/media/EsRNzViXcAA8NFc.png "17% toci v 27% placebo progressed to MV or death at day 14. No difference in 28d mortality: 11% v 12%. Steroid use in a pre-steroid world was 33% toci vs 61% placebo—so less salvage needed. The effect of toci greater with steroids (HR 0.38, 90% CI 0.13-1.11) [22/n]")
Finally, Salvarani et al enrolled pts w/ PaO2/FiO2 200-300 on HFNC in Italy. 60 toci v 66 control—again, such small numbers https://pubmed.ncbi.nlm.nih.gov/33080005/ [23/n]
They excluded pts with old age and comorbidities because those patients were triaged out of ICU care in Italy at that time. A weird endpoint led to 22% of standard care arm receiving toci…so effects/results really unknown [24/n]
Pre-REMAP, toci significantly
clinical deterioration and maybe saved lives. The Veiga data should absolutely be acknowledged but do seem to be an outlier. The REMAP data are compelling and suggest tocilizumab also saves lives when used judiciously for the right patient. [25/n]
![Pre-REMAP, toci significantlyclinical deterioration and maybe saved lives. The Veiga data should absolutely be acknowledged but do seem to be an outlier. The REMAP data are compelling and suggest tocilizumab also saves lives when used judiciously for the right patient. [25/n]](https://pbs.twimg.com/media/EsROPhwXIAAC-Li.png "Pre-REMAP, toci significantlyclinical deterioration and maybe saved lives. The Veiga data should absolutely be acknowledged but do seem to be an outlier. The REMAP data are compelling and suggest tocilizumab also saves lives when used judiciously for the right patient. [25/n]")
clinical deterioration and maybe saved lives. The Veiga data should absolutely be acknowledged but do seem to be an outlier. The REMAP data are compelling and suggest tocilizumab also saves lives when used judiciously for the right patient. [25/n]
What is the right patient? I think we have a lot to learn here considering the complex phenotypes and immune heterogeneity of this disease. Critically important to consider for therapeutic agents… [26/n] https://pubmed.ncbi.nlm.nih.gov/32669297/
… but really hard to only enroll certain patients in pandemic trials. CP—large negative trials, but still hope of efficacy in the “right” patient. MABS—more benefit if seronegative? RDV… wont go there. Steroids—remember when we recommended AGAINST those?? [27/n]
I have admittedly been unenthusiastic about tocilizumab for the entirety of the pandemic because at first it was unknown/weird, then supply chain issues, and finally as things evolved, we just were never sure it was worth it (~4K/dose), especially not when we had steroids. [28/n]
But-- the REMAP-CAP data really made me earnestly evaluate all the existing data in detail and put it all together. The Veiga data don’t change my mind. I think it is reasonable to give tocilizumab to patients with confirmed COVID-19 pneumonia within ~48h of new HFNC/MV. [29/n]
I’d love to hear other thoughts, esp. since I am in the same boat as everyone trying to maintain institutional guideline criteria 
Also, note, RECOVERY is still enrolling (~3,800 pts to date) so my entire opinion may change in a few weeks. Or hours, since, COVID. [30/30].
Also, note, RECOVERY is still enrolling (~3,800 pts to date) so my entire opinion may change in a few weeks. Or hours, since, COVID. [30/30].
