MSB Trading 2021 - paradigm shift, page-1354

Greatpost by @omnisearch on the paper by Morgan et al published in TheLancet. This paper from the UK looked at patients from March 2020,early in the epidemic. Yes, I realize it doesn't specifically breakout intubated patients, but again it underlines some of the problemswith using mortality at 30 days as a primary end-point. At the top ofp599 the authors make this statement regarding the group of patientseligible for “escalation of treatment”, which is the group ofprimary interest to us:

“However,of the 90 patients in this group who received non-invasiveventilation or intubation, only 51 (57%) remained alive at the end offollow-up”

That'sterrible, but it shows the baseline gross mortality rate in this highrisk group was still only 43%. Unfortunately, they don't break outintubated vs non-intubated, but other papers from Europe at the timeshowed mortality rates in the 40% range for intubated patients. Andyes, again, a 40% mortality is terrible, it's horrible. BUT Isuspect the Mesoblast study design was strongly influenced by whatappeared to be an even higher mortality rate in NY City in March, ieover 80% in intubated patients. The moral of the story is thatMesoblast was misled on study design due to the March-April Coviddisaster in New York. That city was unprepared for the influx ofcases and responded poorly. At one point they were loading bodiesinto refrigerated trucks behind the ICU's. Turns out results in NYCmay not have been representative of more global populations. Yes,breaking out CRP levels, white counts and FiO2 offers parameters thatare predictive of poor outcome... but check out Table 1. We see amortality rate of 29% in the 65 patients with a hyper-inflammationprofile and eligible for “escalation of therapy” ie eligible forintubation. Start sprinkling in steroids and otheranti-inflammatory meds and the power of any study based solely on30-day mortality goes out the window. Again, not necessarily due tothe medication rather the study design. If we're trying to show astatistically significant reduction in mortality – and we make thatthe primary endpoint – as a 40% baseline rate (that we thoughtwould be 80%) drops further still due to the introduction of othertreatments/improved ventilator care, we need to enroll more patientsto achieve the same level of statistical significance. That's astrict issue of statistical mathematics, not medication efficacy. And if we make intubation a last resort, we risk selecting out olderpatients likely to linger on the ventilator with a higher mortalitydue to concomitant problems other than ARDS (as the authors allude toby plotting creatinine levels indicative of renal function). So,we're back to the arguments from the greeting card. Acme andHindenburg. We need to look at more than mortality reduction toevaluate the medication in this population. Or else we need to enrolla lot more patients. Take your pick.

Ithink there's a good chance we will progress to a situationcomparable to to the hyper-immune state of GvHD. First, steroids aretried... and in cases where they fail, remestemcel-L is used. Insteroid resistant aGvHD we do see mortality rates that can approach80%. Over time. Lets see what the rate comes in at in the ARDSstudy - at 30 days.

Wiseinvestors will wait for final results before making rash investmentdecisions or getting caught up in the semantics of “studyfutility”. Results may yet surprise – even on mortalityreduction, but more likely on improvement in VFD's and othersecondary end-points, especially those associated with so-called“long-hauler' syndrome, which is marked by a tendency to relapse orto get re-admitted. And which I suspect is a motivator for Novartis.

Lookingforward to results soon, hopefully this quarter.