Instructive, lowish 3x 12mg, ivermectin Randomized Controlled Trial from Egypt (n=164) in mild/moderate, young hospitalized Covid patients struggles for statistical power:
• Length of hosp. stay 9 vs 11 days p=.08
• Death 3 vs 4, non significant.
🚨🚨🚨Remember the extraordinary, sudden eradication of Covid severity in Mexico City following the distribution of medical kits including ivermectin to all outpatients testing positive since December 2020?
"we found a negative and significant effect of the ivermectin kit on the probability of hospitalization.
Depending on subsampling, the effect ranges from 50% to 76% difference in hospitalization odds between treated and untreated patients, statistically significant in all cases"
Another inconclusive, underpowered, aborted outpatients randomized controlled-trial of hydroxychloroquine (n=214 on HCQ, 244 on Kaletra, 227 placebo), with data pointing to efficacy (without statistical significance).
Do not expect the authors to point out that those hospitalization results are entirely and remarkably consistent with the other 4, all identically underpowered, HCQ outpatient RCTs, homogenously pointing to 30%-ish hospitalization benefit.
However, the matter of lack of statistical power is now resolved by the national registry study publish in Iran (n=28,759).
The remarkable match of the HCQ/non-HCQ arms strongly signals quasi-randomization.
Let's start with the sad, usual evidence that peer reviewers' standards have collapsed to a lamentable level: they simply do not do their quality-control job anymore and medical journal editors are mediocre enough to let that happen again and again.
The meta-analysis claims to exclude trials where hydroxychloroquine or chloroquine was given in the control arm (the right thing to do).
A few comments:
• the control arm might be active (16% positive drug effect seen in 6 vitamin C studies so far), optically lowering the efficacy of the other arms
• cluster randomization shows material baseline seropositivity differences: vit C 6.3% HCQ 10 IVM12.4 PVI 14.5
• unclear whether you could be undetected infected at baseline, which could explain the odd infection rate of the IVM arm
• unlike others, IVM was single dose (200 µg/kg with12mg cap), a BIG ask for a 42 day prophylaxis
• >70% adherence only 71% in HCQ arm, lowering efficacy
• Multi-centric, randomized, double-blind, placebo-controlled
• n=590, 294 treated with Proxalutamide
• 300mg/day for 14 days
• hospitalized patients, average age 53
• 97% requiring oxygen, 67% high flux or invasive
After 14 days (all p<.0001):
• 3.7% death vs 47.6% placebo
• Hospital days 5 vs 14
• Intubation 4.4% vs 52.7%
• Off oxygen 92.5% vs 33.3%
• Still hospitalized 32.8% vs 89.1%
• Time to clinical improvement 3 days vs 19
• Serious adverse events: 0
The authors should have stopped there, but being what many infatuated and/or biased scientific researchers are, they decide to overreach, go for a bigger prize and conclude than none of the drugs works in Covid-19.
So what control group are they going to use to demonstrate it?
The only analysis in the discussion part of the report is below.
Unspecified, undocumented, unreferenced.
Absolutely gigantic red flag. 🚨🚨🚨
(it is where I come to despair of peer-reviewers...)