TOGETHER enrolled high risk people with COVID from 12 outpatient clinics in Brazil.
Patients could be enrolled up to 7 days after symptom onset (more on this later).
They were randomized to either placebo or to ivermectin 0.4mg/kg daily for 3 days. (Also more on dose later) 3/
The primary endpoint was a composite of hospitalization or >6hr ED visit. (Not an awesome primary endpoint IMO 🤷).
Secondary endpoints included: time until hospitalization, hospital LOS, need for mechanical ventilation, duration of MV, and death. (All very reasonable 👍)
4/
It enrolled n=1358 people.
The patients were slightly younger (median age 49) with the expected mix of comorbid conditions seen in COVID (DM2, HTN, asthma).
The groups appear well balanced (see my prior thread about the likely fraudulent vitamin C paper for more on this.) 5/
The study was stone cold NEGATIVE.
There was NO statistically significant difference in the primary endpoint or *ANY* of the secondary endpoints.
Beyond the overall negative findings, there was no prespecified subgroup that benefited from ivermectin.
Roughly half the patients (44%) got treatment within 3 days. That’s early. Among the patients who were treated earlier they did… worse!
8/
Also if we look at the patients who *completed* the 3 day course of ivermectin (per-protocol analysis) they actually did *worse* than the intention to treat group.
If ivermectin really worked, you might expect the people who completed a course of it to better. They didn’t. 9/
“BuT mOrTaLiTy wAs LoWeR!”
A common misconception about stats.
Let’s look at the 0.88 mortality effect.
The confidence intervals mean there is a 95% chance that mortality is between 51% less OR 55% more with IVM.
Would *YOU* take a drug that might increase mortality by 55%? 10/
“tHe tRiAl WaS dEsIgNeD tO fAiL bY eViL pHaRmA!”
This was one arm of a *multi-arm study*. Another arm of #TOGETHER found that a repurposed cheap generic med (fluvoxamine) *improved* outcomes in COVID.
How (& why) would an evil cabal sabotage just one arm of a multi arm RCT? 11/
Related dumb criticism:
“ThE vAcCiNeS wOuLd LoSe tHiEr EUA iF iVeRmEcTiN wAs PrOvEn!”
This is nonsense. The EUA for *vaccines* to prevent severe disease has nothing to do with the absence of therapies. Otherwise Dex, Bari, Toci, etc would have already “voided the EUAs” 12/
On the other hand the *only* trials that have found *any* benefit to ivermectin are:
- fraudulent (Surgisphere, Elegazzar, etc)
- flawed observational studies that are likely biased
- tiny studies looking at non patient centered outcomes like viral load
Clinical 🥡 points: #TOGETHER is the largest RCT of ivermectin to date. It found that early high dose ivermectin did NOT prevent hospitalization, mechanical ventilation, or mortality in high risk outpatients with COVID.
All prior (non-fraudulent RCTs) have found the same.
15/
First a question: which of the following snake venoms have been used to develop medical therapies?
(don't worry we’ll go over the answer at the end)
2/
It was long known that workers in banana plantations would collapse due to low blood pressure if they were bitten by the Brazilian Arrow Headed Pit Viper Bothrops jararaca.
A Brazilian graduate student - Sergio Ferreira - thought this could be useful... 3/
To understand the proof you should read Dr Sheldricks post.
To summarize, he observed that in this non-randomized study the baseline characteristics of the pre & post intervention groups are far too perfectly matched. This perfect matching is unlikely to occur by chance.
2/
Specifically, you’d expect to see a range of p values for each baseline characteristic.
(This is especially true in a tiny trial with only n=47 patients)
Instead the range of p values for almost every variable was exactly 1.
This is *extremely* unlikely to occur by chance. 3/
They made a whole lot of money in 2020. This included:
$84.9m in revenue from ERAS
$32.7m from the AMCAS
$27.1m from the MCAT
$14.4m from membership dues
$10.3m from workshops
3/
Here’s a situation many of us have seen in the ICU or ED: “It looked like there was ST elevation on the monitor but when I took a 12 lead it was gone?!”
A STEMI went MIA? Here’s a #tweetorial all about why ST segments look different on monitors.
First, here’s another great example of "disappearing ST elevation", from Dr. Smith’s ECG Blog @smithECGBlog
(If you don’t already you should definitely follow Dr Smith & bookmark his site; hqmeded-ecg.blogspot.com IMO it's the best site for ECGs; you can thank me later) 2/
In order to understand *WHY* the ST segment looks different, we need to know how an ECG works & understand just a little bit of electronics & math.
(Don't worry, I promise no equations or circuit diagrams 🤞)
3/