1/n The COLCORONA clinical trial (Lancet 27 May 2021) concludes that if we treat 70 Covid positive people with colchicine for a month, we prevent 1 hospital admission or death.
How did the researchers estimate the sample size of their study?
2/n Assumptions. They assumed that 7% of those on placebo shall either die or need hospital admission. Oral colchicine shall reduce this number to 5.25.
To detect a difference between the 2 arms of the study, they needed to enrol about 6000 people in the study. Which they did.
3/n The outcome in the colchicine trial was death or hospital admission. The secondary outcome was the need for mechanical ventilation.
Outcomes that are clinically meaningful. Outcomes that matter most to patients or their family.
4/n By contrast, let us look at the clinical trial that tested the efficacy of Favipiravir in mild Covid illness and made this drug so popular that it rocketed to the rank of the largest-selling pharma brand in India in April 2021.
5/n How many patients did favipiravir trialists enrol?
They enrolled 90 mildly ill and 60 moderately ill covid patients. Total, 150 patients. Yes. 150.
What outcome did they count? Viral clearance
What secondary outcome did they measure? Fever and cough.
6/n Researchers often term such outcomes as surrogate outcomes because they are not clinically meaningful.
In the Colchicine study, the researchers wanted to reduce death or hospitalisation.
7/n In the favipiravir study, the researchers were counting the cough and measuring the temperature or looking at viral clearance.
Even patients shrug off such outcomes with “Who cares” and “So what” gestures
8/end
We need large-sized, multi-centric studies that ask a focused question, identify the most meaningful outcomes and set out to enrol thousands of patients to answer the question.
The contrast between Colchicine and Favipiravir could not have been more striking!
• • •
Missing some Tweet in this thread? You can try to
force a refresh
I cringe when Times of India reports results from a COVID19 case-series from Maharashtra- drugs are doing great. Not a word on control arm or standard of care. They should ask—and tell us—who did they compare the drug with. These numbers are confusing and distort science. 1/4
First, #Favipiravir. Maharashtra reports that 5.5% (27/495) mildly sick COVID patients on Favipiravir died. Normally, about 98% of such patients recover. Are these results “encouraging”? Or should we be worried about the association of increased mortality with Favipiravir? 2/4
Second, #Remdesivir. The NEJM RCT showed that 7.1% on Remdesivir died compared to 11.9 % on placebo. The results lacked statistical significance. In Maharashtra, 29% (89/311) patients receiving Remdesivir died. Should we call these results promising? 3/4
My quick comments on the #Favipiravir study results released by #GLENMARK in a press release.
Did Favipiravir succeed in achieving faster viral clearance? No. The difference between the two arms of the study lacks statistical significance. The 95% CIs cross 1. N/1
What about secondary endpoints? Would you pay Rs 12500 for a drug that promises to make you fever-free a day earlier?
Look at the 95% CIs again- they barely float above 1! Not impressive! And were these patients hypoxic? The data does not tell us. 2/2
#Favipiravir group tended to require oxygen late compared to the usual care group.
OK. The trouble is, we do not know, eventually how many required oxygen in each group. 95% Cis touch 1- meaning no difference between the two groups. 3/4