Re: Ivermectin

hari kumar

I said I would provide further materials. This was in reply to a pro-Ivermectin corespondent on PEN-L. I have not made any major changes for here. The two papers referred to are Meta-analyses, that if people want but are blocked from getting, I can provide if they write me off-line.  

"There has been a pro. vs anti. debate on ivermectin thus far. I am sorry to go back to it, but there was an outstanding issue. I hope members of the list can either ignore if bored or overlook in some way re-addressing the same issue. 
However my major prior pressing deadlines are now over and I can now respond to Paul Z's request to review Dr Tess Lawrie.  The first part of this perhaps over-long note is that review. 
1) I preface this by admitting that I did not view the video of Dr. Tess Lawrie, partly time, but more because I feared a reprise of an earlier video referred to in the prior strand, that was frankly propagandistic. However, instead I chose to review her peer published work. No doubt this is the preferable route anyway.
The main one was at this reference: Bryant A, Theresa A. Lawrie, Therese Dowswell, et al; Am J Ther 2021 June 21; 28 (4):e434. This study concluded that "ivermectin reduced risk of death compared with no ivermectin (average risk ratio 0.38, 95% confidence interval 0.190.73; n 2438; I49%; moderate-certainty evidence)."
Unfortunately one of the biggest trials included in that Bryant et al analysis was recently retracted on July 20 2021:
"A large Egyptian study of ivermectin for COVID-19 patients has been retracted over concerns of plagiarism and serious problems with their raw data, the publisher confirmed to MedPage... The study was one of the largest ivermectin trials in the world, and has been included in two recent meta-analyses (Bryant et al. and Hill et al.) that received much attention for their positive results -- particularly the Hill review, which had been anticipated by a U.S. group that has long promoted ivermectin.....
David Boulware, MD, MPH, of the University of Minnesota, told MedPage Todaythat the 400-patient Egyptian trial -- from Ahmed Elgazzar, MD, of Benha University, and colleagues -- was the largest study included in the Hill review and accounted for 20% of the total data...". 
If you now back to the Bryant Meta-analysis, you can see from the forest pot on Figure 3 (death from any cause) - just how powerfully that single trial drives the results of the pooled analysis - for example as shown here for the outcome of death - in the varying sub-groups by severity of COVID. I think you might see that the mean & 95% CI bars for Elgazzar are way over to the left and clear of 'equivalence' or null effect as expressed by the vertical line. This drives the summary - expressed in the diamond at the bottom - over to a positive formal test of statistical significance in the first case; and below that even more dramatically in the second case: 

In table 2 - you see how the Bryant & Lawrie teams assessed by GRADE analysis - this is built to assess how 'good' or 'reliable' the data was. They say the evidence of death from any cause using the"GRADE" assessment in that particular trial was 'Moderate'.
In contrast the "living meta-analysis" in BMJ that I had cited earlier uses GRADE also but more critically. Indeed the authors of that paper are the originators of the GRADE approach:  Jessica J Bartoszko, et al: Prophylaxis against covid-19: living systematic review and network meta-analysis

BMJ 2021;373:n949 


These workers pooled all data as well - even though they assessed all data as low level. But they assessed GRADE more critically as seen in the final estimate "of very low certainty" of -1 (-3 to 68) (Figure 2).  

I think one 'moral' of these two MAs against each other, is to be critical of trials and how they are conducted. We are doing a lot more trials than before (thank goodness) - but HOW a trial is done is also important. 

Hari Kumar

Join to automatically receive all group messages.