Metformin's benefit for #COVID19 is from an anti-viral effect.
3.6-fold greater reduction in #SARSCOV2 viral load over placebo equating to: -0.56 log10 copies/mL (95%CI, -1.05 to -0.06,p=0.027) #IDTwitter medrxiv.org/content/10.110…
We observed no virologic effect of #ivermectin at median of 430 mcg/kg/day x3 days or of low-dose #fluvoxamine at 50mg 2x/d.
Metformin Virologic Effect was generally consistent across subgroups, larger in the unvaccinated at -0.773 log10 copies/mL (95%CI, -1.408 to -.139)
Metformin is a $1 therapy.
At Day 5 antiviral effect over placebo was 0.47 log10 copies/mL for metformin, 0.30 log10 copies/mL for molnupiravir, & 0.80 log10 copies/mL for nirmatrelvir; and at Day 10 was 0.64 log10 copies/mL for metformin and 0.35 log10 copies/mL for nirmatrelvir
That is a bit of an apples to oranges comparison as comparing nirmatelvir starting within 5 days in unvaccinated persons. No EPIC-SR data available for vaccinated pop.
0.80 log for paxlvoid unvaccinated vs 0.77 log reduction with metformin in unvaccinated is same ballpark.
The obvious question is what would happen if you combined the two meds? Nirmatelvir/ritonavir (paxlovid) + metformin or Shonigi's ensitrelvir + metformin.
(or molnupiravir+metformin... would that actually make molnupiravir more active).
If I was developing a novel antiviral for any RNA virus, I would like to see what is the added effect of adding an indirect host directed therapy such as metformin to any new virologic therapy.
Metformin summary:
⬇️ 58% reduction in COVID-related hospitalizations through 28-days
⬇️42% reductin in ER vistis/hospitalizations within 14-days
⬇️41% reduction in #LongCovid
-- No reduction in self-reported hypoxemia with inaccurate pulse oximters
Metformin in the trial was dosed with immediate release 500mg tabs.
Day 1 : 500mg once
Day 2-5: 500mg twice daily
Day 6-14: 500mg morning, 1000mg night.
Major side effect is GI of loose stool/diarrhea with rapidly increasing the dose, hence we titrated up over several days.
Metformin is an interesting drug.
There is an extensive medical literature about metformin's effect for weight loss, which many people also need (myself included could lose 10 lbs / 5kg). pubmed.ncbi.nlm.nih.gov/?term=metformi…
More data coming on Thursday, 6:30pm EDT.
Reduced incidence of #longcovid with acute #COVID19 treatment with metformin in a randomized clincial trial, now published.
One difference between the antiviral effect of #paxlovid vs #metformin is that paxlovid has an effect through day 5 & then with short half-life, fades. In contrast 14 days of metformin, the antiviral effect increases over time.
At day 10, -0.64 log for MET vs. -0.35 for PAX.
This distinction on the duration of antiviral effect might be why metformin had a 41% reduction in long covid symptoms.
What would happen if people with ongoing symptoms at 14 days continued an antiviral? be it metformin or metformin + paxlovid. An interesting question???
For further basic science justification on in vitro antiviral effects of metformin, see this paper:
While in vitro lab data provide an antiviral mechanism, it is nice to see actual human data from a proper RCT to validate the in vitro lab finding.ncbi.nlm.nih.gov/pmc/articles/P…
• • •
Missing some Tweet in this thread? You can try to
force a refresh
@TogetherTrial has completed a sensitivity analysis for its metformin outpatient #COVID19 trial
Receiving metformin was NOT an exclusion criterion for the trial. Gving 1500mg/d on top of existing metformin was not great. Odds Ratio =1.35 for ER/hospital
But Odds Ratio=0.59 if not
The overall updated results of the trial testing metformin (after some data cleaning, rerunning the results), including people taking metformin at baseline who were then randomized to receive +1500mg/d metformin XR or blinded placebo on top of all their existing meds, was:
But the more interesting sensitivity analysis of the TogetherTrial #metformin #COVID19 double-blind randomized trail is only considering those 368 particiapnts NOT already receiving metformin at baseline. #IDTwitter
37% reduction in ER visits/hospitalizations sounds similar...
#Metformin RCT on 41% prevention of #LongCOVID with acute #COVID19 treatment is now published in @TheLancetInfDis at: doi.org/10.1016/S1473-…
N=1126 with long term follow up (~5% lost by 9mo)
10.4% vs. 6.3% LC incidence.
Double blind RCT using identical matched placebo.
Cost $1
The distribution of symptoms reported over time was similar as others have found. The figure (a new additional from the original pre-print) displays the freq of symptoms reported over 9 months among those wih a long covid diagnosis vs. no diagnosis.
#LongCovid symptoms interfered with both the ability to work and affected leisure activities in about half of those with LC. This very, very slowly decreased over time.
Metformin:
Statistically significant 42% ⬇️in ER visits & #COVID19 hospitalizations
Statistically significant 58% ⬇️in 28-day hospitalization
Statistically significant 41% ⬇️in #LongCOVID
Statistically significant -0.56 log10 copies/mL mean ⬇️in viral load at Day 5 &10
@NIHCOVIDTxGuide does not recommend metformin stating that COVID-Out trial did not demonstrate a benefit of metformin in reducing the risk of hospitalization or death in patients with COVID-19.
A very declarative statement. covid19treatmentguidelines.nih.gov/therapies/misc…
Although these are all secondary endpoints of the COVID-Out trial, it is rather a stretch to be overly declarative to say that metformin did not reduce the risk of hospitalization. Guidelines panel do cite the 14-day mITT analysis, ignoring the 28-day data or ITT analysis.
As a combo blinded trial (FLV + inhaler), which one had benefit? Hard to know. In activ-6, we did not observe a benefit of inhaled fluticasone in a double blind RCT, yet UK Principle did see a benefit in open label trial with budesonide.
Importantly the #fluvoxamine dose studied here was 100mg 2x/day.
Covid-out and activ-6 both confirmed a lack of benefit of 50mg 2x/day dose.
If I gave someone 1/2 the effective dose of an HIV med & it didn't work, does that mean the med doesn't work? Or just dose was wrong?
#Paxlovid effect on reducing #COVID19 hospitalizations in vaccinated population finally revealed to the public by FDA. #IDTwitter
0.9% (3/317) paxlovid vs. 2.2% (7/314) placebo fda.gov/media/166197/d…
The relative risk reduction (RRR) of #covid19-related hospitalization was 57.5% (95%CI, -63% to 89%) with #Paxlovid in vaccinated high risk population.
Will point out that the the Hazard Ratio is also 58% for $1 of #metformin.
If one combines the EPIC-SR vaccinated + EPIC-HR seropositive groups together, then risk of #covid19 hospitalization was:
▪️ 0.5% (4/807) with #paxlovid
▪️ 1.9% (15/791) with placebo
Relative RIsk Reduction = 74% (95%CI, 22%-91%)
Number Needed to Treat = 71 (95%CI, 40-300)
Does #metformin reduce hospitalization or not? In a double-blind placebo controlled trial, 28-day data:
NIH guidelines cite the 14-day mITT data where there was not a statistical reduction.
The 14-day and 28-day ITT data also had a statistical reduction in hospitalization.
These numerical complete case data were published in NEJM.
"Through day 28, hospitalization or death occurred in 8 of 596 patients (1.3%) receiving metformin and in 19 of 601 controls (3.2%)"
(2.9% & 1.2% above are by Kaplan-Meier methodology). nejm.org/doi/full/10.10…
While the authors were not allowed to highlight these data or put a p-value in, apparently there are a lot of poeple who can't run a chi-square or fisher's exact test -- as I keep hearing that metformin didn't reduce hospitalization in the COVID-Out trial.