Can Metformin Treat Long COVID?

The diabetes drug metformin isn’t flashy. It is a cheap medication that lowers blood sugar by decreasing the quantity of glucose produced by the liver. It works gradually and incompletely; in most cases, additional approaches are needed for proper blood sugar control.

Because metformin has been a generic drug for over 60 years, you won't see it advertised on billboards or in magazines—and definitely not on television. You won't see any of those classic medication ads—the ones that go on forever depicting normal-looking people doing normal things with joy while a voice-over lists a dozen or so horrible side effects.

Despite its lack of traditional marketing exposure, metformin has nonetheless transformed from a boring diabetes drug to a powerhouse with vast anti-inflammatory and, many believe, anti-aging properties. Metformin has been shown to extend lifespan in mice by 4 to 6 percent and in humans to reduce circulating cytokines, such as interleukin-6, that are implicated in conditions such as acute COVID-19. While patients with Type 2 diabetes fared very poorly if they contracted COVID, those on metformin had a significantly better prognosis.

Beyond this, there is evidence that metformin can modulate virtually all of the cellular hallmarks of aging—including mitochondrial dysfunction, genetic instability, and stem cell dysfunction. [1] No wonder it's being treated as an anti-aging panacea by some Silicon Valley techies.

Not everyone is convinced, though, but that's a longer discussion. Today’s topic is more focused but potentially impactful: Can the so-called metformin miracle extend to the mysterious and befuddling symptoms of long COVID? A new study published today in The Lancet Infectious Diseases can help address that question.

The investigation from Carolyn T. Bramante from the University of Minnesota and the “COVID-OUT” team compared the effectiveness of three medications—metformin, ivermectin, and fluvoxamine (an anti-depressant also used to treat obsessive-compulsive disorder)—taken early in the course of a SARS-CoV-2 infection on two outcomes: 1) acute severe COVID-19 at 14 days and 2) diagnosis of long COVID within 300 days. They restricted the study population to adults aged 30-85 who were overweight or obese (based on body mass index) and had a diagnosis of SARS-CoV-2 infection (documented with a positive test within three days and less than a week of associated symptoms).

Now, you may be wondering about the choice of treatments being tested here. It is important to note that this study began enrollment in late December 2020, around the time that the first vaccines were being administered and well before Paxlovid emerged as an approved treatment.

Also, before we proceed further, a note about this COVID-OUT study team and their novel methodology. This team was exceptionally agile in implementing a shelter-in-place technique for prospective trials. Within weeks after the pandemic began, key members of the team were remotely enrolling patients in trials of hydroxychloroquine (remember when this was the most promising treatment we had for COVID?) as pre- and post-exposure prophylaxis for healthcare workers.*

Dr. David Boulware, a Professor of Medicine at the University of Minnesota and a senior member of the COVID-OUT team, told me that the inspiration materialized over four days earlier in the pandemic. Key members had planned to attend an HIV conference in Africa. After COVID canceled their plans, they hunkered down to figure out an expeditious way to start studying the emerging pandemic.

What they came up with were “decentralized, remotely delivered” trials that utilize an online survey and consent form followed by a prescription, either study drug(s) or placebo(s) manufactured to look exactly the same, sent in the mail, and followed up by a steady stream of online surveys. Super simple and elegant—but not without limitations. (We'll come back to that.)

So, shooting a bit blind, the study team chose to trial metformin, ivermectin, and fluvoxamine. There was actually a plausible basis for each medication having anti-viral and or anti-inflammatory properties.

Metformin, as we know, has many anti-inflammatory mechanisms and there was some preliminary evidence in mice that it might help protect against severe lung injury in COVID-19. Ivermectin had shown some promise in the lab against SARS-CoV-2, although at doses much higher than tolerated in humans. And fluvoxamine, which really seemed like an odd choice to me initially, apparently does have some anti-inflammatory actions mediated by the sigma-1 receptor. [2]

The study used a factorial design in which there were six study groups, each of which received two pills with some combination of study drugs and/or placebo. This meant that some subjects were taking two different study drugs (e.g., ivermectin and metformin).

As before, the investigators had two study outcomes after randomization. For the first one, which was severe COVID-19 disease (e.g., low oxygen levels, hospitalization) in the short term, there was a slight suggestion of a metformin benefit but otherwise no significant findings. The results were published in the New England Journal of Medicine. [3]

Long Covid Essential Reads

Covid-19 Found in People’s Blood Months After Infection

My Son’s Crushing Long-COVID Journey

In the long COVID arm, there was a nine-month follow-up that relied on study subjects to self-report whether they had been diagnosed with long COVID. The survey, which was sent monthly starting at the 9-month mark after their initial COVID diagnosis, asked the following:

Has a medical provider told you that you have "long COVID"?
– If yes: "Approximately when?"
– If yes: "Who told you?"
● My primary care provider;
● A provider who specializes in long COVID;
● A specialist; then branching logic for: cardiologist; neurologist; pulmonologist; other:_______
● A chiropractor;
● Other:_______

If the answer was yes, there was branching logic about symptoms. And that was pretty much it—a couple of questions. No phone calls, chart reviews, independent verification, or calls to any chiropractors out there who might be making long COVID diagnoses.

What's more, as the authors acknowledge, there was also the challenging dynamic of evolving long COVID diagnostic criteria and provider awareness (or lack thereof) during the study time period. Does COVID become long COVID at 28 days or three months? Does some brain fog or fatigue at month two count as long COVID, or could it be attributed to some other cause? Given that the most common of the 30 or so long COVID symptoms reported in the study were “tiredness,” “difficulty focusing” and “difficulty sleeping,” this seems a particularly relevant concern.

Recognizing the potential for reporting bias—both under-reported or over-reported long COVID—a strength of the study methodology was that the follow-up was the same for all arms on the trial and should ideally wash out bias. And there was a significant finding: The incidence of long COVID in the metformin group was 6.2 percent compared to 10.6 percent in blinded controls and 8.0 percent and 10.1 percent in the ivermectin and fluvoxamine groups, respectively. [4] Importantly, those subjects who received metformin and one of the other two study drugs also showed a trend toward lower incidence of long COVID.

Where does this leave us? Could metformin take the “long” out of long COVID? These findings suggest the answer may be yes. Should it be used with Paxlovid for prevention, or on its own? This could be problematic, as Paxlovid interferes with the metabolism of metformin. Dr. Boulware, for one, says he would take metformin if he contracted COVID-19: “I think I would take the trade-off of the possibility of a bit of loose stool for a 40 percent reduction in the chances of having long COVID.”

Can metformin treat long COVID that has already established itself? Maybe, as it does have anti-inflammatory properties, but its anti-viral properties are less likely to make a difference in a post-COVID scenario. And finally, is metformin in fact a miracle drug that can help us all live to age 1,000? We can certainly hope.

*Full disclosure: I was an early enrollee in the post-exposure trial who had some regrets after learning more about the side effect profile of hydroxychloroquine.