Long COVID Treatments: What Medications Show Promise and What Safety Risks Remain

More than five years after the pandemic began, millions of people still struggle with symptoms that won’t go away. Fatigue so deep it feels like walking through wet concrete. Brain fog that makes simple decisions impossible. Heart palpitations after climbing a flight of stairs. These aren’t just lingering side effects-they’re the reality of Long COVID, or Post-Acute Sequelae of SARS-CoV-2 infection (PASC). As of 2025, an estimated 200 million people worldwide live with it, including 15 million in the U.S. alone. And yet, there’s still no FDA-approved treatment for it.

That’s why doctors and researchers are turning to medications already on the market-drugs approved for other conditions-and testing them in people with Long COVID. It’s a race against time, fueled by desperation and hope. But here’s the catch: what works for one person might harm another. And for many of these drugs, we simply don’t know enough about how they’ll behave in this new group of patients.

Baricitinib: A Double-Edged Sword

One of the most studied drugs right now is baricitinib, a JAK inhibitor originally approved for rheumatoid arthritis and alopecia areata. It was one of the first drugs shown to reduce death rates in hospitalized COVID-19 patients during the pandemic. Now, researchers are testing it in people with Long COVID through the NIH-funded REVERSE-LC trial, a large, multi-site study led by Vanderbilt University.

The theory is simple: if inflammation drives Long COVID symptoms, then blocking inflammatory pathways might help. Early signals from smaller studies showed improved energy levels and reduced brain fog in some patients. But baricitinib isn’t harmless. In people with autoimmune diseases, it increases the risk of serious infections-up to 20% of users develop pneumonia, urinary tract infections, or other severe illnesses. It also raises the chance of blood clots, heart attacks, and certain cancers like lymphoma.

Here’s the problem: most Long COVID patients weren’t sick before. They weren’t elderly. They weren’t immunocompromised. So does the safety profile from rheumatoid arthritis patients even apply? Maybe not. But we don’t have better data yet. The REVERSE-LC trial is tracking these risks closely, with results expected in late 2026. Until then, prescribing baricitinib for Long COVID is experimental-and risky.

Metformin: The Unexpected Hero

Then there’s metformin, a cheap, widely used diabetes drug that’s showing surprising promise. In a major 2023 study published in Nature Medicine, researchers found that people who took metformin within the first few days of a COVID-19 infection had a 41% lower chance of developing Long COVID. That’s not just statistically significant-it’s meaningful.

How does it work? The exact mechanism is still unclear, but it may reduce inflammation, improve mitochondrial function, and even interfere with viral replication. Unlike baricitinib, metformin has been used safely for decades in millions of people. But it’s not without side effects. In the study, 35.7% of users had gastrointestinal issues-nausea, diarrhea, stomach cramps. For some, those symptoms were worse than their Long COVID.

Still, because it’s cheap, available, and has a strong safety record in general populations, many doctors are now prescribing it off-label. A 2025 survey of Long COVID support groups found that 32% of respondents had tried metformin. About half said it helped with fatigue or brain fog. But nearly a third said the side effects made them stop. It’s not a cure. But for some, it’s the best option they’ve found so far.

Low-Dose Naltrexone: A Quiet Contender

Then there’s low-dose naltrexone (LDN), a version of a drug normally used to treat opioid addiction-but given at 1 to 5 milligrams instead of 50. At these tiny doses, it appears to modulate the immune system and reduce inflammation in the nervous system.

Observational data from a 2024 study at Nova Southeastern University showed that 62% of Long COVID patients reported improved energy after taking LDN for three months. Another 45% said brain fog lessened. But side effects were common: 28% had trouble sleeping, 19% got headaches, and a few reported vivid dreams or anxiety.

Unlike baricitinib or metformin, LDN isn’t approved for any use related to Long COVID. It’s prescribed off-label, and there are no large randomized trials yet. But because it’s inexpensive and has a relatively mild side effect profile, it’s become a go-to for many patients who’ve tried everything else. Still, the lack of rigorous data means we can’t say for sure if it’s working-or if the improvements are just placebo.

Paxlovid: The Antiviral That Might Not Work

Paxlovid, the antiviral combo of nirmatrelvir and ritonavir, was a game-changer for acute COVID-19. It cuts hospitalization risk by nearly 90% if taken early. So naturally, doctors tried it for Long COVID.

One small 2024 study from UCSF found that 38% of patients improved after a 15-day course. But a much larger NIH trial published in JAMA Internal Medicine in early 2025 found no difference between Paxlovid and placebo. The numbers were almost identical: 34.1% improved on Paxlovid, 32.8% on placebo.

And then there are the side effects. Nearly 80% of users reported a terrible metallic or bitter taste that lasted for days. Ritonavir, the booster in Paxlovid, interferes with dozens of common medications-blood thinners, statins, antidepressants, even some birth control pills. For someone already juggling multiple symptoms, adding drug interactions can be dangerous. For many, the trade-off isn’t worth it.

Drugs That Failed

Not every promising idea works. BC007, a drug designed to neutralize harmful autoantibodies thought to cause Long COVID, was halted in March 2025 after a phase II trial showed no benefit over placebo. Worse, three patients in the treatment group had serious infusion reactions-something that didn’t happen in the placebo group.

Other drugs like AER002 (a long-acting antibody) and polymerized collagen are still being tested, but their safety data is limited to small groups. The collagen pilot, for example, had only 40 participants and no serious side effects-but that’s not enough to say it’s safe for thousands.

The Big Unknowns

Here’s what we still don’t know:

• Who responds to what? Long COVID isn’t one illness. The NIH now recognizes at least four distinct types-some driven by autoimmunity, others by lingering virus, others by nerve damage or metabolic dysfunction. A drug that helps one group might do nothing for another.
• How long should you take it? Should you take metformin for three months? A year? Forever? We don’t know.
• Are there biomarkers? No blood test, scan, or lab result can confirm if someone has Long COVID-or if a treatment is working. That makes clinical trials incredibly hard to design.
• What about long-term safety? Baricitinib’s cancer risk? LDN’s impact on sleep over years? Metformin’s effect on vitamin B12? We’re guessing based on short-term use in other populations.

And then there’s the human factor. A 2025 survey of 15,000+ members in the Body Politic Long COVID group found that 68% had tried at least one medication off-label. But 57% said they didn’t get enough relief. And 41% said the side effects were worse than their original symptoms.

What Should You Do?

If you’re living with Long COVID, you’re probably tired of waiting for a cure. You might be tempted to ask your doctor for baricitinib, metformin, or LDN. Here’s what to consider:

• Metformin is the safest bet if you’re healthy and don’t have kidney problems. Start low (500 mg daily) and go slow. Watch for stomach upset.
• LDN is worth discussing if you’ve tried other options. Ask for a 1.5-3 mg dose at bedtime to minimize sleep disruption.
• Baricitinib should only be considered in a clinical trial. The risks are real, and the benefits are still unproven.
• Paxlovid is unlikely to help unless you still have active viral replication-which is rare in Long COVID. Skip it unless your doctor sees clear signs of lingering virus.

And always tell your doctor what you’re taking-even if it’s over-the-counter or herbal. Many Long COVID patients mix supplements, painkillers, and sleep aids. That’s a recipe for dangerous interactions.

The Road Ahead

The NIH’s RECOVER initiative has poured over $1 billion into Long COVID research. Phase 3 trials for baricitinib and metformin will wrap up in 2026-2027. If they show strong results, the first FDA-approved treatments could arrive by late 2027.

But even then, it won’t be a magic bullet. Long COVID is too complex for one pill to fix. The future likely lies in personalized medicine-matching the right drug to the right patient based on their symptoms, biology, and biomarkers. That’s still years away.

For now, the best advice is this: stay informed, stay cautious, and don’t rush into unproven treatments. The science is moving fast-but your health is worth waiting for.