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

The number of people living with Long COVID continues to climb. As of early 2026, an estimated 200 million people worldwide are still dealing with symptoms that won’t go away - fatigue that crushes your energy, brain fog that makes simple decisions feel impossible, heart palpitations that come out of nowhere, and muscle pain that lingers long after the virus is gone. For many, the acute phase of COVID-19 is over, but the illness isn’t. And here’s the hard truth: there are still no FDA-approved drugs specifically for Long COVID. That means people are turning to medications that were never designed for this - and that’s where the real danger and the real hope lie.

Repurposed Drugs: Hope Built on Old Safety Data

Most of the current treatment efforts are based on repurposing drugs already approved for other conditions. It’s the fastest path forward. But using a drug for a new purpose doesn’t automatically mean it’s safe for that new group of patients. The safety profile of baricitinib, for example, was built in people with rheumatoid arthritis - older adults with weakened immune systems. Long COVID patients? Many are younger, previously healthy, and not immunocompromised. So when you hear about baricitinib being tested in the REVERSE-LC trial, you need to ask: does its known risk of serious infections, blood clots, and cancer make sense for someone who just wants to walk up stairs without collapsing?

Baricitinib showed promise in reducing death during acute COVID-19. Now it’s being tested to see if it can help with Long COVID fatigue and inflammation. The NIH is pouring money into this trial - over $100 million - because the early signals were strong. But the FDA label warns of major cardiovascular events happening in 3.3 to 4.1 cases per 100 patient-years. That’s not a small number. And we don’t yet know if Long COVID patients, who may already have endothelial damage, are at higher risk. The trial results won’t be out until late 2026. Until then, some doctors are prescribing it off-label. Others refuse.

Metformin: The Unexpected Front-Runner

Metformin, a 60-year-old diabetes drug, is now one of the most talked-about treatments. A 2023 study from the University of Minnesota, published in Nature Medicine, found that people who started metformin within five days of a positive COVID test had a 41% lower chance of developing Long COVID. That’s not just statistically significant - it’s dramatic. The effect was so strong, it’s now being tested in larger phase 3 trials.

But here’s the catch: 35.7% of people in the trial had gastrointestinal side effects - nausea, diarrhea, cramps. That’s more than one in three. For someone already battling brain fog and fatigue, adding constant stomach pain might feel worse than the original symptoms. And we still don’t know if metformin works for people who’ve had Long COVID for months or years. The study only looked at early intervention. What about the person who’s been sick for 18 months? No data yet.

Low-Dose Naltrexone: A Quiet Hope with Hidden Costs

Low-dose naltrexone (LDN), usually given at 1-5 mg daily, is another off-label favorite. Originally used to treat opioid addiction at 50 mg, the low dose seems to calm immune overactivity. A 2024 patient survey from Nova Southeastern University found that 62% of Long COVID patients reported improved energy after three months of LDN. That’s powerful. But 28% had worse sleep. 19% got headaches. One person reported severe anxiety. These aren’t rare side effects - they’re common enough to be tracked.

There’s no large, randomized trial yet. Just observational data. That means we don’t know if the improvement was real, or if people felt better because they believed it would work. The placebo effect is strong in Long COVID. And because LDN is cheap and off-patent, there’s no pharmaceutical company funding big studies. So progress is slow. Patients are left to experiment on themselves - often without medical supervision.

Paxlovid: Mixed Signals and a Bitter Taste

Paxlovid - the antiviral combo of nirmatrelvir and ritonavir - was the go-to for early COVID. Now it’s being tested for Long COVID. The results? Confusing. A small study from UCSF in 2024 showed a 38% symptom improvement with a 15-day course. But a 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% vs. 32.8% improvement. So what’s going on?

One thing we do know: 79% of users reported a bitter, metallic taste. Some say it’s unbearable. Others describe it as a constant reminder they’re sick. And ritonavir - the booster in Paxlovid - interacts with dozens of common medications. Blood thinners, statins, anti-anxiety drugs, even some heart medications. For someone already on multiple pills, this isn’t just inconvenient - it’s dangerous. A 2025 review found that 40% of Long COVID patients on Paxlovid had to stop or adjust other meds because of interactions.

Failed Trials and Hidden Risks

Not every drug even makes it to the finish line. The BC007 trial - designed to clear out harmful autoantibodies thought to cause Long COVID - was shut down in March 2025. Why? It didn’t work better than placebo. Worse, three patients in the treatment group had serious infusion reactions. One needed ICU care. That’s the reality of experimental treatments: success is rare. Failure is common.

Other drugs are still in early testing. AER002, a long-acting antibody, showed only mild infusion reactions in 18% of participants. Polymerized collagen, injected into the skin, caused only temporary pain in 12%. These look safe - but we don’t know if they do anything. Safety without efficacy is just a placebo with needles.

The Unknowns Are the Biggest Threat

Here’s what we still don’t know - and why this is so dangerous:

• No biomarkers. We can’t test for Long COVID. No blood test. No scan. No clear signal that someone has it - or that a treatment is working. Doctors are guessing.
• No consensus on subtypes. The NIH now recognizes at least four different forms of Long COVID: one driven by fatigue and brain fog, another by heart rhythm issues, another by immune overdrive, and a fourth by metabolic dysfunction. Treating them all with the same drug is like using one key to open every lock in a city.
• Unknown long-term effects. We don’t know if taking baricitinib for six months will raise cancer risk. We don’t know if LDN for two years will damage the liver. We don’t know if metformin changes gut bacteria permanently. These drugs were studied for months - not years - in different populations.
• Who gets treated? A 2025 survey of 15,000 Long COVID patients found that 68% tried at least one off-label medication. But only 12% had any medical guidance. The rest self-prescribed based on online forums. That’s not healthcare. That’s trial and error.

What Patients Are Really Experiencing

Community feedback paints a clearer picture than clinical trials. The Body Politic support group, with over 15,000 members, reports:

• 32% have tried metformin - but 57% say it didn’t help enough
• 29% tried LDN - 41% said side effects were worse than symptoms
• 24% tried Paxlovid - many quit because of the taste or drug interactions
• 57% of respondents said they felt worse after trying a medication

These aren’t just numbers. These are people who spent their savings on supplements, drove hours to clinics offering unproven treatments, and lost jobs because they couldn’t keep up. And now they’re told: “We don’t know what works.”

What Comes Next?

The NIH’s RECOVER initiative is now testing GLP-1 agonists like tirzepatide - drugs originally for diabetes and weight loss. The theory? Long COVID may involve metabolic dysfunction and brain inflammation. But tirzepatide causes nausea in 20-25% of users. Would someone with chronic fatigue want to add vomiting to their list of problems?

Stellate ganglion blocks - nerve injections used for chronic pain - are also being tested. Early data from pain clinics shows 15% get hoarseness. 5% get a hematoma. But no one has studied whether this helps brain fog or fatigue in Long COVID. It’s a shot in the dark.

And then there’s the new discovery: a compound from WEHI in Australia that prevented Long COVID symptoms in mice. It’s not even in human trials yet. Toxicity studies start in early 2026. That’s two years away. For someone who’s been sick for three years, that’s a lifetime.

Bottom Line: Patience, Caution, and Realistic Expectations

There’s no magic pill. Not yet. And there won’t be until we understand what Long COVID actually is. Right now, we’re treating symptoms with drugs designed for other diseases - and hoping for the best. The safest approach? Avoid unmonitored off-label use. If you’re considering a medication, ask your doctor: “What’s the evidence? What are the risks for someone like me? And what happens if it doesn’t work?”

The next two years will be critical. Results from baricitinib and metformin trials could lead to the first FDA-approved Long COVID treatments by 2027 or 2028. But until then, the biggest risk isn’t doing nothing - it’s doing something unproven without knowing the cost.