Long COVID Treatments: Medication Safety Signals and Unknowns in 2026

Imagine waking up three years after your initial infection, still exhausted, still foggy, and still searching for answers. You are not alone. As of 2025, approximately 200 million people globally live with the lingering effects of SARS-CoV-2, a condition formally known as Post-Acute Sequelae of SARS-CoV-2 infection (PASC), commonly referred to as Long COVID. The frustration is palpable because, despite the massive scale of this health crisis, there are currently no FDA-approved treatments specifically designed for it. We are in a period of intense experimentation, where doctors and researchers are repurposing existing drugs while racing to understand the underlying biology. But here is the catch: every medication carries risks, and in a population that is already vulnerable, understanding those medication safety signals is just as critical as finding what works.

The Reality of Current Treatment Landscapes

We need to be honest about where we stand. There is no magic bullet yet. The current approach relies heavily on clinical trials investigating drug classes originally developed for other conditions. The National Institutes of Health’s RECOVER initiative is leading this charge, having allocated $1.15 billion to research through 2025. This isn't just academic curiosity; it's an urgent effort to treat a multisystem condition that presents with over 200 documented symptoms. From brain fog to persistent fatigue, the variability makes targeted treatment incredibly difficult. When you look at the data from Stanford University’s 2025 symposium, the complexity becomes clear: we are treating a chameleon-like illness with tools built for different diseases. This mismatch creates significant unknowns regarding both efficacy and long-term safety.

Baricitinib: Promise vs. Immune Risks

One of the most closely watched candidates is Baricitinib, a Janus kinase (JAK) inhibitor. Originally approved for rheumatoid arthritis in 2018, it gained fame during the acute phase of the pandemic for reducing mortality. Now, it is being tested in the REVERSE-LC trial, a large phase 3 study coordinated by Vanderbilt University Medical Center. The logic is sound: if it calms the immune system during acute infection, maybe it can help those stuck in chronic inflammation. However, the safety profile raises eyebrows. In rheumatoid arthritis trials, serious infections occurred in 10-20% of patients. There are also concerns about malignancy, particularly lymphoma, and major cardiovascular events. For a Long COVID patient who may already have endothelial dysfunction or immune dysregulation, adding an immunosuppressant is a calculated risk. Dr. Wes Ely argues that its established profile provides a strong foundation, but Dr. Priscilla Hsue cautions that careful monitoring is essential. The results from this trial, expected in late 2026, will be pivotal.

Metformin: The Surprising Contender

If there is one drug generating cautious optimism, it is Metformin. Known primarily for type 2 diabetes, it showed remarkable potential in the University of Minnesota’s STOP COVID trial published in *Nature Medicine* in July 2023. The study demonstrated a 41% reduction in Long COVID incidence when administered early after infection. That is a substantial number. However, the side effects are real. Gastrointestinal issues affected 35.7% of participants compared to 16.3% in the placebo group. For someone already struggling with nausea or digestive problems-a common Long COVID symptom-this trade-off is significant. Community feedback from the Body Politic Long COVID Support Group indicates that 32% of respondents have tried metformin off-label, but many report that the gut distress outweighs the benefits. It remains a promising avenue, but not without its own set of hurdles.

Other Investigational Therapies and Setbacks

The path to effective treatment is rarely straight. Consider Sipavibart, a long-acting monoclonal antibody developed by AstraZeneca. Approved for prophylaxis in Japan and the EU in 2024, it is now being tested for treatment in a trial led by Dr. Nancy Klimas. Early safety data shows injection site reactions in 15.2% of recipients, but we don’t yet know how it affects the complex autoimmunity seen in some Long COVID cases. Then there was BC007, designed to neutralize functional autoantibodies. Its phase II trial was suspended in March 2025 after failing to show superiority over placebo, with higher adverse event rates in the treatment group. This setback highlights the high failure rate in this field. Other therapies like AER002, a long-acting human immunoglobulin, are showing mild infusion reactions in 18% of participants, while polymerized type I collagen has shown favorable safety in small pilot studies. Each represents a thread in a very tangled web.

The Paxlovid Paradox

You might wonder why Nirmatrelvir/ritonavir (Paxlovid), the gold standard for acute COVID, hasn't solved Long COVID. The answer lies in the mixed results. A UCSF study in May 2024 reported symptom improvement in 38% of participants using extended courses. However, a larger NIH RECOVER trial published in *JAMA Internal Medicine* in February 2025 found no significant difference between Paxlovid and placebo. Furthermore, the side effects are notable. Seventy-nine percent of users reported a bitter taste, and 40% faced potential drug interactions due to ritonavir’s inhibition of CYP3A4 enzymes. For patients managing multiple medications for comorbidities, this interaction risk is a major safety signal that cannot be ignored. It serves as a reminder that what works for acute viral replication may not address the post-viral inflammatory or autoimmune mechanisms driving Long COVID.

Critical Unknowns: Biomarkers and Endotypes

Perhaps the biggest hurdle isn't just the drugs themselves, but our lack of understanding of the disease. We still lack validated biomarkers for diagnosis or treatment response. The RECOVER initiative’s analysis of 1.2 million electronic health records in June 2025 failed to identify consistent biomarkers. Without these markers, we are essentially guessing which patient will respond to which drug. The NIH recognizes at least four distinct Long COVID endotypes, suggesting that one size definitely does not fit all. If you have an autoimmune-driven endotype, an antiviral won't help. If you have mitochondrial impairment, an immunosuppressant might make things worse. This heterogeneity complicates clinical trial design and makes personalized medicine a necessity rather than a luxury. Until we can reliably subtype patients, safety signals will remain difficult to interpret across the broader population.

What Patients Are Doing Now

In the absence of approved guidelines, patients are taking matters into their own hands. Data from the Body Politic Long COVID Support Group reveals that 68% of respondents have tried at least one medication off-label. Low-dose naltrexone (LDN) is popular, used by 29% of respondents, with observational studies suggesting fatigue improvement in 62% of patients. However, sleep disturbances occurred in 28% of users. The reality is stark: 57% reported insufficient symptom relief, and 41% experienced side effects that outweighed the benefits. This highlights a dangerous gap in care. Without standardized protocols, practice variation is high, and patients often navigate a minefield of unproven therapies. The American Medical Association acknowledged this lack of consensus in January 2025, urging caution and shared decision-making.

Future Directions: GLP-1 Agonists and Beyond

Looking ahead, the NIH RECOVER-TLC program plans to evaluate new classes of drugs, including GLP-1 receptor agonists like Tirzepatide (Mounjaro). Approved for diabetes and obesity, these drugs are hypothesized to help with metabolic dysregulation and neuroinflammation in Long COVID. While they carry known gastrointestinal side effects (nausea in 20-25%), their potential to address systemic inflammation is intriguing. Additionally, procedural interventions like stellate ganglion blocks are being considered, though safety data specific to Long COVID populations is lacking. On the horizon, WEHI researchers discovered a novel antiviral compound that prevented symptoms in mice in April 2025. Human toxicity studies are scheduled for early 2026. These developments offer hope, but they also underscore the long road ahead. The global market for Long COVID treatments is projected to reach $5.2 billion by 2030, driven by this innovation, but regulatory pathways remain uncertain without validated endpoints.