Peptides for Long-COVID Fatigue & Recovery: What the Evidence Actually Shows

Thymosin alpha-1 ranks first because, of the four, it has the strongest and deepest human evidence base in general — an approved immunomodulator (Zadaxin/thymalfasin) in 30-plus countries with acute-COVID RCTs and meta-analyses — plus the only PASC-specific human study of any peptide here, and it targets a genuine long-COVID phenotype: post-viral T-cell exhaustion and immune dysregulation. But first place still does not mean proven. In the single long-COVID study, blood lymphocytes from 10 previously hospitalized PASC patients were treated in the test tube (50 µg/mL, 48 h); thymosin alpha-1 reduced PD-1+ 'exhausted' CD4/CD8 T cells, lowered IL-6 and IFN-gamma, and raised anti-inflammatory IL-10, with the largest effect in patients who had had severe disease and systemic symptoms including asthenia (fatigue). Crucially, the authors state this is ex vivo with a limited number of patients — no patient was actually dosed and no fatigue scale was measured. Its acute-COVID meta-analyses are heterogeneous: one pooling eight studies found lower mortality in moderate-to-critical patients (RR 0.59) but called for RCTs, while another (9 studies, 5,352 patients) found no overall mortality effect, with benefit only in subgroups over 60 or with severe disease. Tellingly, its highest-quality trial in a related immune indication — sepsis (TESTS) — was negative. Biologically plausible for the immune subtype, but for fatigue it is unproven. Grade C for this condition.

Best for

Understanding the immune-targeted rationale in post-viral fatigue as a hypothesis, not an evidenced therapy

Pricing

Prescription / compounded (gray zone); varies by pharmacy

Source: Int Immunopharmacol 2023 (PASC ex-vivo study, PMC10030336)

Elamipretide ranks second because long COVID has documented mitochondrial dysfunction and SS-31 is the most mature mitochondria-targeted peptide in clinical development — the only candidate here with human data against validated fatigue scales. The problem is that data is in genetic mitochondrial disease, not long COVID, and the definitive trial failed. In primary mitochondrial myopathy, where fatigue is the cardinal symptom, the phase-3 MMPOWER-3 trial (N=218, 24 weeks, 40 mg SC daily) did not meet its co-primary endpoints of six-minute-walk distance or total fatigue on the PMMSA. A pre-specified nuclear-DNA/replisome subgroup improved on the 6MWT (CPEO p=0.0024), motivating the ongoing NuPOWER trial, but the headline result was negative for fatigue. There are no published long-COVID trials of elamipretide, and the leading 2024 mechanistic review of mitochondrial dysfunction in long COVID does not list it among candidate therapeutics — a tell that mito-peptides remain theoretical, not established, PASC therapies. It works by binding cardiolipin, the inner-membrane phospholipid that scaffolds cristae and clusters the oxidative-phosphorylation supercomplexes, reducing electron leak and preserving ATP synthesis in models. Its single FDA approval (Forzinity, Sept 2025) is exclusively for the ultra-rare Barth syndrome on a muscle-strength endpoint in about 12 patients — not for fatigue, energy or long COVID. Notably, fatigue itself was a reported adverse event in the program; it is not a clean stimulant. Genuine mechanism, hypothetical for PASC. Grade C.

Best for

Understanding why the mitochondrial rationale, even at its strongest, has not translated into a fatigue benefit

Pricing

Prescription (Forzinity, Barth syndrome only) / no long-COVID pathway

Source: Karaa et al., Neurology 2023 (MMPOWER-3)

Aviptadil is the cautionary tale and the most over-cited 'long-COVID peptide.' It is constantly listed in 'peptides for long COVID' content and actually has the most human RCT data of the four — but all in acute critical COVID respiratory failure, and the best of those trials failed. It is a textbook example of why 'has human trials' must be read as 'positive, in the right population.' The NIH platform trial ACTIV-3b/TESICO (IV aviptadil versus placebo in COVID hypoxemic respiratory failure) found aviptadil did not improve recovery at day 90, mortality at 90 days, or any other endpoint. The phase-2 I-SPY COVID trial of nebulized aviptadil was stopped early for futility. An earlier company-sponsored phase-2b/3 (NCT04311697) reported a recovery/survival signal in subgroups, but the independent NIH trial did not confirm it — and the FDA never approved aviptadil for COVID. Mechanistically, VIP is concentrated in alveolar type-II cells, inhibits IL-6 and TNF-alpha, and protects the alveolar barrier in models — but that is an acute lung-injury rationale, not a fatigue rationale. There is no registered or published aviptadil trial targeting long-COVID fatigue, cognition or recovery, so any 'aviptadil for long COVID' claim is an extrapolation from negative acute-illness data. It also carries a PASC-specific safety concern: dose-limiting hypotension, flushing and diarrhea were consistently more frequent than placebo — meaningful in a population with frequent orthostatic intolerance and POTS. Grade D for long COVID.

Best for

Recognizing why 'has human trials' is not the same as 'positive human trials in the right disease'

Pricing

Investigational / not FDA-approved; no long-COVID pathway

Source: ACTIV-3b/TESICO, Lancet Respir Med 2023

BPC-157 ranks last because it is heavily marketed for 'post-COVID gut and tissue recovery' yet has no human long-COVID data, no long-COVID-specific animal data, and no completed human RCT for any indication. All of its efficacy claims rest on general rodent injury, anti-inflammatory and angiogenesis models, and the leap from 'heals rat tendon or gut' to 'fixes long-COVID fatigue' is unsupported. Its marketed PASC rationale is generic 'gut-vascular repair and anti-inflammation,' extrapolated from rodent models, with no PASC-specific mechanistic human work behind it. In evidence terms it is a preclinical-only compound (a Grade C base in its primary tissue-repair claims) that collapses to Grade D for long COVID specifically, because there is not even condition-relevant animal data to extrapolate from. On regulation, BPC-157 is not FDA-approved; it was placed in Category 2 ('significant safety concerns,' citing immunogenicity and impurity risks and a lack of human safety data) in 2023, then removed from Category 2 on April 22, 2026 — but not added to the Category-1/503A compounding-permitted list. That is a regulatory gray zone, not an endorsement of safety or efficacy. It is also prohibited in sport by WADA at all times, a real sanction risk for tested athletes. Most 'research-use-only' BPC-157 product is unregulated for purity, sterility and dose. For long-COVID fatigue there is simply nothing to grade above D.

Best for

Recognizing where the evidence is thinnest — marketing extrapolation with no condition-relevant data at all

Pricing

No approved product; 'research-use-only' material is unapproved/unverified

Source: Holt Law, Regulatory Status of Compounded Peptides 2026

This entry is included as context, not as a peptide recommendation, because it defines the evidence bar no peptide has cleared. From a functional, root-cause perspective the honest reading inverts the marketing: long-COVID fatigue is a problem of broken bioenergetics and a mis-calibrated immune system, and the interventions with the best actual PASC fatigue evidence in 2026 are not peptides at all. AXA1125 is a non-peptide amino-acid metabolic modulator — five amino acids plus N-acetylcysteine, not a peptide drug — that in a 41-patient phase-2a pilot missed its primary 31P-MRS endpoint (tauPCr, P=0.24) but improved Chalder fatigue (LSMD -4.30, P=0.0039). Separately, an adaptive RCT found the repurposed SSRI fluvoxamine reduced long-COVID fatigue (day-90 mean difference -0.58). Neither of these is an endorsement — the AXA1125 signal is a small pilot that missed its mechanistic primary, and fluvoxamine's effect is modest — but they show what a real PASC fatigue evidence base looks like, and the peptides above do not have one. Their presence here is deliberate: when people ask which peptide 'works' for long COVID, the honest comparison is that the interventions with genuine PASC fatigue trial data are not peptides, and even they are early. This is the reference point that keeps the peptide discussion honest, and part of why every peptide in this review is capped at Grade C for proven long-COVID efficacy.

Best for

Anchoring expectations: the interventions with real long-COVID fatigue data are non-peptides, and even they are early

Pricing

Investigational (AXA1125) / off-label repurposed drug (fluvoxamine); not FDA-approved for long COVID

Source: AXA1125 in long COVID, eClinicalMedicine 2023 (PMC10102537)