Long COVID Treatment 2026: What Randomized Clinical Trials Actually Show
By OneDayMD Editorial Team · Published September 30, 2026 · Updated October 1, 2026 · Evidence last checked October 1, 2026
.png)
Quick Answer: does anything work for Long COVID?
No specific treatment has been proven to work for Long COVID in general. Randomized trials have found no benefit from extended Paxlovid (959 adults), metformin (396) or short-course hyperbaric oxygen (80). Promising but unconfirmed signals include the SIM01 synbiotic (463 people, Hong Kong) and physical-activity coaching (50). For prevention during acute COVID, Paxlovid (144 people) and metformin show mixed or underpowered results. Care remains individualized and symptom-focused.
- Re-checked every trial against primary sources and corrected details (for example, PANORAMIC Norway: 144 enrolled, 143 followed up).
- Added STOP-PASC and PAX LC to the Paxlovid evidence, COVID-OUT to prevention, and the 24-week STIMULATE-ICP result.
- Added trial limits that were missing (SIM01 single centre; fasting attrition; metformin placebo recovery).
- Updated low-dose naltrexone to the 2026 BMJ Open review; added company-reported ADDRESS-LC topline data.
- Fixed table-of-contents links; added a CEBM key, FAQ and AI-assistant guide.
Table of Contents
- Quick Answer
- How evidence is graded here (CEBM)
- What is Long COVID?
- Evidence grid: every trial at a glance
- Treating established Long COVID
- — Does Paxlovid work?
- — Does the SIM01 synbiotic work?
- — Does intermittent fasting work?
- — Does metformin treat it?
- — Rehabilitation, digital care and activity coaching
- — Does hyperbaric oxygen work?
- — What is new since publication?
- Can anything prevent Long COVID?
- Supplements, nattokinase and low-dose naltrexone
- Bottom line: supported vs experimental
- How to read a treatment claim
- What should patients do with this evidence?
- Using this article with Claude, ChatGPT, Gemini and Perplexity
- Frequently asked questions
- Sources and further reading
How is the evidence graded in this article?
We label each study with its Oxford CEBM (OCEBM 2011) level for treatment questions: L1 systematic review of randomized trials; L2 individual randomized trial; L3 non-randomized controlled cohort; L4 case series or uncontrolled before–after study; L5 mechanistic reasoning. The level describes study design, not whether the result was positive. Size, blinding, endpoint choice, attrition and replication determine how much weight a result deserves, so each entry also carries a certainty note.
What is Long COVID, and how is it diagnosed?
Long COVID, also called post-COVID condition or post-acute sequelae of SARS-CoV-2 infection (PASC), means persistent or recurrent symptoms after COVID-19: commonly fatigue, post-exertional symptom exacerbation (PESE), cognitive problems, breathlessness, sleep disturbance, dysautonomia and pain. The CDC states that no laboratory test can definitively diagnose or rule out Long COVID; diagnosis relies on history, examination and targeted testing to exclude other causes (CDC clinical guidance).
Two points shape how to read every trial below. First, Long COVID is heterogeneous, so a treatment can help one subgroup without helping everyone. Second, prevention (acting during acute COVID) and treatment (acting months later) are separate questions; a prevention signal does not show that a drug treats established disease.
Evidence grid: how strong is the randomized evidence for each Long COVID treatment?
Strong enough to rule out several popular ideas; not strong enough to recommend one general protocol. A 2026 systematic review of 43 randomized trials (2,878 participants) rated certainty as low in all but one study; its searches ended in August 2024, so it predates RECOVER-VITAL, STIMULATE-ICP and the 2026 metformin trial (Tamariz et al.).
| Intervention | Key trial | Design and size | Main finding | CEBM level / certainty |
|---|---|---|---|---|
| Paxlovid, established Long COVID | RECOVER-VITAL, Lancet Infect Dis 2026 | Phase 2, double-blind, placebo-controlled; 959 adults; 69 US sites | No benefit at 15 or 25 days in any of three phenotypes | L2: large, negative |
| SIM01 synbiotic | Lau et al., Lancet Infect Dis 2024 | Double-blind RCT; 463 adults; one Hong Kong centre | More participants had alleviation of five symptoms at 6 months | L2: promising; needs replication |
| Intermittent fasting + no-added-sugar diet | Bunker et al., Sci Rep 2025 | Open-label crossover; 77 started, 58 completed | Larger symptom drop in fasting weeks; frequent flare-ups | L2: preliminary; high bias risk |
| Metformin or UDCA, established | Lim et al., Ann Intern Med 2026 | Double-blind RCT; 396 adults; 14 days | Recovery 63.6% / 68.2% / 68.2% (metformin / UDCA / placebo) | L2: negative |
| MRI-guided care; digital rehabilitation | STIMULATE-ICP, Nat Med 2026 | Cluster-randomized phase 3; 1,152 participants | No added benefit at 12 weeks; small digital-rehab gains at 24 weeks (secondary) | L2: mixed |
| Hyperbaric oxygen (10 sessions) | HOT-LoCO, BMJ Open 2025 | Double-blind, sham-controlled; 80 participants | No difference from sham on primary endpoints | L2: negative, small |
| Physical-activity coaching | Diciolla et al., Sci Rep 2026 | Assessor-blinded pilot; 50 participants | Feasible; better activity, walking and fatigue vs usual care | L2: pilot only |
| Paxlovid, prevention | PANORAMIC Norway, Lancet Infect Dis 2026 | Double-blind; 144 enrolled of 2,000 planned | Long COVID at 3 months: 26% vs 43% (RR 0.60) | L2: underpowered |
| Metformin, prevention | COVID-OUT 2023; ACTIV-6 | Double-blind RCTs in outpatients | COVID-OUT: HR 0.59 (secondary). ACTIV-6: primary endpoint missed threshold | L2: mixed |
| Low-dose naltrexone | BMJ Open 2026 systematic review | No RCTs; four small before–after studies | Possible benefit; low certainty | L4: uncontrolled |
| Nattokinase | None identified | No Long COVID randomized trial found | Mechanistic interest only | L5: insufficient |
| Bezisterim (ADDRESS-LC) | Company topline, Sept 2026 | Phase 2, double-blind; 203 participants | No significant effect on 22 endpoints overall; subgroup signal | Not peer reviewed |
Swipe sideways on mobile to see all columns.
Which treatments have been tested in established Long COVID?
The sections below cover people who already have persistent symptoms. Each opens with the direct answer.
Does Paxlovid (nirmatrelvir–ritonavir) work for established Long COVID?
No. Extended Paxlovid did not improve established Long COVID in RECOVER-VITAL, the largest antiviral trial so far. Baden et al. (Lancet Infect Dis 2026) enrolled 959 adults at 69 US sites between July 2023 and early 2025. Participants were grouped by cognitive, autonomic or exercise-related (post-exertional) phenotype and randomized to 25 days of Paxlovid, 15 days of Paxlovid followed by ritonavir plus placebo, or 25 days of ritonavir plus placebo. The primary endpoint was a clinically meaningful change in patient-reported outcomes at day 90; the authors concluded there was no evidence of benefit in any of the three phenotypes (RECOVER summary).
Two earlier randomized trials agree: STOP-PASC (155 participants, 15 days; JAMA Intern Med 2024) showed no overall treatment effect, and PAX LC (Lancet Infect Dis 2025) found no significant improvement at day 28. These results do not rule out viral persistence in some patients; they show that these regimens did not produce clinical improvement in the populations tested.
Does the SIM01 synbiotic work for Long COVID symptoms?
It helped more participants than placebo in one well-run trial, but it needs replication. In Lau et al. (Lancet Infect Dis 2024), 463 adults in Hong Kong with at least one of 14 post-acute COVID symptoms for four or more weeks were randomized (232 SIM01, 231 placebo) to sachets of 10 billion colony-forming units twice daily for six months. The primary outcome was symptom alleviation at six months. Adverse events were similar between groups.
| Symptom alleviated at 6 months | SIM01 vs placebo (%) | Odds ratio |
|---|---|---|
| Fatigue | 62.8 vs 42.6 | 2.27 |
| Memory loss | 42.0 vs 26.9 | 1.97 |
| Difficulty concentrating | 62.3 vs 38.5 | 2.64 |
| Gastrointestinal / digestive upset | 70.2 vs 54.1 | 1.99 |
| General unwellness | 77.3 vs 59.0 | 2.36 |
Percentages from the trial's 2023 conference abstract.
Limits: a single centre, an all-Chinese study population and a vitamin C placebo. The result belongs to this standardized formulation and should not be generalized to other probiotics, fermented foods or generic microbiome supplements. SIM01 does not show that gut dysbiosis causes Long COVID.
Does intermittent fasting treat Long COVID?
It is a preliminary signal, not a prescription. Bunker et al. (Sci Rep 2025) ran a 10-week open-label crossover trial: after a two-week run-in, participants did four weeks of a no-added-sugar diet with a 10–12 hour eating window (TRE) and four weeks of the same diet with 16:8 fasting plus a weekly water-only fast (median 38 hours). Of 77 who started, 58 (75%) completed and formed the analytic sample. Symptom score fell 10.2 points during fasting vs 2.8 during TRE (p=0.008); symptom count fell 5.0 vs 1.4 (p=0.002). Over all 10 weeks, mean score fell from 37.8 to 18.2.
Limits: no blinding, patient-reported outcomes, no washout, no follow-up after the intervention, enrollment stopped early, and a patient-led, all-volunteer team. Participants with diabetes, pregnancy or an eating-disorder history were excluded. The mechanisms discussed (autophagy, viral clearance) remain hypotheses.
Does metformin treat established post-COVID condition?
No, not in a 2-week course. Lim et al. (Ann Intern Med 2026) randomized 396 adults at two South Korean hospitals (median age 36; 72% women) to metformin (up to 1,500 mg/day), ursodeoxycholic acid (900 mg/day) or placebo for 14 days. Recovery at 8 weeks (PASC index score below 12) was 63.6% with metformin, 68.2% with UDCA and 68.2% with placebo. The high placebo recovery rate shows why uncontrolled improvement after a treatment proves little. Longer courses, other populations and other doses were not tested.
Do rehabilitation, digital care or activity coaching help?
Specialist care helps many people; the specific add-ons tested did not clearly add more at 12 weeks. STIMULATE-ICP (Nat Med 2026) cluster-randomized 1,152 participants across six NHS clinics to usual care, multi-organ MRI, digital rehabilitation, or both. Fatigue improved in all arms, with no significant extra benefit from MRI or digital rehabilitation on the 12-week primary outcome. Digital rehabilitation showed small but statistically significant gains in fatigue and quality of life at 24 weeks, a secondary result (UCL summary). Because all arms improved and none was untreated, natural recovery and clinic care cannot be separated.
Physical-activity coaching showed an encouraging pilot signal. In Diciolla et al. (Sci Rep 2026), 50 people recruited from one patient advocacy group received a 12-week remote coaching programme (weekly sessions, activity tracker, goal setting) or usual care. The primary outcome was feasibility, which was met (89% recruitment, 92% retention, no intervention-related adverse events). At three months coaching improved steps, sedentary time, walking distance, dyspnoea and fatigue; anxiety and depression did not change. Limits: small pilot, baseline imbalances, unblinded participants, and a usual-care group that also wore trackers.
Why pacing matters. Some patients experience PESE, with symptoms worsening hours to days after physical or mental exertion. Rehabilitation should be individualized and symptom-guided rather than a generic "exercise more" prescription; see the CDC clinical guidance and WHO post-COVID-19 condition resources.
Does hyperbaric oxygen therapy work for Long COVID?
Evidence is insufficient for routine use. In HOT-LoCO (Kjellberg et al., BMJ Open 2025), 80 previously healthy adults aged 18–60 were randomized to 10 hyperbaric oxygen sessions or sham over six weeks; the primary endpoint analysis (79 participants) showed no more short-term benefit than sham. An earlier sham-controlled trial using a different protocol reported benefits (Hadanny et al., Sci Rep 2024 follow-up), so the question is protocol-specific and unsettled. This result does not exclude benefit in every subgroup.
What new Long COVID trial results appeared in September 2026?
Company-reported topline data for bezisterim; no peer-reviewed result yet. BioVie announced on September 15, 2026 that its phase 2 ADDRESS-LC trial (203 participants, randomized, double-blind, placebo-controlled, exploratory) did not reach significance on any of 22 endpoints in the full population, with a reported benefit in a subgroup with more severe baseline fatigue, post-exertional malaise and cognitive impairment (company release; The Sick Times). The company notes the FDA has not validated these endpoints as surrogate endpoints. Treat it as hypothesis-generating until peer-reviewed data and a confirmatory trial exist.
Separately, a brain-imaging study reported September 21, 2026 suggested dopamine-neuron involvement in Long COVID (CAMH summary). That is a mechanistic finding, not a treatment trial.
Can anything prevent Long COVID after acute COVID-19?
Possibly, but nothing is proven. Three randomized datasets point in different directions.
| Trial | Intervention and population | Result | Interpretation |
|---|---|---|---|
| PANORAMIC Norway (Lancet Infect Dis 2026) | Paxlovid vs placebo during acute COVID; 144 enrolled (66/78), 2,000 planned | At 3 months, Long COVID in 17 of 66 (26%) vs 33 of 77 (43%); RR 0.60 (95% CI 0.37–0.98) after imputing one missing value | Positive signal, but stopped early; authors say limited size precludes firm conclusions |
| COVID-OUT (Lancet Infect Dis 2023) | Metformin in outpatient adults with overweight or obesity | Cumulative incidence by day 300: 6.3% vs 10.4%; HR 0.59 (0.39–0.89). Ivermectin and fluvoxamine showed no effect | Long COVID was a secondary outcome added during the trial; the primary endpoint was severe COVID by day 14 |
| ACTIV-6 metformin | Metformin up to 1,500 mg/day for 14 days; about 3,000 low-risk outpatients | Primary endpoint (symptoms at day 180): risk ratio 0.79 (95% CrI 0.47–1.23), below the efficacy threshold. Clinician-diagnosed Long COVID (secondary): 0.56% vs 1.17% | Event rates were low; secondary signals support further study, not a conclusion |
Protocol for PANORAMIC Norway: PMC12590622.
Evidence lesson: a secondary or exploratory result can justify a bigger trial, but it should not be presented as an established treatment. Prevention findings also should not be used to claim a drug treats Long COVID that has already developed.
Do supplements, nattokinase or low-dose naltrexone help Long COVID?
Evidence for most supplements is far thinner than for the trials above. The key question is whether the specific product has improved clinically meaningful Long COVID outcomes in adequately controlled human trials. SIM01 is the main supplement-type exception, and only for that standardized formulation.
Nattokinase. Our October 2026 literature check found no completed randomized trial of nattokinase for Long COVID symptoms; claims rest on laboratory spike-protein degradation work and expert opinion (Science Feedback review). Nattokinase has fibrinolytic activity; anyone using blood thinners, with a bleeding disorder, or facing surgery should speak with a clinician first.
Low-dose naltrexone (LDN). A systematic review searching through May 2026 found no randomized controlled trials, only four small before–after studies suggesting possible benefit for fatigue, cognition, sleep and pain at low certainty; three registered trials were ongoing (BMJ Open 2026). Uncontrolled improvement cannot separate drug effect from natural recovery or placebo response. Absence of trials is not proof of ineffectiveness.
What does the evidence support in 2026, and what remains experimental?
| Status | What it includes | Why |
|---|---|---|
| Supported | Individualized, symptom-focused care; pacing for PESE; evaluation for other conditions | Consistent with CDC and WHO guidance; fatigue improved in specialist-clinic care in STIMULATE-ICP (not a controlled comparison) |
| Promising, unconfirmed | SIM01 synbiotic; physical-activity coaching; metformin or Paxlovid for prevention | Randomized signals, but single-centre, pilot-sized, underpowered or mixed |
| Not supported as general treatment | Extended Paxlovid; short-course metformin or UDCA; 10-session hyperbaric oxygen; routine multi-organ MRI | Negative or null primary endpoints in randomized trials |
| Experimental | Intermittent fasting; bezisterim; viral-persistence, microclot, autoimmunity and dopamine hypotheses; biomarker-guided therapy | Small, unblinded, unreviewed or mechanistic evidence only |
| Insufficient evidence | Nattokinase; low-dose naltrexone | No randomized Long COVID trial found (nattokinase) or none published (LDN) |
A negative randomized trial tests a particular drug, dose, timing, population and outcome; it does not erase a biological hypothesis. Equally, a mechanism, physician recommendation, testimonial or observational association is not equivalent to randomized evidence. The field is moving toward asking which treatment works for which biological or symptom-defined subgroup.
How should you read a Long COVID treatment claim?
Symptoms fluctuate naturally, people try several therapies at once, and improvement can occur without a single cause. A real improvement is not the same as proof that a treatment caused it. Ask:
- Was the study randomized, and was there a placebo or sham control?
- How many people finished, and how many dropped out?
- Was the intervention blinded, and what was the primary endpoint?
- Was the result statistically significant and clinically meaningful?
- Was it replicated, and how durable was the benefit?
- Was the finding a primary or only a secondary outcome?
- What adverse events occurred?
- Does it concern prevention or treatment of established disease?
Evidence ladder: mechanism suggests a hypothesis; observation suggests a signal; a randomized trial estimates a treatment effect; replication across high-quality trials builds confidence.
What should patients do with this evidence?
Current evidence supports an individualized approach rather than a universal protocol. Clinical evaluation usually starts with the symptoms causing the most functional impairment and with excluding other diseases that mimic or accompany Long COVID, which may involve cardiovascular, pulmonary, neurological, autonomic, sleep or metabolic assessment. If you experience PESE, ask for pacing and energy-conservation guidance rather than graded exercise alone. Bring the trial names and sample sizes from this article to your appointment.
How can you use this article with Claude, ChatGPT, Gemini and Perplexity?
AI assistants can help you prepare questions and compare trials, but they can be outdated or wrong and are not a substitute for a clinician. Ask for named trials, sample sizes and links to primary sources, then check them.
| Assistant | Useful for | Example prompt |
|---|---|---|
| Claude | Turning this article into questions for your clinician | "Using only the randomized trials in this article, list the evidence for treating [my main symptom] in established Long COVID, and the questions I should ask my doctor." |
| ChatGPT | Side-by-side comparisons and visit-prep sheets | "Compare the SIM01 trial and the physical-activity coaching pilot: size, endpoint, result, limits and what is still unknown." |
| Gemini | Cross-checking against official guidance | "Does current CDC guidance agree with this article's description of post-exertional symptom exacerbation? Cite the CDC page." |
| Perplexity | Finding newer trials with primary sources | "List Long COVID randomized trials published after October 1, 2026, with design, sample size, primary endpoint and a link to the paper." |
Frequently asked questions about Long COVID treatment
What is the best treatment for Long COVID in 2026?
No treatment has been proven effective for Long COVID in general. Randomized trials support individualized, symptom-focused care, and a few approaches show preliminary signals, including the SIM01 synbiotic and physical-activity coaching. Extended Paxlovid, metformin and short-course hyperbaric oxygen did not beat placebo or sham on their primary endpoints. A clinician should tailor care to your main symptoms and exclude other conditions.
Does Paxlovid work for Long COVID?
Not as a treatment for established Long COVID. In RECOVER-VITAL, 959 adults received 15 or 25 days of nirmatrelvir–ritonavir or placebo, and no benefit appeared in any studied phenotype. Giving Paxlovid during acute COVID showed a possible prevention signal in the small PANORAMIC Norway trial (144 people), but that study stopped early and needs confirmation.
Can metformin prevent or treat Long COVID?
Results differ by timing. For established Long COVID, a 396-person randomized trial found no benefit over placebo. For prevention, COVID-OUT reported 41% lower Long COVID incidence with metformin started during acute COVID, while ACTIV-6 missed its primary endpoint but favored metformin on clinician-diagnosed Long COVID. Metformin for prevention remains unproven.
Is intermittent fasting a proven Long COVID treatment?
No. The only randomized trial was small, unblinded and patient-reported: 58 of 77 starters completed it, and symptom scores fell more during fasting weeks. However, 90% of completers reported at least one flare-up, and two hospitalizations were reported, one before fasting began. Prolonged water fasts can be unsafe for some people, so discuss any fasting plan with a clinician.
Do probiotics help Long COVID?
One standardized synbiotic, SIM01, helped more participants than placebo on fatigue, memory, concentration, digestive symptoms and general unwellness at six months in a 463-person Hong Kong trial. That result applies to that formulation, comes from one centre and awaits replication. It should not be generalized to other probiotics or fermented foods.
Does hyperbaric oxygen work for Long COVID?
Evidence is insufficient for routine use. In the sham-controlled HOT-LoCO trial, 10 sessions in 80 people did not outperform sham treatment on the primary endpoints. An earlier trial using a different protocol reported benefits, so effects may depend on protocol and patient selection, and larger confirmatory trials are needed.
Does nattokinase work for Long COVID?
Our October 2026 literature check found no completed randomized trial testing nattokinase for Long COVID symptoms. Laboratory and mechanistic findings are hypothesis-generating only. Nattokinase has fibrinolytic activity, so people taking blood thinners or facing surgery should speak with a clinician first.
Is exercise safe for people with Long COVID?
It depends on whether you have post-exertional symptom exacerbation (PESE), where activity triggers delayed symptom worsening. A 50-person pilot found symptom-guided activity coaching feasible and safe, but exercise should be individualized, paced and monitored. Pushing through symptoms can backfire; a clinician or rehabilitation specialist familiar with Long COVID can help set limits.
Sources and further reading
- Baden L, Shah N, Liu S, et al. Nirmatrelvir–ritonavir targeting viral persistence in post-COVID-19 condition (RECOVER-VITAL). Lancet Infect Dis 2026. DOI
- Geng LN, et al. Nirmatrelvir–ritonavir and symptoms in adults with postacute sequelae of SARS-CoV-2 infection (STOP-PASC). JAMA Intern Med 2024;184:1024. PubMed
- PAX LC: nirmatrelvir–ritonavir vs placebo–ritonavir in long COVID. Lancet Infect Dis 2025. Lancet
- Oppegaard et al. Nirmatrelvir for acute COVID-19 to prevent long COVID (PANORAMIC Norway). Lancet Infect Dis 2026. PubMed
- Lau RI, et al. A synbiotic preparation (SIM01) for post-acute COVID-19 syndrome in Hong Kong (RECOVERY). Lancet Infect Dis 2024;24(3):256–265. PMID 38071990. PubMed
- Bunker T, et al. Intermittent fasting and a no-sugar diet for Long COVID symptoms: a randomized crossover trial. Sci Rep 2025;15:27563. Sci Rep
- Lim SY, et al. Neither metformin nor ursodeoxycholic acid effectively treats postacute sequelae of COVID-19. Ann Intern Med 2026. PubMed
- Bramante CT, et al. Outpatient treatment of COVID-19 and incidence of post-COVID-19 condition over 10 months (COVID-OUT). Lancet Infect Dis 2023;23:1119–1129. PubMed
- ACTIV-6 Study Group. Metformin on the presence of COVID-19 symptoms over 6 months: the ACTIV-6 randomized clinical trial. PubMed
- STIMULATE-ICP: integrated care pathway in individuals with Long COVID, a cluster-randomized phase 3 trial. Nat Med 2026. Nature Medicine
- Kjellberg A, et al. Ten sessions of hyperbaric oxygen versus sham treatment in patients with long covid (HOT-LoCO). BMJ Open 2025;15(4):e094386. BMJ Open
- Hadanny A, et al. Long term outcomes of hyperbaric oxygen therapy in post covid condition: longitudinal follow-up of a randomized controlled trial. Sci Rep 2024;14:3604. DOI
- Diciolla NS, et al. Physical activity coaching programme for people with Long COVID: a pilot randomised clinical trial. Sci Rep 2026;16:14820. Sci Rep
- Tamariz L, et al. Systematic review of randomized clinical trials for the treatment of long COVID syndrome. J Community Hosp Intern Med Perspect 2026;16(4). PMC
- Byambasuren O, et al. Effect of low-dose naltrexone for long covid: a systematic review and meta-analysis. BMJ Open 2026;16(7):e111253. BMJ Open
- Science Feedback. No scientific evidence for the claim that nattokinase can treat long COVID. Science Feedback
- BioVie. Topline results from the phase 2 ADDRESS-LC trial of bezisterim. Press release, September 15, 2026 (not peer reviewed). GlobeNewswire
- CDC. Long COVID clinical guidance. CDC
- World Health Organization. Post COVID-19 condition resources. WHO
Related OneDayMD pages (practitioner-protocol and mechanistic content; lower-tier evidence than the randomized trials above): Post-COVID and post-vaccine spike protein recovery guide · Peter McCullough spike protein detox protocol · How to measure spike protein antibodies · I-Recover protocol and McCullough Base Spike Detox guide · Nutraceuticals and Long COVID
Last updated: October 1, 2026 · Topics: long covid · paxlovid · metformin · probiotics · fasting
Comments
Post a Comment