Debunked COVID Treatments
Hydroxychloroquine / Chloroquine
The claim: Hydroxychloroquine (HCQ), an antimalarial drug, could prevent or treat COVID-19 based on early in vitro studies and anecdotal reports. The claim received prominent political endorsement in the U.S. in March–April 2020.
The evidence: Multiple large, randomized controlled trials — including the WHO SOLIDARITY trial[M1] (enrolling 11,266 patients), the RECOVERY trial[M2] (4,716 patients), and the NIH ORCHID trial — found no benefit of HCQ in reducing mortality, time to discharge, or disease progression in hospitalized COVID-19 patients. The FDA revoked its Emergency Use Authorization for HCQ in June 2020 after safety concerns, including cardiac arrhythmia risk, emerged alongside the null efficacy data.
Why it doesn't work in vivo: Although HCQ inhibited SARS-CoV-2 in cell cultures, the concentrations required exceeded those safely achievable in human lung tissue.
Ivermectin
The claim: Ivermectin, an antiparasitic drug, was promoted by a vocal community of physicians and online advocates as an effective COVID-19 treatment, particularly after early observational studies appeared promising.
The evidence: The TOGETHER trial[M3], a double-blind, randomized, placebo-controlled trial of 1,358 high-risk adults, found no meaningful benefit of ivermectin on hospitalization or prolonged emergency care (14.7% vs. 16.3% placebo; relative risk 0.90, 95% CI 0.70–1.16). The WHO SOLIDARITY PLUS trial similarly found no significant benefit. The Cochrane Collaboration's 2022 systematic review found no evidence supporting ivermectin to prevent COVID-19, and concluded that the trial evidence did not justify using it to treat COVID-19 outside of well-designed randomized studies.[M4] Earlier favorable studies were found to contain data irregularities.
Bleach / Disinfectant Injection
The claim: In April 2020, publicly speculated suggestions were made that injecting or ingesting household disinfectants might kill the virus inside the body.
The evidence: Household disinfectants (bleach, isopropyl alcohol, Lysol) are acutely toxic when ingested or injected. They cause severe chemical burns to mucous membranes, esophagus, and stomach; systemic absorption causes organ failure. Poison control centers in multiple U.S. states reported increased calls about disinfectant ingestion in the days following the public remarks. Under no circumstances should disinfectants be ingested or injected.
Colloidal Silver
The claim: Colloidal silver — microscopic silver particles suspended in liquid — can treat or prevent COVID-19.
The evidence: The FDA and FTC issued warnings to multiple companies making these claims. There is no scientific evidence that colloidal silver is effective against any virus, including SARS-CoV-2. Chronic ingestion causes argyria, a permanent blue-gray discoloration of the skin. The FDA has stated that colloidal silver is not generally recognized as safe or effective for any medical condition.
UV Light Internally
The claim: Shining ultraviolet light inside the body could kill the virus.
The evidence: UV-C light (100–280 nm) is germicidal and is used to disinfect surfaces, water, and air. It cannot be safely applied to internal tissues; UV-C radiation at germicidal wavelengths causes severe cellular damage including DNA strand breaks, and would destroy human tissue along with any virus. External UV lamps do not penetrate below the skin surface. No approved or experimental device exists that safely delivers germicidal UV light to respiratory tissues.
The Numbers Behind "Not Effective"
"The trials found no benefit" is the summary. These are the trials, with what they actually measured.
Hydroxychloroquine: RECOVERY did not just fail to find a benefit
The RECOVERY trial, published in the NEJM, closed enrolment to its hydroxychloroquine arm on June 5, 2020, after an interim analysis found a lack of efficacy. Death within 28 days occurred in 421 patients (27.0%) given hydroxychloroquine and 790 (25.0%) given usual care — rate ratio 1.09, 95% CI 0.97–1.23. That is a difference consistent with chance, and the trial reported consistent results across every prespecified subgroup.
Two secondary findings in the same NEJM report point the other way and are rarely quoted. Patients given hydroxychloroquine were less likely to be discharged alive within 28 days (59.6% vs 62.9%; rate ratio 0.90, 95% CI 0.83–0.98), and among those not ventilated at baseline, more reached invasive mechanical ventilation or death (30.7% vs 26.9%; risk ratio 1.14, 95% CI 1.03–1.27). Both intervals exclude 1.0. The trial reported a small numerical excess of cardiac deaths (0.4 percentage points) but no difference in new major arrhythmia.
These are secondary outcomes from an open-label trial and should not be overstated into a claim that hydroxychloroquine killed people. What they do establish is that the honest summary is not "it didn't help" but "it didn't help, and the measured signals ran in the wrong direction." A drug given to hospitalised patients on the strength of cell-culture results and anecdote carried real downside risk, and that is the part of the story worth remembering.
Ivermectin: a well-powered trial with a tight interval
The TOGETHER trial randomised 3,515 patients across its arms, of whom 679 received ivermectin and 679 placebo. A primary-outcome event — hospital admission for COVID-19 progression, or prolonged emergency-department observation — occurred in 100 of 679 (14.7%) on ivermectin against 111 of 679 (16.3%) on placebo: relative risk 0.90, 95% Bayesian credible interval 0.70 to 1.16. The modified intention-to-treat analysis gave 0.89 (0.69–1.15) and the per-protocol analysis, restricted to patients reporting 100% adherence, gave 0.94 (0.67–1.35). Of the 211 events, 171 (81.0%) were hospital admissions. There were no significant effects on any secondary outcome.
The per-protocol result matters, because the most common defence of ivermectin is that trials used the wrong dose or enrolled patients too late. Among patients who took every dose as assigned, the point estimate moved toward no effect, not away from it. The Cochrane review reached the same conclusion across the literature and noted that several early favourable studies contained data irregularities.
How a treatment gets debunked — and why the sequence matters
Hydroxychloroquine and ivermectin followed an identical path, and recognising it is more durable than memorising either verdict:
- In vitro activity. Both drugs inhibit SARS-CoV-2 in cell culture. This is a weak signal: the concentrations required frequently exceed anything achievable in human tissue at a safe dose, which is exactly what happened with hydroxychloroquine in lung tissue.
- Observational reports. Small case series and retrospective comparisons appear, often without randomisation. Because clinicians choose who gets treated, these studies systematically compare different kinds of patient rather than different treatments.
- Public advocacy outruns the evidence. Both drugs became political positions before any adequately powered randomised trial had reported.
- Randomised trials report. RECOVERY, SOLIDARITY and TOGETHER each enrolled thousands and found no benefit, with intervals tight enough to rule out a clinically meaningful one.
Dexamethasone ran the same gauntlet and came out the other side: RECOVERY found it cut 28-day mortality by roughly a third in ventilated patients. The trials were not biased against repurposed cheap generics — RECOVERY tested dexamethasone and hydroxychloroquine in the same protocol, at the same time, and reported both. That is the strongest available answer to the claim that inexpensive off-patent drugs were never given a fair test.
What is not known
- Whether either drug has any effect too small for these trials to detect. A relative risk of 0.90 with an interval reaching 1.16 does not exclude a very small benefit. It does exclude the large one that was claimed, and a treatment whose effect cannot be measured across thousands of patients is not a treatment worth choosing over ones that can.
- What the harm signals in RECOVERY's hydroxychloroquine arm represent. Secondary outcomes in an open-label trial support caution, not a causal claim.
- How much of the excess mortality attributable to delayed or forgone effective care can be quantified. It is widely discussed and poorly measured, and this page will not put a number on it.
For the treatments that do have randomised evidence behind them, with the same level of detail, see our COVID-19 treatment page. Claims specific to the vaccines are addressed on the anti-vaccine claims page.
References
- Pan H, et al. "Repurposed Antiviral Drugs for Covid-19 — Interim WHO SOLIDARITY Trial Results." NEJM, 2021. doi:10.1056/NEJMoa2023184
- RECOVERY Collaborative Group. "Effect of Hydroxychloroquine in Hospitalized Patients with Covid-19." NEJM, 2020. doi:10.1056/NEJMoa2022926
- Reis G, et al. "Effect of Early Treatment with Ivermectin among Patients with Covid-19." NEJM, 2022. doi:10.1056/NEJMoa2115869
- Popp M, et al. "Ivermectin for preventing and treating COVID-19." Cochrane Database of Systematic Reviews, 2022. doi:10.1002/14651858.CD015017.pub3
