Acetaminophen increases COVID-19 risk: real-time meta-analysis of 28 studies

@CovidAnalysis, October 2026, Version 23V23 (V1 2022)
Meta-analysis28 studies · 540K patientsc19early.org

Acetaminophen for COVID-19

Serious outcome risk
Acetaminophen
Control
Outcome (studies)
Improvement
Relative risk · 95% CI
All studies (28)−29%
Mortality (14)−24%
Ventilation (3)−59%
ICU admission (3)−57%
Hospitalization (5)−31%
Progression (3)−199%
Recovery (3)−62%
Cases (6)−19%
RCTs (2)−50%
Peer-reviewed (26)−26%
Prophylaxis (14)−21%
Early (6)−23%
Late (9)−51%
after exclusions
00.511.5+
← Acetaminophen
reduces riskIncreased
risk →
c19early.org
October 2026

Media coverage of acetaminophen for COVID‑19

Media covered none of the 25 negative studies

No negative acetaminophen studies were covered

0%
Abstract
Meta-analysis shows 24% [9‑40%] higher mortality, and pooled analysis using the most serious outcome reported shows 29% [17‑41%] higher risk.
Meta-analysis28 studies · 540K patientsc19early.org

Acetaminophen for COVID-19

Serious outcome risk
Acetaminophen
Control
Outcome (studies)
Improvement
Relative risk · 95% CI
All studies (28)−29%
Mortality (14)−24%
Ventilation (3)−59%
ICU admission (3)−57%
Hospitalization (5)−31%
Progression (3)−199%
Recovery (3)−62%
Cases (6)−19%
RCTs (2)−50%
Peer-reviewed (26)−26%
Prophylaxis (14)−21%
Early (6)−23%
Late (9)−51%
after exclusions
00.511.5+
← Acetaminophen
reduces riskIncreased
risk →
Concerns have been raised over the use of acetaminophen (paracetamol) for COVID-191,2. Studies show significantly increased risk.
Potential mechanisms of harm include glutathione depletion, fever suppression, liver toxicity, immunosuppression, cytokine disruption, prostaglandin inhibition, COX inhibition, cell/tissue injury, mitochondrial dysfunction, glycine depletion, disruption of redox balance, increased oxidative stress, mineral depletion, microbiome alteration, and endocannabinoid system dysfunction.
All data and sources to reproduce this analysis are in the appendix. Fever has risks that can outweigh the immunological benefits in certain circumstances.
Download PDF
Evolution of COVID-19 clinical evidence Meta-analysis results over time Acetaminophen p=0.00000021 Vitamin D p<0.0000000001 2020 2021 2022 2023 2024 2025 2026 Lowerrisk Higherrisk c19early.org October 2026 50% 0% -50%

Acetaminophen for COVID-19

Highlights
  • 2nd treatment shown harmful in November 2020, now with p = 0.00000021 from 28 studies, but still recommended in 103 countries.
  • Real-time updates and corrections with a consistent protocol for 227 treatments. Outcome specific analysis and combined evidence from all studies including treatment delay, a primary confounding factor.
A
c19early.org
October 2026

Acetaminophen COVID-19 studies

StudyImprovementRR · 95% CIOutcomeTreatmentControlRelative Risk
Rinott−473%5.73 · 0.30–109death3/850/49OT1 Lapi (ES)−15%1.15 · 0.92–1.43death/hosp.n/an/a Sharif−77%1.77 · 0.39–8.09death9/3612/142 Chen−32%1.32 · 0.98–1.78PASC98/23239/122LONG COVID Rahman−23%1.23 · 0.96–1.56hosp.84/244100/356 Vasilkova−60%1.60 · 0.60–4.29ICU8/285/28OT1
Early treatment−23%1.23 · 1.07–1.41202/950146/69723% higher risk
Tau² = 0.00, I² = 0.0%, p = 0.0041
Ravichandran (PSM)−2700%28.00 · 3.91–200oxygen28/721/72OT1 Manjani−220%3.20 · 1.51–6.82death64/3887/136 Lerner−27%1.27 · 0.96–1.68death5,783 (all patients) Ravichandran (RCT)−43%1.43 · 1.14–1.78no recov.77/10752/103OT1 Lapi−75%1.75 · 1.40–2.18death/hosp.n/an/a Abolhassani−56%1.56 · 0.58–4.18death3/68/25 Baldia (ICU)12%0.88 · 0.72–1.07death1,166 (n)1,480 (n)ICU patients Stufano−19%1.19 · 0.70–2.02PASC11/2323/57 Sobhy (DB RCT)−110%2.10 · 1.05–4.20ICU21/9010/90OT1
Late treatment−51%1.51 · 1.19–1.90204/1,852101/1,96351% higher risk
Tau² = 0.07, I² = 69.8%, p = 0.00061
Blanc−51%1.51 · 0.82–2.84cases60 (n)119 (n) Kolin−23%1.23 · 1.05–1.43cases397,064 (all patients) Park (PSM)25%0.75 · 0.35–1.59death12/39716/397OT1 Gálvez-Barrón−47%1.47 · 0.66–3.33death43 (n)60 (n) Reese (PSM)−61%1.61 · 1.40–1.84death20,826 (n)20,826 (n) Chandan (PSM)−18%1.18 · 0.83–1.64death71/8,59579/8,595OT1 CT2 Oh2%0.98 · 0.38–2.49death58 (n)7,655 (n) Leal7%0.93 · 0.91–0.96casesn/an/a Moreno-Martos−29%1.29 · 1.27–1.32hosp. MacFadden−48%1.48 · 1.44–1.51casesn/an/a Campbell (PSW)−1%1.01 · 0.99–1.02death2,074 (n)20,311 (n) Xie−5%1.05 · 0.70–1.56hosp.population-based cohortOT1 Kim (PSM)−71%1.71 · 0.69–4.24death12/1627/162OT1 Ritsinger−21%1.21 · 1.17–1.25death24,641 (n)20,225 (n)
Prophylaxis−21%1.21 · 1.08–1.3795/56,856102/78,35021% higher risk
Tau² = 0.03, I² = 98.0%, p = 0.0013
All studies−29%1.29 · 1.17–1.41501/59,658349/81,01029% higher risk
Tau² = 0.03, I² = 96.1%, p < 0.000100.511.52+
1 OT: comparison with other treatment
2 CT: study uses combined treatment
Rotate screen for more detailsIncrease width for more details
← Acetaminophen
reduces riskAcetaminophen
increases risk →
B
Loading..
Fig. 1. A. Random-effects meta-analysis. This plot shows pooled effects, see the specific outcome analyses for individual outcomes. Analysis validating pooled outcomes for COVID-19 can be found below. Effect extraction is pre-specified, using the most serious outcome reported. For details see the appendix. B. Timeline of results in acetaminophen studies. The marked dates indicate the time when a harmful effect was identified with statistical significance from ≥3 studies for pooled outcomes and one or more specific outcome. Harm based on specific outcomes was delayed by 5.4 months, compared to using pooled outcomes.
Introduction
Concerns have been raised over the use of acetaminophen (paracetamol) for COVID-191,2. Studies show significantly increased risk. Potential mechanisms of harm include glutathione depletion, fever suppression, liver toxicity, immunosuppression, cytokine disruption, prostaglandin inhibition, COX inhibition, cell/tissue injury, mitochondrial dysfunction, glycine depletion, disruption of redox balance, increased oxidative stress, mineral depletion, microbiome alteration, and endocannabinoid system dysfunction.
Other infections
Increased risk with acetaminophen has been shown for rhinovirus3.
Analysis
We analyze all significant controlled studies of acetaminophen for COVID-19. Search methods, inclusion criteria, effect extraction criteria (more serious outcomes have priority), all individual study data, PRISMA answers, and statistical methods are detailed in Appendix 1. We present random-effects meta-analysis results for all studies, studies within each treatment stage, individual outcomes, peer-reviewed studies, Randomized Controlled Trials (RCTs), and higher quality studies.
Treatment timing
Fig. 2 shows stages of possible treatment for COVID-19. Prophylaxis refers to regularly taking medication before becoming sick, in order to prevent or minimize infection. Early treatment refers to treatment immediately or soon after symptoms appear, while late treatment refers to more delayed treatment.
Prophylaxis pre-exposure post-exposure Early treatment Late treatment Treatment delay viral load disease severity exposed symptom onset hospitalized ICU Expected benefit of antiviral therapy −2 0 2 4 6 8 10 Time relative to symptom onset (days) trajectories vary widely immediately or soon after symptoms after disease progression
Fig. 2. Treatment stages.
Mechanisms of Action
Table 1 shows potential mechanisms by which acetaminophen treatment for COVID-19 could be harmful.

Acetaminophen: potential mechanisms of harm for COVID-19

Glutathione depletion
Acetaminophen metabolism relies on glutathione, an antioxidant that helps protect cells from damage. Higher or chronic doses can deplete glutathione levels, which may lead to impaired immune cell function.
Fever suppression
Fever is a natural defense mechanism that helps the body fight off infection. By reducing fever, acetaminophen could potentially prolong infections.
Liver toxicity
Overdose may damage the liver, which plays an important role in immune function and inflammation.
Immunosuppression
Acetaminophen may directly suppress immune cells such as lymphocytes and macrophages, reducing immune defenses.
Cytokine disruption
Acetaminophen exposure has been found to alter cytokine production, such as reducing IL-6 levels. Cytokines regulate immunity, so this could impair immune responses.
Prostaglandin inhibition
Acetaminophen inhibits prostaglandins, signaling molecules that play a role in inflammation and immunity. Reduced prostaglandins may alter immune regulation and make it more difficult for the body to fight off infection.
COX inhibition
Acetaminophen weakly inhibits COX-1/COX-2 enzymes. These generate immune-modulating prostaglandins, so inhibition could alter immunity.
Cell/tissue injury
Acetaminophen is known to cause oxidative injury to cells, even at normal doses. This low-grade damage could potentially trigger inflammatory and immune responses.
Mitochondrial dysfunction
High doses of acetaminophen may impair mitochondrial energy production. Mitochondria play important roles in immune cell activation and function.
Glycine depletion
Conjugating acetaminophen requires glycine, an amino acid involved in important biological processes, including immune function. Depletion of glycine could reduce antioxidant production and have immunomodulatory effects.
Disruption of redox balance
Acetaminophen may disrupt redox balance—the balance between antioxidants and free radicals (unstable molecules that can damage cells). Imbalance may increase inflammation and impair immune cell function.
Increased oxidative stress
Acetaminophen can increase reactive oxygen species and oxidative stress, which may damage cells and trigger inflammation.
Mineral depletion
Acetaminophen increases urinary excretion of minerals involved in immunity like zinc, selenium, and manganese.
Microbiome alteration
Acetaminophen exposure might alter the intestinal microbiota, which plays an important role in immune function.
Endocannabinoid system dysfunction
Acetaminophen may disrupt endocannabinoid signaling, which helps regulate immune function. This could lead to improper immune responses.
Table 1. Mechanisms of action for potential harmful effects with acetaminophen treatment.
Benefits of Fever
Fever is an important component of the acute response to coronavirus infection4. The evolutionary conservation of fever for over 600 million years supports a survival benefit5. Viral particle sensing occurs via pattern recognition receptors, such as toll-like receptors, triggering release of endogenous pyrogens such as interleukin-1. These cytokines induce thermoregulatory centers in the hypothalamus to elevate core temperature setpoints above normal homeostasis. The resulting fever enhances multiple aspects of the innate and adaptive immune systems5, and creates a suboptimal internal environment that impairs SARS-CoV-2 enzyme function and replication. In vitro studies demonstrate reduced viral output at sustained febrile temperatures of 38-39°C compared to basal 37°C conditions. Fever also correlates clinically with heightened interferon-γ, interleukin-6, lymphocyte activation, and antibody production critical for viral clearance.
Los et al. showed that higher temperature enhanced the expression of antiviral genes and reduced SARS-CoV-2 replication in Calu-3 and Caco-2 cells. An in vivo hamster model showed that higher body temperature at the time of infection correlated with lower viral loads.
Zhou et al. showed that SARS-CoV-2 patients with higher fever had lower viral load. Molecular dynamics simulations, surface plasmon resonance experiments, and pseudovirus cell entry assays showed decreased SARS-CoV-2 binding affinity to the human ACE2 receptor at higher temperature (40°C vs. 37°C).
Downing et al. induced hyperthermia (fever-like temperatures) in human volunteers by immersing them in warm water baths. They found that lymphocytes isolated from individuals with core body temperatures elevated to 39°C produced up to 10 times more interferon-γ, as shown in Fig. 3. They also found an increase in suppressor/cytotoxic T cells and natural killer cells. The threshold of 39°C suggests relevance to fever, and the results suggest fever may play a role in boosting antiviral and immunoregulatory activities.
Core Body Temperature (°C)3737.53838.539Interferon-γ (units/ml)4,0003,0002,0001,000From Downing et al., J. Interferon Research, 8:143-150, 1988
Fig. 3. A 10 fold increase in interferon-γ production was seen when core body temperature reached 39°C, from Downing et al.
Herder et al. perform in vitro analysis with a 3D respiratory epithelial model using cells from human donors. Authors showed that elevated temperature (39-40°C) restricts SARS-CoV-2 infection and replication independently of interferon-mediated antiviral defenses. Authors found SARS-CoV-2 can still enter respiratory cells at 40°C but viral transcription and replication are inhibited, limiting the production of infectious virus. This temperature-dependent restriction correlates with altered host gene expression related to antiviral immunity and epigenetic regulation. The results suggest that febrile temperature ranges may confer protection to respiratory tissues by restricting SARS-CoV-2 propagation.
Dominguez-Nicolas et al. induced localized hyperthermia using LF-ThMS applied to the dorsal thorax (up to 44°C externally), resulting in significantly increased peripheral oxygen saturation (SpO2) levels in COVID-19 patients, as shown in Fig. 4.
Fig. 4. Rapidly increasing SpO2 in COVID-19 patients with localized thoracic hyperthermia, from Dominguez-Nicolas et al.
Ramirez et al. compared COVID-19 mortality in Finland and Estonia, where sauna use is part of the culture and is typically practiced at least once a week, with the rest of Europe. Authors found significantly lower mortality with sauna culture, and suggest this may be due to the beneficial effects of hydrothermotherapy.
Ruble et al. compared army hospital vs. sanitarium treatment for the 1918 Spanish influenza, showing lower progression to pneumonia and lower mortality with sanitarium treatment, which involves hydrothermotherapy, sunlight, and fresh air.
Stewart reports on the use of diathermy in the treatment of pneumonia in 1926, with case reports from several physicians covering over 300 patients. Author reports that diathermy had consistent positive effects without significant adverse events, resulted in about half the mortality of the control group, significantly alleviated symptoms such as dyspnea, pain, and cardiac strain, and improved sleep and reduced respiratory rates.
Recent atom-level work strengthens the mechanistic case for fever-mediated viral attenuation. Xie et al. performed 200-ns equilibration followed by replicate 100-ns all-atom MD simulations of the spike RBD-ACE2 peptidase complex across physiologic-to-febrile temperatures. At 315 K the interface lost ~1 hydrogen bond, solvent exposure grew by ~4 Ų, dissociation probability tripled, and MM-PBSA binding free energy became ≈59 kcal mol-¹ less favorable, driven by heat-induced straightening of the ACE2 α1-helix and withdrawal of the β3β4 hairpin that jointly destabilise the two anchor regions. Mild-cool conditions (305 K) had the opposite effect, α1-helix curvature tightened the interface, dissociation dropped eight-fold, and binding free energy became ~21 kcal mol-¹ more favorable. These thermodynamic shifts directly support febrile-range hyperthermia as a barrier to initial viral attachment.
In summary, fever is a key component of the response to infection. Fever enhances immune cell performance, induces cellular stress on pathogens, and may act synergistically with other stressors like iron deprivation. While results show beneficial effects of fever, it is not universally beneficial. Extreme or prolonged cases may be harmful. Fever may be more detrimental for individuals with lower tolerance for the increased metabolic demands.
Fever may also reduce transmissibility. Fever helps clear infection faster by enhancing immune responses and applying cellular stress to pathogens. Faster clearance gives the pathogen less time to amplify within the host to reach contagious levels. Fever may also apply evolutionary pressure resulting in sacrificing replicative fitness at normal temperatures, minimizing infection in other hosts. Further, fever promotes reduced activity, minimizing the opportunity for transmission.
The beneficial effects of fever suggest potential harm from fever-reducing medications in terms of an increased risk of poor outcomes and increased transmission. However, these may be offset by other effects of specific medications, including anticoagulant, anti-inflammatory, or antiviral effects.
Risks of Fever
Fever has risks that can outweigh the immunological benefits in certain circumstances. Elevated body temperature increases metabolic demand, oxygen consumption, heart rate, respiratory rate, and fluid loss, which may be dangerous in people with limited physiological reserve, including some patients with significant cardiac or pulmonary disease, infants, frail elderly adults, and pregnant patients. In pregnancy, high maternal temperature or hyperthermia in very early gestation has been associated with neural tube defects and other fetal risks, although separating the effect of fever itself from the underlying infection, inflammatory illness, or treatments is challenging.
Very high body temperatures can cause direct cellular, neurological, and multiorgan injury, although this is more characteristic of hyperthermia or hyperpyrexia than ordinary regulated infectious fever. Infectious fever is typically physiologically regulated and rarely rises above about 41°C, whereas heatstroke and other hyperthermic states can exceed that range and are more directly dangerous. Fever can also worsen dehydration, tachycardia, delirium, and physiologic stress. Febrile seizures are a particular concern in young children.
Results
Table 2 summarizes the results for all stages combined, for Randomized Controlled Trials, for peer-reviewed studies, after exclusions, and for specific outcomes. Table 3 shows results by treatment stage. Fig. 5 plots individual results by treatment stage. Fig. 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, and 17 show forest plots for random-effects meta-analysis of all studies with pooled effects, mortality results, ventilation, ICU admission, hospitalization, progression, recovery, cases, viral clearance, peer reviewed studies, non-symptomatic vs. symptomatic results, and long COVID.
Table 2. Random-effects meta-analysis for all stages combined, for Randomized Controlled Trials, for peer-reviewed studies, after exclusions, and for specific outcomes. Results show the relative risk with treatment and the 95% confidence interval. * p<0.05  ** p<0.01  *** p<0.001  **** p<0.0001.
Relative Risk Studies Patients
All studies1.29 [1.17‑1.41]****28540K
After exclusions1.26 [1.14‑1.39]****24540K
Peer-reviewedPeer-reviewed1.26 [1.15‑1.39]****26500K
RCTsRCTs1.50 [1.16‑1.93]**2390
Mortality1.24 [1.09‑1.40]**14140K
VentilationVent.1.59 [0.45‑5.54]31,642
ICU admissionICU1.57 [0.86‑2.87]3560
HospitalizationHosp.1.31 [1.18‑1.45]****51,304
Recovery1.62 [1.40‑1.87]****3534
Cases1.19 [0.91‑1.56]6410K
Table 3. Random-effects meta-analysis results by treatment stage. Results show the relative risk with treatment and the 95% confidence interval.treatment and the 95% confidence interval. * p<0.05  ** p<0.01  *** p<0.001  **** p<0.0001.
Early treatment Late treatment Prophylaxis
All studies1.23 [1.07‑1.41]**1.23**
[1.07‑1.41]		1.51 [1.19‑1.90]***1.51***
[1.19‑1.90]		1.21 [1.08‑1.37]**1.21**
[1.08‑1.37]
After exclusions1.22 [1.02‑1.45]*1.22*
[1.02‑1.45]		1.47 [1.12‑1.93]**1.47**
[1.12‑1.93]		1.21 [1.08‑1.37]**1.21**
[1.08‑1.37]
Peer-reviewedPeer-reviewed1.23 [1.07‑1.41]**1.23**
[1.07‑1.41]		1.51 [1.19‑1.90]***1.51***
[1.19‑1.90]		1.17 [1.03‑1.32]*1.17*
[1.03‑1.32]
RCTsRCTs1.50 [1.16‑1.93]**1.50**
[1.16‑1.93]
Mortality2.27 [0.59‑8.75]2.27
[0.59‑8.75]		1.36 [0.91‑2.04]1.36
[0.91‑2.04]		1.21 [1.05‑1.41]*1.21*
[1.05‑1.41]
VentilationVent.5.34 [1.98‑14.45]***5.34***
[1.98‑14.45]		0.87 [0.42‑1.83]0.87
[0.42‑1.83]
ICU admissionICU1.60 [0.60‑4.29]1.60
[0.60‑4.29]		2.10 [1.05‑4.20]*2.10*
[1.05‑4.20]		0.60 [0.15‑2.47]0.60
[0.15‑2.47]
HospitalizationHosp.1.23 [0.96‑1.56]1.23
[0.96‑1.56]		1.59 [1.09‑2.30]*1.59*
[1.09‑2.30]		1.29 [1.20‑1.37]****1.29****
[1.20‑1.37]
Recovery1.62 [1.40‑1.87]****1.62****
[1.40‑1.87]
Cases1.19 [0.91‑1.56]1.19
[0.91‑1.56]
Loading..
Fig. 5. Scatter plot showing the most serious outcome in all studies, and for studies within each stage. Diamonds shows the results of random-effects meta-analysis.
c19early.org
October 2026

Acetaminophen COVID-19 studies

StudyImprovementRR · 95% CIOutcomeTreatmentControlRelative Risk
Rinott−473%5.73 · 0.30–109death3/850/49OT1 Lapi (ES)−15%1.15 · 0.92–1.43death/hosp.n/an/a Sharif−77%1.77 · 0.39–8.09death9/3612/142 Chen−32%1.32 · 0.98–1.78PASC98/23239/122LONG COVID Rahman−23%1.23 · 0.96–1.56hosp.84/244100/356 Vasilkova−60%1.60 · 0.60–4.29ICU8/285/28OT1
Early treatment−23%1.23 · 1.07–1.41202/950146/69723% higher risk
Tau² = 0.00, I² = 0.0%, p = 0.0041
Ravichandran (PSM)−2700%28.00 · 3.91–200oxygen28/721/72OT1 Manjani−220%3.20 · 1.51–6.82death64/3887/136 Lerner−27%1.27 · 0.96–1.68death5,783 (all patients) Ravichandran (RCT)−43%1.43 · 1.14–1.78no recov.77/10752/103OT1 Lapi−75%1.75 · 1.40–2.18death/hosp.n/an/a Abolhassani−56%1.56 · 0.58–4.18death3/68/25 Baldia (ICU)12%0.88 · 0.72–1.07death1,166 (n)1,480 (n)ICU patients Stufano−19%1.19 · 0.70–2.02PASC11/2323/57 Sobhy (DB RCT)−110%2.10 · 1.05–4.20ICU21/9010/90OT1
Late treatment−51%1.51 · 1.19–1.90204/1,852101/1,96351% higher risk
Tau² = 0.07, I² = 69.8%, p = 0.00061
Blanc−51%1.51 · 0.82–2.84cases60 (n)119 (n) Kolin−23%1.23 · 1.05–1.43cases397,064 (all patients) Park (PSM)25%0.75 · 0.35–1.59death12/39716/397OT1 Gálvez-Barrón−47%1.47 · 0.66–3.33death43 (n)60 (n) Reese (PSM)−61%1.61 · 1.40–1.84death20,826 (n)20,826 (n) Chandan (PSM)−18%1.18 · 0.83–1.64death71/8,59579/8,595OT1 CT2 Oh2%0.98 · 0.38–2.49death58 (n)7,655 (n) Leal7%0.93 · 0.91–0.96casesn/an/a Moreno-Martos−29%1.29 · 1.27–1.32hosp. MacFadden−48%1.48 · 1.44–1.51casesn/an/a Campbell (PSW)−1%1.01 · 0.99–1.02death2,074 (n)20,311 (n) Xie−5%1.05 · 0.70–1.56hosp.population-based cohortOT1 Kim (PSM)−71%1.71 · 0.69–4.24death12/1627/162OT1 Ritsinger−21%1.21 · 1.17–1.25death24,641 (n)20,225 (n)
Prophylaxis−21%1.21 · 1.08–1.3795/56,856102/78,35021% higher risk
Tau² = 0.03, I² = 98.0%, p = 0.0013
All studies−29%1.29 · 1.17–1.41501/59,658349/81,01029% higher risk
Tau² = 0.03, I² = 96.1%, p < 0.000100.511.52+
1 OT: comparison with other treatment
2 CT: study uses combined treatment
Rotate screen for more detailsIncrease width for more details
← Acetaminophen
reduces riskAcetaminophen
increases risk →
Fig. 6. Random-effects meta-analysis for all studies. This plot shows pooled effects, see the specific outcome analyses for individual outcomes. Analysis validating pooled outcomes for COVID-19 can be found below. Effect extraction is pre-specified, using the most serious outcome reported. For details see the appendix.
c19early.org
October 2026

Acetaminophen COVID-19 mortality results

StudyImprovementRR · 95% CITreatmentControlRelative Risk
Rinott−473%5.73 · 0.30–1093/850/49OT1 Sharif−77%1.77 · 0.39–8.099/3612/142
Early treatment−127%2.27 · 0.59–8.7512/4462/191127% higher risk
Tau² = 0.00, I² = 0.0%, p = 0.24
Manjani−220%3.20 · 1.51–6.8264/3887/136 Lerner−27%1.27 · 0.96–1.685,783 (all patients) Abolhassani−56%1.56 · 0.58–4.183/68/25 Baldia (ICU)12%0.88 · 0.72–1.071,166 (n)1,480 (n)ICU patients
Late treatment−36%1.36 · 0.91–2.0467/1,56015/1,64136% higher risk
Tau² = 0.10, I² = 67.8%, p = 0.13
Park (PSM)25%0.75 · 0.35–1.5912/39716/397OT1 Gálvez-Barrón−47%1.47 · 0.66–3.3343 (n)60 (n) Reese (PSM)−61%1.61 · 1.40–1.8420,826 (n)20,826 (n) Chandan (PSM)−18%1.18 · 0.83–1.6471/8,59579/8,595OT1 CT2 Oh2%0.98 · 0.38–2.4958 (n)7,655 (n) Campbell (PSW)−1%1.01 · 0.99–1.022,074 (n)20,311 (n) Kim (PSM)−71%1.71 · 0.69–4.2412/1627/162OT1 Ritsinger−21%1.21 · 1.17–1.2524,641 (n)20,225 (n)
Prophylaxis−21%1.21 · 1.05–1.4195/56,796102/78,23121% higher risk
Tau² = 0.02, I² = 93.6%, p = 0.011
All studies−24%1.24 · 1.09–1.40174/58,802119/80,06324% higher risk
Tau² = 0.02, I² = 89.6%, p = 0.001200.511.52+
1 OT: comparison with other treatment
2 CT: study uses combined treatment
Rotate screen for more detailsIncrease width for more details
← Acetaminophen
reduces riskAcetaminophen
increases risk →
Fig. 7. Random-effects meta-analysis for mortality results.
c19early.org
October 2026

Acetaminophen COVID-19 mechanical ventilation results

StudyImprovementRR · 95% CITreatmentControlRelative Risk
Manjani−434%5.34 · 1.98–14.45388 (n)136 (n)
Late treatment−434%5.34 · 1.98–14.45388 (n)136 (n)434% higher risk
Tau² = 0.00, I² = 0.0%, p = 0.001
Park (PSM)38%0.62 · 0.19–1.895/3978/397OT1 Kim (PSM)−14%1.14 · 0.42–3.088/1627/162OT1
Prophylaxis13%0.87 · 0.42–1.8313/55915/55913% lower risk
Tau² = 0.00, I² = 0.0%, p = 0.73
All studies−59%1.59 · 0.45–5.5413/94715/69559% higher risk
Tau² = 0.94, I² = 77.4%, p = 0.4800.511.52+
1 OT: comparison with other treatment
Rotate screen for more detailsIncrease width for more details
← Acetaminophen
reduces riskAcetaminophen
increases risk →
Fig. 8. Random-effects meta-analysis for ventilation.
c19early.org
October 2026

Acetaminophen COVID-19 ICU results

StudyImprovementRR · 95% CITreatmentControlRelative Risk
Vasilkova−60%1.60 · 0.60–4.298/285/28OT1
Early treatment−60%1.60 · 0.60–4.298/285/2860% higher risk
Tau² = 0.00, I² = 0.0%, p = 0.36
Sobhy (DB RCT)−110%2.10 · 1.05–4.2021/9010/90OT1
Late treatment−110%2.10 · 1.05–4.2021/9010/90110% higher risk
Tau² = 0.00, I² = 0.0%, p = 0.036
Kim (PSM)40%0.60 · 0.15–2.473/1625/162OT1
Prophylaxis40%0.60 · 0.15–2.473/1625/16240% lower risk
Tau² = 0.00, I² = 0.0%, p = 0.49
All studies−57%1.57 · 0.86–2.8732/28020/28057% higher risk
Tau² = 0.05, I² = 17.7%, p = 0.1400.511.52+
1 OT: comparison with other treatment
Rotate screen for more detailsIncrease width for more details
← Acetaminophen
reduces riskAcetaminophen
increases risk →
Fig. 9. Random-effects meta-analysis for ICU admission.
c19early.org
October 2026

Acetaminophen COVID-19 hospitalization results

StudyImprovementRR · 95% CIOutcomeTreatmentControlRelative Risk
Rahman−23%1.23 · 0.96–1.56hosp.84/244100/356
Early treatment−23%1.23 · 0.96–1.5684/244100/35623% higher risk
Tau² = 0.00, I² = 0.0%, p = 0.096
Manjani−100%2.00 · 1.33–3.02hosp. time388 (n)136 (n) Sobhy (DB RCT)−36%1.36 · 1.07–1.73hosp. time90 (n)90 (n)OT1
Late treatment−59%1.59 · 1.09–2.30478 (n)226 (n)59% higher risk
Tau² = 0.05, I² = 60.7%, p = 0.015
Moreno-Martos−29%1.29 · 1.27–1.32hosp. Xie−5%1.05 · 0.70–1.56hosp.population-based cohortOT1
Prophylaxis−29%1.29 · 1.20–1.3729% higher risk
Tau² = 0.00, I² = 4.8%, p < 0.0001
All studies−31%1.31 · 1.18–1.4584/722100/58231% higher risk
Tau² = 0.00, I² = 30.1%, p < 0.000100.511.52+
1 OT: comparison with other treatment
Rotate screen for more detailsIncrease width for more details
← Acetaminophen
reduces riskAcetaminophen
increases risk →
Fig. 10. Random-effects meta-analysis for hospitalization.
c19early.org
October 2026

Acetaminophen COVID-19 progression results

StudyImprovementRR · 95% CITreatmentControlRelative Risk
Vasilkova−19%1.19 · 0.90–1.5628 (n)28 (n)OT1
Early treatment−19%1.19 · 0.90–1.5628 (n)28 (n)19% higher risk
Tau² = 0.00, I² = 0.0%, p = 0.22
Manjani−244%3.44 · 1.49–8.54132 (n)136 (n) Ravichandran (RCT)−3925%40.25 · 2.47–65720/1070/103OT1
Late treatment−710%8.10 · 0.81–80.5420/2390/239710% higher risk
Tau² = 1.91, I² = 63.3%, p = 0.074
All studies−199%2.99 · 0.84–10.6120/2670/267199% higher risk
Tau² = 0.88, I² = 81.9%, p = 0.09100.511.52+
1 OT: comparison with other treatment
Rotate screen for more detailsIncrease width for more details
← Acetaminophen
reduces riskAcetaminophen
increases risk →
Fig. 11. Random-effects meta-analysis for progression.
c19early.org
October 2026

Acetaminophen COVID-19 recovery results

StudyImprovementRR · 95% CIOutcomeTreatmentControlRelative Risk
Ravichandran−75%1.75 · 1.48–2.06recov. time72 (n)72 (n)OT1 Ravichandran (RCT)−43%1.43 · 1.14–1.78no recov.77/10752/103OT1 Sobhy (DB RCT)−33%1.33 · 0.31–5.79no recov.4/903/90OT1
Late treatment−62%1.62 · 1.40–1.8781/26955/26562% higher risk
Tau² = 0.00, I² = 6.8%, p < 0.0001
All studies−62%1.62 · 1.40–1.8781/26955/26562% higher risk
Tau² = 0.00, I² = 6.8%, p < 0.000100.511.52+
1 OT: comparison with other treatment
Rotate screen for more detailsIncrease width for more details
← Acetaminophen
reduces riskAcetaminophen
increases risk →
Fig. 12. Random-effects meta-analysis for recovery.
c19early.org
October 2026

Acetaminophen COVID-19 case results

StudyImprovementRR · 95% CIOutcomeTreatmentControlRelative Risk
Blanc−51%1.51 · 0.82–2.84cases60 (n)119 (n) Kolin−23%1.23 · 1.05–1.43cases397,064 (all patients) Chandan (PSM)−27%1.27 · 0.90–1.75cases8,595 (n)8,595 (n)OT1 CT2 Leal7%0.93 · 0.91–0.96casesn/an/a MacFadden−48%1.48 · 1.44–1.51casesn/an/a Xie3%0.97 · 0.72–1.29casespopulation-based cohortOT1
Prophylaxis−19%1.19 · 0.91–1.568,655 (n)8,714 (n)19% higher risk
Tau² = 0.09, I² = 98.2%, p = 0.21
All studies−19%1.19 · 0.91–1.568,655 (n)8,714 (n)19% higher risk
Tau² = 0.09, I² = 98.2%, p = 0.2100.511.52+
1 OT: comparison with other treatment
2 CT: study uses combined treatment
Rotate screen for more detailsIncrease width for more details
← Acetaminophen
reduces riskAcetaminophen
increases risk →
Fig. 13. Random-effects meta-analysis for cases.
c19early.org
October 2026

Acetaminophen COVID-19 viral clearance results

StudyImprovementRR · 95% CIOutcomeTreatmentControlRelative Risk
Ravichandran (RCT)−20%1.20 · 0.93–1.56viral+43/6037/62OT1
Late treatment−20%1.20 · 0.93–1.5643/6037/6220% higher risk
Tau² = 0.00, I² = 0.0%, p = 0.17
All studies−20%1.20 · 0.93–1.5643/6037/6220% higher risk
Tau² = 0.00, I² = 0.0%, p = 0.1700.511.52+
1 OT: comparison with other treatment
Rotate screen for more detailsIncrease width for more details
← Acetaminophen
reduces riskAcetaminophen
increases risk →
Fig. 14. Random-effects meta-analysis for viral clearance.
c19early.org
October 2026

Acetaminophen COVID-19 peer reviewed studies

StudyImprovementRR · 95% CIOutcomeTreatmentControlRelative Risk
Rinott−473%5.73 · 0.30–109death3/850/49OT1 Lapi (ES)−15%1.15 · 0.92–1.43death/hosp.n/an/a Sharif−77%1.77 · 0.39–8.09death9/3612/142 Chen−32%1.32 · 0.98–1.78PASC98/23239/122LONG COVID Rahman−23%1.23 · 0.96–1.56hosp.84/244100/356 Vasilkova−60%1.60 · 0.60–4.29ICU8/285/28OT1
Early treatment−23%1.23 · 1.07–1.41202/950146/69723% higher risk
Tau² = 0.00, I² = 0.0%, p = 0.0041
Ravichandran (PSM)−2700%28.00 · 3.91–200oxygen28/721/72OT1 Manjani−220%3.20 · 1.51–6.82death64/3887/136 Lerner−27%1.27 · 0.96–1.68death5,783 (all patients) Ravichandran (RCT)−43%1.43 · 1.14–1.78no recov.77/10752/103OT1 Lapi−75%1.75 · 1.40–2.18death/hosp.n/an/a Abolhassani−56%1.56 · 0.58–4.18death3/68/25 Baldia (ICU)12%0.88 · 0.72–1.07death1,166 (n)1,480 (n)ICU patients Stufano−19%1.19 · 0.70–2.02PASC11/2323/57 Sobhy (DB RCT)−110%2.10 · 1.05–4.20ICU21/9010/90OT1
Late treatment−51%1.51 · 1.19–1.90204/1,852101/1,96351% higher risk
Tau² = 0.07, I² = 69.8%, p = 0.00061
Kolin−23%1.23 · 1.05–1.43cases397,064 (all patients) Park (PSM)25%0.75 · 0.35–1.59death12/39716/397OT1 Gálvez-Barrón−47%1.47 · 0.66–3.33death43 (n)60 (n) Chandan (PSM)−18%1.18 · 0.83–1.64death71/8,59579/8,595OT1 CT2 Oh2%0.98 · 0.38–2.49death58 (n)7,655 (n) Leal7%0.93 · 0.91–0.96casesn/an/a Moreno-Martos−29%1.29 · 1.27–1.32hosp. MacFadden−48%1.48 · 1.44–1.51casesn/an/a Campbell (PSW)−1%1.01 · 0.99–1.02death2,074 (n)20,311 (n) Xie−5%1.05 · 0.70–1.56hosp.population-based cohortOT1 Kim (PSM)−71%1.71 · 0.69–4.24death12/1627/162OT1 Ritsinger−21%1.21 · 1.17–1.25death24,641 (n)20,225 (n)
Prophylaxis−17%1.17 · 1.03–1.3295/35,970102/57,40517% higher risk
Tau² = 0.03, I² = 98.3%, p = 0.017
All studies−26%1.26 · 1.15–1.39501/38,772349/60,06526% higher risk
Tau² = 0.03, I² = 96.3%, p < 0.000100.511.52+
1 OT: comparison with other treatment
2 CT: study uses combined treatment
Rotate screen for more detailsIncrease width for more details
← Acetaminophen
reduces riskAcetaminophen
increases risk →
Fig. 15. Random-effects meta-analysis for peer reviewed studies. Zeraatkar et al. analyze 356 COVID-19 trials, finding no significant evidence that preprint results are inconsistent with peer-reviewed studies. They also show extremely long peer-review delays, with a median of 6 months to journal publication. A six month delay was equivalent to around 1.5 million deaths during the first two years of the pandemic. Authors recommend using preprint evidence, with appropriate checks for potential falsified data, which provides higher certainty much earlier. Davidson et al. also showed no important difference between meta-analysis results of preprints and peer-reviewed publications for COVID-19, based on 37 meta-analyses including 114 trials. Effect extraction is pre-specified, using the most serious outcome reported, see the appendix for details. Analysis validating pooled outcomes for COVID-19 can be found below.
c19early.org
October 2026

Acetaminophen COVID-19 symptomatic vs. case outcomes

StudyImprovementRR · 95% CIOutcomeTreatmentControlRelative Risk
Rinott−473%5.73 · 0.30–109death3/850/49OT1 Park (PSM)25%0.75 · 0.35–1.59death12/39716/397OT1 Gálvez-Barrón−47%1.47 · 0.66–3.33death43 (n)60 (n) Reese (PSM)−61%1.61 · 1.40–1.84death20,826 (n)20,826 (n) Chandan (PSM)−18%1.18 · 0.83–1.64death71/8,59579/8,595OT1 CT2 Oh2%0.98 · 0.38–2.49death58 (n)7,655 (n) Ravichandran (PSM)−2700%28.00 · 3.91–200oxygen28/721/72OT1 Manjani−220%3.20 · 1.51–6.82death64/3887/136 Moreno-Martos−29%1.29 · 1.27–1.32hosp. Lerner−27%1.27 · 0.96–1.68death5,783 (all patients) Ravichandran (RCT)−43%1.43 · 1.14–1.78no recov.77/10752/103OT1 Campbell (PSW)−1%1.01 · 0.99–1.02death2,074 (n)20,311 (n) Xie−5%1.05 · 0.70–1.56hosp.population-based cohortOT1 Lapi (ES)−15%1.15 · 0.92–1.43death/hosp.n/an/a Lapi−75%1.75 · 1.40–2.18death/hosp.n/an/a Abolhassani−56%1.56 · 0.58–4.18death3/68/25 Sharif−77%1.77 · 0.39–8.09death9/3612/142 Baldia (ICU)12%0.88 · 0.72–1.07death1,166 (n)1,480 (n)ICU patients Kim (PSM)−71%1.71 · 0.69–4.24death12/1627/162OT1 Stufano−19%1.19 · 0.70–2.02PASC11/2323/57 Sobhy (DB RCT)−110%2.10 · 1.05–4.20ICU21/9010/90OT1 Ritsinger−21%1.21 · 1.17–1.25death24,641 (n)20,225 (n) Chen−32%1.32 · 0.98–1.78PASC98/23239/122LONG COVID Rahman−23%1.23 · 0.96–1.56hosp.84/244100/356 Vasilkova−60%1.60 · 0.60–4.29ICU8/285/28OT1
Symptomatic−29%1.29 · 1.18–1.41501/59,598349/80,89129% higher risk
Tau² = 0.02, I² = 91.9%, p < 0.0001
Blanc−51%1.51 · 0.82–2.84cases60 (n)119 (n) Kolin−23%1.23 · 1.05–1.43cases397,064 (all patients) Chandan (PSM)−27%1.27 · 0.90–1.75cases8,595 (n)8,595 (n)OT1 CT2 Leal7%0.93 · 0.91–0.96casesn/an/a MacFadden−48%1.48 · 1.44–1.51casesn/an/a Ravichandran (RCT)−20%1.20 · 0.93–1.56viral+43/6037/62OT1 Xie3%0.97 · 0.72–1.29casespopulation-based cohortOT1
Cases−19%1.19 · 0.93–1.5243/8,71537/8,77619% higher risk
Tau² = 0.09, I² = 97.9%, p = 0.16
All studies−27%1.27 · 1.16–1.38544/68,313386/89,66727% higher risk
Tau² = 0.03, I² = 95.6%, p < 0.000100.511.52+
1 OT: comparison with other treatment
2 CT: study uses combined treatment
Rotate screen for more detailsIncrease width for more details
← Acetaminophen
reduces riskAcetaminophen
increases risk →
Fig. 16. Random-effects meta-analysis for non-symptomatic vs. symptomatic results. Effect extraction is pre-specified, using the most serious outcome reported, see the appendix for details. Analysis validating pooled outcomes for COVID-19 can be found below.
c19early.org
October 2026

Acetaminophen COVID-19 long COVID results

StudyImprovementRR · 95% CIOutcomeTreatmentControlRelative Risk
Chen−32%1.32 · 0.98–1.78PASC98/23239/122
Early treatment−32%1.32 · 0.98–1.7898/23239/12232% higher risk
Tau² = 0.00, I² = 0.0%, p = 0.068
Stufano−19%1.19 · 0.70–2.02PASC11/2323/57
Late treatment−19%1.19 · 0.70–2.0211/2323/5719% higher risk
Tau² = 0.00, I² = 0.0%, p = 0.54
All studies−29%1.29 · 0.99–1.67109/25562/17929% higher risk
Tau² = 0.00, I² = 0.0%, p = 0.05700.511.52+
Rotate screen for more detailsIncrease width for more details
← Acetaminophen
reduces riskAcetaminophen
increases risk →
Fig. 17. Random-effects meta-analysis for long COVID. Effect extraction is pre-specified, using the most serious outcome reported, see the appendix for details. Analysis validating pooled outcomes for COVID-19 can be found below.
Randomized Controlled Trials (RCTs)
Fig. 18 shows a comparison of results for RCTs and observational studies. Fig. 19 shows a forest plot for random-effects meta-analysis of all Randomized Controlled Trials. RCT results are included in Table 2 and Table 3.
Loading..
Fig. 18. Results for RCTs and observational studies.
RCTs have many potential biases
RCTs help to make study groups more similar and can provide a higher level of evidence, however they are subject to many biases17, and analysis of double-blind RCTs has identified extreme levels of bias18. For COVID-19, the overhead may delay treatment, dramatically compromising efficacy; they may encourage monotherapy for simplicity at the cost of efficacy which may rely on combined or synergistic effects; the participants that sign up may not reflect real world usage or the population that benefits most in terms of age, comorbidities, severity of illness, or other factors; standard of care may be compromised and unable to evolve quickly based on emerging research for new diseases; errors may be made in randomization and medication delivery; and investigators may have hidden agendas or vested interests influencing design, operation, analysis, reporting, and the potential for fraud. All of these biases have been observed with COVID-19 RCTs. There is no guarantee that a specific RCT provides a higher level of evidence.
Conflicts of interest for COVID-19 RCTs
RCTs are expensive and many RCTs are funded by pharmaceutical companies or other organizations with conflicts of interest, for example governments that previously denied treatment with the study drug. For COVID-19, this creates an incentive to show efficacy for patented commercial products, and an incentive to show a lack of efficacy for inexpensive treatments. The bias is expected to be significant, for example Als-Nielsen et al. analyzed 370 RCTs from Cochrane reviews, showing that trials funded by for-profit organizations were 5 times more likely to recommend the experimental drug compared with those funded by nonprofit organizations. Bekelman et al. and Lundh et al. show that industry-sponsored studies are more likely to be favorable. For COVID-19, some major philanthropic organizations are largely funded by investments with extreme conflicts of interest for and against specific COVID-19 interventions.
RCTs for novel acute diseases requiring rapid treatment
High quality RCTs for novel acute diseases are more challenging, with increased ethical issues due to the urgency of treatment, increased risk due to enrollment delays, and more difficult design with a rapidly evolving evidence base. For COVID-19, the most common site of initial infection is the upper respiratory tract. Immediate treatment is likely to be most successful and may prevent or slow progression to other parts of the body. For a non-prophylaxis RCT, it makes sense to provide treatment in advance and instruct patients to use it immediately on symptoms, just as some governments have done by providing medication kits in advance. Unfortunately, no RCTs have been done in this way. Every treatment RCT to date involves delayed treatment. Among the 227 treatments we have analyzed, 67% of RCTs involve very late treatment 5+ days after onset. No non-prophylaxis COVID-19 RCTs match the potential real-world use of early treatments. They may more accurately represent results for treatments that require visiting a medical facility, e.g., those requiring intravenous administration.
RCT bias for widely available treatments
RCTs have a bias against finding an effect for interventions that are widely available—patients that believe they need the intervention are more likely to decline participation and take the intervention. RCTs for acetaminophen are more likely to enroll low-risk participants that do not need treatment to recover, making the results less applicable to clinical practice. This bias is likely to be greater for widely known treatments, and may be greater when the risk of a serious outcome is overstated. This bias does not apply to the typical pharmaceutical trial of a new drug that is otherwise unavailable.
RCTs may be less reliable
Concato et al. report a paradoxical finding—RCT results had higher variability, and only RCTs were found to sometimes report significant results the opposite of the overall result. The same trend is seen for the most popular (most politicized) COVID-19 treatments—considering all statistically significant results reported in studies, RCTs are slightly more likely to report a result in the opposite direction. In other words, for these COVID-19 treatments and for the topics covered by Concato et al., assuming causality from a single study is more likely to result in an incorrect conclusion for RCTs.
Increased risk of inconsistent results for RCTs suggests higher prevalence of bias, which may arise due to many issues including design bias, conflicts of interest, treatment differences by physicians aware of allocation, attrition bias, ascertainment bias, randomization failures, errors, or fraud.
Using all studies identifies efficacy 8+ months faster (9+ months for low-cost treatments)
Currently, 59 of the treatments we analyze show statistically significant efficacy or harm, defined as ≥10% decreased risk or >0% increased risk from ≥3 studies. Of these, 54% have been confirmed in RCTs, with a mean delay of 7.8 months (62% with 8.7 months delay for low-cost treatments). The remaining treatments either have no RCTs, or the point estimate is consistent.
All studies must be carefully analyzed
Neither observational studies nor RCTs prove causation—any study can be flawed or fraudulent. We need much more, for example a combination of results from many independent teams, detailed understanding of each study, knowledge of conflicts/team reliability, dose-response relationships, delay-response relationships, logical results across outcomes, or details consistent with preclinical expectations.
All studies must be evaluated individually. RCTs for a given medication and disease may be more reliable, however they may also be less reliable. For off-patent medications, very high conflict of interest trials may be more likely to be RCTs, and more likely to be large trials that dominate meta-analyses.
c19early.org
October 2026

Acetaminophen COVID-19 Randomized Controlled Trials

StudyImprovementRR · 95% CIOutcomeTreatmentControlRelative Risk
Ravichandran (RCT)−43%1.43 · 1.14–1.78no recov.77/10752/103OT1 Sobhy (DB RCT)−110%2.10 · 1.05–4.20ICU21/9010/90OT1
Late treatment−50%1.50 · 1.16–1.9398/19762/19350% higher risk
Tau² = 0.01, I² = 7.7%, p = 0.0018
All studies−50%1.50 · 1.16–1.9398/19762/19350% higher risk
Tau² = 0.01, I² = 7.7%, p = 0.001800.511.52+
1 OT: comparison with other treatment
Rotate screen for more detailsIncrease width for more details
← Acetaminophen
reduces riskAcetaminophen
increases risk →
Fig. 19. Random-effects meta-analysis for all Randomized Controlled Trials. This plot shows pooled effects, see the specific outcome analyses for individual outcomes. Analysis validating pooled outcomes for COVID-19 can be found below. Effect extraction is pre-specified, using the most serious outcome reported. For details see the appendix.
Exclusions
To avoid bias in the selection of studies, we analyze all non-retracted studies. Here we show the results after excluding studies with major issues likely to alter results, non-standard studies, and studies where very minimal detail is currently available. Our bias evaluation is based on analysis of each study and identifying when there is a significant chance that limitations will substantially change the outcome of the study. We believe this can be more valuable than checklist-based approaches such as Cochrane GRADE, which can be easily influenced by potential bias, may ignore or underemphasize serious issues not captured in the checklists, and may overemphasize issues unlikely to alter outcomes in specific cases (for example certain specifics of randomization with a very large effect size and well-matched baseline characteristics).
The studies excluded are as below. Fig. 20 shows a forest plot for random-effects meta-analysis of all studies after exclusions.
Lapi, substantial unadjusted confounding by indication likely.
Rahman, unadjusted results with no group details; significant unadjusted confounding possible.
Rinott, unadjusted differences between groups.
Sharif, unadjusted results with no group details.
c19early.org
October 2026

Acetaminophen COVID-19 studies after exclusions

StudyImprovementRR · 95% CIOutcomeTreatmentControlRelative Risk
Lapi (ES)−15%1.15 · 0.92–1.43death/hosp.n/an/a Chen−32%1.32 · 0.98–1.78PASC98/23239/122LONG COVID Vasilkova−60%1.60 · 0.60–4.29ICU8/285/28OT1
Early treatment−22%1.22 · 1.02–1.45106/26044/15022% higher risk
Tau² = 0.00, I² = 0.0%, p = 0.027
Ravichandran (PSM)−2700%28.00 · 3.91–200oxygen28/721/72OT1 Manjani−220%3.20 · 1.51–6.82death64/3887/136 Lerner−27%1.27 · 0.96–1.68death5,783 (all patients) Ravichandran (RCT)−43%1.43 · 1.14–1.78no recov.77/10752/103OT1 Abolhassani−56%1.56 · 0.58–4.18death3/68/25 Baldia (ICU)12%0.88 · 0.72–1.07death1,166 (n)1,480 (n)ICU patients Stufano−19%1.19 · 0.70–2.02PASC11/2323/57 Sobhy (DB RCT)−110%2.10 · 1.05–4.20ICU21/9010/90OT1
Late treatment−47%1.47 · 1.12–1.93204/1,852101/1,96347% higher risk
Tau² = 0.07, I² = 67.4%, p = 0.005
Blanc−51%1.51 · 0.82–2.84cases60 (n)119 (n) Kolin−23%1.23 · 1.05–1.43cases397,064 (all patients) Park (PSM)25%0.75 · 0.35–1.59death12/39716/397OT1 Gálvez-Barrón−47%1.47 · 0.66–3.33death43 (n)60 (n) Reese (PSM)−61%1.61 · 1.40–1.84death20,826 (n)20,826 (n) Chandan (PSM)−18%1.18 · 0.83–1.64death71/8,59579/8,595OT1 CT2 Oh2%0.98 · 0.38–2.49death58 (n)7,655 (n) Leal7%0.93 · 0.91–0.96casesn/an/a Moreno-Martos−29%1.29 · 1.27–1.32hosp. MacFadden−48%1.48 · 1.44–1.51casesn/an/a Campbell (PSW)−1%1.01 · 0.99–1.02death2,074 (n)20,311 (n) Xie−5%1.05 · 0.70–1.56hosp.population-based cohortOT1 Kim (PSM)−71%1.71 · 0.69–4.24death12/1627/162OT1 Ritsinger−21%1.21 · 1.17–1.25death24,641 (n)20,225 (n)
Prophylaxis−21%1.21 · 1.08–1.3795/56,856102/78,35021% higher risk
Tau² = 0.03, I² = 98.0%, p = 0.0013
All studies−26%1.26 · 1.14–1.39405/58,968247/80,46326% higher risk
Tau² = 0.03, I² = 96.6%, p < 0.000100.511.52+
1 OT: comparison with other treatment
2 CT: study uses combined treatment
Rotate screen for more detailsIncrease width for more details
← Acetaminophen
reduces riskAcetaminophen
increases risk →
Fig. 20. Random-effects meta-analysis for all studies after exclusions. This plot shows pooled effects, see the specific outcome analyses for individual outcomes. Analysis validating pooled outcomes for COVID-19 can be found below. Effect extraction is pre-specified, using the most serious outcome reported. For details see the appendix.
Media Bias
Low-cost treatments were subject to bias and censorship during the pandemic. Scientific bias is seen in the design, analysis, presentation, and selective reporting of studies, which often favored negative results. A similar bias is seen in the media coverage for low-cost treatments. While broadly seen, bias was particularly notable for ivermectin and hydroxychloroquine, e.g., Scott Alexander noted that "if you say anything in favor of ivermectin you will be cast out of civilization and thrown into the circle of social hell reserved for Klan members and 1/6 insurrectionists. All the health officials in the world will shout 'horse dewormer!' at you and compare you to Josef Mengele."27.
We analyze media coverage for the 227 treatments we cover using Altmetric28 and PlumX Metrics29. Altmetric covers ~14,000 news outlets30 and PlumX covers ~55,000 news sources31. The two services are complementary with coverage of different news sources and different methods for identifying articles not containing a DOI. We consider studies to have received significant media coverage if they were covered by at least 100 articles in either Altmetric of PlumX31,32. The most widespread study was covered by 1,047 outlets. Studies covered by major syndicated sources shows hundreds of news articles, while studies covered by few outlets typically appear in niche medical news sources or press releases. Fig. 21 and 22 show bias towards negative results for low-cost treatments, in contrast to the opposite bias for high-profit treatments.
For acetaminophen, studies show higher risk. These results are relevant to many people and for resolution of the pandemic, however the mainstream media did not cover any of the negative studies as shown in Fig. 23.
c19early.org
October 2026

Media coverage of low-cost COVID‑19 treatments

Media covered only 0.8% of studies showing significantly lower COVID-19 risk

Only 5 positive studies were covered

0.8%

one each: budesonide, cannabidiol, fluvoxamine, HCQ, vitamin D

34 negative studies were covered

4%

HCQ 12, ivermectin 6, lopinavir/ritonavir 4, azithromycin 3, vitamin D 3, fluvoxamine 2, one each: ibuprofen, metformin, vitamin C, zinc

5.4×
Negative bias for low-cost treatments
negative studies were 5.4 times more likely to be covered

Data from Altmetric and PlumX: studies receiving significant mainstream media coverage from 6,000+ studies for 227 treatments, defined as ≥100 news outlets in either Altmetric or PlumX, which cover ~14,000 and ~55,000 outlets respectively. Media coverage was significantly higher for negative studies, p < 0.001 (Fisher's exact test). Studies during the pandemic (until May 5, 2023).

Fig. 21. Mainstream media coverage was biased against positive results for low-cost treatments. When restricting to RCTs, the bias was even larger, with negative RCTs 6.9× more likely to be covered.
c19early.org
October 2026

Media coverage of high-profit COVID‑19 treatments

Media was 3.8× more likely to cover positive studies

19 positive studies were covered

11%

paxlovid 5, remdesivir 3, casirivimab/imdevimab 2, convalescent plasma 2, molnupiravir 2, one each: antiandrogens, peginterferon lambda, sargramostim, tixagevimab/cilgavimab, tocilizumab

Only 6 negative studies were covered

3%

remdesivir 3, one each: bamlanivimab/etesevimab, bebtelovimab, sotrovimab

3.8×
Positive bias for high-profit treatments
positive studies were 3.8 times more likely to be covered

Data from Altmetric and PlumX: studies receiving significant mainstream media coverage from 6,000+ studies for 227 treatments, defined as ≥100 news outlets in either Altmetric or PlumX, which cover ~14,000 and ~55,000 outlets respectively. Media coverage was significantly higher for positive studies, p = 0.002 (Fisher's exact test). Studies during the pandemic (until May 5, 2023).

Fig. 22. For high-profit treatments, mainstream media coverage was biased towards positive results. When restricting to RCTs, the bias was even larger, with positive RCTs 12.0× more likely to be covered.
c19early.org
October 2026

Media coverage of acetaminophen for COVID‑19

Media covered none of the 25 negative studies

No negative acetaminophen studies were covered

0%

Data from Altmetric and PlumX: studies receiving significant mainstream media coverage from 28 studies for acetaminophen treatment, defined as ≥100 news outlets in either Altmetric or PlumX, which cover ~14,000 and ~55,000 outlets respectively.

Fig. 23. Mainstream media did not cover any of the studies for acetaminophen.
A combination of factors may have led to the very low media coverage of low-cost treatments:
•
Politicization
led to a media environment where coverage was often framed to support a political narrative rather than to provide objective scientific information. As Scott Alexander said:
"if you say anything in favor of ivermectin you will be cast out of civilization and thrown into the circle of social hell reserved for Klan members and 1/6 insurrectionists. All the health officials in the world will shout 'horse dewormer!' at you and compare you to Josef Mengele."
There was strong social pressure to discredit low-cost treatments.
•
Censorship
of information conflicting with selected authorities. For example, individuals and organizations presenting conflicting science were often shadowbanned or completely banned on Twitter and YouTube if the science they were reporting conflicted with the opinions of specific authorities selected by those companies, including when the authorities had not even reviewed the scientific results in question.
•
FDA requires "no adequate, approved, and available alternatives"
in order to grant an EUA for novel high-profit interventions, creating an incentive for authorities to ignore or downplay existing low-cost treatments.
•
Regulatory capture
biases authorities towards high-profit interventions.
•
Authorities ignored most evidence for low-cost treatments
, for example the NIH references only 2% of studies in delayed, rarely-updated, biased commentaries with no quantitive analysis.
•
Media coverage of science is often not very accurate
, e.g., misunderstanding confounding issues. For example the media widely considered the RECOVERY HCQ RCT to be conclusive on efficacy, but very late treatment of late-stage patients (mostly on oxygen already) with an excessive toxic dose (shown dangerous in a dose comparison RCT) provides no information on the recommended early/prophylactic treatment. With difficulting in understanding basic confounders like treatment delay and dose, the media may favor deferring to authorities. Many studies for low-cost treatments require greater expertise to analyze.
•
Substantial funding from pharmaceutical advertising
biases editorial decisions towards high-profit interventions.
•
PR power
- companies/teams with strong PR presence are favored in the media, which correlates with high-profit and high conflict of interest studies.
•
The media was very negative in general
, inflating risk, and elevating fear and anxiety. A negative bias may improve ratings and revenue, increasing motivation to continue watching coverage.
Authority Delay
26 low-cost treatments were approved in one or more countries, yet many countries approved no low-cost treatments. The countries that did adopt low-cost treatments analyzed the evidence early and made timely approvals. With few exceptions, authorities did not change their initial views, regardless of how much evidence accumulated showing either efficacy or harm. Why?
The harms of smoking here hidden for 25 yearsA. Authorities did not analyze the data in real-time, failing to act when harm was known. Widespread acknowledgement of harm came only after attempts by two new surgeon generals, along with pressure from health advocates and a new president, and a review of 7,000 studies.
Similarly for COVID-19, most authorities and experts did not proactively analyze data in real-time. This guarantees delayed recognition of efficacy or harm, by which time moral, legal, career, and reputational liabilities strongly disincentivize any admission of error. Claims of no efficacy (for effective treatments) or safety (for harmful treatments) were often made prior to strong data being available. Correction would require admitting to errors that increased mortality, which is unlikely with the same generation of officials.
Analysis of potential treatments was rarely done, and when done these were typically minimal efforts. For example, NIH reviews were highly delayed, cover only a tiny fraction of treatments, reference only 2% of studies for the treatments covered, and include no quantitative analysis. They appear as rarely updated side projects from external panels implicitly tasked with justifying prior failures. As with smoking, the thousands of studies could (and should) have been analyzed and acted on in real-time.
A key structural improvement, applicable to all current and future diseases, is for authorities to implement real-time proactive analysis of clinical evidence. This does not remove all bias, but does make it possible to act on evidence, whereas delayed action may be unlikely due to moral, legal, career, and reputational liabilities.
c19early.org
Delayed public health acknowledgments
Official acknowledgment of efficacy or harm is often delayed—legal, career, and status risks disincentivize admission of error.
Evidence Official Acknowledgment Approx. Delay
Citrus Fruit (vitamin C) for Scurvy (effectiveness) 1747: James Lind conducted one of the first-ever controlled clinical trials, proving that oranges and lemons cured scurvy in sailors. 1795: The British Royal Navy finally made a daily ration of lemon juice a standard issue for all its sailors, effectively eliminating the disease. 48 years
Handwashing (lower mortality) 1847: Dr. Ignaz Semmelweis provided conclusive proof that having doctors wash their hands with a chlorine solution before delivering babies reduced maternal mortality rates from over 18% to around 1%. ~1870s: Semmelweis's findings were rejected and he was ridiculed. His work was only validated decades later (after his death). ~20+ years
Helicobacter pylori (bacteria causes ulcers) 1982-1984: Marshall and Warren discovered that Helicobacter pylori bacteria causes ulcers, confirmed via direct exposure. Officials maintained that ulcers were caused by stress and spicy food. 1994: The US NIH released a consensus statement officially recommending antibiotics as the standard treatment for peptic ulcers, overturning decades of acid-suppression therapy. ~12 years
Asbestos (causes asbestosis & cancer) 1924: The British Medical Journal published the first case study of a death from "asbestosis." By 1918, U.S. insurance companies had stopped selling life insurance to asbestos workers. 1971 (US): The Occupational Safety and Health Administration (OSHA) was formed and began regulating asbestos as a carcinogen, setting the first federal workplace safety standards for it. ~47 years
Leaded Gasoline (neurotoxicity) ~1924: Dangers of low-level lead exposure were known. Experts like Alice Hamilton warned the U.S. Surgeon General that adding lead to gasoline would cause widespread public poisoning. 1973 (US): The Environmental Protection Agency (EPA) ordered the first phasedown of lead in gasoline, following the Clean Air Act of 1970. A full ban for on-road vehicles took effect in 1996. ~49 years
Harms of Smoking (causes lung cancer) 1939: Franz Müller (Germany) published the first case-control epidemiological study strongly linking tobacco smoking to lung cancer. This was followed by major U.S. & U.K. studies in the 1950s. 1964 (US): The U.S. Surgeon General's report, "Smoking and Health," was released. It was the first U.S. government report to definitively link smoking to lung cancer and heart disease. 25 years
Heterogeneity
Heterogeneity in COVID-19 studies arises from many factors including:
Treatment delay
The time between infection or the onset of symptoms and treatment may critically affect how well a treatment works. For example an antiviral may be very effective when used early but may not be effective in late stage disease, and may even be harmful. Oseltamivir, for example, is generally only considered effective for influenza when used within 0-36 or 0-48 hours33,34. Baloxavir marboxil studies for influenza also show that treatment delay is critical — Ikematsu et al. report an 86% reduction in cases for post-exposure prophylaxis, Hayden et al. show a 33 hour reduction in the time to alleviation of symptoms for treatment within 24 hours and a reduction of 13 hours for treatment within 24-48 hours, and Kumar et al. report only 2.5 hours improvement for inpatient treatment.
Table 4. Studies of baloxavir marboxil for influenza show that early treatment is more effective.
Treatment delayResult
Post-exposure prophylaxis86% fewer cases35
<24 hours-33 hours symptoms36
24-48 hours-13 hours symptoms36
Inpatients-2.5 hours to improvement37
Fig. 24 shows a mixed-effects meta-regression for efficacy as a function of treatment delay in COVID-19 studies from 227 treatments, showing that efficacy declines rapidly with treatment delay. Early treatment is critical for COVID-19.
Loading..
Fig. 24. Early treatment is more effective. Meta-regression showing efficacy as a function of treatment delay in COVID-19 studies from 227 treatments.
Patient demographics
Details of the patient population including age and comorbidities may critically affect how well a treatment works. For example, many COVID-19 studies with relatively young low-comorbidity patients show all patients recovering quickly with or without treatment. In such cases, there is little room for an effective treatment to improve results, for example as in López-Medina et al.
SARS-CoV-2 variants
Efficacy may depend critically on the distribution of SARS-CoV-2 variants encountered by patients. Risk varies significantly across variants39, for example the Gamma variant shows significantly different characteristics40-43. Different mechanisms of action may be more or less effective depending on variants, for example the degree to which TMPRSS2 contributes to viral entry can differ across variants44,45.
Treatment regimen
Effectiveness may depend strongly on the dosage and treatment regimen.
Medication quality
The quality of medications may vary significantly between manufacturers and production batches, which may significantly affect efficacy and safety. Williams et al. analyze ivermectin from 11 different sources, showing highly variable antiparasitic efficacy across different manufacturers. Xu et al. analyze a treatment from two different manufacturers, showing 9 different impurities, with significantly different concentrations for each manufacturer.
Other treatments
The use of other treatments may significantly affect outcomes, including supplements, other medications, or other interventions such as prone positioning. Treatments may be synergistic48-72, therefore efficacy may depend strongly on combined treatments.
Effect measured
Across all studies there is a strong association between different outcomes, for example improved recovery is strongly associated with lower mortality. However, efficacy may differ depending on the effect measured, for example a treatment may be more effective against secondary complications and have minimal effect on viral clearance.
Meta-analysis
The distribution of studies will alter the outcome of a meta-analysis. Consider a simplified example where everything is equal except for the treatment delay, and effectiveness decreases to zero or below with increasing delay. If there are many studies using very late treatment, the outcome may be negative, even though early treatment is very effective. All meta-analyses combine heterogeneous studies, varying in population, variants, and potentially all factors above, and therefore may obscure efficacy by including studies where treatment is less effective. Generally, we expect the estimated effect size from meta-analysis to be less than that for the optimal case. Looking at all studies is valuable for providing an overview of all research, important to avoid cherry-picking, and informative when a positive result is found despite combining less-optimal situations. However, the resulting estimate does not apply to specific cases such as early treatment in high-risk populations. While we present results for all studies, we also present treatment time and individual outcome analyses, which may be more informative for specific use cases.
Pooled Effects
Pooled effects are no longer required to show harm as of April 2021
This section validates the use of pooled effects for COVID-19, which enables earlier detection of harm, however pooled effects are no longer required for acetaminophen as of April 2021. Harm is now known based on specific outcomes. Harm based on specific outcomes was delayed by 5.4 months compared to using pooled outcomes.
Combining studies is required
For COVID-19, delay in clinical results translates into additional death and morbidity, as well as additional economic and societal damage. Combining the results of studies reporting different outcomes is required. There may be no mortality in a trial with low-risk patients, however a reduction in severity or improved viral clearance may translate into lower mortality in a high-risk population. Different studies may report lower severity, improved recovery, and lower mortality, and the significance may be very high when combining the results. "The studies reported different outcomes" is not a good reason for disregarding results. Pooling the results of studies reporting different outcomes allows us to use more of the available information. Logically we should, and do, use additional information when evaluating treatments—for example dose-response and treatment delay-response relationships provide additional evidence of efficacy that is considered when reviewing the evidence for a treatment.
Specific outcome and pooled analyses
We present both specific outcome and pooled analyses. In order to combine the results of studies reporting different outcomes we use the most serious outcome reported in each study, based on the thesis that improvement in the most serious outcome provides comparable measures of efficacy for a treatment. A critical advantage of this approach is simplicity and transparency. There are many other ways to combine evidence for different outcomes, along with additional evidence such as dose-response relationships, however these increase complexity.
Ethical and practical issues limit high-risk trials
Trials with high-risk patients may be restricted due to ethics for treatments that are known or expected to be effective, and they increase difficulty for recruiting. Using less severe outcomes as a proxy for more serious outcomes allows faster and safer collection of evidence.
Validating pooled outcome analysis for COVID-19
For many COVID-19 treatments, a reduction in mortality logically follows from a reduction in hospitalization, which follows from a reduction in symptomatic cases, which follows from a reduction in PCR positivity. We can directly test this for COVID-19.
Analysis of the the association between different outcomes across studies from all 227 treatments we cover confirms the validity of pooled outcome analysis for COVID-19. Fig. 25 shows that lower hospitalization is very strongly associated with lower mortality (p < 0.0000000001). Similarly, Fig. 26 shows that improved recovery is very strongly associated with lower mortality (p < 0.0000000001). Considering the extremes, Singh et al. show an association between viral clearance and hospitalization or death, with p = 0.003 after excluding one large outlier from a mutagenic treatment, and based on 44 RCTs including 52,384 patients. Fig. 27 shows that improved viral clearance is strongly associated with fewer serious outcomes. The association is very similar to Singh et al., with higher confidence due to the larger number of studies. As with Singh et al., the confidence increases when excluding the outlier treatment, from p = 0.0000000074 to p = 0.00000000015.
Loading..
Fig. 25. Lower hospitalization is associated with lower mortality, supporting pooled outcome analysis.
Loading..
Fig. 26. Improved recovery is associated with lower mortality, supporting pooled outcome analysis.
Loading..
Fig. 25. Improved viral clearance is associated with fewer serious outcomes, supporting pooled outcome analysis.
Pooled outcomes identify efficacy 5 months faster (7 months for RCTs)
Currently, 59 of the treatments we analyze show statistically significant efficacy or harm, defined as ≥10% decreased risk or >0% increased risk from ≥3 studies. 85% of these have been confirmed with one or more specific outcomes, with a mean delay of 4.6 months. When restricting to RCTs only, 53% of treatments showing statistically significant efficacy/harm with pooled effects have been confirmed with one or more specific outcomes, with a mean delay of 7.5 months. Fig. 28 shows when treatments were found effective during the pandemic. Pooled outcomes often resulted in earlier detection of efficacy.
Loading..
Loading..
Fig. 28. The time when studies showed that treatments were effective, defined as statistically significant improvement of ≥10% from ≥3 studies. Pooled results typically show efficacy earlier than specific outcome results. Results from all studies often shows efficacy much earlier than when restricting to RCTs. Results reflect conditions as used in trials to date, these depend on the population treated, treatment delay, and treatment regimen.
Limitations
Pooled analysis could hide efficacy, for example a treatment that is beneficial for late stage patients but has no effect on viral clearance may show no efficacy if most studies only examine viral clearance. In practice, it is rare for a non-antiviral treatment to report viral clearance and to not report clinical outcomes; and in practice other sources of heterogeneity such as differences in treatment delay are more likely to hide efficacy.
Summary
Analysis validates the use of pooled effects and shows significantly faster detection of efficacy on average. However, as with all meta-analyses, it is important to review the different studies included. We also present individual outcome analyses, which may be more informative for specific use cases.
Discussion
Results for other infections
Increased risk with acetaminophen has also been shown for rhinovirus3.
Publication bias
Publishing is often biased towards positive results, however evidence suggests that there may be a negative bias for inexpensive treatments for COVID-19. Both negative and positive results are very important for COVID-19, media in many countries prioritizes negative results for inexpensive treatments (inverting the typical incentive for scientists that value media recognition), and there are many reports of difficulty publishing positive results74-77. For acetaminophen, there is currently not enough data to evaluate publication bias with high confidence.
Funnel plots have traditionally been used for analyzing publication bias. This is invalid for COVID-19 acute treatment trials — the underlying assumptions are invalid, which we can demonstrate with a simple example. Consider a set of hypothetical perfect trials with no bias. Fig. 29 plot A shows a funnel plot for a simulation of 80 perfect trials, with random group sizes, and each patient's outcome randomly sampled (10% control event probability, and a 30% effect size for treatment). Analysis shows no asymmetry (p > 0.05). In plot B, we add a single typical variation in COVID-19 treatment trials — treatment delay. Consider that efficacy varies from 90% for treatment within 24 hours, reducing to 10% when treatment is delayed 3 days. In plot B, each trial's treatment delay is randomly selected. Analysis now shows highly significant asymmetry, p < 0.0001, with six variants of Egger's test all showing p < 0.0578-85. Note that these tests fail even though treatment delay is uniformly distributed. In reality treatment delay is more complex — each trial has a different distribution of delays across patients, and the distribution across trials may be biased (e.g., late treatment trials may be more common). Similarly, many other variations in trials may produce asymmetry, including dose, administration, duration of treatment, differences in SOC, comorbidities, age, variants, and bias in design, implementation, analysis, and reporting.
Log Risk Ratio Standard Error 1.406 1.055 0.703 0.352 0 -3 -2 -1 0 1 2 A: Simulated perfect trials p > 0.05 Log Risk Ratio Standard Error 1.433 1.074 0.716 0.358 0 -4 -3 -2 -1 0 1 2 B: Simulated perfect trials with varying treatment delay p < 0.0001
Fig. 29. Example funnel plot analysis for simulated perfect trials.
Conflicts of interest
Pharmaceutical drug trials often have conflicts of interest whereby sponsors or trial staff have a financial interest in the outcome being positive. Acetaminophen for COVID-19 lacks this because it is off-patent, has multiple manufacturers, and is very low cost.
Limitations
Summary statistics from meta-analysis necessarily lose information. As with all meta-analyses, studies are heterogeneous, with differences in treatment delay, treatment regimen, patient demographics, variants, conflicts of interest, standard of care, and other factors. We provide analyses for specific outcomes and by treatment delay, and we aim to identify key characteristics in the forest plots and summaries. Results should be viewed in the context of study characteristics.
Some analyses classify treatment based on early or late administration, as done here, while others distinguish between mild, moderate, and severe cases. Viral load does not indicate degree of symptoms — for example patients may have a high viral load while being asymptomatic. With regard to treatments that have antiviral properties, timing of treatment is critical — late administration may be less helpful regardless of severity.
Details of treatment delay per patient is often not available. For example, a study may treat 90% of patients relatively early, but the events driving the outcome may come from 10% of patients treated very late. Our 5 day cutoff for early treatment may be too conservative, 5 days may be too late in many cases.
Comparison across treatments is confounded by differences in the studies performed, for example dose, variants, and conflicts of interest. Trials with conflicts of interest may use designs better suited to the preferred outcome.
In some cases, the most serious outcome has very few events, resulting in lower confidence results being used in pooled analysis, however the method is simpler and more transparent. This is less critical as the number of studies increases. Restriction to outcomes with sufficient power may be beneficial in pooled analysis and improve accuracy when there are few studies, however we maintain our pre-specified method to avoid any retrospective changes.
Studies show that combinations of treatments can be highly synergistic and may result in many times greater efficacy than individual treatments alone48-72. Therefore standard of care may be critical and benefits may diminish or disappear if standard of care does not include certain treatments.
This real-time analysis is constantly updated based on submissions. Accuracy benefits from widespread review and submission of updates and corrections from reviewers. Less popular treatments may receive fewer reviews.
No treatment or intervention is 100% available and effective for all current and future variants. Efficacy may vary significantly with different variants and within different populations. All treatments have potential side effects. Propensity to experience side effects may be predicted in advance by qualified physicians. We do not provide medical advice. Before taking any medication, consult a qualified physician who can compare all options, provide personalized advice, and provide details of risks and benefits based on individual medical history and situations.
Notes
9 of the 28 studies compare against other treatments, which may reduce the effect seen. 1 of 28 studies combine treatments. The results of acetaminophen alone may differ. None of the RCTs use combined treatment.
Reviews
Multiple reviews cover acetaminophen for COVID-19, presenting additional background on mechanisms and related results, including1,2,86-88.
Perspective
Results compared with other treatments
SARS-CoV-2 infection and replication involves a complex interplay of 500+ host and viral proteins and other factors89-96, providing many therapeutic targets. Over 12,000 compounds have been predicted to reduce COVID-19 risk97, either by directly minimizing infection or replication, by supporting immune system function, or by minimizing secondary complications. Fig. 30 shows an overview of the results for acetaminophen in the context of multiple COVID-19 treatments, and Fig. 31 shows a plot of efficacy vs. cost for COVID-19 treatments.
Loading..
Fig. 30. Scatter plot showing results within the context of multiple COVID-19 treatments. Diamonds shows the results of random-effects meta-analysis. 0.5% of 12,000+ proposed treatments show efficacy98.
Loading..
Loading..
Fig. 31. Efficacy vs. cost for COVID-19 treatments.
Conclusion
Meta-analysis shows 24% [9‑40%] higher mortality, and pooled analysis using the most serious outcome reported shows 29% [17‑41%] higher risk.
Potential mechanisms of harm include glutathione depletion, fever suppression, liver toxicity, immunosuppression, cytokine disruption, prostaglandin inhibition, COX inhibition, cell/tissue injury, mitochondrial dysfunction, glycine depletion, disruption of redox balance, increased oxidative stress, mineral depletion, microbiome alteration, and endocannabinoid system dysfunction.
Concerns have been raised over the use of acetaminophen (paracetamol) for COVID-191,2. Studies show significantly increased risk.
Increased risk with acetaminophen has also been shown for rhinovirus3.
Fever has risks that can outweigh the immunological benefits in certain circumstances.
 
Contact.
Contact us on X at @CovidAnalysis.
Funding.
We have received no funding or compensation in any form, and do not accept donations. This is entirely volunteer work.
Conflicts of interest.
We have no conflicts of interest. We have no affiliation with any pharmaceutical companies, supplement companies, governments, political parties, or advocacy organizations.
AI.
We use AI models (Gemini, Grok, Claude, and ChatGPT) tasked with functioning as additional peer-reviewers to check for errors, suggest improvements, and review spelling and grammar. Any corrections are manually verified. Our preference for em dashes is independent of AI.
Updates.
Our COVID-19 meta-analyses involve the extraction of over 229,000 datapoints from thousands of papers for 227 treatments. We thank the thousands of scientists, physicians, and other contributors that have provided updates, suggestions, feedback, and corrections. These are all welcome and can be submitted at https://c19early.org/acemeta.html.
Dedication.
This work is dedicated to top evidence-based physicians that worked tirelessly to analyze evidence and greatly reduce mortality and morbidity during the pandemic. In alphabetical order: Dr. Thomas J. Borody, Dr. Mary Talley Bowden, Dr. Flavio Cadegiani, Dr. Shankara Chetty, Dr. Ryan Cole, Dr. George Fareed, Dr. Sabine Hazan, Dr. Pierre Kory, Dr. Tess Lawrie, Dr. Robert Malone, Dr. Paul Marik, Dr. Peter McCullough, Dr. Didier Raoult, Dr. Harvey Risch, Dr. Jackie Stone, Dr. Brian Tyson, Dr. Joseph Varon, and Dr. Vladimir Zelenko.
Public domain.
This is a public domain work distributed in accordance with the Creative Commons CC0 1.0 Universal license, which dedicates the work to the public domain by waiving all rights worldwide under copyright law. You can distribute, remix, adapt, and build upon this work in any medium or format, including for commercial purposes, without asking permission. Referenced material and third-party images retain any original copyrights or restrictions. See: https://creativecommons.org/publicdomain/zero/1.0/.
Study Notes
Retrospective 31 hospitalized patients ≤19 with pre-existing inborn errors of immunity, showing no significant difference in mortality with acetaminophen. Only 6 patients were treated with acetaminophen. Submit Corrections or Updates.
ICU patientsProspective · NCT04321265 · 2,646 patients · multinational
Acetaminophen for COVID-19
Baldia et al. BMC Geriatrics · December 2022
Outcome
Improvement
Relative Risk · 95% CI
Mortality, day 90
12%0.88 · 0.72–1.07
Mortality, day 30
14%0.86 · 0.72–1.03
00.511.52+
← Acetaminophen
reduces riskAcetaminophen
increases risk →
Lower mortality with acetaminophen
Not statistically significant · p = 0.2
Prospective study of 2,646 ICU patients ≥70 years old, showing no significant difference in mortality with acetaminophen use in the 10 days prior to ICU admission. Submit Corrections or Updates.
Retrospective 179 elderly patients in France, showing higher risk of COVID-19 cases with acetaminophen use, without statistical significance. Submit Corrections or Updates.
Retrospective 28,856 COVID-19 patients in the USA, showing no significant difference in mortality for chronic acetaminophen use vs. sporadic NSAID use. Since acetaminophen is available OTC and authors only tracked prescriptions, many patients classified as sporadic users may have been chronic users. Submit Corrections or Updates.
ProphylaxisPSM retrospective · 17,190 patients · UK · Jan – Jul 2020
Acetaminophen for COVID-19
Chandan et al. Arthritis & Rheumatology · April 2021
Outcome
Improvement
Relative Risk · 95% CI
Mortality
-18%1.18 · 0.83–1.64
Case
-27%1.27 · 0.90–1.75
00.511.52+
← Acetaminophen
reduces riskAcetaminophen
increases risk →
Higher mortality and more cases
Not statistically significant · p = 0.35 (mortality) and p = 0.17 (more cases) Study compares with NSAIDs, results vs. placebo may differ
Retrospective 12,457 patients prescribed paracetamol with codeine/dihydrocodeine and 13,202 prescribed NSAIDs, showing no significant differences in cases and mortality. Patients prescribed codeine/dihydrocodeine may have different susceptibility to COVID-19. Submit Corrections or Updates.
Early treatmentProspective · 354 patients · USA · May 2020 – Jun 2021
Acetaminophen for COVID-19
Chen et al. Frontiers in Medicine · October 2023
Outcome
Improvement
Relative Risk · 95% CI
Long COVID
-32%1.32 · 0.98–1.78
Long COVID, ibuprofen
-1%1.01 · 0.67–1.52
00.511.52+
← Acetaminophen
reduces riskAcetaminophen
increases risk →
Higher long COVID with acetaminophen
Not statistically significant · p = 0.067
Prospective study of 494 COVID-19 patients showing higher risk of PASC with acetaminophen use in unadjusted results, without reaching statistical significance (p=0.07). Higher risk is also seen for dexamethasone and remdesivir (statistically significant for dexamethasone), however confounding by indication may be significant for these treatments, with increased use for more severe patients. While details of treatment timing and dose are not available, the result for acetaminophen can be compared with ibuprofen, with comparable indication for use. Notably there is no increased risk with ibuprofen, suggesting higher risk with acetaminophen, consistent with the higher risk seen in meta analysis99. Submit Corrections or Updates.
ProphylaxisRetrospective · 103 patients · Spain · Mar – May 2020
Acetaminophen for COVID-19
Gálvez-Barrón et al. Gerontology · April 2021
Outcome
Improvement
Relative Risk · 95% CI
Mortality
-47%1.47 · 0.66–3.33
Severe case
23%0.77 · 0.35–1.71
00.511.52+
← Acetaminophen
reduces riskAcetaminophen
increases risk →
Higher mortality and lower severe cases
Not statistically significant · p = 0.42 (mortality) and p = 0.55 (severe cases)
Analysis of 103 elderly hospitalized COVID-19 patients in Spain, showing higher mortality with acetaminophen, without statistical significance. Submit Corrections or Updates.
Early treatmentRetrospective · South Korea · Jan – Apr 2020
Acetaminophen for COVID-19
Jeong et al. Clinical Infectious Diseases · July 2020
Outcome
Improvement
Relative Risk · 95% CI
Cardiovascular complications
-15%1.15 · 0.69–1.92
Renal failure
-92%1.92 · 0.27 to >9
00.511.52+
← Acetaminophen
reduces riskAcetaminophen
increases risk →
Study compares with NSAIDs, results vs. placebo may differ
Retrospective 1,824 hospitalized COVID-19 patients in South Korea, showing higher progression to combined death, ICU, ventilation, or sepsis (4% versus 0%, group sizes not provided) with paracetamol vs. NSAIDs. Treatment time may vary - exposure was defined as 7 days before and including cohort entry in hospitalized COVID-19 patients. Submit Corrections or Updates.
ProphylaxisPSM retrospective · 324 patients · South Korea
Acetaminophen for COVID-19
Kim et al. Int. J. Environmental Research and Public Health · Feb 2023
Outcome
Improvement
Relative Risk · 95% CI
Mortality
-71%1.71 · 0.69–4.24
Ventilation
-14%1.14 · 0.42–3.08
ICU admission
40%0.60 · 0.15–2.47
Oxygen therapy
-9%1.09 · 0.64–1.86
00.511.52+
← Acetaminophen
reduces riskAcetaminophen
increases risk →
Higher mortality with acetaminophen
Not statistically significant · p = 0.34 Study compares with NSAIDs, results vs. placebo may differ
PSM retrospective in South Korea, showing no significant differences in outcomes with acetaminophen use vs. NSAID use. Adherence and dosage are unknown. Submit Corrections or Updates.
397,064 patient UK Biobank retrospective showing higher risk of COVID-19 with acetaminophen use. Submit Corrections or Updates.
Retrospective paracetamol use with a primary care database in Italy, showing no significant difference in hospitalization/death for use 0-3 and 4-7 days from diagnosis, and significantly higher risk for use >7 days from diagnosis. Confounding by indication may have a greater effect on late usage. Submit Corrections or Updates.
UK Biobank retrospective showing lower cases with acetaminophen use. Submit Corrections or Updates.
Retrospective 5,783 hospitalized patients in France, showing higher mortality with paracetamol use, without statistical significance. Submit Corrections or Updates.
Retrospective 26,121 cases and 2,369,020 controls ≥65yo in Canada, showing higher cases with chronic use of acetaminophen. Submit Corrections or Updates.
Late treatmentRetrospective · 524 patients · USA · Feb – Jun 2020
Acetaminophen for COVID-19
Manjani et al. Chest · October 2021
Higher mortality and ventilation
p = 0.001 (mortality) and p = 0.001 (ventilation)
Retrospective 524 hospitalized patients in the USA, showing higher mortality and progression with acetaminophen use. Submit Corrections or Updates.
ProphylaxisRetrospective · 99,890 patients · multinational · Jan – Jun 2020
Acetaminophen for COVID-19
Moreno-Martos et al. Wellcome Open Research · January 2022
Higher hospitalization with acetaminophen
p < 0.000001
Aanlysis of prescriptions in multiple databases showing higher risk of COVID-19 hospitalization with acetaminophen use for COPD patients. Acetaminophen use was more prevalent in hospitalized patients compared to diagnosed patients (data from tables 1, 5, and S3). Submit Corrections or Updates.
Retrospective 7,713 COVID-19 patients in Korea, showing no significant difference in mortality with paracetamol use. Submit Corrections or Updates.
ProphylaxisPSM retrospective · 794 patients · South Korea
Acetaminophen for COVID-19
Park et al. Scientific Reports · March 2021
Outcome
Improvement
Relative Risk · 95% CI
Mortality
25%0.75 · 0.35–1.59
Ventilation
38%0.62 · 0.19–1.89
00.511.52+
← Acetaminophen
reduces riskAcetaminophen
increases risk →
Lower mortality and ventilation
Not statistically significant · p = 0.46 (mortality) and p = 0.42 (ventilation) Study compares with NSAIDs, results vs. placebo may differ
Retrospective 2,365 patients prescribed acetaminophen and 398 prescribed NSAIDs in South Korea, showing no significant differences. Submit Corrections or Updates.
Retrospective 416 non-hospitalized and 184 hospitalized COVID-19 patients in Bangladesh, showing higher acetaminophen and lower vitamin C usage for hospitalized patients. Confounding may be significant and baseline details per treatment group are not provided, however fever and symptomatic patients were more common in the non-hospitalized group. Note there is an alignment mismatch in Table 1. Submit Corrections or Updates.
Late treatmentRCT · CTRI/2021/05/033544 · 210 patients · India
Acetaminophen for COVID-19
Ravichandran et al. Scientific Reports · April 2022
Worse recovery and higher progression
p = 0.0018 (recovery) and p < 0.0001 (progression) Trial compares with indomethacin, results vs. placebo may differ
RCT with 107 paracetamol and 103 indomethacin patients, showing higher progression and worse recovery with paracetamol. Submit Corrections or Updates.
Late treatmentRetrospective · ISRCTN11970082 · 144 patients · India
Acetaminophen for COVID-19
Ravichandran et al. J. Indian Medical Association · July 2021
Outcome
Improvement
Relative Risk · 95% CI
Oxygen therapy
-2700%28.0 · 3.91 to >9
Recovery time
-75%1.75 · 1.48–2.06
Recovery time (b)
-117%2.17 · 1.72–2.72
Recovery time (c)
-167%2.67 · 2.00–3.56
00.511.52+
← Acetaminophen
reduces riskAcetaminophen
increases risk →
Higher need for oxygen therapy and slower recovery
p < 0.0001 (need for oxygen) and p < 0.0001 (slower recovery) Study compares with indomethacin, results vs. placebo may differ
Is late treatment with acetaminophen beneficial for COVID-19?
c19early.org
PSM retrospective 72 indomethacin and 72 paracetamol patients in India, showing higher progression and worse recovery with acetaminophen. Submit Corrections or Updates.
N3C retrospective 250,533 patients showing significantly higher mortality with acetaminophen use. Note that acetaminophen results were not included in the journal version or v2 of this preprint, which focuses on NSAID analysis. Submit Corrections or Updates.
Early treatmentRetrospective · 134 patients · Israel · Mar – Apr 2020
Acetaminophen for COVID-19
Rinott et al. Clinical Microbiology and Infection · Sep 2020
Outcome
Improvement
Relative Risk · 95% CI
Mortality
-473%5.73 · 0.30 to >9
Oxygen therapy
-534%6.34 · 0.84 to >9
00.511.52+
← Acetaminophen
reduces riskAcetaminophen
increases risk →
Higher mortality and higher oxygen therapy
Not statistically significant · p = 0.3 (mortality) and p = 0.055 (oxygen) Study compares with ibuprofen, results vs. placebo may differ
Retrospective 89 febrile COVID-19 patients in Israel taking paracetamol and 49 taking ibuprofen, showing higher need for respiratory support with paracetamol. Although not statistically significant, patients in the paracetamol group were older. Submit Corrections or Updates.
Retrospective 44,866 hospitalized COVID-19 patients in Sweden, showing higher mortality with vitamin D deficiency and with acetaminophen use.

The study focuses on cardiorenal disease, finding higher risk of mortality with CRD. Authors also show that COVID-19 mortality was about 1.5x higher when compared with influenza in the first two pandemic waves, but there was no significant difference in the third wave (HR 1.53 [1.45-1.62] and 1.52 [1.44-1.61] in the first two waves and 1.07 [0.99-1.14] in the third). Submit Corrections or Updates.
Retrospective COVID-19 patients in Bangladesh, showing higher mortality with acetaminophen use in unadjusted results. Submit Corrections or Updates.
Late treatmentRCT · 180 patients · Egypt · Jan – May 2022
Acetaminophen for COVID-19
Sobhy et al. Open Anesthesia J. · April 2023 · PACTR202202880140319
Higher ICU admission and higher oxygen therapy
p = 0.047 (ICU admission) and p = 0.047 (oxygen) Trial compares with ibuprofen, results vs. placebo may differ
RCT 180 moderate hospitalized COVID-19 patients in Egypt, showing higher ICU admission and longer hospitalization with acetaminophen compared with ibuprofen. Submit Corrections or Updates.
Retrospective 80 mild COVID-19 patients in Italy, showing no significant difference in long COVID with acetaminophen use during infection. Submit Corrections or Updates.
Early treatmentRetrospective · 56 patients · Russia
Acetaminophen for COVID-19
Vasilkova et al. Pharmacology & Pharmacotherapy · June 2026
Outcome
Improvement
Relative Risk · 95% CI
ICU admission
-60%1.60 · 0.60–4.29
Mean case severity score
-19%1.19 · 0.90–1.56
00.511.52+
← Acetaminophen
reduces riskAcetaminophen
increases risk →
Higher ICU admission and progression
Not statistically significant · p = 0.53 (ICU admission) and p = 0.22 (progression) Study compares with ibuprofen, results vs. placebo may differ
Retrospective 86 infants under 1 year old comparing pre-hospital antipyretic therapy with Viburcol, paracetamol, and ibuprofen. No statistically significant differences were found between groups for hospitalization duration, disease severity, ICU transfer rate, oxygen saturation, or CT changes. There was a non-significant trend toward higher severity and ICU transfers in the paracetamol group. Viburcol is marketed differently and may have significant confounding by indication, with paracetamol and ibuprofen more likely used for more serious cases. Therefore the comparison between paracetamol and ibuprofen may be most informative, however potential confounding remains - ibuprofen is not licensed for the youngest group that may have higher risk (favors ibuprofen), and ibuprofen may be used as a second line treatment when paracetamol fails (favors paracetamol). Submit Corrections or Updates.
ProphylaxisRetrospective · 1,370,600 patients · USA · Feb – Oct 2020
Acetaminophen for COVID-19
Xie et al. Drugs · July 2022
Outcome
Improvement
Relative Risk · 95% CI
Hospitalization
-5%1.05 · 0.70–1.56
Case
3%0.97 · 0.72–1.29
00.511.52+
← Acetaminophen
reduces riskAcetaminophen
increases risk →
No significant difference in outcomes seen
Study compares with ibuprofen, results vs. placebo may differ
PSM retrospective 1,370,600 osteoarthritis or back pain patients in the US, showing no significant differences in COVID-19 cases and hospitalization for paracetamol vs. ibuprofen. Submit Corrections or Updates.
Appendix 1. Methods and Data
Search methods
We perform ongoing searches of PubMed, medRxiv, Europe PMC, ClinicalTrials.gov, The Cochrane Library, Google Scholar, Research Square, ScienceDirect, Oxford University Press, the reference lists of other studies and meta-analyses, and submissions to the site c19early.org, which regularly receives notification of studies upon publication. Search terms are acetaminophen and COVID-19 or SARS-CoV-2. Automated searches are performed twice daily, with all matches reviewed for inclusion. All studies regarding the use of acetaminophen for COVID-19 that report a comparison with a control group are included in the main analysis. Sensitivity analysis is performed, excluding studies with major issues, epidemiological studies, and studies with minimal available information. Studies with major unexplained data issues, for example major outcome data that is impossible to be correct with no response from the authors, are excluded.
Fig. 32. Mid-recovery results can more accurately reflect efficacy when almost all patients recover. Mateja et al. confirm that intermediate viral load results more accurately reflect hospitalization/death.
Effect extraction
We extracted effect sizes and associated data from all studies. If studies report multiple kinds of effects then the most serious outcome is used in pooled analysis, while other outcomes are included in the outcome-specific analyses. For example, if effects for mortality and cases are reported then they are both used in specific outcome analyses, while mortality is used for pooled analysis. If symptomatic results are reported at multiple times, we use the latest time, for example if mortality results are provided at 14 days and 28 days, the results at 28 days have preference. Mortality alone is preferred over combined outcomes. Outcomes with zero events in both arms are not used, the next most serious outcome with one or more events is used. For example, in low-risk populations with no mortality, a reduction in mortality with treatment is not possible, however a reduction in hospitalization, for example, is still valuable. Clinical outcomes are considered more important than viral outcomes. When basically all patients recover in both treatment and control groups, preference for viral clearance and recovery is given to results mid-recovery where available. After most or all patients have recovered there is little or no room for an effective treatment to do better, however faster recovery is valuable. An IPD meta-analysis confirms that intermediate viral load reduction is more closely associated with hospitalization/death than later viral load reduction100. If only individual symptom data is available, the most serious symptom has priority, for example difficulty breathing or low SpO2 is more important than cough.
Statistical methods
Forest plots are computed using PythonMeta101 with the DerSimonian and Laird random-effects model (the fixed effect assumption is not plausible in this case) and inverse variance weighting. Results are presented with 95% confidence intervals. Heterogeneity among studies was assessed using the I2 statistic. When results provide an odds ratio, we compute the relative risk when possible, or convert to a relative risk according to Zhang et al. Reported confidence intervals and p-values are used when available, and adjusted values are used when provided. If multiple types of adjustments are reported propensity score matching and multivariable regression has preference over propensity score matching or weighting, which has preference over multivariable regression. Adjusted results have preference over unadjusted results for a more serious outcome when the adjustments significantly alter results. When needed, conversion between reported p-values and confidence intervals followed Altman, Altman (B), and Fisher's exact test was used to calculate p-values for event data. If continuity correction for zero values is required, we use the reciprocal of the opposite arm with the sum of the correction factors equal to 1105. Results are expressed with RR < 1.0 favoring treatment, and using the risk of a negative outcome when applicable (for example, the risk of death rather than the risk of survival). If studies only report relative continuous values such as relative times, the ratio of the time for the treatment group versus the time for the control group is used. Calculations are done in Python (3.15.0) with scipy (1.18.1), pythonmeta (1.26), numpy (2.5.4), statsmodels (0.15.0), and plotly (6.9.0). Mixed-effects meta-regression results are computed with R (4.4.0) using the metafor (4.6-0) and rms (6.8-0) packages, and using the most serious sufficiently powered outcome. For all statistical tests, a p-value less than 0.05 was considered statistically significant. Grobid 0.8.2 is used to parse PDF documents.
When evaluating potential effect modification across groups, we use an interaction test as described by Altman (C) et al. We compared the log-transformed relative risks using a z-test, deriving the standard error of the difference from the 95% confidence intervals. A two-sided interaction p-value of < 0.05 was considered a statistically significant difference in treatment effect between the groups.
Quality evaluation
Cochrane RoB 2/ROBINS-I are often used to evaluate studies, and have the advantage of providing standardized rules that can be applied with minimal understanding of the domain and study. However, the rules do not account for many real-world issues, often overemphasize or underemphasize others, and studies show low inter-rater reliability113. Certain domains are more applicable for these tools, however the time-sensitive nature of a pandemic, with significant mortality for every day of delay in evidence assessment, and the characteristics of COVID-19 make them inappropriate for this domain. This can be demonstrated with examples where expert RoB 2/ROBINS-I ratings do not match reality for COVID-19. Popp et al. use RoB 2 to classify Reis et al. as low risk of bias, however this is the opposite of reality—the trial not only has very high risk of bias, but has very high actual known bias, refusing to release data despite pledging to, reporting multiple impossible numbers, having blinding and randomization failure, and many other issues115. Axfors et al. use RoB 2 to classify Horby et al. as low risk of bias, however this is the opposite of reality—the very late treatment and excessive dosage used produces results with no relevance to recommended usage. HCQ shows poor results with late treatment and excessive dosage, and the combination shows harmB. Hempenius et al. use ROBINS-I to classify 33 studies for HCQ. The two rated as having the lowest risk of bias111,112 are far from the most informative. Both involve very late treatment, providing no information on recommended usage, and ROBINS-I does a very poor job of accounting for the impact of confounding factorsC.
Our quality evaluation focuses on known issues and bias, and the potential impact on outcomes, rather than just the risk of bias. The estimated potential impact of each confounding factor, and the direction of the impact is considered. For example, consider a study that shows significantly lower risk, the value of the study varies significantly if confounding points to an underestimate or an overestimate of efficacy. In one case, the real effect may be null, while the other case provides stronger evidence of efficacy (which may be greater than the study shows). Analysis focusing on the risk of bias, while simpler, may penalize studies for theoretical or technical issues that have no or minimal impact on outcomes. Analysis also depends on the outcome, for example certain issues are less relevant for objective outcomes such as mortality. Inaccurate penalization, and inaccurate high-quality evaluation in the face of known major issues affecting outcomes, increases in significance during a pandemic when immediate recognition of new evidence is critical, and when considering all global studies, as required during a pandemic. Investigators in other countries may have different customs for design, analysis, and reporting, and different English language skills, however they may not be less diligent or have greater bias. Investigators in lower-pharmaceutical-profit countries may have lower bias towards profitable interventions.
Treatment time
We have classified studies as early treatment if most patients are not already at a severe stage at the time of treatment (for example based on oxygen status or lung involvement), and treatment started within 5 days of the onset of symptoms. If studies contain a mix of early treatment and late treatment patients, we consider the treatment time of patients contributing most to the events (for example, consider a study where most patients are treated early but late treatment patients are included, and all mortality events were observed with late treatment patients). We note that a shorter time may be preferable. Antivirals are typically only considered effective when used within a shorter timeframe, for example 0-36 or 0-48 hours for oseltamivir, with longer delays not being effective33,34.
Living analysis
This is a living analysis and is updated regularly. Submit updates or corrections with the form below. We received no funding, this research is done in our spare time. We have no affiliation with any pharmaceutical companies, supplement companies, governments, political parties, or advocacy organizations.
A summary of study results is below. Please submit updates and corrections at the bottom of this page.
A summary of study results is below. Please submit updates and corrections at https://c19early.org/acemeta.html.
Effect extraction follows pre-specified rules as detailed above and gives priority to more serious outcomes. For pooled analyses, the first (most serious) outcome is used, which may differ from the effect a paper focuses on. Other outcomes are used in outcome specific analyses.
Chen (B), 10/16/2023, prospective, USA, peer-reviewed, 17 authors, study period May 2020 - June 2021.
Submit Corrections or Updates.
 risk of long COVID, 32.1% higher, RR 1.32, p = 0.07, treatment 98 of 232 (42.2%), control 39 of 122 (32.0%).
risk of long COVID, 0.9% higher, RR 1.01, p = 1.00, treatment 16 of 41 (39.0%), control 121 of 313 (38.7%), ibuprofen.
Lapi, 7/30/2022, retrospective, Italy, peer-reviewed, 8 authors, early treatment subset.
Submit Corrections or Updates.
 risk of death/hospitalization, 15.0% higher, OR 1.15, p = 0.22, adjusted per study, early use, RR approximated with OR.
risk of death/hospitalization, 29.0% higher, OR 1.29, p = 0.52, adjusted per study, mid-term use, RR approximated with OR.
Rahman, 11/8/2023, retrospective, Bangladesh, peer-reviewed, 5 authors, excluded in exclusion analyses: unadjusted results with no group details; significant unadjusted confounding possible.
Submit Corrections or Updates.
 risk of hospitalization, 22.6% higher, RR 1.23, p = 0.11, treatment 84 of 244 (34.4%), control 100 of 356 (28.1%).
Rinott, 9/30/2020, retrospective, Israel, peer-reviewed, median age 45.0, 5 authors, study period 15 March, 2020 - 15 April, 2020, this trial compares with another treatment - results may be better when compared to placebo, excluded in exclusion analyses: unadjusted differences between groups.
Submit Corrections or Updates.
 risk of death, 472.9% higher, RR 5.73, p = 0.30, treatment 3 of 85 (3.5%), control 0 of 49 (0.0%), continuity correction due to zero event (with reciprocal of the contrasting arm).
risk of oxygen therapy, 534.1% higher, RR 6.34, p = 0.06, treatment 11 of 85 (12.9%), control 1 of 49 (2.0%).
Sharif, 11/26/2022, retrospective, Bangladesh, peer-reviewed, 14 authors, study period 13 December, 2020 - 4 February, 2021, excluded in exclusion analyses: unadjusted results with no group details.
Submit Corrections or Updates.
 risk of death, 77.0% higher, RR 1.77, p = 0.74, treatment 9 of 361 (2.5%), control 2 of 142 (1.4%), unadjusted, ACE.
Vasilkova, 6/2/2026, retrospective, Russia, peer-reviewed, 2 authors, this trial compares with another treatment - results may be better when compared to placebo.
Submit Corrections or Updates.
 risk of ICU admission, 60.0% higher, RR 1.60, p = 0.53, treatment 8 of 28 (28.6%), control 5 of 28 (17.9%).
mean case severity score, 18.8% higher, relative time 1.19, p = 0.22, treatment mean 1.71 (±0.81) n=28, control mean 1.44 (±0.81) n=28.
Effect extraction follows pre-specified rules as detailed above and gives priority to more serious outcomes. For pooled analyses, the first (most serious) outcome is used, which may differ from the effect a paper focuses on. Other outcomes are used in outcome specific analyses.
Abolhassani, 9/13/2022, retrospective, Iran, peer-reviewed, 23 authors.
Submit Corrections or Updates.
 risk of death, 56.2% higher, RR 1.56, p = 0.64, treatment 3 of 6 (50.0%), control 8 of 25 (32.0%).
Baldia, 12/27/2022, prospective, multiple countries, peer-reviewed, median age 75.0, 26 authors, trial NCT04321265 (history).
Submit Corrections or Updates.
 risk of death, 12.0% lower, OR 0.88, p = 0.20, treatment 1,166, control 1,480, adjusted per study, multivariable, day 90, RR approximated with OR.
risk of death, 14.0% lower, OR 0.86, p = 0.20, treatment 1,166, control 1,480, adjusted per study, multivariable, day 30, RR approximated with OR.
Lapi, 7/30/2022, retrospective, Italy, peer-reviewed, 8 authors, excluded in exclusion analyses: substantial unadjusted confounding by indication likely.
Submit Corrections or Updates.
 risk of death/hospitalization, 75.0% higher, OR 1.75, p < 0.001, adjusted per study, late use, RR approximated with OR, late treatment result.
Lerner, 3/30/2022, retrospective, France, peer-reviewed, median age 69.2, 7 authors, study period 1 February, 2020 - 15 June, 2021.
Submit Corrections or Updates.
 risk of death, 26.9% higher, RR 1.27, p = 0.10, odds ratio converted to relative risk, weighted and trimmed, day 28, control prevalance approximated with overall prevalence.
Manjani, 10/31/2021, retrospective, USA, peer-reviewed, 6 authors, study period February 2020 - June 2020.
Submit Corrections or Updates.
 risk of death, 220.5% higher, RR 3.20, p = 0.001, treatment 64 of 388 (16.5%), control 7 of 136 (5.1%).
risk of mechanical ventilation, 434.5% higher, RR 5.34, p < 0.001, treatment 388, control 136.
risk of progression, 244.0% higher, OR 3.44, p < 0.005, treatment 132, control 136, triaged to higher level of care, high exposure, RR approximated with OR.
risk of progression, 201.0% higher, OR 3.01, p < 0.007, treatment 256, control 136, triaged to higher level of care, moderate exposure, RR approximated with OR.
hospitalization time, 100% higher, relative time 2.00, p < 0.001, treatment 388, control 136.
Ravichandran, 4/19/2022, Randomized Controlled Trial, India, peer-reviewed, 8 authors, this trial compares with another treatment - results may be better when compared to placebo, trial CTRI/2021/05/033544.
Submit Corrections or Updates.
 risk of no recovery, 42.5% higher, RR 1.43, p = 0.002, treatment 77 of 107 (72.0%), control 52 of 103 (50.5%), day 14.
risk of progression, 3925.2% higher, RR 40.25, p < 0.001, treatment 20 of 107 (18.7%), control 0 of 103 (0.0%), continuity correction due to zero event (with reciprocal of the contrasting arm), SpO2 ≤93.
recovery time, 133.3% higher, relative time 2.33, p < 0.001, treatment median 7.0 IQR 2.75 n=107, control median 3.0 IQR 1.0 n=103, fever.
recovery time, 75.0% higher, relative time 1.75, p < 0.001, treatment median 7.0 IQR 2.0 n=107, control median 4.0 IQR 2.0 n=103, myalgia.
recovery time, 75.0% higher, relative time 1.75, p < 0.001, treatment median 7.0 IQR 3.0 n=107, control median 4.0 IQR 1.0 n=103, cough.
risk of no viral clearance, 20.1% higher, RR 1.20, p = 0.19, treatment 43 of 60 (71.7%), control 37 of 62 (59.7%), day 7.
Ravichandran (B), 7/31/2021, retrospective, India, peer-reviewed, 6 authors, this trial compares with another treatment - results may be better when compared to placebo, trial ISRCTN11970082.
Submit Corrections or Updates.
 risk of oxygen therapy, 2700.0% higher, RR 28.00, p < 0.001, treatment 28 of 72 (38.9%), control 1 of 72 (1.4%), propensity score matching.
recovery time, 75.0% higher, relative time 1.75, p < 0.001, treatment median 7.0 IQR 1.0 n=72, control median 4.0 IQR 1.0 n=72, fever.
recovery time, 116.7% higher, relative time 2.17, p < 0.001, treatment median 6.5 IQR 3.25 n=72, control median 3.0 IQR 2.0 n=72, myalgia.
recovery time, 166.7% higher, relative time 2.67, p < 0.001, treatment median 8.0 IQR 2.0 n=72, control median 3.0 IQR 2.0 n=72, cough.
Sobhy, 4/19/2023, Double Blind Randomized Controlled Trial, Egypt, peer-reviewed, 6 authors, study period January 2022 - May 2022, this trial compares with another treatment - results may be better when compared to placebo, trial PACTR202202880140319.
Submit Corrections or Updates.
 risk of ICU admission, 110.0% higher, RR 2.10, p = 0.047, treatment 21 of 90 (23.3%), control 10 of 90 (11.1%).
risk of oxygen therapy, 110.0% higher, RR 2.10, p = 0.047, treatment 21 of 90 (23.3%), control 10 of 90 (11.1%).
hospitalization time, 35.7% higher, relative time 1.36, p = 0.01, treatment 90, control 90.
risk of no recovery, 33.3% higher, RR 1.33, p = 1.00, treatment 4 of 90 (4.4%), control 3 of 90 (3.3%), day 4, dyspnea.
risk of no recovery, 75.0% higher, RR 1.75, p = 0.25, treatment 14 of 90 (15.6%), control 8 of 90 (8.9%), day 4, fever.
risk of no recovery, 92.3% higher, RR 1.92, p = 0.04, treatment 25 of 90 (27.8%), control 13 of 90 (14.4%), day 4, lymphopenia.
risk of no recovery, 70.0% higher, RR 1.70, p = 0.03, treatment 34 of 90 (37.8%), control 20 of 90 (22.2%), day 4, cough.
Stufano, 4/18/2023, retrospective, Italy, peer-reviewed, 7 authors.
Submit Corrections or Updates.
 risk of long COVID, 18.5% higher, RR 1.19, p = 0.62, treatment 11 of 23 (47.8%), control 23 of 57 (40.4%).
Effect extraction follows pre-specified rules as detailed above and gives priority to more serious outcomes. For pooled analyses, the first (most serious) outcome is used, which may differ from the effect a paper focuses on. Other outcomes are used in outcome specific analyses.
Blanc, 5/2/2020, retrospective, France, preprint, mean age 84.1, 22 authors, study period 2 March, 2020 - 8 April, 2020.
Submit Corrections or Updates.
 risk of case, 51.4% higher, OR 1.51, p = 0.19, treatment 60, control 119, RR approximated with OR.
Campbell, 5/5/2022, retrospective, USA, peer-reviewed, 4 authors, study period 2 March, 2020 - 14 December, 2020.
Submit Corrections or Updates.
 risk of death, 1.0% higher, OR 1.01, p = 0.43, treatment 2,074, control 20,311, adjusted per study, propensity score weighting, multivariable, day 60, RR approximated with OR.
risk of death, no change, OR 1.00, p = 0.86, treatment 2,074, control 20,311, adjusted per study, propensity score weighting, multivariable, day 30, RR approximated with OR.
Chandan, 4/29/2021, retrospective, United Kingdom, peer-reviewed, mean age 65.4, 24 authors, study period 30 January, 2020 - 31 July, 2020, this trial compares with another treatment - results may be better when compared to placebo, this trial uses multiple treatments in the treatment arm (combined with codeine or dihydrocodeine) - results of individual treatments may vary.
Submit Corrections or Updates.
 risk of death, 17.6% higher, HR 1.18, p = 0.35, treatment 71 of 8,595 (0.8%), control 79 of 8,595 (0.9%), adjusted per study, inverted to make HR<1 favor treatment, propensity score matching, multivariable.
risk of case, 26.6% higher, HR 1.27, p = 0.17, treatment 8,595, control 8,595, adjusted per study, inverted to make HR<1 favor treatment, propensity score matching, multivariable.
Gálvez-Barrón, 4/14/2021, retrospective, Spain, peer-reviewed, mean age 86.8, 13 authors, study period 12 March, 2020 - 2 May, 2020.
Submit Corrections or Updates.
 risk of death, 47.0% higher, OR 1.47, p = 0.42, treatment 43, control 60, RR approximated with OR.
risk of severe case, 23.0% lower, OR 0.77, p = 0.55, treatment 43, control 60, RR approximated with OR.
Kim, 2/21/2023, retrospective, South Korea, peer-reviewed, mean age 55.8, 4 authors, this trial compares with another treatment - results may be better when compared to placebo.
Submit Corrections or Updates.
 risk of death, 71.4% higher, RR 1.71, p = 0.34, treatment 12 of 162 (7.4%), control 7 of 162 (4.3%), propensity score matching.
risk of mechanical ventilation, 14.3% higher, RR 1.14, p = 1.00, treatment 8 of 162 (4.9%), control 7 of 162 (4.3%), propensity score matching.
risk of ICU admission, 40.0% lower, RR 0.60, p = 0.72, treatment 3 of 162 (1.9%), control 5 of 162 (3.1%), NNT 81, propensity score matching.
risk of oxygen therapy, 9.1% higher, RR 1.09, p = 0.87, treatment 24 of 162 (14.8%), control 22 of 162 (13.6%), propensity score matching.
Kolin, 11/17/2020, retrospective, United Kingdom, peer-reviewed, 4 authors.
Submit Corrections or Updates.
 risk of case, 23.0% higher, RR 1.23, p = 0.009.
Leal, 8/16/2021, retrospective, United Kingdom, peer-reviewed, 5 authors, study period 16 March, 2020 - 1 February, 2021.
Submit Corrections or Updates.
 risk of case, 7.0% lower, OR 0.93, p = 0.004, RR approximated with OR.
MacFadden, 3/29/2022, retrospective, Canada, peer-reviewed, 9 authors, study period 15 January, 2020 - 31 December, 2020.
Submit Corrections or Updates.
 risk of case, 48.0% higher, OR 1.48, p < 0.001, RR approximated with OR.
Moreno-Martos, 1/24/2022, retrospective, multiple countries, peer-reviewed, 24 authors, study period January 2020 - June 2020.
Submit Corrections or Updates.
 risk of hospitalization, 29.3% higher, RR 1.29, p < 0.001, treatment 10,367, control 89,523, meta analysis of all databases combined.
risk of hospitalization, 51.7% higher, RR 1.52, p < 0.001, treatment 103 of 178 (57.9%), control 196 of 514 (38.1%), US CU-AMC.
risk of hospitalization, 5.1% higher, RR 1.05, p = 0.57, treatment 87 of 144 (60.4%), control 360 of 626 (57.5%), US CUIMC.
risk of hospitalization, 21.8% lower, RR 0.78, p = 0.02, treatment 64 of 319 (20.1%), control 1,585 of 6,181 (25.6%), NNT 18, US HealthVerity.
risk of hospitalization, 16.5% higher, RR 1.16, p < 0.001, treatment 2,597 of 4,983 (52.1%), control 28,320 of 63,279 (44.8%), US IQVIA OpenClaims.
risk of hospitalization, 57.0% higher, RR 1.57, p < 0.001, treatment 1,090 of 1,868 (58.4%), control 3,414 of 9,188 (37.2%), US Optum EHR.
risk of hospitalization, 47.2% higher, RR 1.47, p < 0.001, treatment 1,397 of 2,875 (48.6%), control 3,214 of 9,735 (33.0%), US VA-OMOP.
Oh, 6/24/2021, retrospective, South Korea, peer-reviewed, 5 authors, study period 1 January, 2020 - 4 June, 2020.
Submit Corrections or Updates.
 risk of death, 1.9% lower, RR 0.98, p = 0.97, treatment 58, control 7,655, adjusted per study, odds ratio converted to relative risk, multivariable, control prevalance approximated with overall prevalence.
Park, 3/3/2021, retrospective, South Korea, peer-reviewed, 5 authors, this trial compares with another treatment - results may be better when compared to placebo.
Submit Corrections or Updates.
 risk of death, 24.8% lower, HR 0.75, p = 0.46, treatment 12 of 397 (3.0%), control 16 of 397 (4.0%), NNT 99, inverted to make HR<1 favor treatment, propensity score matching.
risk of mechanical ventilation, 37.5% lower, HR 0.62, p = 0.42, treatment 5 of 397 (1.3%), control 8 of 397 (2.0%), NNT 132, inverted to make HR<1 favor treatment, propensity score matching.
Reese, 4/20/2021, retrospective, USA, preprint, 23 authors.
Submit Corrections or Updates.
 risk of death, 61.0% higher, HR 1.61, p < 0.001, treatment 20,826, control 20,826, propensity score matching, Cox proportional hazards, Table S58.
risk of severe case, 816.0% higher, OR 9.16, p < 0.001, treatment 20,826, control 20,826, propensity score matching, Table S50, RR approximated with OR.
Ritsinger, 4/28/2023, retrospective, Sweden, peer-reviewed, mean age 79.8, 8 authors, study period 1 January, 2020 - 9 September, 2021.
Submit Corrections or Updates.
 risk of death, 21.0% higher, HR 1.21, p < 0.001, treatment 24,641, control 20,225.
Xie (B), 7/13/2022, retrospective, USA, peer-reviewed, 9 authors, study period 1 February, 2020 - 31 October, 2020, this trial compares with another treatment - results may be better when compared to placebo.
Submit Corrections or Updates.
 risk of hospitalization, 4.8% higher, HR 1.05, p = 0.83, inverted to make HR<1 favor treatment, Open Claims, PharMetrics Plus, both periods combined.
risk of case, 3.5% lower, HR 0.97, p = 0.82, inverted to make HR<1 favor treatment, Open Claims, PharMetrics Plus, both periods combined.
Footnotes
a.
Smoking was known to cause lung cancer since at least 1939, but this was not widely recognized in the US until 1964, 25 years later. Surgeon general Leroy Burney tried publicizing the danger starting in 1957, with limited success. Surgeon general Luther Terry, appointed in 1961, prompted by President Kennedy in 1962 amid pressure from health advocates, finally got recognition in 1964. The 1964 report reviewed 7,000+ studies, but these could (and should) have been reviewed and acted upon in real-time as they were published. Historians attribute the 25 year delay to an industry campaign to manufacture doubt and controversy, through tactics like funding biased and fraudulent research from "independent" organizations, attacking scientists, and political lobbying. The success of the industry campaign is only possible because officials did not analyze the research in detail. Fraudulent industry research supported prior failures, but would have been called out as fraudulent by officials that analyzed and understood the research in real-time.
b.
When administered late in infection, HCQ may enhance viral egress by further increasing lysosomal pH beyond the effect of ORF3a's water channel activity, thereby promoting lysosomal exocytosis, inactivating degradative enzymes, and facilitating the release of SARS-CoV-2 particles into the extracellular environment107,108. Research also suggests potential cardioprotective effects at lower doses, but cardiotoxicity with excessive dosage109. Bobrowski et al. also indicate negative effects if HCQ and remdesivir are combined.
c.
Peters (B) et al. is subject to confounding by calendar-time (SOC evolved rapidly early in the pandemic, the linear covariate does not reflect non-linear SOC changes and hospital specific effects), hospital type (non-treatment hospitals were tertiary university centers), confounding by indication (4/7 hospitals initiated treatment on deterioration), immortal-time bias for as-treated (exposure assigned after baseline), significant differences for other experimental treatments, potential overadjustment from collider bias (steroid use and indication bias), limited baseline severity information, differences in hospice referral propensity across hospitals, unadjusted difference in time from onset to admission, difference in PCR positivity, and other factors. Mahévas et al. is subject to confounding by hospital (treatment highly dependent on the hospital, different SOC/ICU transfer practices, not included in PS), immortal time (only partly addressed in sensitivity analysis), co-treatment differences, calendar-time (SOC evolved rapidly early in the pandemic), binary coding for age (age ≥65 despite steep age-risk gradient), residual imbalance (variables dropped from PS), a composite outcome dependent on hospital triage/capacity, and other factors.
References