Hydroxychloroquine

What it is

Hydroxychloroquine is a synthetic 4-aminoquinoline, a chemical relative of chloroquine and of the antimalarial quinine family, taken as a tablet. It is a weak base that concentrates in acidic compartments inside cells, has a very large volume of distribution and an unusually long terminal half-life of 40 to 50 days in whole blood. It is metabolised mainly by CYP2C8, CYP3A4 and CYP2D6. In the United States it is approved for malaria, rheumatoid arthritis, systemic lupus erythematosus and chronic discoid lupus erythematosus.

What the research says

In lupus the evidence is good. A randomised withdrawal trial in 47 people with quiescent lupus found 16 of 22 on placebo flared versus 9 of 25 who stayed on the drug, and a systematic review of 95 articles graded the evidence for preventing flares and improving long-term survival as high. In rheumatoid arthritis a Cochrane review of four placebo-controlled trials found a moderate benefit. For COVID-19 the evidence is large and clearly negative: the RECOVERY trial in 4716 hospitalised patients found no mortality benefit and more progression to ventilation or death, a 28-trial meta-analysis found increased mortality (odds ratio 1.11, 95% CI 1.02 to 1.20), and a Cochrane review rated the no-effect-on-death conclusion high certainty while finding adverse events tripled. The signature long-term harm is irreversible retinal damage, found in 7.5% of people who had taken it for at least five years.

Evidence grade: Well established.

How it works

Drug class: 4-aminoquinoline antimalarial and antirheumatic (disease-modifying antirheumatic drug)

Hydroxychloroquine is a weak base that accumulates inside lysosomes, the cell's acidic recycling compartments, and raises their pH. That interferes with lysosomal activity and autophagy, with membrane stability, and with signalling that depends on acidic compartments, so less inflammatory cytokine is made and some immune co-stimulatory signals are dampened. The authors of a 2020 pharmacology review in Nature Reviews Rheumatology are explicit that the mechanism is still being worked out, and the Plaquenil label says the mechanism in rheumatic disease is not fully known. (Source 1)

What it is used for

• A randomised withdrawal trial found flares in 16 of 22 people switched to placebo versus 9 of 25 who continued hydroxychloroquine, and a systematic review graded the evidence for preventing flares and increasing long-term survival as high. Evidence for an effect on severe lupus activity, lipids and subclinical atherosclerosis was graded weak by the same review. Evidence: established. (Source 2)
• A Cochrane review of four placebo-controlled trials in 592 patients found a statistically significant but moderate benefit at six months (standardised mean differences -0.33 to -0.52) with no excess of withdrawals for toxicity. Evidence: established. (Source 3)
• This is an approved indication on the FDA-approved label, version effective 2026-04-09, which states the drug is indicated for chronic discoid lupus erythematosus in adults. That is a label position and its date, not evidence of benefit: the systematic review of antimalarials in lupus that we read covers systemic lupus rather than isolated discoid disease, and we reached no trial or review specific to this indication. Evidence: unknown. (Source 4)
• An approved indication of long standing on the FDA-approved label, version effective 2026-04-09, restricted by that label to uncomplicated malaria and to areas where chloroquine resistance is not reported. What we hold for this use is the label's position on that date plus its dosing instructions; we reached no randomised trial or systematic review of hydroxychloroquine for malaria, so this is not evidence of benefit. Evidence: unknown. (Source 5)
• RECOVERY randomised 1561 patients to hydroxychloroquine and 3155 to usual care: 28-day death was 27.0% versus 25.0%, and among those not ventilated at baseline, progression to ventilation or death was more common on the drug (30.7% vs 26.9%). A meta-analysis of 26 trials found an odds ratio for death of 1.11 (95% CI 1.02-1.20) and Cochrane rated the no-benefit finding high certainty. Evidence: not-supported. (Source 6)
• A double-blind placebo-controlled trial in 821 exposed adults found illness compatible with COVID-19 in 11.8% on hydroxychloroquine versus 14.3% on placebo, an absolute difference of -2.4 percentage points (95% CI -7.0 to 2.2, P=0.35), while side effects more than doubled (40.1% vs 16.8%). Evidence: not-supported. (Source 7)
• A UK Biobank cohort of 6793 people with rheumatoid arthritis or lupus followed a median 13.78 years found a lower rate of new diabetes among hydroxychloroquine users (fully adjusted hazard ratio 0.87, 95% CI 0.79-0.96). This is observational and cannot establish cause; the drug's known ability to cause hypoglycaemia is the plausible mechanism and also a harm. Evidence: limited. (Source 8)

Interactions

• Antacids and kaolin (including magnesium- and aluminium-containing antacid supplements) (label): Antacids and kaolin reduce the absorption of chloroquine, and the label says the same interaction with hydroxychloroquine cannot be ruled out, advising at least four hours between doses. The direct evidence is for chloroquine, not hydroxychloroquine. (Source 9)
• Food and milk (label): The label directs taking hydroxychloroquine with food or milk. Absolute oral bioavailability in fasting conditions is 79%, and absorption in rheumatoid arthritis patients varied widely (30 to 100% of the dose). (Source 5)
• Insulin and other antidiabetic drugs (label): Hydroxychloroquine lowers blood glucose in its own right, so it adds to insulin and oral diabetes drugs and can cause hypoglycaemia; the label says a dose reduction of the diabetes drug may be needed. (Source 10)
• Other QT-prolonging or arrhythmogenic drugs (label): Hydroxychloroquine itself lengthens the QT interval, so combining it with another QT-prolonging drug raises the risk of ventricular arrhythmia; the label says it is not recommended in that situation. (Source 10)
• Digoxin (label): Taking the two together can raise digoxin blood levels, so digoxin levels need watching. (Source 9)
• Methotrexate (label): The two are often prescribed together in rheumatoid arthritis; the label notes the combination may increase the incidence of adverse reactions. (Source 9)
• Tamoxifen (pharmacokinetic study): Concurrent tamoxifen was one of the strongest risk factors for hydroxychloroquine retinal toxicity in a 2361-patient case-control study, with an odds ratio of 4.59. (Source 11)
• Cimetidine (pharmacokinetic study): Cimetidine doubled exposure to chloroquine, and the label says an interaction with hydroxychloroquine cannot be ruled out and advises avoiding the combination. Note that the label's cimetidine section ends without a full stop, so on the label the instruction to avoid cimetidine runs straight into the next heading; it is reproduced here as printed. (Source 9)
• Rifampicin (case reports): Loss of hydroxychloroquine efficacy has been reported when rifampicin was given at the same time; the label advises avoiding the combination. (Source 9)

Stopping it

• There is direct randomised evidence about stopping: in 47 people with quiescent lupus, switching to placebo for 24 weeks produced flares in 16 of 22, versus 9 of 25 who stayed on the drug, and the time to flare was shorter. (Source 2)
• The drug leaves the body extremely slowly, so stopping is not an on-off switch: the terminal half-life in whole blood is 40 to 50 days, meaning measurable drug persists for months. (Source 12)
• Retinal damage is not reversed by stopping, and the label warns changes can keep progressing after the drug is withdrawn. (Source 13)
• Among 78 people in the cardiac case review who were taken off the drug, 44.9% recovered normal heart function, 12.9% had irreversible damage and 30.8% died, so stopping early matters for cardiac toxicity. (Source 14)
• The authors of the cardiac review argue that withdrawal is required once cardiac manifestations appear, because the damage may be permanent. (Source 15)
• There is no dependence or withdrawal syndrome described for hydroxychloroquine in the sources read; what happens on stopping is the return of the underlying disease. Stopping is nonetheless common in practice, which is what prompted the lupus systematic review in the first place. (Source 16)

What goes wrong

Among hospitalised COVID-19 patients not on a ventilator at baseline, hydroxychloroquine increased progression to invasive ventilation or death. (Source 6)

• Randomized trial, High certainty.
• Size: Subgroup not ventilated at baseline within the 4716-patient comparison.
• Who: Adults hospitalised with COVID-19 and not receiving invasive mechanical ventilation at randomisation.
• How long: 28 days.
• Result: Invasive mechanical ventilation or death 30.7% versus 26.9% (risk ratio 1.14, 95% CI 1.03-1.27), an absolute excess of 3.8 percentage points, about 26 people treated per extra event.
• Funding: UK Research and Innovation and the National Institute for Health Research.

those in the hydroxychloroquine group had a higher frequency of invasive mechanical ventilation or death (30.7% vs. 26.9%; risk ratio, 1.14; 95% CI, 1.03 to 1.27)

A meta-analysis of published and unpublished COVID-19 trials found hydroxychloroquine was associated with increased all-cause mortality. (Source 17)

• Meta-analysis, Moderate certainty.
• Size: 10,012 patients across 26 trials (14 of them previously unpublished)
• Who: Mostly hospitalised adults with COVID-19; results dominated by RECOVERY and WHO SOLIDARITY (67% of the sample)
• How long: Trial-specific.
• Result: Odds ratio for all-cause mortality 1.11 (95% CI 1.02 to 1.20), heterogeneity I-squared 0%; chloroquine odds ratio 1.77 with a very wide interval (95% CI 0.15-21.13) in 4 trials and 307 patients.
• Funding: Not stated in the abstract read; investigator collaboration that requested unpublished data.

The combined OR on all-cause mortality for hydroxychloroquine is 1.11 (95% CI: 1.02, 1.20; I² = 0%; 26 trials; 10,012 patients)

In COVID-19 trials hydroxychloroquine roughly tripled adverse events, though serious adverse events were not clearly increased. (Source 18)

• Systematic review, Moderate certainty.
• Size: 1394 participants across 6 trials for any adverse event; 1004 across 6 trials for serious adverse events.
• Who: Adults with COVID-19.
• How long: Trial-specific.
• Result: Any adverse event relative risk 2.90 (95% CI 1.49 to 5.64), moderate certainty; serious adverse events RR 0.82 (95% CI 0.37 to 1.79), low certainty.
• Funding: Cochrane Infectious Diseases Group (independent)

HCQ probably results in an almost three-fold increased risk of adverse events (RR 2.90, 95% CI 1.49 to 5.64; 1394 participants; 6 trials; moderate-certainty evidence)

In the post-exposure prophylaxis trial, side effects were more than twice as common on hydroxychloroquine as on placebo. (Source 7)

• Randomized trial, Moderate certainty.
• Size: 821 participants.
• Who: Healthy exposed adults, 66% healthcare workers, mean age about 40.
• How long: 5 days of drug, 14 days of follow-up.
• Result: Side effects 40.1% versus 16.8%, an absolute excess of 23.3 percentage points, roughly 4 people treated per extra person with a side effect; no serious adverse reactions reported.
• Funding: Private philanthropy, not industry.

Side effects were more common with hydroxychloroquine than with placebo (40.1% vs. 16.8%), but no serious adverse reactions were reported.

Among people who had taken hydroxychloroquine for at least five years, 7.5% had retinal toxicity, and risk rose steeply with dose and duration. (Source 11)

• Case-control study, Moderate certainty.
• Size: 2361 patients screened within a health system of about 3.4 million members.
• Who: Adults who had used hydroxychloroquine continuously for at least five years, assessed with visual fields or spectral-domain OCT.
• How long: At least 5 years of use; some beyond 20 years.
• Result: Overall prevalence 7.5%; odds ratio 5.67 (95% CI 4.14-7.79) for more than 5.0 mg/kg daily and 3.22 (2.20-4.70) for more than 10 years of use; kidney disease odds ratio 2.08 (1.44-3.01); concurrent tamoxifen odds ratio 4.59 (2.05-10.27)
• Funding: Not stated in the abstract read (Kaiser Permanente Northern California)

The overall prevalence of hydroxychloroquine retinopathy was 7.5% but varied with daily consumption (odds ratio, 5.67; 95% CI, 4.14-7.79 for >5.0 mg/kg) and with duration of use (odds ratio, 3.22; 95% CI, 2.20-4.70 for >10 years).

Even at a dose of 4 to 5 mg/kg, retinal toxicity reached almost 20% after twenty years of use. (Source 11)

• Case-control study, Moderate certainty.
• Size: 2361 long-term users.
• Who: Adults on hydroxychloroquine for at least five years.
• How long: Up to and beyond 20 years.
• Result: Prevalence under 2% in the first 10 years at 4.0-5.0 mg/kg daily, rising to almost 20% after 20 years.
• Funding: Not stated in the abstract read.

For daily consumption of 4.0 to 5.0 mg/kg, the prevalence of retinal toxicity remained less than 2% within the first 10 years of use but rose to almost 20% after 20 years of use.

The FDA label states retinal damage from hydroxychloroquine is irreversible and related to cumulative dose and duration, and names risk factors. (Source 13)

• Official position, Moderate certainty.
• Size: Not applicable (regulatory warning)
• Who: People on long-term hydroxychloroquine, with extra risk from dosages of 5 mg/kg or more of actual body weight, more than five years of use, renal impairment, concomitant tamoxifen and existing macular disease.
• How long: Long-term use.
• Result: No rate given in the label; it directs a baseline eye examination in the first year and annual examinations for higher-risk patients.
• Funding: FDA-approved labelling (position of the regulator, label version effective 2026-04-09)

Irreversible retinal damage was observed in some patients treated with hydroxychloroquine sulfate and it is related to cumulative dosage and treatment duration.

A systematic review of published cardiac cases found conduction disorders in 85% of 127 reported patients, usually after years of treatment, but could not quantify how often this happens. (Source 14)

• Systematic review, Very low certainty.
• Size: 86 articles reporting individual cases or short series, covering 127 patients (65.4% female); 39.4% on hydroxychloroquine, 58.3% on chloroquine.
• Who: People treated with chloroquine or hydroxychloroquine, median 7 years of treatment and high cumulative dose.
• How long: Median 7 years (range 3 days to 35 years)
• Result: Conduction disorders in 85%; heart failure 26.8%; ventricular hypertrophy 22%; valvular dysfunction 7.1%; pulmonary arterial hypertension 3.9%. Of 78 withdrawn from treatment, 44.9% recovered normal heart function, 12.9% had irreversible damage and 30.8% died.
• Funding: Not stated in the abstract read.

Limit of this finding: These percentages are shares of 127 patients whose cases were written up and published, not rates in people taking the drug. They cannot tell you how likely heart damage is: the review's own authors say the risk could not be quantified.

Conduction disorders were the main side effect reported, affecting 85% of patients.

The FDA label records fatal and life-threatening cardiotoxicity, QT prolongation and torsades de pointes with hydroxychloroquine. (Source 19)

• Official position, Low certainty.
• Size: Not applicable (regulatory warning drawing on post-marketing reports)
• Who: People on hydroxychloroquine, with particular risk where QT is already prolonged, in bradycardia below 50 bpm, or with uncorrected hypokalaemia or hypomagnesaemia.
• How long: Acute and chronic treatment.
• Result: No rates given; the label advises avoiding the drug with congenital or acquired QT prolongation and correcting electrolytes first.
• Funding: FDA-approved labelling (position of the regulator, 2026-04-09)

Fatal and life-threatening cases of cardiotoxicity, including cardiomyopathy, have been reported in patients treated with PLAQUENIL.

Hydroxychloroquine can cause severe, life-threatening hypoglycaemia even in people not taking diabetes medicines. (Source 20)

• Official position, Low certainty.
• Size: Not applicable (regulatory warning)
• Who: Anyone taking hydroxychloroquine, with or without antidiabetic drugs.
• How long: Any point in treatment.
• Result: No rate given; the label directs measuring blood glucose in anyone with symptoms suggesting hypoglycaemia.
• Funding: FDA-approved labelling (position of the regulator, 2026-04-09)

PLAQUENIL can cause severe and potentially life-threatening hypoglycemia, in the presence or absence of antidiabetic agents

Post-marketing reports attributed to 4-aminoquinolines including hydroxychloroquine cover bone marrow suppression, fulminant hepatic failure, proximal myopathy, seizures and neuropsychiatric reactions, with no reliable frequency. (Source 21)

• Official position, Very low certainty.
• Size: Spontaneous reports from a population of uncertain size.
• Who: People taking hydroxychloroquine or related 4-aminoquinolines.
• How long: Post-approval use.
• Result: No frequencies can be estimated; the label states this explicitly.
• Funding: FDA-approved labelling (position of the regulator, 2026-04-09)

Because these reactions are reported voluntarily from a population of uncertain size, it is not always possible to reliably estimate their frequency or establish a causal relationship to drug exposure

Post-marketing reports for hydroxychloroquine include seizures, movement disorders and neuropsychiatric reactions up to psychosis, suicidal ideation and suicidal behaviour, as well as severe skin reactions including Stevens-Johnson syndrome and toxic epidermal necrolysis. (Source 22)

• Case series, Very low certainty.
• Size: Spontaneous reports from a population of uncertain size.
• Who: People treated with 4-aminoquinoline drugs including hydroxychloroquine.
• How long: Not stated.
• Result: No frequencies are given. The label lists, among nervous system disorders, ataxia, dizziness, headache, seizure and extrapyramidal disorders; among neuropsychiatric disorders, psychosis, suicidal ideation, suicidal behavior, depression, hallucinations, mania and sleep disorders; and among skin disorders, DRESS syndrome, Stevens-Johnson syndrome and toxic epidermal necrolysis.
• Funding: FDA-approved labelling.

Limit of this finding: These are voluntary reports with no denominator, so they show what has been reported, not how often any of it happens. The label says so itself. The label also carries a separate warning section on neuropsychiatric reactions including suicidality.

-Neuropsychiatric disorders: Affect/emotional lability, irritability, nervousness, psychosis, suicidal ideation, suicidal behavior, depression, hallucinations, anxiety, agitation, confusion, delusions, paranoia, mania, and sleep disorders (insomnia, night terrors, nightmares)

What the evidence supports

Continuing hydroxychloroquine in people with quiescent lupus roughly halved the chance of a clinical flare compared with switching to placebo. (Source 2)

• Randomized trial, Moderate certainty.
• Size: 47 patients (25 continued hydroxychloroquine, 22 placebo)
• Who: Adults with clinically stable systemic lupus erythematosus; 10 in each group also on prednisone.
• How long: 24 weeks.
• Result: Flares in 16 of 22 on placebo (72.7%) versus 9 of 25 continuing the drug (36.0%); relative risk 2.5 (95% CI 1.08-5.58) for placebo; absolute difference about 37 percentage points, around 3 people treated per flare prevented. Severe exacerbation requiring withdrawal 5 of 22 versus 1 of 25, relative risk 6.1 with a very wide interval (95% CI 0.72-52.44)
• Funding: Not stated in the abstract read (Canadian Hydroxychloroquine Study Group)

Limit of this finding: The raw counts in the paper (16 of 22 versus 9 of 25) work out to a ratio of about 2.0, while the published relative risk is 2.5. The 2.5 comes from the trial's own analysis, which accounts for how long people went before flaring; both figures are from the same paper.

The relative risk of a clinical flare-up, defined as the development of specific clinical manifestations of systemic lupus erythematosus or an increase in their severity, was 2.5 times higher (95 percent confidence interval, 1.08 to 5.58) in the patients taking placebo than in those continuing to take hydroxychloroquine (16 of 22 patients vs. 9 of 25 had flare-ups)

A systematic review using GRADE found high-level evidence that antimalarials prevent lupus flares and increase long-term survival in lupus. (Source 23)

• Systematic review, High certainty.
• Size: 95 articles including randomised trials and observational studies.
• Who: People with systemic lupus erythematosus, including pregnant women.
• How long: Literature 1982 to 2007.
• Result: Graded high for flare prevention and long-term survival; moderate for protection against irreversible organ damage, thrombosis and bone mass loss.
• Funding: Not stated in the abstract read.

High levels of evidence were found that AMs prevent lupus flares and increase long-term survival of patients with SLE; moderate evidence of protection against irreversible organ damage, thrombosis and bone mass loss.

In rheumatoid arthritis, hydroxychloroquine beat placebo on joint counts, pain and global assessment at six months, with a moderate effect size. (Source 3)

• Systematic review, Moderate certainty.
• Size: 592 patients (300 hydroxychloroquine, 292 placebo) across four trials.
• Who: Adults with rheumatoid arthritis.
• How long: Six-month endpoint.
• Result: Standardised mean differences across outcome measures -0.33 to -0.52, all statistically significant; ESR also improved; withdrawals for lack of efficacy more frequent on placebo; no difference in withdrawals for toxicity.
• Funding: Cochrane Musculoskeletal Group (independent)

A statistically significant benefit was observed when hydroxychloroquine was compared to placebo. The standardized mean differences for the various outcome measures ranged from -0.33 to -0.52, and were statistically significant.

Among people with rheumatoid arthritis or systemic lupus erythematosus, hydroxychloroquine use was associated with a lower rate of new type 2 diabetes, an observational association that cannot establish cause. (Source 8)

• Cohort study, Low certainty.
• Size: 6793 participants with rheumatoid arthritis or lupus at baseline, drawn from 502,392 UK Biobank participants; 537 developed diabetes.
• Who: Adults in the UK Biobank with rheumatoid arthritis or systemic lupus erythematosus, diabetic at baseline excluded.
• How long: Median 13.78 years.
• Result: New diabetes in 33 of hydroxychloroquine users (6.36%) versus 504 of non-users (8.03%); fully adjusted hazard ratio 0.87 (95% CI 0.79-0.96, P=0.005)
• Funding: Not stated in the article read.

Limit of this finding: Two things in the paper do not line up. It gives 13.78 years as both the mean and the median follow-up, so one of those labels is wrong. And although everyone in the cohort had either rheumatoid arthritis or lupus, the paper's own conclusion narrows the result to rheumatoid arthritis patients, which the quoted results do not support either way. Only 33 of the people who developed diabetes were taking hydroxychloroquine, so the estimate is fragile, and nothing here applies to people without these conditions.

In complete multivariate model hazard ratio for hydroxychloroquine was 0.87 (95% confidence interval, 0.79- 0.96, P=0.005).

Decades of use and published epidemiological studies have not identified a drug-associated risk of major birth defects or miscarriage with hydroxychloroquine, and untreated lupus itself carries pregnancy risk. (Source 24)

• Official position, Low certainty.
• Size: Not applicable (regulatory summary of published epidemiologic and clinical studies)
• Who: Pregnant women taking hydroxychloroquine.
• How long: Pregnancy.
• Result: No effect estimate given; the label notes US background risks of 2-4% for major birth defects and 15-20% for miscarriage, and that animal reproduction studies were not conducted.
• Funding: FDA-approved labelling (position of the regulator, 2026-04-09)

Prolonged clinical experience over decades of use and available data from published epidemiologic and clinical studies with PLAQUENIL use in pregnant women have not identified a drug-associated risk of major birth defects, miscarriage, or adverse maternal, or fetal outcomes

What the evidence does not support

The same systematic review graded the evidence for an effect on severe lupus activity, lipid levels and subclinical atherosclerosis as weak. (Source 23)

• Systematic review, Low certainty.
• Size: 95 articles.
• Who: People with systemic lupus erythematosus.
• How long: Literature 1982 to 2007.
• Result: Evidence described as weak for severe lupus activity, lipids and subclinical atherosclerosis.
• Funding: Not stated in the abstract read.

By contrast, evidence supporting an effect on severe lupus activity, lipid levels and subclinical atherosclerosis was weak.

Hydroxychloroquine did not reduce 28-day death in hospitalised COVID-19 patients. (Source 6)

• Randomized trial, High certainty.
• Size: 4716 patients (1561 hydroxychloroquine, 3155 usual care)
• Who: Adults hospitalised with COVID-19 in the UK.
• How long: 28 days.
• Result: Death 27.0% versus 25.0% (rate ratio 1.09, 95% CI 0.97-1.23, P=0.15); discharge alive by 28 days 59.6% versus 62.9% (rate ratio 0.90, 95% CI 0.83-0.98)
• Funding: UK Research and Innovation and the National Institute for Health Research (not industry)

Death within 28 days occurred in 421 patients (27.0%) in the hydroxychloroquine group and in 790 (25.0%) in the usual-care group (rate ratio, 1.09; 95% confidence interval [CI], 0.97 to 1.23; P = 0.15).

Cochrane rated as high certainty the conclusion that hydroxychloroquine makes little or no difference to death in COVID-19. (Source 25)

• Systematic review, High certainty.
• Size: 8208 participants across 9 trials for the mortality outcome; 12 trials and 8569 participants overall.
• Who: Adults with COVID-19, nine trials in hospital and three in ambulatory care.
• How long: Trial-specific, searches to 15 September 2020.
• Result: Death from any cause relative risk 1.09 (95% CI 0.99 to 1.19), high-certainty evidence; progression to mechanical ventilation RR 1.11 (0.91-1.37), moderate certainty.
• Funding: Cochrane Infectious Diseases Group (independent)

HCQ makes little or no difference to death due to any cause (RR 1.09, 95% CI 0.99 to 1.19; 8208 participants; 9 trials; high-certainty evidence).

Hydroxychloroquine did not prevent COVID-19 when started within four days of a high-risk exposure. (Source 7)

• Randomized trial, Moderate certainty.
• Size: 821 asymptomatic participants (414 hydroxychloroquine, 407 placebo)
• Who: Adults with household or occupational exposure to a confirmed case; 87.6% high-risk exposures.
• How long: 14 days.
• Result: New illness compatible with COVID-19 in 11.8% versus 14.3%; absolute difference -2.4 percentage points (95% CI -7.0 to 2.2, P=0.35)
• Funding: Private philanthropy (David Baszucki and Jan Ellison Baszucki and others), not industry.

The incidence of new illness compatible with Covid-19 did not differ significantly between participants receiving hydroxychloroquine (49 of 414 [11.8%]) and those receiving placebo (58 of 407 [14.3%]); the absolute difference was -2.4 percentage points (95% confidence interval, -7.0 to 2.2; P = 0.35).

Where the research disagrees

How common hydroxychloroquine retinopathy really is

• Melles and Marmor (JAMA Ophthalmology 2014), retrospective case-control study of 2361 patients screened with visual fields or spectral-domain OCT, which found 7.5% prevalence: Prior estimations of risk were low but were based largely on short-term users or severe retinal toxicity (bull's eye maculopathy). The risk may be much higher because retinopathy can be detected earlier when using more sensitive screening techniques. (Source 26)
• Ruiz-Irastorza and colleagues (Annals of the Rheumatic Diseases 2010 systematic review), GRADE-rated systematic review of 95 articles published up to 2007, before sensitive OCT screening was routine: Toxicity related to AMs is infrequent, mild and usually reversible, with HCQ having a safer profile. (Source 23)

Whether hydroxychloroquine merely fails in COVID-19 or actively increases mortality

• Cochrane (Singh and colleagues, 2021), systematic review with GRADE of 12 trials and 8569 participants, searches to September 2020: HCQ makes little or no difference to death due to any cause (RR 1.09, 95% CI 0.99 to 1.19; 8208 participants; 9 trials; high-certainty evidence). (Source 25)
• Axfors and colleagues (Nature Communications 2021), meta-analysis of 26 trials including 14 unpublished, with mortality data requested directly from investigators: The combined OR on all-cause mortality for hydroxychloroquine is 1.11 (95% CI: 1.02, 1.20; I² = 0%; 26 trials; 10,012 patients) (Source 17)

How much

• Reference intake: There is no dietary reference intake for a prescription medicine; the dose is set by the prescriber. As a position, the FDA-approved label (version effective 2026-04-09) gives an initial dose of 400 to 600 mg daily for rheumatoid arthritis with a chronic dose of 200 to 400 mg daily, and 200 mg once daily or 400 mg daily for systemic lupus erythematosus. (Source 27)
• Upper limit: The label does not state a single maximum dose but ties dose to body weight for safety: it states that daily doses above 5 mg/kg of actual weight increase the incidence of retinopathy. The retinopathy case-control study is the evidence behind that: the odds ratio for retinal toxicity above 5.0 mg/kg daily was 5.67 (95% CI 4.14-7.79). (Source 27)
• Studied: The lupus withdrawal trial kept 25 clinically stable patients on their existing hydroxychloroquine dose and switched 22 to placebo for 24 weeks. (Source 28)
• Studied: The COVID-19 post-exposure prophylaxis trial gave 800 mg once, then 600 mg 6 to 8 hours later, then 600 mg daily for four more days. (Source 29)
• Studied: For malaria prophylaxis the label's position is 400 mg once a week in adults, started two weeks before travel and continued for four weeks after leaving. (Source 5)

A common belief, and what the research shows

The belief: Hydroxychloroquine works against COVID-19.

What the research shows: It does not, and the body of evidence is large rather than thin. The RECOVERY trial stopped its hydroxychloroquine arm for futility after 4716 patients: "Death within 28 days occurred in 421 patients (27.0%) in the hydroxychloroquine group and in 790 (25.0%) in the usual-care group (rate ratio, 1.09; 95% confidence interval [CI], 0.97 to 1.23; P = 0.15)." Cochrane rated the no-benefit conclusion high certainty and added that "HCQ probably results in an almost three-fold increased risk of adverse events (RR 2.90, 95% CI 1.49 to 5.64; 1394 participants; 6 trials; moderate-certainty evidence)". A meta-analysis of 26 trials found mortality went the wrong way. None of this bears on its established value in lupus and rheumatoid arthritis, which rests on separate trials.

Questions and answers

What is it?

Hydroxychloroquine is a prescription tablet, a synthetic 4-aminoquinoline closely related to chloroquine. It is a weak base with a very large volume of distribution and a half-life of around 50 days, so it builds up slowly in tissues over weeks. It is used both as an antimalarial and as a disease-modifying drug in lupus and rheumatoid arthritis. (Source 1)

What does it do in the body?

It gets inside cells and into lysosomes, the acidic recycling compartments, and interferes with how they work. That dampens autophagy, alters membrane stability and changes signalling, which reduces production of inflammatory cytokines and modulates some immune co-stimulatory molecules. The honest answer is that the mechanism is still being worked out; the FDA label says the mechanism in rheumatic disease is not fully known. (Source 1)

Is it good or bad for you?

Both, depending on the condition and the duration. In lupus it is one of the better-supported drugs: a systematic review graded the evidence for preventing flares and increasing long-term survival as high. In rheumatoid arthritis the benefit is moderate. In COVID-19 it does not work and increased progression to ventilation or death. Taken for many years it causes irreversible retinal damage in a substantial minority, and it can cause severe hypoglycaemia and QT prolongation at any point. (Source 23)

How do you get more of it?

It is prescription-only and is not present in food or supplements, so there is nothing to eat or take to get more of it. The dose is set by a prescriber and is deliberately capped by body weight because higher doses cause retinal damage. The label's stated range is 400 to 600 mg daily initially in rheumatoid arthritis, dropping to 200 to 400 mg daily, and 200 to 400 mg daily in lupus. Absorption is improved by taking the tablet with food or milk. (Source 5)

If it is harmful, what reduces it?

Clearance is slow: the terminal half-life in whole blood is 40 to 50 days, so the drug persists for months after the last tablet and there is no way to speed that up. If toxicity appears the drug is withdrawn. For the eye that may not be enough, because retinal changes can keep progressing after the drug is stopped. For the heart, withdrawal recovered normal function in 44.9% of 78 reported cases but 12.9% had irreversible damage. (Source 12)

Why might someone be low in it or missing it?

Someone may be off it, or on less than the trials used, because of retinal findings on screening, because of cardiac conduction problems or QT prolongation, because of hypoglycaemia, or because it is deliberately dose-capped at 5 mg/kg of actual body weight to protect the retina. Absorption is also genuinely variable: in rheumatoid arthritis patients the fraction of the dose absorbed ranged from 30 to 100%, and mean blood levels were higher in those with less active disease. (Source 30)

Which whole foods contain it or feed it?

No whole food contains hydroxychloroquine. Food matters in two ways in the sources we read: the label directs taking the tablet with food or milk, and antacids or kaolin, which includes magnesium- and aluminium-containing antacid products, can reduce absorption of the closely related drug chloroquine, with at least four hours advised between them. We found no documented interaction with grapefruit or alcohol in the label or the reviews we read. (Source 9)

What happens if you do not have it?

The withdrawal trial is the cleanest answer for lupus: over 24 weeks off the drug, 16 of 22 people with previously stable lupus flared, versus 9 of 25 who stayed on it, and five of the 22 had a severe exacerbation requiring withdrawal from the study versus one of 25. In rheumatoid arthritis, withdrawals for lack of efficacy were significantly more frequent in the placebo groups of the pooled trials. (Source 2)

How can you test for it?

Blood levels of hydroxychloroquine can be measured but are not part of routine care. What is monitored is harm: the label calls for a baseline eye examination in the first year with best corrected visual acuity, an automated threshold visual field of the central 10 degrees and spectral-domain OCT, then annual examinations for higher-risk patients and from five years for everyone else. Those sensitive tests are exactly why measured retinopathy rates rose from historical estimates to 7.5%, so the test is reliable enough to change the risk estimate. Blood glucose is measured if hypoglycaemia is suspected, and an ECG is relevant where QT prolongation is a concern. (Source 13)

References

1. Nature reviews. Rheumatology. Mechanisms of action of hydroxychloroquine and chloroquine: implications for rheumatology. 2020. PMID 32034323, DOI 10.1038/s41584-020-0372-x. Read the source
2. The New England journal of medicine. A randomized study of the effect of withdrawing hydroxychloroquine sulfate in systemic lupus erythematosus. 1991. PMID 1984192, DOI 10.1056/nejm199101173240303. Read the source
3. The Cochrane database of systematic reviews. Antimalarials for treating rheumatoid arthritis. 2000. PMID 11034691, DOI 10.1002/14651858.cd000959. Read the source
4. DailyMed / Advanz Pharma (US) Corp. (label version effective 2026-04-09). PLAQUENIL (hydroxychloroquine sulfate) tablet - FDA prescribing information, 1.2 Rheumatoid Arthritis, 1.3 Systemic Lupus Erythematosus and 1.4 Chronic Discoid Lupus Erythematosus. 2026. Read the source
5. DailyMed / Advanz Pharma (US) Corp. (label version effective 2026-04-09). PLAQUENIL (hydroxychloroquine sulfate) tablet - FDA prescribing information, 2.1 Important Administration Instructions and 2.2 Dosage for Malaria in Adult and Pediatric Patients. 2026. Read the source
6. The New England journal of medicine. Effect of Hydroxychloroquine in Hospitalized Patients with Covid-19. 2020. PMID 33031652, DOI 10.1056/nejmoa2022926. Read the source
7. The New England journal of medicine. A Randomized Trial of Hydroxychloroquine as Postexposure Prophylaxis for Covid-19. 2020. PMID 32492293, DOI 10.1056/nejmoa2016638. Read the source
8. Frontiers in endocrinology. Association between hydroxychloroquine use and risk of diabetes mellitus in systemic lupus erythematosus and rheumatoid arthritis: a UK Biobank-based study. 2024. PMID 39568816, DOI 10.3389/fendo.2024.1381321. Read the source
9. DailyMed / Advanz Pharma (US) Corp. (label version effective 2026-04-09). PLAQUENIL (hydroxychloroquine sulfate) tablet - FDA prescribing information, 7.5 Methotrexate, 7.6 Cyclosporine, 7.7 Digoxin, 7.8 Cimetidine, 7.9 Rifampicin, 7.10 Praziquantel and 7.11 Antacids and kaolin. 2026. Read the source
10. DailyMed / Advanz Pharma (US) Corp. (label version effective 2026-04-09). PLAQUENIL (hydroxychloroquine sulfate) tablet - FDA prescribing information, 7.1 Drugs Prolonging QT Interval and Other Arrhythmogenic Drugs, 7.2 Insulin or Other Antidiabetic Drugs, 7.3 Drugs that Lower the Seizure Threshold and 7.4 Antiepileptics. 2026. Read the source
11. JAMA ophthalmology. The risk of toxic retinopathy in patients on long-term hydroxychloroquine therapy. 2014. PMID 25275721, DOI 10.1001/jamaophthalmol.2014.3459. Read the source
12. DailyMed / Advanz Pharma (US) Corp. (label version effective 2026-04-09). PLAQUENIL (hydroxychloroquine sulfate) tablet - FDA prescribing information, 12.3 Pharmacokinetics. 2026. Read the source
13. DailyMed / Advanz Pharma (US) Corp. (label version effective 2026-04-09). PLAQUENIL (hydroxychloroquine sulfate) tablet - FDA prescribing information, 5.2 Retinal Toxicity. 2026. Read the source
14. Drug safety. Cardiac Complications Attributed to Chloroquine and Hydroxychloroquine: A Systematic Review of the Literature. 2018. PMID 29858838, DOI 10.1007/s40264-018-0689-4. Read the source
15. Drug safety. Cardiac Complications Attributed to Chloroquine and Hydroxychloroquine: A Systematic Review of the Literature - authors’ concluding recommendation. 2018. PMID 29858838, DOI 10.1007/s40264-018-0689-4. Read the source
16. Annals of the rheumatic diseases. Clinical efficacy and side effects of antimalarials in systemic lupus erythematosus: a systematic review. 2010. PMID 19103632, DOI 10.1136/ard.2008.101766. Read the source
17. Nature communications. Mortality outcomes with hydroxychloroquine and chloroquine in COVID-19 from an international collaborative meta-analysis of randomized trials. 2021. PMID 33859192, DOI 10.1038/s41467-021-22446-z. Read the source
18. The Cochrane database of systematic reviews. Chloroquine or hydroxychloroquine for prevention and treatment of COVID-19. 2021. PMID 33624299, DOI 10.1002/14651858.cd013587.pub2. Read the source
19. DailyMed / Advanz Pharma (US) Corp. (label version effective 2026-04-09). PLAQUENIL (hydroxychloroquine sulfate) tablet - FDA prescribing information, 5.1 Cardiomyopathy and Ventricular Arrhythmias. 2026. Read the source
20. DailyMed / Advanz Pharma (US) Corp. (label version effective 2026-04-09). PLAQUENIL (hydroxychloroquine sulfate) tablet - FDA prescribing information, 5.10 Hypoglycemia. 2026. Read the source
21. DailyMed / Advanz Pharma (US) Corp. (label version effective 2026-04-09). PLAQUENIL (hydroxychloroquine sulfate) tablet - FDA prescribing information, 6 ADVERSE REACTIONS (post-approval reports: blood, cardiac, ear, eye, gastrointestinal, general, hepatobiliary, immune, metabolic and musculoskeletal disorders). 2026. Read the source
22. DailyMed / Advanz Pharma (US) Corp. (label version effective 2026-04-09). PLAQUENIL (hydroxychloroquine sulfate) tablet - FDA prescribing information, 6 ADVERSE REACTIONS (post-approval reports continued: nervous system, neuropsychiatric and skin disorders). 2026. Read the source
23. Annals of the rheumatic diseases. Clinical efficacy and side effects of antimalarials in systemic lupus erythematosus: a systematic review. 2010. PMID 19103632, DOI 10.1136/ard.2008.101766. Read the source
24. DailyMed / Advanz Pharma (US) Corp. (label version effective 2026-04-09). PLAQUENIL (hydroxychloroquine sulfate) tablet - FDA prescribing information, 8.1 Pregnancy. 2026. Read the source
25. The Cochrane database of systematic reviews. Chloroquine or hydroxychloroquine for prevention and treatment of COVID-19. 2021. PMID 33624299, DOI 10.1002/14651858.cd013587.pub2. Read the source
26. JAMA ophthalmology. The risk of toxic retinopathy in patients on long-term hydroxychloroquine therapy. 2014. PMID 25275721, DOI 10.1001/jamaophthalmol.2014.3459. Read the source
27. DailyMed / Advanz Pharma (US) Corp. (label version effective 2026-04-09). PLAQUENIL (hydroxychloroquine sulfate) tablet - FDA prescribing information, 2.3 Dosage for Rheumatoid Arthritis in Adults, 2.4 Dosage for Systemic Lupus Erythematosus in Adults and 2.5 Dosage for Chronic Discoid Lupus Erythematosus in Adults. 2026. Read the source
28. The New England journal of medicine. A randomized study of the effect of withdrawing hydroxychloroquine sulfate in systemic lupus erythematosus. 1991. PMID 1984192, DOI 10.1056/nejm199101173240303. Read the source
29. The New England journal of medicine. A Randomized Trial of Hydroxychloroquine as Postexposure Prophylaxis for Covid-19. 2020. PMID 32492293, DOI 10.1056/nejmoa2016638. Read the source
30. DailyMed / Advanz Pharma (US) Corp. (label version effective 2026-04-09). PLAQUENIL (hydroxychloroquine sulfate) tablet - FDA prescribing information, 12.3 Pharmacokinetics (Absorption). 2026. Read the source