Dose of Reality? Hydroxychloroquine and Chloroquine Again Disappoint in COVID-19

Two large analyses reinforce the view that the antimalarial medications raise the risk of death and arrhythmia.

Dose of Reality? Hydroxychloroquine and Chloroquine Again Disappoint in COVID-19

Two new papers, representing the largest number of patients studied to date, reinforce the view that the use of hydroxychloroquine and chloroquine is toxic to the heart and likely does more harm than good as a treatment for COVID-19.

NOTE: On June 2, 2020, the Lancet issued an “Expression of Concern” to alert readers that questions have been raised about the validity of the data reported by Mehra et al. The journal states that an independent audit of the data has been commissioned by the authors—those not affiliated with Surgisphere—and that they will update the notice when they have further information.


In the first study, a large global analysis of nearly 100,000 patients with COVID-19 published in the Lancet, hydroxychloroquine or chloroquine use was associated with a significantly higher risk of in-hospital mortality either when given alone or with a macrolide, such as azithromycin, and with an increased risk of new ventricular arrhythmia during hospitalization.

More of TCTMD's coverage on our COVID-19 hub.
More of TCTMD's coverage on our COVID-19 hub.

In the second, an analysis of the World Health Organization (WHO) pharmacovigilance database published in Circulation, researchers found evidence of prolonged QT and ventricular tachycardia for both hydroxychloroquine and azithromycin, while hydroxychloroquine was significantly linked with the development of conduction disorders and heart failure. This particular series looked at all patients taking these medications, for any indication, up until March 1, 2020.

According to Mandeep Mehra, MD (Brigham and Women’s Hospital, Boston, MA), lead investigator on the Lancet paper, the findings should help to douse some of the rampant enthusiasm for these agents.

“Not only was there absolutely no shift towards benefit any which way we sliced the data, but there was clear evidence of harm with a hazard that ranged anywhere from a 30% to 45% increased risk of mortality after adjustments,” Mehra told TCTMD. “Without a shadow of doubt, a hospitalized patient with COVID-19 should not be treated with these drugs.”

Small, uncontrolled clinical trials, as well as anecdotal reports out of China and Europe, suggested early on in the pandemic that use of hydroxychloroquine, possibly in combination with azithromycin, might be a treatment option for patients with COVID-19. On March 28, 2020, the US Food and Drug Administration issued an emergency use authorization for oral formulations of hydroxychloroquine and chloroquine in patients hospitalized where clinical trials weren’t available or participation wasn’t feasible. In the ensuing weeks and months, US President Donald Trump has repeatedly trumpeted hydroxychloroquine during media briefings as a “real game changer” for COVID-19, stating last week that he himself was taking the agent. In Brazil, which has the second highest numbers of infections and deaths following the United States, President Jair Bolsonaro also has championed the antimalarials. But as reports of adverse events increased, the FDA issued a warning against the use of the antimalarial agents outside of a hospital or a clinical trial due to the risk of adverse events, including QT interval prolongation, ventricular tachycardia, ventricular fibrillation, and even death. Last week, the American College of Physicians warned people not to use chloroquine or hydroxychloroquine either alone or in combination with azithromycin for prevention or treatment of COVID-19 outside of a research setting.

Given the uncertainty and fear that SARS-CoV-2 infections have sparked, Mehra is sympathetic to the enthusiasm surrounding hydroxychloroquine and chloroquine but stressed that, during times of medical crises, one of the most important roles for scientific and medical leadership is to remain calm and provide clear direction based on the best available evidence.

“In that spirit, the initial evangelism to try to treat a desperate disease with perhaps a desperate measure was undertaken and I can actually understand that,” he said. “What I can’t understand is unfettered persistence when the evidence starts to shift in an opposite direction. That does cause some degree of concern.”

The WHO has responded to these latest findings by placing a temporary pause on the hydroxychloroquine arm of the SOLIDARITY trial while the trial’s data safety and monitoring board reviews the results collected thus far. SOLIDARITY is testing four treatments—remdesivir (Gilead Sciences), the HIV drugs lopinavir/ritonavir (Kaletra; AbbVie) alone or with interferon beta 1a, which is normally used to treat multiple sclerosis, and hydroxychloroquine—in patients hospitalized with COVID-19.

Emergency Use Authorization Granted Early

This Lancet analysis focused on 96,032 patients (mean age 53 years; 46.3% women) hospitalized with COVID-19 between December 20, 2019, and April 14, 2020, at 671 hospitals from six continents. Patients who received one of four treatments—hydroxychloroquine alone, chloroquine alone, hydroxychloroquine with a macrolide, chloroquine with a macrolide—within 48 hours of a positive COVID-19 diagnosis were included. In total, 14,888 patients received one of the treatments—nearly half received hydroxychloroquine plus a macrolide—and 81,144 patients served as controls. Patients on mechanical ventilation and those who received remdesivir were excluded.

Without a shadow of doubt, a hospitalized patient with COVID-19 should not be treated with these drugs. Mandeep Mehra

Given its size and scope, “this is as real-world as an observational study gets,” said Mehra. “We looked at four separate regimens, because across the world in some places hydroxychloroquine isn’t available and they’re using chloroquine. In some places azithromycin isn’t available and they use clarithromycin.”

In-hospital mortality was 9.3% in the control group, 18.3% in the hydroxychloroquine group, 23.8% in those who received hydroxychloroquine with a macrolide, 16.4% in the chloroquine arm, and 22.2% among those who received chloroquine plus a macrolide. After adjusting for multiple confounding variables, including age, body mass index, and comorbidities, treatment with hydroxychloroquine or chloroquine—either alone or with a macrolide--was associated with significantly increased risks of death and de novo ventricular arrhythmias.

Outcomes of COVID-19: HR (95% CI)


In-hospital Mortality

Ventricular Arrhythmia During Hospitalization


1.365 (1.218-1.531)

3.561 (2.760-4.596)


1.335 (1.223-1.457)

2.369 (1.935-2.90)

Chloroquine + Macrolide

1.368 (1.273-1.469)

4.011 (3.344-4.812)

Hydroxychloroquine + Macrolide

1.447 (1.368-1.531)

5.106 (4.357-5.983)

In a propensity score-matched analysis, the same signal of harm was evident, and the results were similar when the researchers analyzed the effects of treatment across the continents.

The group also performed a “tipping point” analysis, which was done to assess the effects of an unmeasured confounding variable on the significance of the findings with hydroxychloroquine or chloroquine. Assuming they missed an unobserved confounder, they found it would need to be prevalent in roughly 50% of the exposed patients to tip the results to nonsignificance. Following this calculation, “we became even more certain that our analysis was quite robust,” said Mehra.

To TCTMD, Mehra said that even if the agents are used for their approved indications, these new data should give physicians pause for thought, particularly when using them in patients with underlying cardiovascular disease.

In an editorial, Christian Funck-Brentano, MD, PhD, and Joe-Elie Salem, MD, PhD (Sorbonne Université/Pitié-Salpêtrière Hospital, Paris, France), state that while it might be tempting to attribute the higher number of deaths with treatment to the higher incidence of ventricular arrythmias, the researchers did not study this relationship and the causes of death were not adjudicated. “Nevertheless, the increased incidence of ventricular arrhythmias is intriguing,” they write. “Chloroquine, hydroxychloroquine, and azithromycin have sodium channel blocking properties that might contribute to proarrhythmia and heart failure in the context of myocardial injury and hypoxia present in COVID-19.”

Evidence of Harm in WHO Database

Funck-Brentano and Salem are actually co-authors on the Circulation paper, led by Lee Nguyen, MD (Sorbonne Université/Pitié-Salpêtrière Hospital). The WHO’s VigiBase includes more than 21 million adverse event drug reports from 130 countries and using this, Nguyen and colleagues identified 76,822 adverse drug reactions with hydroxychloroquine alone, 89,692 with azithromycin alone, and 607 with the combination of both drugs. Hydroxychloroquine was believed to be the cause of the reaction in 28.4% of cases, as opposed to a concomitant medication, and azithromycin the trigger in 60.8% of cases.

There were more reports of QT interval prolongation and/or ventricular tachycardia with each drug used individually, while hydroxychloroquine was associated with conduction disorders and heart failure. Compared with hydroxychloroquine monotherapy, azithromycin used alone was associated with a higher risk of long QT and/or ventricular tachycardia, including torsades de pointes (TdP; reporting OR 2.36; 95% CI 2.05-2.71). Using azithromycin with hydroxychloroquine together was associated with a higher incidence of long QT and/or ventricular tachycardia and TdP than either drug used alone (reporting OR 2.48: 95% CI 1.28-4.79).

In total, 8.4% of patients treated with hydroxychloroquine who developed ventricular tachycardia and/or TdP died compared with 20.2% treated with azithromycin who developed the arrhythmia. The mortality rate was 20.7% for the patients treated with hydroxychloroquine who developed heart failure, and the risk of mortality was higher when exposure to the drug was prolonged over several months.

The main limitation of this analysis is that the VigiBase does not include the numbers of patients exposed to these medications: it is impossible to assess incidence without this denominator, the authors note. The analysis was also not specific to COVID-19 patients. “However, our results are consistent with the facts that these cardiovascular adverse drug reactions are found in the FDA's labels of hydroxychloroquine and azithromycin, and that both drugs are referenced as known-risk of TdP at CredibleMeds website.”

President Trump has since announced that he is no longer taking the medication, according to recent media reports, something that Mehra said he was pleased to hear.

Michael O’Riordan is the Associate Managing Editor for TCTMD and a Senior Journalist. He completed his undergraduate degrees at Queen’s…

Read Full Bio
  • None of the study authors report relevant conflicts of interest.