Initial Anecdotal Evidence

In a recent Alert, we discussed FDA’s March 28 issuance of an Emergency Use Authorization (EUA) for oral formulations of the antimalarial drugs chloroquine (CQ) and hydroxychloroquine (HCQ), released from the Strategic National Stockpile, for the treatment of COVID-19. The Agency cited the drugs’ current use in some other countries based on very limited in vitro testing and some anecdotal clinical data from case reports. The latter consisted of mostly preliminary, unconnected clinical trials in China, performed in 10 or more hospitals and designed to test safety and efficacy in the treatment of COVID-19-associated pneumonia. The results from more than 100 patients apparently showed chloroquine was superior to the standard treatment of care. This information was reported in a letter by Gao et al. to a Japanese scientific journal in early March but, somewhat astoundingly, the authors admitted that their source was actually a news briefing (and not hard data).

Questionable Clinical Trials in France

The only other early, apparently positive results, from two trials of HCQ in combination with the antibiotic azithromycin (AZ), had come from a group headed by Didier Raoult in Marseille, France; their preprint results were posted online in the International Journal of Antimicrobial Agents on March 20, and in Travel Medicine and Infectious Disease in early April. However, according to a review in the journal Science, these studies have been widely criticized for their limitations and problems with methodology. Thus, in the first study in 36 patients, they were not randomized – Raoult himself selected the 20 patients who would get the test drugs and those who would be controls and get the placebo.

The second study didn’t even have a control group (all 80 mildly infected patients were treated), so patients may well have recovered anyway, with or without the test drugs. According to Science, Raoult dismissed any criticism and complained about scientists who insist on randomization and controls – key attributes of any credible clinical trial. Raoult garnered significant political support for his HCQ-AZ regimen in France, including a former Minister of Health and some prominent members of the medical profession, and was even visited in Marseilles by President Macron. However, the board of the society that produces the International Journal of Antimicrobial Agents soon issued a statement that the article did not meet the Society’s expected standards, following which the publisher (Elsevier) indicated that the paper was being re-assessed and that a “correction to the scientific record” might be considered. Nevertheless, the early reports from France and China appear to have provided the impetus for a high-level promotion of CQ, HCQ, and AZ in the United States.

Recent Clinical Data Not Encouraging

Since then, the results from more methodical clinical trials of HCQ (sometimes in combination with AZ) and CQ, both inside and outside China, have been reported in not-yet-peer-reviewed preprints, and they have not been encouraging.

China. The results of a multi-center, parallel, open-label, randomized, controlled clinical trial of HCQ in 150 patients hospitalized with COVID-19 were reported by Tang et al. in a preprint published on April 14 in medRxiv. Half the patients received standard of care (SOC) alone, and half received SOC plus HCQ (1200 mg/day for three days followed by 800 mg/day for the remainder of the treatment period – two weeks for mild/moderate cases, or three weeks for severe cases). The authors found no difference in the 28-day negative conversion rate (the rate at which patients go from testing positive to testing negative for the virus) between the SOC-only group and the SOC + HCQ group. The negative conversion rates at days 4, 7, 10, 14 and 21 were also similar between the two groups. The authors also reported no difference in the rate of alleviation of symptoms at 28 days between the two groups, although HCQ treatment was associated with a more rapid decline in C-reactive protein (CRP) levels and recovery from lymphopenia ( abnormal reduction in lymphocytes). HCQ treatment was associated with a significant increase in adverse effects, primarily diarrhea, but not serious adverse effects. The authors noted that retinopathy and adverse gastrointestinal and cardiac effects from the use of HCQ or CQ in the treatment of malaria and rheumatic diseases are well documented, although HCQ is generally preferred to CQ because of its lower toxicity, particularly retinal toxicity. Overall, they believed the results did not support the clinical use of HCQ as an antiviral agent in spite of the earlier promise of in vitro studies; nevertheless, it did appear to have beneficial anti-inflammatory properties in patients who had gone on to develop increased CRP and lymphopenia.

Brazil. In a parallel, double-blind, randomized phase IIb clinical trial in the city of Manaus, an initial set of 81 patients with severe acute respiratory distress syndrome (ARDS) due to COVID-19 were divided into two arms: a high dose arm received 12 g CQ over 10 days, and a low dose arm received 2.7 g CQ over 5 days and placebo for a further 5 days. The study, by Borba et al., was posted in medRxiv on April 16. No comparison was made with the additional use of AZ since all patients were already receiving it per hospital protocol. Unfortunately, official national guidelines prevented the use of a placebo-based control arm. More adverse cardiac effects (prolongation of QTc) were seen in the high dose group than in the low dose group, to the point that by day 13, a total of 16 of 41 patients had died in the high dose group, but just 6 of 40 patients in the low dose group. Because of the lack of a control group, results were compared to those of a cohort of similarly critically ill patients in Italy who had not received CQ; overall, fatality in the Brazilian patients treated with CQ appeared to be higher than in Italian controls. The authors, therefore, stopped enrolling patients for the high dose group, and strongly recommended that this high dose (based on the dose proposed by Chinese authorities) not be used anywhere in the world. However, they decided to continue enrolling patients in the low dose group, even though they had so far seen no apparent benefit in using CQ to reduce lethality in seriously ill patients. They also recommended clinical trials evaluating CQ’s role as a potential prophylactic drug, and in preventing progression to the severity in patients with mild or moderate symptoms.

United States. In a preprint posted on medRxiv on April 23, Magagnoli et al. presented a retrospective analysis of data from a total of 368 male COVID-19 patients hospitalized in VA medical centers throughout the United States. The authors categorized patients based on whether they had been treated with no HCQ (n= 158), with HQC alone (n=97), or with HQC + AZ (n=113). Mortality was 11.4% in the no-HCQ group, 27.8% in the HCQ group, and 22.1% in the HCQ + AZ group. Statistical analysis showed that HQC with or without AZ did not improve mortality or reduce the need for mechanical ventilation; in fact, HQC alone showed an increased risk of mortality compared to no HCQ (standard care), even after adjusting for the propensity of the more severe cases to be treated with the drug. The authors noted that the study provides the largest dataset yet reported of the outcomes of COVID-19 patients treated with HCQ, with or without AZ, anywhere in the world.

Adverse Cardiac Effects

As already noted, the use of CQ or HCQ has been associated with adverse cardiac effects, primarily prolongation of QTc and arrhythmia. Following the initial burst of enthusiasm for these drugs that appeared to be based on sparse clinical evidence, a group of cardiac and other specialists from the Mayo Clinic published on-line an “Urgent Guidance for Navigating the QTc-Prolonging and Torsadogenic Potential of Possible Pharmacotherapies for Coronavirus Disease19 (COVID-19)” in Mayo Clinic Proceedings in late March. (“Torsadogenic” refers to the condition torsades de pointes, a specific type of ventricular arrhythmia that can lead to sudden cardiac death.) The authors noted that while CQ and HCQ have relatively favorable safety profiles, in high-risk individuals they can precipitate drug-induced torsades de pointes or even drug-induced sudden cardiac death, especially with long-term use or when used in combination with AZ, which is reported to have its own potential for these cardiac effects. The guidance stressed the need to identify the relatively small number of individuals most at risk by monitoring their baseline QTc status prior to medication with HCQ or CQ, either by ECG or preferably (considering the risk of contagion) with the use of a smartphone-enabled mobile QTc meter, and described how best to mitigate the risk once medication had commenced.

Shortly after the guidance appeared, MedRxiv posted on April 3 a preprint of a study by physicians from the New York University School of Medicine (Chorin et al.) describing changes in QTc interval in 84 COVID-19 patients treated with HCQ + AZ. They reported that average QTc values increased on a daily basis, and 11% of patients developed new, severe CTc prolongation of >500 ms, “a known marker for high risk of malignant arrhythmia and sudden cardiac death.” They, therefore, concluded that baseline QTc values alone were not a reliable predictor of subsequent severe QTc prolongation and that QTc should be monitored constantly in COVID-19 patients receiving HCQ + AZ, especially in those with renal failure (a common complication of the disease).

Taking a different approach, an international group of researchers representing the Observational Health Data Sciences and Informatics (OHDSI) COVID-19 consortium reviewed 20 years of safety data from multiple studies from Western Europe, Japan and the USA involving rheumatoid arthritis patients who had been newly treated with HCQ, sulfasalazine (an anti-rheumatic drug), or HCQ + AZ. A preprint of their report (Lane et al.) was posted on April 10 in medRxiv. There were more than 900,000 patients in the HCQ group and more than 300,000 patients in each of the other two groups. The authors found no excess risk of severe adverse events when 30-day use of HCQ or sulfasalazine were compared, but an increased risk of cardiovascular mortality (more than 2x), angina, and heart failure for 30-day use of HCQ + AZ when compared to HCQ + another antibiotic, amoxicillin. They, therefore, concluded that the addition of AZ to HCQ treatment may induce heart failure and cardiovascular mortality, potentially due to synergistic effects on QTc length, and urged caution when using this combination in the treatment of COVID-19.

New Warnings From NIH Expert Panel and From FDA

On April 21, NIH published extensive guidelines for COVID-19 developed by a large Expert Panel convened by the National Institute of Allergy and Infectious Diseases (NIAID) and intended for healthcare providers. After reviewing all available clinical data, the Panel stated that it did not recommend the use of any agents (including HCQ and CQ) for SARS-CoV-2 pre-exposure or post-exposure prophylaxis outside the setting of a clinical trial, and also stated that there are insufficient data to recommend either for or against the use of any antiviral therapy in patients with COVID-19 who have a mild, moderate, severe, or critical illness. It specifically recommended against the use of the combination of HCQ and AZ, except in the context of a clinical trial, because of the potential for toxicities. All of these recommendations were categorized as AIII (strong recommendation based on Expert Opinion).

Three days later, on April 24, FDA issued a warning, cautioning against the use of HCQ or CQ for COVID-19 outside of the hospital setting or a clinical trial, due to the risk of heart rhythm problems. The Agency stated:

Hydroxychloroquine and chloroquine have not been shown to be safe and effective for treating or preventing COVID-19… [they] can cause abnormal heart rhythms such as QT interval prolongation and a dangerously rapid heart rate called ventricular tachycardia. These risks may increase when these medicines are combined with other medicines known to prolong the QT interval, including the antibiotic azithromycin, which is also being used in some COVID-19 patients without FDA approval for this condition. Patients who also have other health issues such as heart and kidney disease are likely to be at increased risk of these heart problems when receiving these medicines… The EUA was based upon limited evidence that the medicines may provide benefit, and for this reason, we authorized their use only in hospitalized patients under careful heart monitoring.

Overall, there is an urgent need for additional data from large, well-designed clinical trials of these antimalarials performed in a controlled hospital setting where patients can be continuously monitored. Two such trials of HCQ have recently been announced by Novartis and by NIH’s National Heart, Lung, and Blood Institute. However, the latest safety and efficacy data do not look encouraging, and certainly do not appear to warrant the early burst of enthusiasm for, and promotion of, these drugs for the treatment of COVID-19.