COVID-19: the rise and fall of hydroxychloroquine

THE United States Strategic National Stockpile is laden with 63 million doses of hydroxychloroquine (HCQ) that have no clinical use in coronavirus disease 2019 (COVID-19). This has occurred as the US Food and Drug Administration (FDA) has revoked the emergency use authorisation for the use of HCQ in patients with COVID-19 admitted to hospital for whom a clinical trial is unavailable or not feasible. It is worth reflecting how things ever got to this situation.

Developing drugs for clinical use is expensive and time consuming, and the costs and time are opportunity costs for other areas of clinical service and research in COVID-19. Thus, following the removal of HCQ from large randomised clinical trials and from the FDA emergency register, it is time to question the assumptions and decisions that were made in selecting HCQ for repurposing in COVID-19, such as the use of non-standard methods to the standard physiological, pharmacological and regulatory processes for taking repurposed drugs at appropriate dose and timing into clinical trials.

Specific questions should be asked regarding how HCQ was chosen for clinical trials, what the biological rationale was, and how the choice was made to study antiviral therapy over “treating the host” response – the factor that causes morbidity and mortality in humans in COVID-19. Were there adequate pre-clinical in vitro studies showing efficacy? Were target concentrations of HCQ even achievable in relevant human compartments, and was posology optimisation considered to ensure appropriate dosing and timing of therapy prior to human trials? Importantly, was the drug concentration required for efficacy in COVID-19 likely to cause unacceptable toxicity?

It seems that these questions were not adequately answered prior to the commencement of clinical trials. What followed was an example of the piecemeal and fragmented approach to a global therapeutics research platform, including studying HCQ, which, based on pharmacology and physiology, was not only unlikely to be effective but also had complicated pharmacology and significant toxicities. It has meant an opportunity cost for clinical research using other biologically plausible research platforms and obviously also a cost to patients, who at this point still have no therapeutic options. We consider these aspects and suggest for future pandemic planning a national research group spanning all relevant disciplines, including clinical drug and therapeutics experts, be convened to develop a research platform.

Biological plausibility and in vitro studies

Results from in vitro studies for a potential repurposed drug play a critical role for the “go/no-go” decision to commence follow-on clinical trials. Following the initial outbreak of COVID-19, in vitro antiviral studies of chloroquine and HCQ were undertaken; however, there were significant limitations with these. For example, the in vitro HCQ EC50 (the drug concentration required to obtain a 50% effect) should be compared with in vivo HCQ concentration within an appropriate dosing regimen for COVID-19, as opposed to using doses effective for other diseases. Specifically, if the in vivo concentration is larger or much larger than the in vitro EC50, those doses of HCQ might be effective for patients with COVID-19. The in vitro HCQ EC50 is effectively extracellular drug concentration and it should be compared with the in vivo HCQ concentration (free plasma HCQ concentration). This comparison seems obvious, but a number of recent studies (here, here, here, here, here and here) compared the in vitro EC50 not with the free plasma HCQ concentration, but with the intracellular HCQ concentrations. This turns out to be an important error, because with HCQ, the intracellular HCQ concentrations are significantly higher (about 1000-fold) than the corresponding extracellular or free plasma HCQ concentration, due to the entrapment of HCQ in the cell as a consequence of the intracellular acidic environment (here and here). Once intracellular HCQ concentration is compared with the determined in vitro EC50, it is clear that the intracellular concentration is larger than the in vitro EC50 — this is certain because the in vivo HCQ concentration was “artificially” made about 1000-fold larger from its corresponding free plasma HCQ concentration. Consequently, clinical investigators starting studies may have received incorrect information that HCQ might be effective at standard doses and moved to the clinical trial phase. This error was actually described and presented by the FDA before removing the emergency use of HCQ. As well as HCQ, these errors in pre-clinical trialling of a repurposed drug for COVID-19 were seen with other drugs studied for this disease.

Enough efficacy to start a clinical trial?

On 15 February 2020, a Chinese news briefing claimed chloroquine had been used successfully to enhance viral clearance and reduce disease progression in an unpublished case series of 100 patients followed by the development of Chinese treatment guidelines. Subsequently, a small (20 cases treated with HCQ), open label, non-randomised trial, with a high proportion of loss to follow-up in the treatment group (six of the initial 26 enrolled) emerged, reporting induction of viral clearance with HCQ but published before peer review. The US president infamously promoted HCQ during a media briefing based on this report, labelling HCQ as “a powerful drug” and later on Twitter as a “game-changer”. Later, on 28 March, the US FDA provided emergency use authorisation for HCQ, allowing its use in patients with COVID-19 admitted to hospital.

Cognitive dissonance – additional trials are subsequently ethically approved

A significant quantity of poor quality trials then followed during April and May 2020, often published prior to peer review, with predominantly negative or equivocal results. A second observational, non-comparative trial evaluated a group of 80 cases who received HCQ and azithromycin with relatively mild COVID-19. The researchers reported only one death during the study period, lower than the death rate France was experiencing at that time. This group also published a retrospective descriptive analysis of 1061 cases of COVID-19 treated with HCQ and azithromycin suggesting that the therapeutic combination appeared safe and associated with a low mortality rate. A small, prospective trial conducted in Wuhan randomised 62 admitted patients with COVID-19 to receive HCQ in addition to standard treatment (therapies such as other antiviral agents, immunoglobulin and corticosteroids at the discretion of the treating physician) or standard treatment alone. The non-validated end point chosen in this trial was “time to clinical recovery”, defined as the return of body temperature to normal range and relief of cough for more than 72 hours. The HCQ group had a shorter “time to clinical recovery” – 3.2 days in the control and 2.2. days in the HCQ group. Another retrospective Chinese trial concluded that HCQ use was associated with decreased mortality in critically ill patients. In observational case reports of 568 patients in Wuhan who presented with severe COVID-19, 48 received HCQ. Although there was a significant difference in mortality between the two groups, favouring HCQ, as with the many other studies quoted above, the patients who received HCQ were likely to be the less unwell patients.

But there’s more

Multiple negative higher quality trials were subsequently published. Jun et al conducted a small prospective trial in which 30 patients with moderate COVID-19 were randomised to receive either HCQ plus supportive treatment or supportive treatment alone. There was no difference between the groups in the primary outcome of negative conversion of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) nucleic acid detection in nasopharyngeal swab after 7 days, nor in another small, retrospective study with SARS-CoV-2 pneumonia. Additionally, a retrospective analysis of data from patients with COVID-19 admitted to hospital in all US Veterans Health Administration medical centres categorised patients based on their exposure to HCQ and azithromycin. The use of HCQ did not reduce the risk of the primary outcome of death and need for mechanical ventilation. In fact, the authors reported an association with increased mortality in patients treated with HCQ. Similarly, an observational study in New York analysing COVID-19 cases admitted to hospital also failed to show a reduction in risk of intubation or death. An observational study in France showed no survival benefit of HCQ use in COVID-19 pneumonia requiring oxygen.

Furthermore, a large, multinational registry analysis did not show benefit of HCQ or chloroquine on in-hospital outcomes for COVID-19, yet an association with decreased in-hospital survival. However, the article was later infamously retracted, not because the conclusions were incorrect (nevertheless, concerns had been raised with respect to the veracity of the data and analyses, prompting the attempt at an independent third-party peer review), but because the company with the database refused auditing.

Although large, randomised clinical trial data from groups around the world are still pending publication (eg, SOLIDARITY and RECOVERY), trial investigators have now removed the HCQ arm of these trials based on preliminary data.

How did we get here? The HCQ journey timeline

Overall, there have been several great collaborations and excellent public health leadership in many areas of the world during the COVID-19 pandemic. What we lacked was a cohesive vision and strategy for drug development in the repurposing area and implementation in the therapeutics area. The choice of HCQ and then the dose for COVID-19 treatment were areas that experts skilled in pharmacology, clinical pharmacology, pharmacokinetic/pharmacodynamic modelling and pre-clinical drug development could have provided guidance on. A global drug coordination by means of the World Health Organization would have enabled local experts to contribute knowledge to a more streamlined program; a program that could have given advice at each one of the steps in the HCQ journey. Also important to note is that additional HCQ trials continued despite equipoise being less likely as time went on.

What can we do better next time?

As previously noted, an international approach to rapidly identify drugs that treat the host to permit time for vaccine and antiviral development is needed. Importantly, coordination of the repurposing of existing drugs based on the principles of Fedson’s “treating the host” and repurposing existing drugs for new indications using existing pharmacology and physiology knowledge and observational administrative data are important to ensure time is not wasted, and that the most likely therapeutic targets based on physiology and pharmacology are chosen for clinical trials. Lastly, as we have seen, the work of epidemiology and public health staff is key to our broader response to worldwide pandemics.

Jennifer Martin is a physician and clinical pharmacologist co-directing the national Centre for Drug Repurposing and Medicines Research. Together with colleagues Nikola Bowden and Richard Head she has proposed a platform for repurposing drugs for COVID-19 which includes an international approach to rapidly identify drugs that treat the host in a pandemic.

Dr Zheng Liu received his PhD in Process Modeling in Chemical Engineering. He is now focusing on pharmacokinetic and pharmacodynamic modeling and simulation, and clinical trial design as a pharmacometrician.

Dr Steven Bollipo is the Director of Gastroenterology at John Hunter Hospital in Newcastle. He has authored the GESA statement on management of liver disease during the COVID-19 pandemic in Australia.

Dr Joshua McCarthy is a medical registrar in the John Hunter Hospital in Newcastle. He is currently studying a Master of Public Health and Tropical Medicine through James Cook University.

The statements or opinions expressed in this article reflect the views of the authors and do not represent the official policy of the AMA, the MJA or InSight+ unless so stated.