6. Future Outlook
Reviewing the safety of potential COVID-19 treatments (table 1) is complex due to the fast-moving pace of research in this field. For example, chloroquine and hydroxychloroquine with or without an accompanying macrolide antibiotic, have consistently been at the forefront of COVID-19 research efforts since the outbreak began. However, the astonishing developments over a week or so have led to retraction of a highly publicised paper, and results from a post-exposure prophylaxis trial and a treatment trial (RECOVERY), both of which have shown no beneficial effect of hydroxychloroquine (Boulware et al., 2020; Horby & Landray, 2020; Mehra, Ruschitzka, & Patel, 2020). This highlights that the rapid rate of discoveries surrounding COVID-19 therapies generates the need to update this perspective frequently, in order to ensure that the safety of any newly repositioned therapies, novel developmental compounds, or new therapeutic combinations are investigated. For example, the potential use of heparin in novel forms, including nebulised therapy (clinical trial identifier NCT04397510), as an antiviral agent is currently the subject of several investigational trials. In addition, the potential utility of nitazoxanide is currently the subject of several clinical trials (Clinical Trials.gov, 2020a; Pepperrell, Pilkington, Owen, Wang, & Hill, 2020; Rajoli et al., 2020).
It is clearly essential that the harm:benefit ratio of any pharmaceuticals being considered for use in the treatment of COVID-19 are thoroughly considered. This ratio changes dependent upon the disease stage and is correlated to potential mortality. For example, a higher risk may be accepted for patients in the later stage of severe disease than the same therapeutic agent administered in mild disease. This difference in harm-benefit analysis becomes even more striking when considering the use of such agents to prevent infection. As is the case for many highly contagious viruses, prevention by prophylaxis would be incredibly valuable. Some of the agents described in this review, including chloroquine and ritonavir have been suggested as potential prophylactic agents, but to date, data on efficacy have been disappointing (Rathi, Ish, Kalantri, & Kalantri, 2020; Spinelli, Ceccarelli, Di Franco, & Conti, 2020). Clearly, treatment duration for prophylaxis is expected to be longer than for treatment of COVID-19, and this may further alter the harm-benefit ratio, reinforcing the need for safety considerations at the outset of any clinical trials.
Similarly, the evaluation of therapy risk also applies to long-term recovery. As the current pandemic progresses, it is becoming apparent that being discharged from hospital does not necessarily mean that patients are free from COVID-19 symptoms. Large numbers of patients who have survived severe SARS-CoV-2 infection may have incurred long-term health problems, including some permanent loss of lung and kidney function (Foundation, 2020; Su et al., 2020; Summers, 2020). Consequently, it is probable that long-term therapies will be required for many patients to maintain, or ideally restore, normal physiological organ function. It is vital that therapies which will be used to treat patients during their long-term recovery are also undergoing evaluation for their safety, particularly as many of these agents may need to be administered over much longer periods of time than initial COVID-19 treatments.
The identification and characterisation of biomarkers of disease and safety will be invaluable in the further development and deployment of therapies for COVID-19. Disease biomarkers, for example of lung injury or the hyperinflammatory reponse, may allow the stratification of therapy in order to select the agent best suited to the stage of disease. Moreover, biomarkers should be considered to monitor patient safety in cases of known AEs. For example, the manufacturer’s guidelines for remdesivir recommend daily liver function tests due to the risk of transaminase elevations (Gilead, 2020). These tests are essential, particularly with regards to COVID-19 where increased ALT levels are reported to be common amongst hospitalised patients (Bangash et al., 2020; L. Zhang et al., 2009). Looking to the future, improvements in the specificity, predictivity and reliability of drug-induced organ damage, through academic-industry partnerships such as the Biomarker Qualification Program in the Critical Path Institute in the US, and the European Innovative Medicines Initiative consortium Transbioline, will help improve clinical assessment of COVID-19 drug safety issues.
Continued enhancements in the speed, predictivity, and human translation of safety assessment for toxicity of anti-viral compounds is clearly warranted, and this may include animal models of SARS-CoV-2 as well asin vitro models, in order to assess efficacy alongside safety. Such a full understanding for individual therapies will indicate the combinations that can have the potential to provide the best synergy for benefit, while forewarning of the potential for increased risk/harm through pharmacokinetic or toxicodynamic interaction.
Although outside the scope of this review, a vaccine for COVID-19 remains the greatest hope to end the pandemic and protect the population. As of 6th June 2020, according to WHO there are 10 vaccines in clinical trial stages and 123 in preclinical stages of evaluation (World Health Organisation, 2020b). Currently, potential vaccines are only just beginning to be tested for efficacy in humans in early phase studies, and therefore safety data will begin to emerge as larger numbers of individuals are administered the vaccine. Safety data regarding preliminary vaccinations against SARS and MERS are limited, but the available information may be useful during the development of COVID-19 vaccines due to the similarities between the coronavirus strains (Padron-Regalado, 2020). One safety concern relevant to coronaviruses is the potential for the induction of antibody-dependent enhancement (ADE), a phenomenon which was observed in cats vaccinated against feline infectious peritonitis coronavirus, and has also been seen in patients vaccinated against Zika virus and Dengue virus (Khandia et al., 2018; Padron-Regalado, 2020; Vennema et al., 1990). ADE can occur when non-neutralising antibodies bind to virus particles and increase their uptake into host cells, instead of rendering them non-infectious (Padron-Regalado, 2020; Tirado & Yoon, 2003). This caused concern in initial SARS vaccine development, but can reportedly be avoided by using truncated versions of the viral S glycoproteins (He et al., 2004). Acknowledging safety concerns such as this, as well as the ways they can be attenuated, may be paramount in the timely development of a vaccine against COVID-19.
In conclusion, although expanding extremely rapidly, the field of therapies to treat COVID-19 remains in its infancy. Safety will continue to play a major role in therapeutic success, as apparent with recent reports of increased cardiac toxicity associated with the use of chloroquine/hydroxychloroquine in the treatment of COVID-19, despite its long history of use as an antimalarial. Above all, this review has exemplified the need to view safety concerns in the context of the individual and specific phase of disease in order to formulate a comprehensive harm-benefit balance. Importantly, an awareness of potential safety concerns will support the development of the next stage of therapy targeting prophylaxis and recovery post-COVID infection. It is imperative that safety scientists look to rise to the challenge of COVID-19 by utilising their expertise in mechanistic understanding, biomarker development and toxicokinetic modelling in order to support the development of COVID-19 therapies that can be used effectively and safely.