key: cord-0986959-zhfrgaxf authors: Osborne, V.; Davies, M.; Dhanda, S.; Roy, D.; Lane, S.; Evans, A.; Shakir, S. A. title: Systematic benefit-risk assessment for the use of chloroquine or hydroxychloroquine as a treatment for COVID-19: Establishing a framework for rapid decision-making date: 2020-05-12 journal: nan DOI: 10.1101/2020.05.07.20093989 sha: a85cd1c0cb14b3ce3b8cec89e7b36d9f7e14ef47 doc_id: 986959 cord_uid: zhfrgaxf Objectives: Given the current pandemic, there is an urgent need to identify effective, safe treatments for COVID-19 (coronavirus disease). A systematic benefit-risk assessment was designed and conducted to strengthen the ongoing monitoring of the benefit-risk balance for chloroquine and hydroxychloroquine in COVID-19 treatment. Methods: The overall benefit-risk of the use of chloroquine or hydroxychloroquine as a treatment for COVID-19 compared to standard of care, placebo or other treatments was assessed using the Benefit-Risk Action Team (BRAT) framework. We searched PubMed and Google Scholar to identify literature reporting clinical outcomes in patients taking chloroquine or hydroxychloroquine for COVID-19. A value tree was constructed and key benefits and risks were ranked by two clinicians in order of considered importance. Results: Several potential key benefits and risks were identified for use of hydroxychloroquine or chloroquine in COVID-19 treatment. For the benefit of virological clearance, three studies were identified. A significant risk difference (RD) between hydroxychloroquine and the comparator group (standard of care) was found for only one study (RD=0.58, 95% CI: 0.17, 0.98). The risk difference was not significant for the other two studies (RD=-0.07, 95% CI:-0.75, 0.61 and RD=0.08, 95% CI:-0.74, 0.91). In addition, no significant risk difference between hydroxychloroquine and the comparator group (standard of care) was identified for the risk of abnormal liver function tests (LFTs) (RD=0.07, 95% CI: -0.28, 0.41). Conclusions: Overall, no conclusion can be made on the benefit-risk profile of hydroxychloroquine or chloroquine in the treatment of COVID-19 compared to standard of care, placebo or other treatments at this time. Whilst the availability of comparative data are limited, the current framework summarises the key anticipated benefits and risks. As further data from clinical trials and real world use on these benefits and risks becomes available, this can be incorporated into the framework for an ongoing benefit-risk assessment. Coronaviruses have circulated among humans and animals for many years, of which several strains are highly transmissible and pathogenic in humans [1] . Severe Acute Respiratory Syndrome (SARS) emerged in 2002 and 2003, while Middle East Respiratory Syndrome coronavirus (MERS-CoV) emerged 10 years later [1] . In December 2019, a novel coronavirus emerged in Wuhan, China [2] , subsequently called Severe Acute Respiratory Syndrome coronavirus 2 (SARS-CoV-2) [3] . SARS-CoV-2 causes coronavirus disease Coronaviruses predominantly cause respiratory tract infections in humans [1] . Specifically, the main symptoms of COVID-19 have been reported as fever, cough and shortness of breath [5] , with a less abrupt onset of symptoms compared to SARS [6, 7] . Data are still emerging regarding the epidemiology of COVID-19, though initial reports estimate a transmission rate (basic reproduction number, R0) of 2.2 [8] and a case fatality rate that increases among older adults [9] . Given the current pandemic, there is an urgent need to identify effective, safe treatments for COVID-19. Two such proposed treatments are chloroquine and hydroxychloroquine, which are well-established medications predominantly used to treat malaria, lupus and rheumatoid arthritis. In vitro studies have shown that chloroquine and hydroxychloroquine are effective at inhibiting SARS-Cov-2 infection, with the latter appearing to have more potent antiviral activity [10, 11] . Thus, repositioning of these drugs as antiviral therapies for COVID-19 is of global interest, however clinical data are limited and inconclusive. Currently, there are multiple ongoing clinical trials for use of these treatments in COVID-19, while the US Food and Drug Administration (FDA) has issued an Emergency Use Authorization (EUA) for oral formulations of chloroquine phosphate and hydroxychloroquine sulfate for patients hospitalised with COVID- 19 [12] . To date, whilst there have been many publications which have described the main effectiveness and safety concerns with these treatments, there has not been a systematic benefit-risk assessment on the use of chloroquine or hydroxychloroquine for COVID-19 treatment using a structured descriptive framework. A systematic benefit-risk assessment strengthens the ongoing monitoring of the benefit-risk balance for chloroquine or hydroxychloroquine in COVID-19 treatment. For this assessment, the Benefit-Risk Action Team (BRAT) framework is highly applicable as it allows identification of the key benefits and risks of a product in a defined disease context within a structured descriptive framework; further quantitative assessments can then be applied and conducted according to the availability of relevant data at that time [13] . The BRAT framework was also specifically designed to assist communication with regulatory authorities [14] . The decision-. CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 12, 2020. . https://doi.org/10.1101/2020.05.07.20093989 doi: medRxiv preprint making process is transparent due to the framework design, while any assumptions can be explored further by sensitivity analysis through a quantitative component [15] . Given the public health urgency with the COVID-19 pandemic, this benefit-risk assessment has been conducted based on publicly available information to date (data-lock April 7 th 2020). It is however acknowledged that there is extremely limited data available from ongoing clinical trials at this timepoint. To inform the debate expeditiously the benefit-risk assessment has been designed to be implemented regardless of the quantity of data available. The intention is that the framework will subsequently be readily available to repeat the assessment as further data arise, e.g. results from new clinical trials, allowing for rapid decision-making. To examine the benefit-risk profile of chloroquine or hydroxychloroquine in COVID-19 patients compared to standard of care, placebo or other treatments. The overall benefit-risk of the use of chloroquine or hydroxychloroquine as a treatment for COVID-19 compared to standard of care, placebo or other treatments was assessed using the BRAT framework. BRAT uses a six step iterative process to support the decision and communication of a Benefit-Risk Assessment: define decision context, identify outcomes, identify data sources, customise framework, assess outcome importance, and display and interpret key Benefit-Risk metrics [14, 15] . Three settings of interest were identified for use of these treatments in COVID-19; treatment for severe disease, treatment of milder disease in the community, and prevention in health care professionals exposed to the virus. For the purposes of this benefit risk assessment, we have focused on the use of chloroquine or hydroxychloroquine for the treatment of severe COVID-19 disease. The population of interest were patients with severe COVID-19, while the exposure of interest was hydroxychloroquine or chloroquine. The comparators of interest were standard of care, placebo or other treatments for COVID-19. . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 12, 2020. . https://doi.org/10.1101/2020.05.07.20093989 doi: medRxiv preprint Initially, all potential benefits and risks related to hydroxychloroquine or chloroquine, regardless of importance, were identified. From these the key benefits and risks associated with hydroxychloroquine and chloroquine use were selected by clinician judgement. Key benefits and risks were those which were considered to drive the benefit-risk balance of the drug. A value tree was constructed using these key benefits and risks, all of which were ranked in order of considered importance. We searched PubMed and Google Scholar to identify suitable data for inclusion. In both databases, we searched for papers on: (((((((((chloroquine* AND SARS*)) OR (chloroquine* and covid*)) OR (chloroquine* AND coronavirus))))) OR (((chloroquine* AND 2019-NCov)))) (((((((((hydroxychloroquine* AND SARS*)) OR (hydroxychloroquine* and covid*)) OR (hydroxychloroquine* AND coronavirus))))) OR (((hydroxychloroquine* AND 2019-NCov)))) Papers were included if they reported quantitative data on effectiveness and/or safety of chloroquine or hydroxychloroquine in patients with severe COVID-19. Case reports were excluded. Results were restricted to English language only (abstracts in English language were acceptable where sufficient data provided) and peer-reviewed publications since 2019 to 7 th April 2020. Data were extracted for each benefit and risk, for hydroxychloroquine, chloroquine and the comparator (standard of care, placebo or other treatments), where available. EudraVigilance (up to 4 th April 2020) and FAERS spontaneous reporting data (up to 31 st December 2019) for hydroxycholoroquine and chloroquine were also examined. A summary benefit-risk table was created to allow visualisation of the magnitude of each benefit and risk. Risk differences and corresponding 95% confidence intervals (CI) were calculated for each outcome where both numerator (number of events) and denominator (number of patients at risk) were available for both the treatment group (hydroxychloroquine or chloroquine) and comparator group. No appropriate comparator groups were identified in EudraViligance for hydroxychloroquine or chloroquine because there are no established treatments for COVID-19 at this time. Consequently, spontaneous reports are not included in the benefit-risk table and are presented in the text only. . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 12, 2020. . https://doi.org/10.1101/2020.05.07.20093989 doi: medRxiv preprint Due to lack of data, a fully quantitative assessment was not undertaken. However, the outcomes identified in the value tree were ranked so that swing weighting can be applied in future assessments. The weighted net clinical benefit (wNCB) can subsequently be calculated using these weights [15] [16] [17] . We would propose using the Sutton et al method, where benefits have a positive contribution to the wNCB and risks have a negative contribution [17] ; the overall wNCB would be considered positive (benefit outweighs the risk) where wNCB >0. A sensitivity analysis can also be used to examine the robustness of the assigned weights and whether significant changes would alter the benefit-risk profile [13] . The value tree reflecting the key benefits and risks related to hydroxychloroquine or chloroquine treatment in COVID-19 is displayed in Fig 1. Data for these outcomes are presented in the data extraction table and key benefit-risk summary table (Tables 1 and 2, respectively). From literature searching we identified 71 papers from PubMed and 379 results from Google Scholar for chloroquine. We also identified 29 papers from PubMed and 158 results from Google Scholar for hydroxychloroquine. After initial review and removal of duplicates, 15 papers were reviewed further to determine if they met all inclusion criteria; 3 papers were included in the final benefit-risk assessment. A reduction in the risk of death from COVID-19 is considered to be a key benefit of treatment with chloroquine or hydroxychloroquine compared to standard of care, placebo or other treatments. We found no comparative data were available for this key benefit. A reduction in the risk of ICU admission was identified as a key benefit, though no comparative data were found in the existing literature. Another key benefit identified was reduction in the risk of non-invasive ventilation. No data comparing hydroxychloroquine or chloroquine to standard of care, placebo or other treatments was identified. . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 12, 2020. . https://doi.org/10.1101/2020.05.07.20093989 doi: medRxiv preprint Secondary clinical outcomes were also considered key benefits as they refer to anticipated clinical endpoints as a result of chloroquine or hydroxychloroquine treatment, which reflect potential reductions in disease progression, such as development of pneumonia or acute respiratory distress syndrome (ARDS). No comparative data were found in the existing literature for these outcomes. A further key benefit identified was a reduction in the number of patients requiring supplemental/high flow of oxygen. No comparative data was identified. Viral load parameters were considered a surrogate endpoint in this benefit risk assessment context, with outcomes such as virological clearance reflecting the benefit of recovery from COVID-19. In the study by Gautret et al [18] , the authors examined virological clearance . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 12, 2020. No comparative data were identified for cardiac toxicity. In EudraVigilance, there were 13 reports of QT prolongation, one report of ventricular arrythmia, one report of Atrioventricular (AV) block, one fatal cardiac arrest and one non-fatal cardiac arrest in patients using hydroxychloroquine for coronavirus infection. In addition, there were 6 reports of QT prolongation, one report of tachyarrhythmia, and one report of ventricular tachycardia in patients using chloroquine for coronavirus infection. No comparative data were identified on immune reactions among patients using hydroxychloroquine or chloroquine for COVID-19 treatment. No comparative data on ocular events among patients using hydroxychloroquine or chloroquine for COVID-19 treatment were identified. No comparative data were identified on serious skin reactions among patients using hydroxychloroquine or chloroquine for COVID-19 treatment. In the study by Chen J et al [19] , abnormal liver function was reported for four patients in the hydroxychloroquine group (risk=0.27) and three patients in the comparator group (risk=0.20). There was no significant difference in risk between the two groups (RD=0.07, 95% CI: -0.28, 0.41). In EudraVigilance, there were two reports of liver injury and one report of hepatocellular injury in patients using hydroxychloroquine for coronavirus infection. In addition, there was one report of hepatocellular injury in a patient using chloroquine for coronavirus infection. No comparative data were identified on severe hypoglycaemia among patients using hydroxychloroquine or chloroquine for COVID-19 treatment. . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 12, 2020. . https://doi.org/10.1101/2020.05.07.20093989 doi: medRxiv preprint No comparative data were identified for adverse haematological events. In EudraVigilance, there were two reports of neutropenia in patients using hydroxychloroquine for coronavirus infection. No comparative data were identified on drug interactions with chloroquine or hydroxychloroquine. In EudraVigilance, one potentiating drug interaction was reported in a patient using chloroquine for coronavirus infection; the patient experienced QT prolongation and was also taking lithium carbonate and quetiapine fumarate. Several potential key benefits were identified with use of hydroxychloroquine or chloroquine in COVID-19 treatment: reduction in risk of death, reduction in ICU admission, reduction in non-invasive ventilation, reduction in secondary clinical outcomes, reduction in use of oxygen and improved viral load parameters (e.g. virological clearance). The key benefits were outcomes highlighted from currently ongoing clinical trials. In addition, several potential key risks were also identified; cardiac events, immune reactions, ocular events, skin events, liver events, severe hypoglycaemia, blood events and drug interactions. Cardiac toxicity is of particular importance for COVID-19 patients. Both potentially the disease itself and anticipated treatment strategies with chloroquine or hydroxychloroquine pose significant risk of cardiac arrhythmias [21] . Short term use, as expected for the treatment of COVID-19, is likely to pose a lower risk of cardiac toxicity, nevertheless the risk cannot be overlooked as patients are expected to be on higher doses, possibly concomitantly taking other QT prolonging agents, in addition to having a potentially elevated risk due to the disease itself. One of the most serious toxic effects of hydroxychloroquine/ chloroquine are ocular side effects, notably retinopathy [22] [23] [24] [25] [26] , although the risk for hydroxychloroquine is considered to be lower [27] . Both duration of use and weight-based dosing (dose per kg) are important parameters for the risk of developing retinopathy [26] . The risk of retinal damage over a short time period may be negligible even with high doses [25] . However, given the seriousness of the outcome we have identified this as a key risk in the benefit-risk framework. In addition, whilst the most robust evidence for safety concerns associated with chloroquine and hydroxychloroquine is with longer term use, although rare, there have been case reports after . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 12, 2020. . https://doi.org/10.1101/2020.05.07.20093989 doi: medRxiv preprint weeks as opposed to longer duration of use, with respect to adverse hepatic and haematological effects, and hypoglycaemia [28] [29] [30] . Very few studies were identified that compared benefits and risks between those using hydroxychloroquine or chloroquine and those receiving standard of care, placebo or other treatments. Data were only identified for one benefit (viral load parameters) and one risk (liver events). For the benefit of virological clearance, three studies with very small sample sizes were identified. These studies examined virological clearance at different time points as their primary outcome, meaning results from these studies could not be pooled. Only one study revealed a significant risk difference between hydroxychloroquine and the comparator group (standard of care), though given the small sample size and biases in the study design this should be interpreted with caution. For the risk of abnormal LFTs, only one studied was identified with a small sample size. No significant risk difference between hydroxychloroquine and the comparator group (standard of care). Whilst there is a paucity of comparative data from the literature search at this stage, we identified a number of reports in Eudravigilance. The following data was obtained from EudraVigilance for patients taking chloroquine or hydroxychloroquine: QT prolongation (n=19), ventricular arrythmia (n=1), atrioventricular (AV) block (n=1), non-fatal cardiac arrest (n=1), fatal cardiac arrest (n=1), tachyarrhythmia (n=1), ventricular tachycardia (n=1), liver injury (n=2), hepatocellular injury (n=2), neutropenia (n=2), drug interaction (n=1). These data can be used in future benefit-risk assessments once an appropriate treatment comparator has been identified, which should be an established treatment for COVID-19. Given the paucity of available data on benefits and risks with hydroxychloroquine or chloroquine compared to standard of care, placebo or other treatments, we chose not to undertake a fully quantitative assessment of the benefit-risk balance at this time. Instead, all available comparative data for key benefits and risks are presented in a summary table. Further data from clinical trials and observational studies are required in order to determine whether the benefit-risk profile of hydroxychloroquine or chloroquine in treating COVID-19 is favourable. An important point for discussion regardless of the benefits and risks of hydroxychloroquine and chloroquine for COVID-19 treatment is the availability of these medicines. Inadequate . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 12, 2020. . supply and shortages have already been reported [31] , meaning availability is not guaranteed for those using these treatments for licensed indications or COVID-19. This must be taken into consideration in addition to the benefit-risk assessment, since there can be no benefit from a medication which is unobtainable. Sample sizes for each outcome were limited to those available in the original studies and may not have adequate power to detect differences in risk between groups, especially where the outcomes examined were not the primary outcome of interest. The benefit-risk assessment is limited by the availability of data in the published literature. However, this assessment can be subsequently updated once further data from clinical trials are available. In addition, given the public health urgency of the COVID-19 pandemic, it is important to provide a systematic assessment of the benefits and risks of hydroxychloroquine and chloroquine treatments with evidence available to date and create a framework which can be used to rapidly update the assessment when further data are available. Data quality is not reflected in this benefit-risk assessment, though all data included were extracted from peer-reviewed manuscripts. Of note, a statement was issued by the International Society of Antimicrobial Chemotherapy [32] regarding the Gautret et al paper published in the international journal of antimicrobial agents [18] . The paper was not considered to meet the society's expected standards and though it was peer-reviewed, the editor-in-chief was not involved in this process [32] . Confirmation of causality was not a requirement for inclusion of data in the BRAT assessment. Patients may have been on other concomitant medications or had other medical conditions at the time of hydroxychloroquine or chloroquine treatment. Finally, we considered hospitalisation of patients to reflect severe COVID-19, but we acknowledge that severity of disease may vary regardless of hospitalisation. Overall, no conclusion can be made on the benefit-risk profile of hydroxychloroquine or chloroquine in the treatment of COVID-19 compared to standard of care, placebo or other treatments at this time. Whilst the availability of comparative data are limited, the current framework summarises the key anticipated benefits and risks. As further data from clinical trials and real world use on these benefits and risks becomes available, this can be incorporated into the framework for an ongoing benefit-risk assessment. . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 12, 2020. No data - Origin and evolution of pathogenic coronaviruses A novel coronavirus outbreak of global health concern COVID-19) and the virus that causes it Epidemiological and clinical characteristics of 99 cases of 2019 novel coronavirus pneumonia in Wuhan, China: a descriptive study COVID-19: what is next for public health? A familial cluster of pneumonia associated with the 2019 novel coronavirus indicating person-to-person transmission: a study of a family cluster Early Transmission Dynamics in Wuhan, China, of Novel Coronavirus-Infected Pneumonia Estimates of the severity of coronavirus disease 2019: a model-based analysis In Vitro Antiviral Activity and Projection of Optimized Dosing Design of Hydroxychloroquine for the Treatment of Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) Remdesivir and chloroquine effectively inhibit the recently emerged novel coronavirus (2019-nCoV) in vitro Systematic benefit-risk assessment for buprenorphine implant: a semiquantitative method to support risk management A case study using the PrOACT-URL and BRAT frameworks for structured benefit risk assessment Benefit-risk analysis: a proposal using quantitative methods A Bayesian approach to evaluating net clinical benefit allowed for parameter uncertainty Hydroxychloroquine and azithromycin as a treatment of COVID-19: results of an open-label non-randomized clinical trial A pilot study of hydroxychloroquine in treatment of patients with common coronavirus disease-19 (COVID-19) Treating COVID-19 with Chloroquine Clinical Characteristics of 138 Hospitalized Patients With 2019 Novel Coronavirus-Infected Pneumonia in Wuhan, China Hydroxychloroquine and chloroquine retinopathy: screening for drug toxicity Hydroxychloroquine dosing in immunemediated diseases: implications for patient safety Effects of chloroquine on viral infections: an old drug against today's diseases? COVID-19 and Chloroquine/Hydroxychloroquine: is there Ophthalmological Concern? Recommendations on Screening for Chloroquine and Hydroxychloroquine Retinopathy -2016 COVID-19: a recommendation to examine the effect of hydroxychloroquine in preventing infection and progression Hypoglycaemia induced by hydroxychloroquine in a non-diabetic patient treated for RA Fulminant hepatic failure secondary to hydroxychloroquine A case of thrombocytopenia associated with the use of hydroxychloroquine following open heart surgery A Rush to Judgment? Rapid Reporting and Dissemination of Results and Its Consequences Regarding the Use of Hydroxychloroquine for COVID-19 Statement on IJAA paper | International Society of Antimicrobial Chemotherapy