Adverse drug reactions of hydroxychloroquine: Analysis of French pre-pandemic SARS-CoV2 pharmacovigilance data

Abbreviations

ADRs

adverse drug reactions

HCQ

hydroxychloroquine

SL

systemic lupus

WHO

World Health Organization

Introduction

Hydroxychloroquine (HCQ) is a 4-aminoquinoline derivative indicated in systemic lupus (SL), lucites and rheumatoid arthritis [1]. Its main adverse drug reactions (ADRs) described in the summary of product characteristics are ocular damage, particularly retinopathy during long-term exposure with high doses, cardiac damage with QT interval prolongation, cardiomyopathy during chronic treatment, potentially severe skin damage, neuromyopathy, haematological damage such as bone marrow aplasia and agranulocytosis.

Several publications have focused on specific ADRs, particularly cardiac, auditory and ocular ones [2], [3], [4] but to date, to our knowledge, there are no published global pharmacovigilance data on ADRs reported in routine use with this drug.

Methods

We performed a descriptive analysis of spontaneously reported ADRs. These ADRs were collected, analysed and recorded by the network of 31 French regional pharmacovigilance centres between January 1, 1985 (beginning of the French pharmacovigilance database) and December 31, 2019, i.e. before the use of this drug in the context of SARS-CoV2 pandemic. All reports where HCQ was coded as “suspect” or “interaction” were extracted from the French pharmacovigilance database. Causality assessment was performed according to the French causality assessment method [5], [6].

Results

A total of 1717 reports were included, with 985 (57.4%) where HCQ was the only causal drug. These 1717 cases were reported over a period of 35 years (mean value of 50 reports per year), with more reports notified since 2012, reaching approximately 100 cases per year since 2016. Reporters were 90% physicians, 73% working in a health establishment. Less than 2% of reporters were patients or patient associations. The number of “serious” reports, according to World Health Organization (WHO) criteria, was 1022 (59.5%). The M/F sex ratio was 0.29 and mean age 50.8 ± 18.4 years.

After excluding reports during pregnancy, 12 deaths were found. Five involved overdose and/or pharmacodependence to other products. The 7 other reports were 2 septic shocks, 1 pancytopenia, 4 cardiac ADRs (1 cardiac arrest at home after 4 years of exposure, 1 global cardiac failure after 1 year of treatment, 1 ventricular tachycardia on the 3rd day of treatment and 1 hypertrophic cardiomyopathy after 24 years of treatment). There was no overdose for these 4 cardiac ADRs.

These 1717 reports described 2727 ADRs, mainly cutaneo-mucosal or immuno-allergic ones (32%). Other ADRs were found in less than 10% of reports: neurological (9%) [central and peripheral], haematological (8.5%), ocular (7.6%), digestive (6.7%) and cardiac (3.7%) disorders.

However, when looking the 985 reports where HCQ was the only “suspected” drug, the frequency order was different, with the majority (34%) always involving mucocutaneous or immunoallergic disorders, followed by ocular (11%), neurological (10%), digestive (7.5%), hematological (5.1%) and cardiac (4.2%) ADRs.

During the 1985–2019 period, there were 93 reports of cardiac ADRs (57 with HCQ the only “suspected” drug), 85% “serious”. Sex ratio was 0.29 and mean age 53 ± 16 years. Approximately half of reports described development or aggravation of left ventricular failure, suggesting a sodium channel blocking effect of hydroxychloroquine. In some reports, regression or even disappearance of the cardiac ADR was observed after drug discontinuation. Approximately 22% of reports described symptoms suggestive of cardiac arrest or ventricular arrhythmia, half of which involved complex conduction disorders (9/21). This observation suggests that HCQ could block calcium but also especially voltage dependent potassium channels. Finally, rhythm and conduction disorders were found in 20% of reports in the French pharmacovigilance database.

Discussion

Analysis of French pharmacovigilance data makes it possible to define a safety profile of HCQ in patients treated for autoimmune diseases (only some isolated reports with off-label use were found).

This paper underlines that the ADRs are mainly cutaneous, mucocutaneous and immuno-allergic. However, the detailed analysis found “serious” cardiac ADRs, that we related to HCQ after careful clinical analysis, taking into account suggestive chronology and semiology. As with any pharmacovigilance data, the figures given should not be minimised in view of the generally low reporting rate of ADRs. It is admitted that only 5–10% of ADRs are reported [7]. Spontaneous reporting depends on drug's age, perception of risk by healthcare professional or patient, seriousness of the reaction, lack of knowledge of some ADRs, diagnostic difficulties, etc. [6], [8] In our work, “seriousness” of the reports probably means that the much larger number of mild to moderate cases were never reported (and probably never detected).

Therefore, these data shall be taken into account in risks-benefits balance of HCQ, in particular in situations where the clinical benefit is not proven.

Disclosure of interest

The authors declare that they have no competing interest.