Hydroxychloroquine and chloroquine block hERG at doses relevant to COVID-19.
Recent in vitro studies demonstrated EC50s for inhibition of viral replication at 24/48 hours of 0.72 µM/6.14 µM for hydroxychloroquine (1) and suggested that these concentrations could be achieved with a dosing regimen of 2 X 400 mg loading/2 X 200 mg daily. However, other trials have proposed significantly higher dosing. For example, the Norwegian arm of the WHO’s SOLIDARITY trial had a dosing regimen of 2 X 800 mg loading and then 2 X 400 mg daily (24). Based on Phase I trials in patients with melanoma this would be expected to result in an average peak blood Cmax of 2000-3000 ng/ml, with some patients reaching 4000–6000 ng/ml (25). Assuming ~50 % serum protein binding (26), this would correspond to an average peak plasma Cmax of 1000-1500 ng/ml (3–4.5 µM) but up to 4000–6000 ng/ml (6-9 µM) in extreme cases. In comparison, our analysis showed IC50 values for hERG block of 3.7 µM for hydroxychloroquine at physiological temperature (Figure 1).
For chloroquine, Yao et al. measured an EC50, for inhibition of viral replication at 24/48 hours, of 5.47 µM/23.9 µM and suggested that these concentrations could be reached with a daily 500 mg dose (1). Various PK/PD studies broadly support this with 450 mg daily dose for three days yielding a median Cmax of 376 µg/L (1.2 µM) in one study (27) and long term 500 mg daily dose shown to equilibrate to >10 µM plasma concentration (equal to ~3000 µg/L) in a second study (28). However, initial guidelines issued for use of chloroquine in COVID-19 patients recommended twice this dose (29) and this is reflected in proposed trials (30). For comparison, we measured IC50 for hERG block of 1.5 µM for chloroquine (Figure 1). Together, these data suggest significant hERG block, and hence proarrhythmic risk, at doses relevant to COVID-19. For example, a plasma Cmax of ~ 3 µM (a realistic estimate for either drug based on the above), would cause 43 and 67 % block of hERG for hydroxychloroquine and chloroquine. This use case therefore falls well short of the 30 – 45 fold safety margins (Cmax:hERG IC50) that have been proposed as a safe threshold in preclinical assays (31, 32). Furthermore, our analysis does not take into account that both chloroquine and hydroxychloroquine have been reported to accumulate in the heart (33), meaning cardiac hERG channels could be exposed to much higher concentrations. Our data therefore support the conclusion that chloroquine and hydroxychloroquine carry cardiac safety liabilities that are of significant concern at the doses being proposed for treatment of COVID-19.