hERG potency data in new drug applications submitted to the U.S. Food and Drug Administration between 2008 and 2024

Abstract

Since the implementation of ICH S7B by the FDA in 2005, sponsors are required to test human pharmaceuticals at hERG before Phase 1 clinical trials. ICH S7B stipulated that, ‘Ascending concentrations should be tested until a concentration-response curve has been characterized or physicochemical effects become concentration-limiting.’ Thus subsequent New Drug Applications (NDAs) contained hERG potency data. By interrogating the Drugs@FDA database, we have investigated hERG potency in NDAs submitted 2008–2024 inclusive.

We developed an R-based text-mining pipeline to automate data extraction from machine-readable NDA documents in the Drugs@FDA database, using search terms related to hERG/IKr inhibition. The code screened ~11,000 NDAs and identified 1567 with potentially relevant content. After excluding 1000 non-relevant NDAs (e.g., ophthalmic, topical, diagnostic, and large-molecule drugs >1000 Da), hERG data were expertly curated from the remaining 567 documents (corresponding to 288 drugs) and classified into one of three categories based on the data provided: (1) IC₅₀ value (160 drugs); (2) percent inhibition at specified concentration (65 drugs); (3) no observed block at specified concentration (63 drugs). For category 2 compounds, IC₅₀ values were estimated using the Hill equation (slope = 1) if inhibition was ≥25% (11 drugs); otherwise, compounds were assumed to exhibit no hERG block at the tested concentration (54 drugs). To enable us to use the compounds showing no hERG block, we calculated a ‘worst case’ IC₅₀ by multiplying the quoted test concentration by 100.

Using only IC₅₀ values from category 1 or those estimated from category 2, no clear temporal trend was apparent. Nor did one emerge by adding ‘worst case’ IC₅₀ values to the dataset. This was the case whether the data were assessed by year or by grouping the data into 2–4 year ‘bins’. No trends emerged when we split the dataset into oncology and non-oncology drugs. Clearly, this analysis is blind to efficacious free drug concentrations and therefore to safety margins. Also, it does not assess the possibility that there may be a trend in hERG potency in INDs that is not reflected in NDAs owing to drug failures in development.