Prediction of Human Lethality of Psychoactive Drugs From Rodent LD50 Values

Introduction

A recent crime survey (Crime Survey for England and Wales, 2016/17) revealed that 19.2% of adults aged 16 to 24 had taken an illicit drug in the past year. According to the Office for National Statistics (Office for National Statistics), deaths related to drug poisoning in England and Wales are increasing year on year, from 2597 in 2012 to 3756 in 2017, an increase of 44.6%. Hence there is a need for the determination of lethal toxicity, based usually on animal experiments, for both medicines and drugs of abuse in order to obtain an estimation of safety and toxicity in humans.

It is acknowledged (Gable, 1993) that the determination of the human lethal dose (HLD) of a psychoactive substance is very difficult for a number of reasons, for example whether the substance is taken on its own or together with other substances, whether the person is a new or habitual user, whether the person is alone or in the company of others, and because of the sometimes marked interpersonal variability of rates of metabolism. Gable (1993) stated that ‘the “best-guess lethal dose” for an average adult human who has not developed tolerance to the substance is probably the LD₅₀ extrapolated from a broad range of laboratory animal studies that falls within the range of lethality cited in clinical or forensic reports’.

The animal LD₅₀ test was developed by Trevan (1927) for the biological standardization of drugs. Like HLD values, it cannot be considered as a biological constant, for it has been pointed out (Zbinden & Flury-Roversi, 1981) that it can vary with animal species, age, sex, weight, health, genetic variability, diet, method of administration, time of assessment after administration, ambient temperature, housing conditions (e.g. isolated or aggregated), time of day/night and time of year. It is therefore not surprising that correlations between animal LD₅₀ and human toxicity values are generally poor (Abbott, 2005), although a few publications have reported modest results. For example, Hoffmann et al. (2010) found that human acute lethal doses of 30 chemicals correlated with rat oral LD₅₀ values (coefficient of determination (r²) = 0.571); Ekwall et al. (1998) found that human acute lethal doses of 50 chemicals (including a few psychoactive drugs) correlated reasonably well with rat and mouse oral LD₅₀ values (r² = 0.607 and 0.653 respectively); Jover et al. (1992) reported correlations between HLDs of 10 chemicals with rat and mouse oral LD₅₀ values (prediction errors 1.04 and 0.68 log unit respectively). Gable (2004) reported median HLD and range values for 20 commonly abused psychoactive substances; for example, he gave the median HLD for methadone as 100 mg, and the lethal range as 20 – 400 mg. He did not, however, examine possible correlations between human and animal lethal doses.

By far the most widely used animals in LD₅₀ testing are rodents, especially rats and mice. It was therefore decided to examine correlations between HLDs and (following the suggestion of Gable (1993)) a range of rat and mouse LD₅₀ values, with the aim of obtaining one or more valid animal models of human lethality of psychoactive drugs. Such models could then be used to predict the human lethality of new ‘designer’ drugs as they become available. The use of mean LD₅₀ values in the present work hopefully means that many experimental errors and variations in their measurement were cancelled out.