Historically, animal models have been relied upon to provide preliminary efficacy data for therapeutics against infectious diseases in support and justification of subsequent definitive efficacy studies in human participants to obtain regulatory approval by the Food and Drug Administration (FDA). Because the preclinical data would be evaluated in the context of knowledge from human studies, any deficiencies in data correlation and extrapolation from the animal models to the human condition would presumably be compensated for by the actual data collected during the human studies. Biothreats represent a special problem in that efficacy studies before an actual event are unlikely to take place. In addition, the actual risk of a biothreat attack is difficult to ascertain. These difficulties are even more pronounced in the case of the “unknown-unknowns”.6
Comparing the evaluation process for bioterrorism countermeasures following the preclinical development stage with that for drugs for which human efficacy studies are possible puts in better perspective the regulatory challenges with which the countermeasure development for TMT (or other biodefense) products is beset. Under optimal circumstances, the current process from drug discovery to FDA approval takes an average of 10 to 15 years and costs more than $1 billion (Tamimi and Ellis 2009). According to some estimates the developmental cost of a single drug has soared from $1.1 billion in 1995 to $1.7 billion in 2002, factoring in the costs of failed prospective drugs (Crawford 2004; Mundae and Östör 2010). Those figures apply equally to biopharmaceuticals and small molecules (DiMasi and Grabowski 2007). To date only about 8% of drugs that successfully enter phase 1 studies eventually are granted market approval by the FDA as compared with 14% in the 1980s. The success rate of pharmaceuticals from the first phase 1 study in humans to market is less than 10% (DiMasi et al. 2010).
The main causes of failure in the clinical trial setting are safety problems, which account for about 20% of the attrition rate, and lack of effectiveness, which accounts for about 40% (Kola and Landis 2004; Peck 2007). Inability to predict these failures before human testing or early in clinical trials dramatically escalates costs. In the infectious disease arena, data from the 10 largest pharmaceutical corporations in the period of 1991-2000 showed a success rate of about 15%, while the average success rate for all indications was 11% (Gilbert et al. 2003). Similarly, DiMasi and colleagues (2010) showed a success rate for systemic infectious disease of 15.6% during 1994 and 2003. It is useful to note that from 1981 to 1992 the success rate of anti-infective drugs was 28.1% and that large biopharmaceutical companies appeared to have a higher success rate of 30.2% for all indications (DiMasi 2001). A key
3 Lack of data sharing further compounds differences in methods or lack of reproducibility of results across models (see chapter 5 for further discussion).
4 The limitations of animal models for other disease indications (in addition to those encountered in emerging infectious diseases or biothreats research) have been documented in a number of meta-analyses (see Macleod 2011; Perel et al. 2007; Suntharalingam et al. 2006; van der Worp et al. 2010).
5 As discussed in Developing Animal Models for Use in Animal Rule Licensure: The NIAID Approach (Appendix C, p 111-112), developing animal models in biocontainment requires substantial financial and infrastructure investment.
6 As defined in the introduction, the term “unknown-unknown(s)” refers to pathogen(s) that may not be known or knowable because they currently may not exist. Due to the current or future possibility that they may exist, they are considered potential threats (e.g., a novel, genetically engineered, or created pathogen).