How important are concurrent vehicle control groups in (sub)chronic non-human primate toxicity studies conducted in pharmaceutical development? An opportunity to reduce animal numbers

Safety assessment of human pharmaceuticals demands extensive animal experiments before a compound can be tested in patients or released on the market. Such experiments typically include concurrent vehicle control groups. Reconsidering the need for concurrent controls could support the strive to reduce the use of animals for scientific purposes. We reviewed reports from 20 (sub)chronic toxicity studies that were conducted in non-human primates (NHP) to characterize hazards of novel human pharmaceuticals. Firstly, we determined the toxicological endpoints that were identified to characterize the hazard. Secondly, we evaluated if the hazard could have been identified without reference to the concurrent controls. Thirdly, we employed an alternative statistical method to test for any significant change related to dose level or time. We found that toxicologically relevant hazards were identifiable without reference to concurrent controls, because individual measurements could be compared with pre-dosing values or because individual measurements could be compared to historical reference data. Effects that could not be evaluated without reference to concurrent controls were clinical observations and organ weights for which appropriate historical reference data was not available, or immune responses that could not be compared to pre-dosing measurements because their magnitude would change over time. Our investigation indicates that concurrent control groups in (sub)chronic NHP toxicity studies are of limited relevance for reaching the study objective. Under certain conditions, regulatory (sub)chronic NHP toxicity studies represent a good starting point to implement virtual control groups rather than concurrent control groups in nonclinical safety testing.