An FDA overview of rodent carcinogenicity studies of angiotensin II AT-1 receptor blockers: pulmonary adenomas and carcinomas.

Sipahi et al. (2010) performed a meta-analysis of 5 clinical trials (n=68,402) of 3 Angiotensin II (AngII) receptor subtype AT-1 blockers (ARBs) in cardiovascular disease. It revealed excess new lung cancer diagnoses in the cohorts treated with an ARB and background therapy (0.9% vs. 0.7% in non-ARB control; RR: 1.25; CI: 1.05-1.49; p=0.01). The FDA responded with a larger meta-analysis of 31 clinical trials (n=155,816) of ARBs that found no evidence of any excess of site-specific cancer (lung, breast, prostate), solid/skin cancer or cancer death (FDA safety communication, 3 June 2011). The FDA then re-visited the 19 rodent carcinogenicity assays of 9 ARBs, starting with those for Losartan in 1994, for any evidence of dosage-related lung tumorigenicity in this class. Assays were performed in 5 strains of rats and 5 strains of wild-type and transgenic mice per protocols and dosages sanctioned by FDA's executive carcinogenicity assessment committee (eCAC). Duration was lifetime except for 26-week assays of azilsartan and olmesartan in transgenic Tg rasH2 mice, and an assay of olmesartan in p53(+/-) transgenic mice. The dosages provided exposures approximating, and in most cases up to 20-300times greater than, that in patients. Depending on strain, up to 35% of untreated mice spontaneously developed lung tumors. Regression analysis of placebo-corrected mouse lung tumor incidence collapsed across strains, gender, and ARBs vs. multiples of human exposure revealed no excess lung neoplasia. The R(2) of <0.001 reflected the virtually identical number of treated cohorts with more tumors than its control cohort vs. those with less. Regardless of strain, both control and medicated rats were essentially devoid of lung tumors in the lifetime trials. Accordingly, there was neither promotion of background lung tumors in the mouse, nor initiation of de novo lung tumors in the rat. The negative lung findings in the mouse Tg rasH2 strain are also noteworthy given that, historically, the most prevalent spontaneous tumors in 26week trials in that model are lung adenomas and carcinomas. The negative results of the 19, mostly lifetime, assays for cancer viewed en masse add to the results of the meta-analysis of the shorter clinical trials of ARBs that were benign regardless of statistical method used (random vs. fixed effect), comparator arm (with or without ACE-inhibitors) and major co-factors (smoking or cancer history).