What is the translational efficacy of chemotherapeutic drug research in neuro-oncology? A systematic review and meta-analysis of the efficacy of BCNU and CCNU in animal models of glioma

Introduction The translational value of experimental therapeutic neuroscience research to clinical practice is highly variable. This has been particularly well demonstrated in the field of neuroprotective agents following either head injury or stroke. In this study we evaluate the efficacy of systemic BCNU and CCNU in experimental glioma models and how the experimental data has translated into clinical practice. Methods A systematic review of the efficacy of BCNU and CCNU, against experimental rodent and murine in vivo glioma models was conducted. Selected articles were graded on a 15 point scale for scientific methodology. A stratified meta-analysis based on median-survival data and effect sizes was performed to generate global-efficacy estimates for BCNU and CCNU, and to produce 'weighted-mean effect-sizes' for individual sub-categories of selected study-characteristics. Results Fourteen papers satisfied search criteria and encompassed 231 treatment comparisons in 2256 animals. The median methodology score was 9 (range 7-12/15). Global-efficacy estimates were BCNU 0.194 (95% CI -0.538 to 0.927) and CCNU 0.432 (95% CI -0.392 to 1.256), with CCNU being significantly more effective than BCNU. Because of these wide confidence intervals a beneficial or detrimental effect of either agent could not be confirmed. Most selected study-design characteristics (e.g. glioma cell line, drug dosage, drug scheduling, mode of drug administration, timing of therapy after glioma implantation but not animal used) significantly influenced the efficacy-results obtained. The methodological score did not influence efficacy-estimate. Conclusion This review has found (i) experimental-design influenced the efficacy-data obtained and (ii) that there is highly variable outcome data for the efficacy of both BCNU and CCNU in experimental in vivo rodent and murine glioma models. In many ways these findings are analagous to the use of nitrosoureas in human malignant glioma. The statistically significant small beneficial effect of nitrosoureas in combination with other chemotherapeutic agents in human glioma was only noted after a meta-analysis of human randomized controlled trials.