Plant Foods versus Compounds in Carcinogenesis: Observational versus Experimental Human Studies

The protective role of plant foods and its constituents in cancer prevention is under renewed debate since the results of recent observational studies on colorectal cancer as well as large-scale human experimental studies on colorectal adenoma recurrence are disappointing. However, most short-term experimental human studies do show that plant foods favourably modulate potential cancer-preventive mechanisms. Which methodological pitfalls may explain the inconsistencies within and between different study designs? What are the advantages and limitations of the different study approaches? Observational studies do have the advantage to study the population at large with ultimate disease as the study endpoint. These studies are limited by the difficulty to estimate intake of individual compounds by questionnaires and the lack of biological markers of relevant exposure. Controlled experimental short-term studies in humans rely on biological markers of disease as intermediate endpoints. Relatively low sensitivity and specificity of these markers may complicate extrapolation of results. In the case of long-term and large-scale human intervention studies with disease endpoints, issues such as time, dose and duration of intervention, compliance and choice of the study population influence the interpretation of results. An integrated approach combining designs, and implementing new techniques to identify biomarkers, may clarify the role of plant foods in carcinogenesis.