Mendelian randomization: a new approach to studying epidemiology in ESRD.

Associations between environmental exposures and disease in epidemiological studies often are confounded and may result in erroneous interpretations. The random assortment of genes from parents to offspring at gamete formation--Mendelian randomization--is emerging as a useful method for studying the nature (causal or not) of environmental exposures. This occurs because the association between a disease and a polymorphism that mimics the biological link between a given exposure and the same disease is unaffected by the reverse causation that may complicate the interpretation of observational studies. Thus, similarly to randomized trials, association studies between gene polymorphisms with a well-established function may be useful for excluding confounding as an explanation for a given epidemiological relationship. The rationale behind this concept is that transmission of genes occurs in a random way; therefore, offspring have an equal chance of inheriting either of the 2 alleles that their parents have at any particular locus, a phenomenon independent from environmental factors. This is similar to the construct of randomized trials, in which randomization is expected to produce similar distributions of measured and unmeasured confounders between treated and untreated (control) groups. The equivalence between random assortment of alleles and random assignment of interventions in randomized studies is particularly useful because it may give a clue for interpreting associations that may be produced as both the effect of a gene or result of an environmental exposure. Examples are provided about the use of this concept to elucidate paradoxical inverse associations in epidemiological studies in the general and end-stage renal disease populations.