Portal de Periódicos da CAPES

Heme Oxygenase-1 Polymorphisms and Renal Transplantation Outcomes: Balancing at the Detection Limit of Allelic Association Studies

2008; Elsevier BV; Volume: 8; Issue: 5 Linguagem: Inglês

10.1111/j.1600-6143.2008.02207.x

1600-6143

Petra Hřibová, Petra Reinke, Jan Petrášek, Irena Brabcová, Jaroslav A. Hubáček, Ondřej Viklický,

Eicosanoids and Hypertension Pharmacology

To the Editor: Heme oxygenase-1 (HOX1), a cytoprotective factor with anti-inflammatory and antioxidant effects, seems to attenuate the ischemia/reperfusion injury in organ transplantation. Baan et al. (1) described an association of the dinucleotide (GT)n tandem repeat polymorphism in the HOX1 promoter with the outcome of deceased donor renal transplantation (RTx). They found a reduced survival of grafts carrying the HOX1 (GT)n L-allele, which decreases the gene expression, and proposed that in patients receiving grafts with this allele the cold-ischemia time should be kept at minimum. Their finding was not confirmed by Courtney et al. (2), who rejected the effect of the polymorphism on graft survival. However, as polymorphisms act only as disease modifiers, the results of Courtney et al. (2) should be interpreted with caution. We investigated the association of the HOX1−413 A/T promoter polymorphism (rs2071746) with the outcome of RTx in 343 Caucasian donor/recipient pairs, of whom 147 grafts had failed within 5 years after RTx. This polymorphism is in a linkage disequilibrium with the HOX1 (GT)n locus and increases HOX1 expression as well (3). Unfortunately, we did not find any association of the −413 A/T polymorphism in donors and recipients with the outcome of RTx. Nonetheless, we cannot conclude that the polymorphism does not influence the RTx outcome. The explanation relies on an inherent limitation of most allelic associated studies, that is the lack of power to detect an effect of polymorphisms with a minor contribution to the phenotype. The role of single gene variants in complex human diseases is often weaker than that of environmental factors (4), as shown in the article of Baan et al. (1), where delayed graft function (p = 0.004) and donor age (p = 0.005) were stronger predictors of renal allograft failure than the HOX1 (GT)n L-allele (p = 0.03). Also, smaller studies overestimate the true effect size of polymorphisms and frequently report associations that are false positive (type-I error) and nonreplicable, as documented by the article of Baan et al. (1), which was not confirmed by a larger study of Courtney et al. (2). Apart from type-I error, the key question is whether these studies allow for an assessment of polymorphisms with small effect. The effect sizes of most polymorphisms in complex diseases, expressed as odds ratios (OR), range between 1.2 and 1.6 (5). Our calculations, performed for 90% power, showed that the number of cases and controls needed to detect an effect of the HOX1 (GT)n and −413 A/T polymorphisms is 2538 and 382 for the OR of 1.2 and 1.6, respectively. Odds ratios of this magnitude were detectable in neither of the studies of Baan et al. (1), Courtney et al. (2) and of our group, because only 73, 324 and 147 patients with failed grafts were included, making them incapable to detect OR lower than 2.4, 1.8 and 2.1, respectively. Although the study of Courtney et al. (2) surpassed other studies in its size, it was still below sufficient power and therefore might have missed a true genetic association (type-II error). Genetic predisposition to renal allograft failure seems to be determined by multiple polymorphisms with a low individual contribution to the phenotype. Their identification warrants future well-powered studies compliant with the standards for polymorphism-disease association studies (4). Supported by Sixth Framework Program of EU RISET.