Equipoise on the Use of Steroids in Systemic Inflammatory Response Syndrome?*

    loading  Checking for direct PDF access through Ovid

Excerpt

The sequencing of the human genome was heralded with a great deal of hope that a better understanding of our genetic makeup would usher in a new era of “personalized medicine.” The study by Wong et al (1) in this issue of Pediatric Critical Care Medicine is a well-conducted investigation that examines how genetic variation can alter a patient’s response to steroid treatment of the systemic inflammatory response syndrome (SIRS). This investigation illustrates the challenge of conducting genetic studies in the world of pediatric critical care, in which patient numbers are usually much smaller than in comparable adult studies. The authors found no association between steroid treatment and outcome in patients with less common genetic variants. Instead, they found that patients with the most commonly encountered genotype fared worse when their SIRS was treated with steroids. This result may hold significant implications for those who care for critically ill children.
The authors designed this investigation to examine the influence of three single nucleotide polymorphisms (SNPs) on the outcome of patients with SIRS who were treated with steroids. This problem can be approached through radically different study designs, which have contrasting advantages and disadvantages. At one end of the spectrum, a genome-wide association study can be used to examine hundreds of thousands of SNPs across the entire human genome. At the other end of the spectrum is the approach that the authors selected, in which a small group of candidate SNPs is investigated. Genome-wide association studies are attractive, because they survey the entire genome and may reveal unexpected associations between genetic variation and outcome. This was the case when the first published genome-wide association study revealed an unanticipated relationship between age-related macular degeneration and two SNPs in the complement factor H gene (2). Although such unexpected findings are dazzling and accelerate our understanding of the biology of an illness, genome-wide association studies are plagued by low power. Because these studies examine the influence of hundreds of thousands of SNPs, their power is limited to detect an association between any given SNP and the outcome of interest. Each of the thousands of SNPs examined constitutes a separate test of the hypothesis of association between that SNP and the outcome. During the statistical analysis, adjustment must be made for the testing of so many hypotheses, with the result that the power to detect modest statistical associations is greatly diminished. Although measures exist to ameliorate this penalty (3, 4), it remains substantial and can obscure meaningful associations that fail to achieve an exceptionally high level of statistical significance.
The approach selected by the authors is at the other end of this spectrum and almost completely avoids the problem of multiple hypothesis testing, so it is more suited for studies with relatively small number of patients. Instead of examining thousands of SNPs, the authors selected three (candidate) SNPs that have been shown to have functional significance (5) in previous investigations. This time-honored approach greatly boosts the power of the investigation with one important caveat: most of the genetic variation is not examined. In the glucocorticoid receptor (GCR) gene alone, there are over 7,000 SNPs (6). Additionally, there are almost 90 regulatory elements (enhancers and repressors) that act upon the GCR gene (7). At least four microRNAs have been shown to exert regulatory effects on the expression of the GCR gene (8). Genetic variation in any of these elements has the potential to affect outcome. Other types of genetic variation such as insertions, deletions, and transpositions will be undetected by this approach.

Related Topics

    loading  Loading Related Articles