Response to Letter by Grond-Ginsbach et al
We appreciate the comments and interest of Grond-Ginsbach and colleagues. The difference between their results and ours highlights the importance of specificity and precision in defining clinical phenotypes when performing genetic association studies. There are 3 obvious differences in phenotype that may account for the discrepant outcome between their analysis and ours (however, we should caution the reader that without presentation of detailed information in a rigorously peer-reviewed publication, we cannot be certain). First, in their general analysis, these investigators genotyped all patients with acute ischemic stroke without consideration of stroke subtypes. By contrast, because stroke in the young is notoriously heterogeneous in etiology, we specifically excluded those with a known underlying cause, such as cardioembolism, carotid dissection, and hematologic disorders, as well as those with other systemic disorders that contributed to the stroke syndrome. Second, both populations presented in our study were strictly matched to controls according to age, gender, and ethnicity. No information about matching is available in this brief letter from Grond-Ginsbach and colleagues. In fact, we do not know precisely what the age range of their study populations is and whether they included children. Furthermore, other factors that the authors themselves acknowledge, such as ethnicity, may contribute to the difference in results between their study and ours. As we reported, there is clear ethnic variability in GPx-3 haplotype frequencies. Third, we performed our association analysis taking into consideration other potential genetic determinants of (thrombotic) stroke risk. Thus, without a detailed presentation of all of the demographic information (including other nonheritable risk factors) and, equally important, without the benefit of rigorous peer review of the full study, it is difficult to know what to make of their results. Clearly, detailed and comparable studies should be performed in other populations to ascertain the generalizability of our findings. The data of Grond-Ginsbach and colleagues do not provide the necessary detail nor valid comparability to do so. We do, however, look forward to seeing a full publication of their peer-reviewed results.