© 2006 American Thoracic Society
Interleukin-18–607 Promoter Polymorphism in Sarcoidosis: Ignoring "Negative" ResultsTo the Editor:We read with interest the recent article by Kelly and colleagues, which demonstrated that gram-negative bacteria may contribute to helper T-cell type 1 driven responses in sarcoidosis by up-regulating interleukin (IL)-18 expression (1). As part of this study, Kelly and coworkers genotyped 94 Irish patients with sarcoidosis and 97 control subjects for the IL-18–607A/C promoter polymorphism. They found significantly different allele and genotype frequency distributions. As they mentioned in their introduction, a similar association has been found by Takada and coworkers in Japanese patients with sarcoidosis from Niigata (2). A study performed by our group demonstrated no difference between Dutch patients with sarcoidosis and control subjects (3). Our study, however, was not cited. In addition, Zhou and coworkers' study, which was not mentioned either, showed no difference in Japanese patients from Hokkaido (4). Since these studies were dealing with the same subject, namely the IL-18–607A/C polymorphism in sarcoidosis, it is rather surprising that Kelly and coworkers have not mentioned these negative-result articles. Therefore, we provide a brief comparison between the four studies (Table 1).
No difference was found between control subjects from the Netherlands and Ireland (1, 3). A significant difference, however, was found between patients with sarcoidosis from the Netherlands and Ireland (1, 3). It is remarkable that a significant difference in IL-18–607A/C allele and genotype frequency distributions was found between Japanese control subjects from Niigata and Hokkaido, but not between the groups of Japanese patients with sarcoidosis (2, 4). It is of note that the Japanese Niigata control population was almost out of Hardy-Weinberg equilibrium (p = 0.08) (2). Therefore, we question the validity of Takada and coworkers' control data from Niigata (2). The Irish data could be based on a true difference or a mere coincidence, since their p values (p > 0.02) were quite high for a genetic-association study, and their study group was relatively small (1). Kelly and coworkers' genetic data would have been more convincing if they had been replicated in a second cohort of patients. Publication bias is based on the fact that scientists are more likely to submit "positive" results, and journals are less likely to publish "negative" results. Kelly and coworkers' article might illustrate another bias, the citation bias: "positive" articles are more likely to be referred to than "negative" articles. We do not know whether Kelly and coworkers did not perform an adequate PubMed search, or did not discuss the Dutch and Hokkaido data because these did not support their own findings. However, these "negative" references are essential to position the Irish data in the context of the current literature on the genetic background of sarcoidosis.
St. Antonius Hospital, Nieuwegein, The Netherlands FOOTNOTES Conflict of Interest Statement: None of the authors have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. REFERENCES
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
