To the Editor: Jesper Ekelund, M.D., and colleagues (1) recently reported an association between the dopamine D₄ receptor 48-base-pair-repeat polymorphism (DRD4) and the personality trait of novelty seeking in Finnish subjects. They observed a significant (p=0.007) difference in DRD4 allele frequencies between subjects with high novelty-seeking scores and those with low scores and concluded that "these results confirm the original findings of an association between the DRD4 gene and novelty seeking."

Several caveats should be considered, however, in the interpretation of these data. First, for the purpose of statistical analysis, the authors grouped the DRD4 7-repeat allele with the 8-repeat allele (because of the low frequency of the 8-repeat allele in this population). As we have commented elsewhere (2), this kind of arbitrary grouping reduces the degrees of freedom in the analysis and increases the potential for false positive results. An analysis of these data without the collapsing of these two alleles resulted in a marginally significant result (p=0.012). Moreover, the authors also conducted a "presence or absence" analysis despite the lack of data implicating a codominant mode of action for these alleles but did not correct for the other potential groupings—again inflating the potential for a false positive result.

It is also striking that this study yielded a result that is in opposition to that of the original studies, which showed a relation between DRD4 and novelty seeking. Benjamin et al. (3) and Ebstein et al. (4) found that the DRD4 7-repeat allele was associated with higher levels of novelty seeking. Dr. Ekelund and colleagues reported the reverse; in their cohort, the 7-repeat allele was overrepresented in individuals with low novelty-seeking scores. To explain this discrepancy, the authors proposed that another variant in linkage disequilibrium with the DRD4 48-base-pair-repeat polymorphism influences behavior. It is somewhat surprising, therefore, that these investigators did not genotype any additional DRD4 variants to test this hypothesis.

Although the findings of the article by Dr. Ekelund et al. did not replicate the results of the original reports of Benjamin et al. and Ebstein et al., they were consistent with the results of the only previous study of this subject in Finnish subjects (5). In a study of 138 alcoholic Finns, we also observed an association between the 7-repeat allele and scores for low novelty seeking. At the time, we interpreted these data as a failure to replicate the previous reports of an association of the DRD4 gene and novelty seeking (because the association was in the opposite direction). However, since then another group (6) has also reported an association between the 7-repeat allele and low novelty-seeking scores in substance abusers.

In the study by Dr. Ekelund et al., subjects were selected for extreme novelty seeking (more than one standard deviation from the mean). Because there are considerable data that suggest that substance abusers have higher levels of novelty seeking (5), it is possible that the sample of Dr. Ekelund et al. contained a higher proportion of substance abusers than the general population. If so, the study by Dr. Ekelund et al. may be the third study indicating an association between the 7-repeat allele and levels of low novelty seeking in substance abusers, and two of these associations have been observed in Finns.