I have read with interest the recent report by Liu and coworkers on the role of paraoxonase (PON) genes in ischemic stroke . Based on 28 studies, the authors conclude to a risk-enhancing effect of a PON1 missense substitution, Q192R, and to nonsignificant effects of other paraoxonase gene variants. It would appear, however, that the data assembled do not reflect the existing evidence and, therefore, offer a distorted view of the role played by PON genes in the phenotype under study. Liu et al. do not specify the data freeze for including studies in their analysis but a rapid search disclosed multiple investigations that were omitted, e.g., for Q192R, Topić et al. , Voetsch et al. , and Ranade et al. , to name only a few. Intriguingly, a number of other studies were first retrieved by Liu et al. but were then overlooked in the analysis, e.g., for Q192R, references 34 and 36 were dropped. Likewise, for PON1 L55M, data are missing from references 41 and 46. For PON2 G148A, data are missing from reference 53 which identifies G as the major allele leading to a substantial shift in the overall risk. The figures given for PON2 C311S are grossly in error in that the protective and at-risk alleles were muddled for references 46 and 52 in table 2. Moreover, several large investigations are again missing, e.g., Wang et al. . On the whole, the above flaws call for a cautious stand on the proposed contribution made by PON genes to augmenting the risk for ischaemic stroke.
To the editor,