Health food products or supplements are expected to promote health and to prevent and treat diseases; consequently their use has recently increased. Therefore, it is necessary to accurately and efficiently examine adverse events associated with them. Many algorithms for the evaluation and classification of causal relationships of adverse events related to medications have been proposed, however, there is no algorithm concerning health food products or supplements. In this study, considering these circumstances, we designed algorithms for the evaluation and classification of causal relationships of adverse events with health food products or supplements and evaluated their reliability. The algorithms previously proposed for the evaluation and classification of causal relationships of adverse events with medication (Naranjo CA,
et al . Clin Pharmacol Ther 1981; 30: 239-45, Jones JK. Fam Community Health 1982; 5: 58-67) were modified for the present study. We assessed the effects of the ginkgo biloba extract, a common health food supplement, and reviewed the literature using databases such as MEDLINE. We found 29 cases of adverse events related to intake of ginkgo biloba extracts. Using the modified algorithms, the causal relationship of ginkgo biloba extracts with adverse events in the cases was independently evaluated by 4 raters (a clinician and 3 pharmacists), and the inter-rater reliability was tested. The coefficient of intra-class correlation calculated by the total score of Naranjo
et al . was 0.65 and the 95% confidence interval was [0.48, 0.79]. The κ coefficient of multi-rater reliability was 0.27 in the algorithm of Naranjo
et al., while it was, 0.26 when 5 studies were excluded using the algorithm of Jones. In conclusion, the modified algorithm of Naranjo
et al. may be recommended for evaluating causal relationships of adverse events with health food products or supplement consumption.
View full abstract