LETTER TO THE EDITOR Herbal medicines during pregnancy and childhood cancers: an analysis of data from a pregnancy cohort study To the Editor Pregnant women using herbs is common in Taiwan and mainland China, 1,2,3 but relatively little is known about the safety of taking herbal medicines during pregnancy. Our previous study showed a possible link between the use of specific herbal medicines during the first trimester of pregnancy and an increased risk of specific congenital malformations. 4 Thus, this analysis was conducted to explore the relationship between herbal medicine use during pregnancy and the risk of childhood cancer in resulting offspring. Pregnant women of 26 weeks or more of gestation who came to the Taipei Municipal Maternal and Child Hospital in Taiwan for prenatal care in 1984–1987 were enrolled in the study. A total of 14551 liveborn infants born to them were followed-up after birth. Exposure data were collected prospectively before delivery. Cancer incidence data was obtained from multiple sources. The risk of cancer related to the use of herbal medicine during pregnancy was estimated with Cox proportional hazard regression. The mean follow-up of the cohort was 14.9 years. Thirty-two cancer cases were diagnosed during 216 873 person years of follow-up. Taking Coptidis Rhizoma during pregnancy was found to be associated with an increased risk of childhood cancer (adjusted hazard ratio [HR] 2.23, 95% confidence interval [CI] 1.02 to 4.88). Within the diagnostic groups, only brain cancer had sufficient numbers (n ¼ 11) for a separate analysis, and taking Coptidis Rhizoma during preg- nancy was found to be associated with an increased risk (adjusted HR 4.79, 95% CI 1.28–17.91), Table 1. We found evidence for a possible link between the use of Coptidis Rhizoma during pregnancy and increased the risk of childhood cancer, but the finding may be due to the chance because of testing for multiple outcomes. Thus, further research is warrant. REFERENCES 1. Chuang CH, Chang PJ, Hsieh WS, Tsai YJ, Lin SJ, Chen PC. Chinese herbal medicine use in Taiwan during pregnancy and the postpartum period: a population based cohort study. Int J Nurs Stud 2009; 46: 787– 795. 2. Chuang CH, Hsieh WS, Guo YL, Tsai YJ, Chang PJ, Lin. SJ, Chen PC. Chinese herbal medicines used in pregnancy: a population-based survey in Taiwan. Pharmacoepidemiol Drug Saf 2007; 16: 464–468. pharmacoepidemiology and drug safety 2009; 18: 1119–1120 Published online in Wiley InterScience (www.interscience.wiley.com) DOI: 10.1002/pds.1835 Table 1. Hazard ratios and 95% confidence intervals for herbal medicines used during pregnancy and childhood cancers Herbal medicines use Total cancers (n ¼ 32) Brain tumours (n ¼ 11) Other cancers (n ¼ 21) No. of cases Total person-years HR (95% CI) No. of cases Total person-years HR (95% CI) No. of cases Total person-years HR (95% CI) Never used 17 128 945 1.00 (referent) 4 128,992 1.00 (referent) 13 128 951 1.00 (referent) Coptidis Rhizoma alone 10 36 091 2.23 a (1.02, 4.88)  5 36,097 4.79 b (1.28, 17.91)  5 36 102 1.37 (0.49, 3.85) An-Tai-Yin alone 3 22 525 1.03 a (0.30, 3.51) 1 22,536 1.47 b (0.16, 13.13) 2 22 537 0.88 (0.20, 3.90) Ginseng alone 1 4963 1.50 a (0.20, 11.29) 0 4,965 — 1 4963 2.00 (0.26, 15.28) Coptidis Rhizoma & An-Tai-Yin 1 9506 0.86 a (0.11, 6.48) 1 9,506 3.83 b (0.43, 34.34) 0 9508 — Other herbs 0 12 397 — 0 12,397 — 0 12 397 — Abbreviations:  p < 0.05; HR, hazard ratio; CI, confidence interval. A 10% change-in-estimate method was used to assess potential confounders. a Adjusted for maternal education. b Adjusted for paternal education. Other categories of herbal medicines used without cancer case are not shown in the table. Copyright # 2009 John Wiley & Sons, Ltd.