The Authors Respond

We read with great interest the Davis commentary1 and the seven letters2–8 in response to our article.9 We thank the editors for the opportunity to respond to them. Due to the space limitation, we herein discuss four points raised, using data from the non-Fukushima exposed populations10–14 as well as the unexposed.15–19 The first point is about minimum latency. Our inference on minimum latency was not solely based on the Centers for Disease Control-World Trade Center report20 but also on the Chernobyl data showing the excess thyroid cancer cases within 3 years after the 1986 Chernobyl accident.10–13 The excess was mainly in teenagers within 1980s, in contradiction to the statement, “many patients were younger than 10 years old.”4 Second, we discuss the issue of screening effect. In contrast to the second round, sensitivity analysis was performed by employing the latent duration in the first round between the date when thyroid cancer became detectable by screening (i.e., more than 5.1 mm in diameter) and the date when it could be diagnosed clinically. Latent duration assigned in our article was 4 years,9 which may be long for the particular hypothesis tested. One can assign any number of years, even 20, in the sensitivity analysis and still observe the remarkable excesses. The effect of preclinical tumors can be quantitatively estimated using an assigned proportion of preclinical tumors among the detected cancer cases (eAppendix; https://links.lww.com/EDE/B28). According to the pathological details of the 96 thyroid cancer cases operated at Fukushima Medical University, released on August 31, 2015,21 only eight cases (8%) were free of lymph node metastasis, extrathyroidal extension, and/or distant metastasis. Although suggested by some letters,2–5,7,8 a screening effect due to preclinical or slow-growing tumors can be excluded as a plausible interpretation. Third, we discuss the issues of dose–response relationship and individual radiation doses. As for the relatively low risk estimates in the nearest area to the accident, we should consider the effect of the length of time elapsed between the accident and timing of screening. We mentioned as the second limitation of our article,9 that this potential confounding leads to an underestimation. As a sensitivity analysis, we assigned the time from the accident to screening in each area to adjust for the elapsed time: 1 year for the nearest, 2 years for the middle, and 3 years for the least contaminated areas (Table). The adjusted prevalence odds ratios indicated a dose–response relationship. In general, individual-level data can only be derived for ingestion in field settings, and we employed areas and districts as a surrogate for individual radiation dose measurements.9 As we mentioned in our article,9 this potential nondifferential exposure misclassification would lead to underestimation in our findings.TABLE: Results of Internal Comparison from the First Round of Screening in Each Area or District up to June 30, 2015, When the Effect of the Length of Time Elapsed Between the Accident and Timing of Screening Was ConsideredFinally, we respond to concerns about information regarding the unexposed population. When comparing the Ukraine experience14,22 with that in Fukushima, Wakeford et al.5 cited Jacob et al.22 for the expected number of thyroid cancer cases in the first screening (i.e., first 4 years) in Fukushima. This was estimated from data recorded 12–14 years after the Chernobyl accident among 13,127 residents aged 18 or younger at the accident, through extrapolation to the unexposed from “heavily contaminated areas in Ukraine” using a linear excess relative risk model.14 Given the substantially larger sample size, however, direct estimation from ultrasound screening data among 47,203 examinees in the unexposed or relatively low contaminated areas in Ukraine would be more appropriate, where no cancer cases were detected (95% confidence interval: 0–78 per million examinees),16–19 as shown in eTable 1 of our article.9 Furthermore, although disregarded by some of the letters,2–5 comparability, for example by age and diagnostic criteria, should be considered when using the findings from South Korea.15 Screening in South Korea was conducted among adults with different diagnostic criteria from Fukushima, where one quarter of surgical patients had tumors less than 5.0 mm in diameter,15 whereas no cancers in this size range were detected in Fukushima. Takamura4 presented another example of inappropriate comparison with the all-school screening program started at Okayama University, Japan in 2012. Although the Okayama study did detect three thyroid cancer cases by palpation among 2,307 freshmen (ages 18 or older) in 2012, no other cases were detected among the total of 36,927 students enrolled between 2012 and 2015. Davis1 concludes his commentary by stating, “Nobody could imagine that a tsunami would ever be taller than the protective wall.” However, it was public knowledge long before 2011—also well recognized internally by Tokyo Electric Power Company—that a huge earthquake could potentially cause a tsunami much taller than any existing protective walls.23 Nevertheless, no corrective action was taken for the Fukushima Daiichi Nuclear Power Plant. When such knowledge is not shared, damage can ensue and residents suffer: this, regrettably, is the present state of public health in Fukushima Prefecture. According to the latest data available,24,25 the numbers of thyroid cancer cases increased to 113 (including one benign tumor case) and 39 in the first and second round of screening, respectively. ACKNOWLEDGMENT The authors are grateful to Professor Toshiro Ogura (Health Service Center of Okayama University) for providing data, and Dr. Yuri Hiranuma for her thoughtful suggestions. Toshihide Tsuda Department of Human Ecology Graduate School of Environmental and Life Science Okayama University Okayama, Japan [email protected] Akiko Tokinobu Department of Epidemiology Graduate School of Medicine Dentistry and Pharmaceutical Sciences Okayama University Okayama, Japan Eiji Yamamoto Department of Information Science Faculty of Informatics Okayama University of Science Okayama, Japan Etsuji Suzuki Department of Epidemiology Graduate School of Medicine Dentistry and Pharmaceutical Sciences Okayama University Okayama, Japan