三重野 牧子

Background: Our recent study showed that a low lipoproteinemia(a) [Lp(a)] level was a risk factor for cancer and all-cause deaths. The purpose of this study was to verify the role of the Lp(a) level on cancer among consecutive autopsy cases. Methods: The subjects consisted of 1354 cases (775 men and 579 women). The average age at death was 79.9 years. Hypolipoproteinemia(a) was defined as an Lp(a) level of below 80 mg/L. Overall, 62.3% of the subjects had suffered from at least one to a maximum of five malignancies throughout their lives. The most frequent type of malignancy was gastric cancer, followed by leukemia, lung cancer, and colon cancer. Results: The cancer-bearing status decreased linearly according to the Lp(a) level in both men and women (P = 0.01 and P < 0.001, respectively). The median Lp(a) level was significantly lower among the cases with hepato-biliary-pancreatic cancers or hematopoietic malignancy, but was higher among cases with lung cancer, especially lung adenocarcinoma. Hypolipoproteinemia(a) was a significant risk factor for any origins of cancer, with an odds ratio of 1.94 (95% CI, 1.45-2.60; P < 0.001). It was also a risk factor for hepato-biliary cancers and leukemia, but it was a protective factor for lung cancer. Conclusions: Our findings suggested hypolipoproteinemia(a) would be a significant risk factor for cancer except lung cancer. This study complements our previous study showing that hypolipoproteinemia(a) would increase the lifetime risk of cancer other than lung cancer. (C) 2014 Elsevier Ltd. All rights reserved.

Although radiotherapy is recognized as an established risk factor for second malignant neoplasms (SMNs), the dose response of SMNs following radiotherapy has not been well characterized. In our previous meta-analysis of the risks of SMNs occurring among children who have received radiotherapy, the small number of eligible studies precluded a detailed evaluation. Therefore, to increase the number of eligible studies, we developed a method of calculating excess relative risk (ERR) per Gy estimates from studies for which the relative risk estimates for several dose categories were available. Comparing the calculated ERR with that described in several original papers validated the proposed method. This enabled us to increase the number of studies, which we used to conduct a meta-analysis. The overall ERR per Gy estimate of radiotherapy over 26 relevant studies was 0.60 (95% CI: 0.30-1.20), which is smaller than the corresponding estimate for atomic bomb survivors exposed to radiation as young children (1.7; 95% CI: 1.1-2.5). A significant decrease in ERR per Gy with increase in age at exposure (0.85 times per annual increase) was observed in the meta-regression. Heterogeneity was suggested by Cochran's Q statistic (P < 0.001), which may be partly accounted for by age at exposure.

The present study aimed at assessing the relationship between exposure to pyrethroid insecticides and semen quality in 323 university students recruited in a population-based manner in Metropolitan Tokyo. Urinary concentrations of pyrethroid insecticide metabolite, 3-phenoxybenzoic acid (3-PBA), were measured by LC/MS/MS and semen parameters were measured by following internationally harmonized protocols. Median urinary 3-PBA concentration was 0.641 ng/mL (specific gravity-adjusted, n = 322). Median values of semen volume, sperm concentration, motility, total number of sperm, and total number of motile sperm were 2.5 mL, 56 x 10(6)/mL, 61%, 141 x 10(6), and 82 x 10(6), respectively. Urinary concentration of 3-PBA was not selected as significant in multiple regression models indicating, in contrast to previous findings, that environmental exposure to pyrethroid insecticides did not affect semen quality. This inconsistency may be related to exposure to different pyrethroid insecticides and/or levels of exposure as well as to survey design (hospital- vs population-based subject recruitment). (C) 2013 Elsevier Inc. All rights reserved.

Background: The Ryanodine receptor 3 gene (RYR3) encodes an intracellular calcium channel that mediates the efflux of Ca2+ from intracellular stores. Two single-nucleotide polymorphisms (SNPs) in the RYR3 gene have been shown to associate with stroke (rs877087) and carotid intima-media thickness (rs2229116) in two independent genome-wide association studies (GWAS) in Caucasian. We investigated the effect of these two SNPs as well as the 31.1 kilobases spanning region on atherosclerosis in Japanese population. Methods: Atherosclerotic severity was assessed by carotid artery (n = 1374) and pathological atherosclerosis index (PAI) (n = 1262), which is a macroscopic examination of the luminal surfaces of 8 systemic arteries in consecutive autopsy samples. 4 tag SNPs in the 31.1 Kb region, rs877087, rs2132207, rs658750 and rs2229116, were genotyped and haplotypes were inferred to study the association with atherosclerotic indices. Results: rs877087 and rs2229116 were associated with PAI (OR = 2.07 [1.04-4.12] (95% CI), p = 0.038; and OR = 1.38 [1.02-1.86], p = 0.035, respectively). rs2229116 was also associated with common carotid atherosclerosis (OR = 1.45 [1.13-1.86], p = 0.003). The risk allele of rs2229116 was opposite from the original report. The haplotype block of this 31.1 Kb region was different between Caucasian and Japanese. Haplotype analysis revealed that only TAGG haplotype was associated with PAI (OR = 0.67 [0.48-0.94], p = 0.020) and atherosclerosis of common carotid artery (OR = 0.75 [0.58-0.98], p = 0.034). Conclusion: rs877087 and rs2229116 of RYR3 gene are associated with atherosclerosis severity in Japanese. The functional difference caused by rs2229116 needs to be investigated.

Objectives: Acute respiratory distress syndrome is characterized by diffuse alveolar damage and increased extravascular lung water levels. However, there is no threshold extravascular lung water level that can indicate diffuse alveolar damage in lungs. We aimed to determine the threshold extravascular lung water level that discriminates between normal lungs and lungs affected with diffuse alveolar damage. Design: A retrospective analysis of normal lungs and lungs affected with diffuse alveolar damage was performed. Setting: Normal lung cases were taken from published data. Lung cases with diffuse alveolar damage were taken from a nationwide autopsy database. All cases of autopsy followed hospital deaths in Japan from more than 800 hospitals between 2004 and 2009; complete autopsies with histopathologic examinations were performed by board-certified pathologists authorized by the Japanese Society of Pathology. Patients: Normal lungs: 534; lungs with diffuse alveolar damage: 1,688. Interventions: We compared the postmortem weights of both lungs between the two groups. These lung weights were converted to extravascular lung water values using a validated equation. Finally, the extravascular lung water value that indicated diffuse alveolar damage was estimated using receiver operating characteristic analysis. Measurements and Main Results: The extravascular lung water values of the lungs showing diffuse alveolar damage were approximately two-fold higher than those of normal lungs (normal group, 7.3 +/- 2.8 mL/kg vs diffuse alveolar damage group 13.7 +/- 4.5 mL/kg; p < 0.001). An extravascular lung water level of 9.8 mL/kg allowed the diagnosis of diffuse alveolar damage to be established with a sensitivity of 81.3% and a specificity of 81.2% (area under the curve, 0.90; 95% CI, 0.88-0.91). An extravascular lung water level of 14.6 mL/kg represented a 99% positive predictive value. Conclusions: This study may provide the first validated quantitative bedside diagnostic tool for diffuse alveolar damage. Extravascular lung water may allow the detection of diffuse alveolar damage and may support the clinical diagnosis of acute respiratory distress syndrome. The best extravascular lung water cut-off value to discriminate between normal lungs and lungs with diffuse alveolar damage is around 10 mL/kg.

We previously reported 2 osteoporosis-susceptibility genes-formiminotransferase N-terminal sub-domain containing gene (FONG) and thrombospondin, type 1, domain-containing 7A (THSD7A)-in which we identified two common single-nucleotide polymorphisms, rs7605378 (FONG) and rs12673692 (THSD7A). The former was associated with a predisposition to osteoporosis and the latter with bone mineral density. To further elucidate the importance of these polymorphisms in the pathogenesis of osteoporosis, we examined their association with the incidence of vertebral fracture. DNA extracted from the renal cortex of 2427 consecutive Japanese autopsies (1331 men, mean age: 79 years; 1096 women, mean age: 82 years) were examined in this study. The presence or absence of vertebral fracture during each subject's lifetime was determined by a thorough examination of the clinical records, as well as autopsy reports. After adjustments for sex and age at autopsy, logistic regression analysis revealed that homozygotes for the risk alleles of rs7605378 (A-allele) or rs12673629 (A-allele) possess an increased risk of vertebral fracture. The subjects simultaneously homozygous for both the risk alleles of rs7605378 (AA genotype) and rs12673629 (AA genotype) showed significantly higher risk of vertebral fracture (odds ratio 2.401, 95% confidence interval 1.305-4.416, P = 0.0048) than those who had at least one non-risk allele of either rs7605378 (AC/CC genotypes) or rs12673629 (AG/GG genotypes). The results suggest that Japanese subjects homozygous for the risk alleles of rs7605378 and rs12673629 have a higher risk of vertebral fracture. Journal of Human Genetics (2013) 58, 109-112; doi:10.1038/jhg.2012.145; published online 10 January 2013

Objectives: To establish a base line for future studies on temporal trends, to describe potential geographical differences in semen quality and reference values for studies of men from the general population. Design: Cross-sectional study of fertile men from four areas in Japan. Inclusion criteria were: age 20-45 years at the time of invitation, and both the man and his mother had to be born in Japan. Additionally, the current pregnancy of the female partner had to be achieved by normal sexual relations without any fertility treatment. Setting: Four Japanese study centres at urban areas located in Sapporo, Osaka, Kanazawa and Fukuoka. Participants: 792 men, median age 31.4 years, included from 1999 to 2002. Outcome measures: Semen volume, sperm concentration, total sperm count, sperm motility and sperm morphology. Results: Semen volumes, percentages of motile spermatozoa and morphologically normal spermatozoa differed slightly between the four groups, whereas no differences in sperm concentrations or total sperm counts were found. In total, 1.2% of men had a sperm concentration below 5 million/ml, 2.1% below 10 million/ml, 3.5% below 15 million/ml and 16.3% below 40 million/ml. For morphology, 14.7% had less than 5% normal spermatozoa. Reproductive hormone levels varied significantly, however, only little from a biological point of view. Conclusions: This is the first cross-sectional study on semen quality covering fertile men from the major regions of Japan. It showed that semen quality of fertile Japanese men is comparable to that of the best in European regions. The results may serve as reference values for studies of men from the general population.

Objectives To provide information of semen quality among normal young Japanese men and indicate the frequency of reduced semen quality. Design Cross-sectional, coordinated studies of Japanese young men included from university areas. The men had to be 18-24years, and both the man and his mother had to be born in Japan. Background information was obtained from questionnaires. Standardised and quality-controlled semen analyses were performed, reproductive hormones analysed centrally and results adjusted for confounding factors. Setting Four study centres in Japan (Kawasaki, Osaka, Kanazawa and Nagasaki). Participants 1559 men, median age 21.1years, included during 1999-2003. Outcome measures Semen volume, sperm concentration, total sperm count, sperm motility, sperm morphology and reproductive hormone levels. Results Median sperm concentration was 59 (95% CI 52 to 68) million/ml, and 9% and 31.9% had less than 15 and 40 million/ml, respectively. Median percentage of morphologically normal spermatozoa was 9.6 (8.8 to 10.3)%. Small, but statistically significant, differences were detected for both semen and reproductive hormone variables between men from the four cities. Overall, the semen values were lower than those of a reference population of 792 fertile Japanese men. Conclusions Assuming that the investigated men were representative for young Japanese men, a significant proportion of the population had suboptimal semen quality with reduced fertility potential, and as a group they had lower semen quality than fertile men. However, the definitive roleif anyof low semen quality for subfertility and low fertility rates remain to be investigated.

Surfactant protein D (SFTPD) is a lung-specific anti-inflammatory factor that antagonizes inflammation by inhibiting oxidative stress and stimulating innate immunity. Variations in SFTPA2 and SFTPB, genes for other surfactant proteins, have been associated with lung cancer. We therefore investigated associations between SFTPD variations and lung cancer as well as emphysema and interstitial pneumonia, which are characterized by chronic inflammation from which lung cancer often arises. DNA from 1342 autopsy samples, including those from 140 subjects with lung cancer, was investigated. The single nucleotide polymorphism (SNP) rs721917, which results in methionine being exchanged for threonine at amino acid 11 (the Met11Thr variation), tended to be associated with emphysema and was associated with interstitial pneumonia and lung cancer. A haplotype analysis revealed that the haplotypes associated with emphysema and lung cancer differed from that associated with interstitial pneumonia, suggesting a differential role for SFTPD in the development of these diseases. A mediating analysis did not reveal a mediating effect exerted by emphysema or interstitial pneumonia on lung cancer. Our results suggested that SFTPD plays a role in the development of lung cancer and that the role for lung cancer may differ from that for interstitial pneumonia.

Incomplete Kawasaki disease (KD) is associated with delayed diagnosis and treatment, which in turn can lead to the development of coronary artery lesions (CALs). The aim of this study was to determine the epidemiological features of incomplete KD compared with complete KD and to identify risk factors for CALs from incomplete KD patients using data from a nationwide survey of 2007-2008 in Japan. A total of 23,263 patients were classified according to the number of principal clinical signs: 80% (n=18,620) had complete forms of KD, 14.2% had four principal signs, 4.6% had three signs, and 1.2% had only one or two signs. In comparison with complete KD cases, the prevalence of CAL development tended to be larger and the proportion receiving initial intravenous immunoglobulin (IVIG) treatment were significantly smaller in patients with incomplete forms. In addition, hospital attendance after 7 days of illness or later was significantly associated with CAL development in all incomplete groups (OR: 2.52 in total patients with incomplete KD, 3.26 in those with one or two principal signs, 2.94 in those with three signs, 2.35 in those with four signs). Conclusion The higher prevalence of CALs in incomplete KD reflects difficulties in diagnosis and delays in treatment. More timely diagnosis and treatment of incomplete KD patients could further prevent the development of cardiac lesions.

Background: Experimental studies support the anti-neoplastic effect of apo(a), but several clinical studies have reported contradictory results. The purpose of this study was to determine whether a low lipoprotein(a) [Lp(a)] concentration is related to mortality from major causes of death, especially cancer. Methods: The subjects were 10,413 participants (4,005 men and 6,408 women) from a multi-center population-based cohort study in Japan (The Jichi Medical School cohort study). The average age at registration was 55.0 years, and the median observation period was 4,559 days. As the estimated hazard ratio was high for both the low and very high Lp(a) levels, we defined two Lp(a) groups: a low Lp(a) group [Lp(a)<80 mg/L] and an intermediate-to-high Lp(a) group [Lp(a)>= 80]. Participants who died from malignant neoplasms (n=316), cardiovascular disease (202), or other causes (312) during the observation period were examined. Results: Cumulative incidence plots showed higher cumulative death rates for the low Lp(a) group than for the intermediate-to-high Lp(a) group for all-cause, cancer, and miscellaneous-cause deaths (p<0.001, p=0.03, and p=0.03, respectively). Cox proportional hazards analyses with the sex and age of the participants, body mass index, and smoking and drinking histories as covariates showed that a low Lp(a) level was a significant risk for all-cause, cancer, and miscellaneous-cause deaths (p<0.001, p=0.003, and p=0.01, respectively). The hazard ratio (95% CI) [1.48, 1.15-1.92] of a low Lp(a) level for cancer deaths was almost the same as that for a male sex (1.46, 1.00-2.13). Conclusions: This is the first report to describe the association between a low Lp(a) level and all-cause or cancer death, supporting the anti-neoplastic effect of Lp(a). Further epidemiological studies are needed to confirm the present results.

Surfactant protein D (SFTPD) induces emphysema in knockout mice, but the association of SFTPD with chronic obstructive pulmonary disease (COPD) and emphysema in humans is unclear. Therefore, we aimed to determine the association between genetic variations in SFTPD and susceptibility to COPD and emphysema. Two populations were studied: population A comprised 270 smokers, including 188 COPD and 82 at-risk subjects, and population B comprised 1131 autopsy cases including 160 cases with emphysema. Six single-nucleotide polymorphisms (SNPs) that tagged the linkage disequilibrium blocks on the entire SFTPD gene were genotyped; the associations of the genotypes with COPD, pulmonary function, percentage of the low-attenuation area (LAA%), and percentage of the airway wall area (WA%) were determined in population A. In population B, the associations of the genotypes with emphysema were assessed. A C allele at SNP rs721917 that results in the replacement of Met with Thr at position 11 in SFTPD was positively correlated with the LAA% in the upper lung (P=1.1x10(-5)) and overall LAA% (P=1.0x10(-4)), and negatively correlated with the serum concentration of SFTPD (P=7x10(-11)) in the population A. The C/C (rs721917/rs10887199) haplotype was associated with emphysema in both the populations. Subjects with a C allele at rs721917 have a lower serum SFTPD concentration and are more susceptible to emphysema. This suggests a protective effect of SFTPD against COPD and emphysema. European Journal of Human Genetics (2012) 20, 230-235; doi:10.1038/ejhg.2011.183; published online 21 September 2011

The fucosyltransferase 8 gene (FUT8) encodes an enzyme that transfers fucose to the innermost N-acetylglucosamine unit of N-glycan chains. Recent study showed that fut8-deficient mice develop pathological and physiological phenotypes resembling pulmonary emphysema (PE). The role of FUT8 in human PE is not known. A non-synonymous single-nucleotide polymorphism at the amino-acid position of 267 in FUT8 (rs35949016; C/A, C allele = Thr, A allele = Lys) was genotyped in a total of 1149 consecutive autopsies of elderly Japanese. A following study included 182 outpatients with chronic obstructive pulmonary disease, whose emphysematous changes were assessed quantitatively as the percentage of low attenuation area (%LAA) by high-resolution computed tomography. PE was detected in 163 of 1149 autopsy subjects (14.2%). Comparison of patient with vs without PE indicated that the FUT8 A allele was associated with PE (AA+AC vs CC; odds ratio 1.74, 95% confidence intervals 1.19-2.56, P = 0.005). In the clinical study, presence of the FUT8 A allele significantly correlated with %LAA after adjustment (AA+AC vs CC = 37.5 +/- 14.7 vs 32.7 +/- 13.9, P = 0.02). The FUT8 gene Thr267Lys polymorphism is associated with human PE, and the Lys allele is the risk. The core fucosylation might be involved in the molecular pathogenesis of human PE. Journal of Human Genetics (2011) 56, 857-860; doi:10.1038/jhg.2011.118; published online 20 October 2011

Background. Coinfection by Hepatitis B virus (HBV) and hepatitis C virus (HCV) has been reported to increase risk of graft failure for liver transplant recipients. But other studies have controverted that finding. The aim of this study was to determine whether-after adjustments for other important predictors-HBV/HCV coinfection was associated with worse liver graft survival than HBV or HCV mono-infection.Methods. A retrospective cohort study examined Organ Procurement and Transplantation Network/United Network Organ Sharing data for 48,654 deceased-donor primary liver-only transplants that were performed on adults between January 1, 1995, and August 31, 2009, and that included recipient and donor HBV/HCV status. Recipients were classified into four groups: the HBV/HCV coinfected [B(+)/C(+)]; HBV mono-infected [B(+)/C(-)]; HCV mono-infected [B(-)/C(+)]; and hepatitis uninfected [B(-)/C(-)]. Kaplan-Meier methods were used to calculate liver graft survival rates, Cox proportional hazard models were used to estimate the effect of hepatitis virus infection, and adjusted for potential confounders.Results. Graft survival rates were highest with B(+)/C(+): 85.3% 1-year survival and 63.0% 10-year survival. Graft survival with B(+)/C(+) was superior to survival with B(-)/C(+): 83.5% 1-year survival and 53.6% 10-year survival vs. B(-)/C(+): 82.9% 1-year survival and 46.1% 10-year survival. Survival with B(-)/C(-): 83.6% 1-year survival and 56.6% 10-year survival was superior to survival with B(+)/C(+) (shown above). After adjustment for confounders, and with the coinfected as reference, B(-)/C(+) recipients had a higher risk of graft loss (hazard ratio, 1.35; 95% CI, 1.10-1.66); the other two groups had a lower risk.Conclusions. Our results suggested-despite reports to the contrary-statistically better graft outcomes with HBV/HCV coinfection than with HCV mono-infection.

Cancer risks among childhood cancer survivors following radiotherapy have not yet been well characterised in terms of radiation dose. A meta-analysis of studies on the excess relative risk per gray (ERR) of second cancer was conducted previously; unfortunately, the small number of eligible studies restricted quantitative evaluations. To solve this problem, a statistical method to calculate ERR estimates from other estimates was developed, and a meta-analysis was conducted again. The PubMed database was searched and 26 relevant studies were identified. ERR estimates were available in 15 studies, and for the other 11 studies, the regression-based model was used to calculate ERR estimates from other estimates. The overall ERR estimate was 0.40, which was much lower than that of atomic bomb survivors exposed as young children. Heterogeneity of the risk among studies was suggested, and a further study is needed to explore the heterogeneity among studies.

Background: In the randomized study of interferon beta-1b (IFN beta-1b) for multiple sclerosis (MS), it has usually been evaluated the simple annual relapse rate as the study endpoint. This study aimed to investigate the performance of various regression models using information regarding the time to each recurrent event and considering the MS specific data generation process, and to estimate the treatment effect of a MS clinical trial data. Methods: We conducted a simulation study with consideration of the pathological characteristics of MS, and applied alternative efficacy estimation methods to real clinical trial data, including 5 extended Cox regression models for time-to-event analysis, a Poisson regression model and a Poisson regression model with Generalized Estimating Equations (GEE). We adjusted for other important covariates that may have affected the outcome. Results: We compared the simulation results for each model. The hazard ratios of real data were estimated for each model including the effects of other covariates. The results (hazard ratios of high-dose to low-dose) of all models were approximately 0.7 (range, 0.613 - 0.769), whereas the annual relapse rate ratio was 0.714. Conclusions: The precision of the treatment estimation was increased by application of the alternative models. This suggests that the use of alternative models that include recurrence event data may provide better analyses.

Aim: We previously reported significant associations between mitochondrial single nucleotide polymorphisms (mtSNPs) and myocardial infarction, atherothrombotic cerebral infarction, metabolic syndrome and type 2 diabetes. Here, we assessed the hypothesis that mtSNPs may confer a risk for atherosclerosis, the most important intermediate phenotype of ischemic cardiovascular events. Methods: The subjects were 1,536 consecutive autopsy cases (827 men and 709 women). The average age at death was 80 years. The severity of coronary atherosclerosis was semi-quantitatively examined on cut sections. We examined 149 mtSNPs using the PCR-Luminex method, with a success rate of 97%. Phylogenetic tree analysis yielded 36 haplogroups. Multiple logistic regression analysis was performed after adjustments for sex, age, and conventional cardiovascular risk factors. Results: Among the 45 mtSNPs with minor genotype frequencies > 0.05, 6 mtSNPs were associated with coronary atherosclerosis. Among 10 haplogroups with frequencies > 0.04, haplogroups A and M7a were significantly associated with coronary atherosclerosis, with odds ratios (95% confidence intervals) of 1.80 (1.09-2.97; p = 0.023) and 1.92 (1.23-3.01; p = 0.004), respectively. Haplogroup D4a, which was previously reported to be associated with extreme longevity in a Japanese population, was associated with pathological myocardial infarction in men with an odds ratio of 2.05 (1.01-4.14; p = 0.046). Conclusions: The mitochondrial haplogroups A and M7a confer a significant risk for coronary atherosclerosis in the Japanese. The mitochondrial haplogroup may contribute some genetic risk for coronary heart disease.

Aim: Aortic dilatation is a well-known phenomenon in the elderly. We therefore aimed to study the pathobiological determinants of aortic dilatation. Methods: Retrospective chart review. The subjects were 833 consecutive autopsy cases (616 men and 217 women) of community deaths. The age at death ranged from 20 to 94 years, with an average of 59.2 years. We measured the internal circumference of the aortic root, arch, descending portion, abdominal portion, and bifurcation in unfixed opened aorta at the time of autopsy. Results: The simple correlation between age and aortic circumference was strongest for the descending portion, followed by the arch, abdominal portion, root, and bifurcation. The simple correlation coefficient reached 0.836 for the descending portion (p < 0.001). The circumference of the descending portion increased significantly as the severity of aortic atherosclerosis increased (p for trend < 0.001). Multiple regression analysis showed that age, sex, and body height were significantly correlated with the aortic circumference at all five measurement sites, while severe atherosclerosis was correlated with the aortic circumference at the root, and descending and abdominal portions. Six contributing factors (age, sex, body height, smoking history, hypertension, and severe atherosclerosis) explained 68.5% of the variance in circumference in the descending portion; age explained 57.5%; sex 8.4%; body height 1%; and severe atherosclerosis 0.8%. Conclusion: The contribution of atherosclerosis to aortic dilation was very weak, representing less than one seventieth of the contribution of age. The aortic circumference, especially in the descending portion, serves as an excellent age-related marker.

In the light of notable advances made in childhood cancer therapies, an understanding of the late effects of treatment is important for continued medical care. We conducted a meta-analysis of studies on the excess relative risk (ERR) of second malignant neoplasm (SMN) among childhood cancer survivors treated with radiotherapy. Relevant studies were retrieved by searching the PubMed database, supplemented by hand-searching of reference lists of already retrieved papers. Nine studies were identified and overall ERR estimates were calculated using a fixed effects model and a random effects model. The overall ERR per Gy (absorbed dose of ionising radiation) estimates of radiotherapy by a fixed effect model and a random effects model were 0.50 [95% CI 0.20, 1.21] and 0.53 [95% CI 0.22, 1.31] respectively. Heterogeneity among studies was suggested by Cochran's Q statistic (Q = 40.4, d.f. = 8, P < 0.001). The estimate obtained using a random effects model was far smaller than the corresponding estimate of 1.7 [95% CI 1.1, 2.5] from the study on atomic bomb survivors exposed as young children, suggesting underestimation of ERR estimates among the nine studies compared with the estimates from the study of atomic bomb survivors. In view of the heterogeneity and underestimation in ERR estimates, more studies concerning the risk of SMN among childhood cancer survivors are still needed for further understanding of the carcinogenic effects of radiotherapy on children.