10.6084/m9.figshare.7195004.v1 Leonard Nunney Leonard Nunney Suppl. Table 1: The risk of specific cancers with height and with BMI; Suppl. Table 2: Geographical comparisons of the influence of height on the risk of specific cancers.; Suppl. Figure 1: The lack of correlation between the effect of height and of obesity on the risk of specific cancers. from Size matters: height, cell number and a person's risk of cancer The Royal Society 2018 multistage carcinogenesis cancer and height cell number Peto's paradox cancer and BMI GWAS and cancer 2018-10-11 07:47:22 Journal contribution https://rs.figshare.com/articles/journal_contribution/Suppl_Table_1_The_risk_of_specific_cancers_with_height_and_with_BMI_Suppl_Table_2_Geographical_comparisons_of_the_influence_of_height_on_the_risk_of_specific_cancers_Suppl_Figure_1_The_lack_of_correlation_between_the_effect_of_height_and_of_obesity_on_the/7195004 Hazard ratio (HR) defining the increase in cancer risk per 10cm incease in height (and 95% confidence interval), plus, for males and females, the relative cancer risk (RR) per 5kg/m<sup>2</sup> increase in BMI. Bolded values are significantly different from 1.0 (p<0.05). Height HR data are from the references indicated. BMI RR data from Renehan et al (2008)<sup>1</sup>, noting that specific RR values are: (a) post-menopausal only, (b) averaged across the sexes, and (c) adenocarcinoma only.; Pairwise comparison of hazard ratios between the geographical populations used in each of the four studies [22–25] across cancer types. Significant comparisons (at 5%) are bolded, based on a z-test with a critical value for p<0.05 of ±3.46 (women) (94 tests) & ±3.24 (men) (42 tests),using a Bonferroni correction. References: UK [22]; USA [23]; Norway, Sweden & Austria [24], and Korea [25].; The lack of correlation between the effect of height and of obesity on the risk of specific cancers in women (circles; r=−0.024, df=16, ns) and men (triangles; r=0.328, df=14, ns). The height effect is measured by HR10 and the obesity effect by the change in relative risk per 5 units of BMI. Data are from Suppl. Table 1 and the points are enclosed by the 95% confidence ellipses. Specific cancers represented are: 1=lung (smoker); 2=lung (never smoker); 3=stomach; 4=pancreas; 5=leukemia; 6=multiple myeloma; 7=rectum; 8=non-Hodgkin's lymphoma; 9=liver; 10=colon; 11=melanoma; 12=urinary tract/kidney; 13=thyroid; 14=biliary tract; 15=oesophagus; in females only, 16f=endometrium; 17f=ovary; 18f=breast; and in men only, 19m=prostate.