Table 3.
Comparison of revised MR estimates on 25(OH)D and cancer risk against previous findings.
| Previous IVW MR findings | IVW MR estimates based on UKB 25(OH)D instruments | ||||||
|---|---|---|---|---|---|---|---|
| Cancers | OR (95% CI) | Unit | P-value | OR (95% CI) | Unit | P-value | Scaling from per SD increase) |
| BEEA63 | – | – | – | 0.98 (0.85–1.14) | 20 nmol/L | 0.82 | Multiply by 1 |
| BE | 1.21 (0.7–1.92) | 20 nmol/L | 0.41 | 1.00 (0.84–1.18) | 20 nmol/L | 0.97 | Multiply by 1 |
| EA | 0.68 (0.39–1.92) | 20 nmol/L | 0.18 | 0.97 (0.78–1.20) | 20 nmol/L | 0.76 | Multiply by 1 |
| Breast cancer29 | 1.02 (0.97–1.08) | 25 nmol/L | 0.47 | 1.03 (0.93–1.13) | 25 nmol/L | 0.60 | Multiply by 1.25 |
| ER + breast cancer | 1.00 (0.94–1.07) | 25 nmol/L | 0.99 | 1.04 (0.93–1.16) | 25 nmol/L | 0.51 | Multiply by 1.25 |
| ER- breast cancer | 1.02 (0.90–1.16) | 25 nmol/L | 0.75 | 1.00 (0.88–1.15) | 25 nmol/L | 0.94 | Multiply by 1.25 |
| Endometrial cancer | – | – | – | 0.93 (0.80–1.07) | 20 nmol/L | 0.32 | Multiply by 1 |
| EC | – | – | – | 0.92 (0.79–1.08) | 20 nmol/L | 0.32 | Multiply by 1 |
| NEEC | – | – | – | 1.01 (0.73–1.41) | 20 nmol/L | 0.94 | Multiply by 1 |
| Lung Cancer53 | 0.98 (0.70–1.38) | 25 nmol/L | 0.93 | 0.94 (0.78–1.13) | 25 nmol/L | 0.50 | Multiply by 1.25 |
| LAC | 0.93 (0.55–1.58) | 25 nmol/L | 0.81 | 0.91 (0.67–1.18) | 25 nmol/L | 0.46 | Multiply by 1.25 |
| SCLC | 1.02 (0.60–1.73) | 25 nmol/L | 0.95 | 0.97 (0.76–1.26) | 25 nmol/L | 0.84 | Multiply by 1.25 |
| Melanoma30 | 1.06 (0.95–1.19) | 20 nmol/L | 0.3 | 1.09 (0.92–1.28) | 20 nmol/L | 0.31 | Multiply by 1 |
| NMSC28 | 1.11 (0.91–1.35) | 20 nmol/L | 0.38 | – | – | – | – |
| SCC | – | – | – | 1.02 (0.88–1.19) | 20 nmol/L | 0.77 | Multiply by 1 |
| BCC | – | – | – | 1.18 (1.05–1.33) | 20 nmol/L | 0.01 | Multiply by 1 |
| Neuroblastoma53 | 0.77 (0.22–2.70) | 25 nmol/L | 0.69 | 0.74 (0.42–1.29) | 25 nmol/L | 0.29 | Multiply by 1.25 |
| Epithelial ovarian cancer10 | 1.02 (0.72–1.44) | Per unit log(25OHD) | 0.93 | 0.78 (0.63–0.96) | Per unit log(25OHD) | 0.03 | Use log(OH)D betas |
| Clear cell | 1.75 (0.77–3.99) | Per unit log(25OHD) | 0.18 | 0.67 (0.34–1.34) | Per unit log(25OHD) | 0.36 | Use log(OH)D betas |
| Endometrioid | 0.83 (0.46–1.50) | Per unit log(25OHD) | 0.54 | 0.87 (0.54–1.41) | Per unit log(25OHD) | 0.56 | Use log(OH)D betas |
| High-grade serous | 0.98 (0.61–1.56) | Per unit log(25OHD) | 0.92 | 0.82 (0.63–1.07) | Per unit log(25OHD) | 0.15 | Use log(OH)D betas |
| Low-grade serous | 0.64 (0.24–1.75) | Per unit log(25OHD) | 0.39 | 0.95 (0.44–2.08) | Per unit log(25OHD) | 0.99 | Use log(OH)D betas |
| Mucinous | 1.27 (0.56–2.87) | Per unit log(25OHD) | 0.57 | 0.86 (0.50–1.50) | Per unit log(25OHD) | 0.58 | Use log(OH)D betas |
| Pancreatic cancer53 | 1.31 (0.75–2.33) | 25 nmol/L | 0.35 | 0.93 (0.46–1.92) | 25 nmol/L | 0.99 | Multiply by 1.25 |
| Prostate cancer29 | 1.00 (0.93–1.07) | 25 nmol/L | 0.99 | 1.07 (0.89–1.29) | 25 nmol/L | 0.46 | Multiply by 1.25 |
ER refers to oestrogen receptor status. EC and NEEC endometrial cancer refer to the endometrioid and non-endometrioid endometrial cancer subtypes. The linear conversion factor is a scaling parameter to align estimates to the same scale used in previously reported estimates. Here we assume a one SD change in 25(OH)D is approximately equivalent to a 20-nmol/L change based on the distribution of 25(OH)D in the UK Biobank cohort. For ovarian cancer histotypes, we used the beta estimate obtained from the log(25(OH)D GWAS (instead of rank-transformed 25(OH)D) to draw a direct comparison with previous findings reported for per unit increase in log(25(OH)D). All P-values derived from z-scores are two-sided and unadjusted for multiple comparison unless otherwise stated.
BE Barrett’s oesophagus, EA oesophageal adenocarcinoma, SCC squamous cell carcinoma, BCC basal cell carcinoma, LAC lung adenomacarcinoma, SCLC squamous cell lung cancer, NMSC non-melanoma skin cancer.