Triangulating Evidence on Serum Uric Acid and Cancer Risk: Consistent Inverse Associations With Lung Cancer.

Abstract

We examined the association between serum uric acid (UA) levels and cancer incidence through a conventional observational and two-sample Mendelian randomization (MR) study in the same population (triangulation). We followed 13,695 individuals aged 35-69 years for cancer incidence (total, gastric, colorectal, lung, prostate, and breast cancers) over a median period of 10.6 years. The conventional observational dataset consisted of two subsets sampled from the full cohort using different sampling procedures. Odds ratios were computed using logistic regression for one type of single-nucleotide polymorphism (SNP) array, and hazard ratios were computed using the Cox proportional hazards model for another type of SNP array, combining them to calculate conventional relative risks (RRs). For two-sample MR, 42 UA-related SNPs were identified from Biobank Japan genome-wide association study data (109,029 participants). Associations between these SNPs and cancer were estimated using J-MICC Study data. MR RRs of cancer incidence per one standard deviation (SD) increase in genetically predicted UA were predicted using the inverse variance weighted (IVW) method and other MR approaches. In the conventional analysis, the highest UA concentration group (≥ 6.4 mg/dL) showed a decreased risk of lung cancer (RR: 0.64, 95% confidence interval [CI]: 0.41-0.98) compared with the lowest UA concentration group (< 4.0 mg/dL). The two-sample MR (IVW method) also suggested that UA was inversely (although marginally) associated with lung cancer risk (RR per one SD: 0.65, 95% CI: 0.40-1.05). In conclusion, lower, rather than higher, UA concentrations may be associated with an increased risk of lung cancer.