Objective
The relationship between blood glucose levels and osteoarthritis (OA) is unclear. This study aimed to investigate the genetic causal relationship between blood glucose-related traits and OA.
Methods
We first performed univariate Mendelian randomization (UVMR) analyses using published genome-wide association study (GWAS) datasets with fasting glucose (FG), 2 h-glucose post-challenge glucose (2hGlu), and glycosylated hemoglobin (HbA1c) as exposures, and hip osteoarthritis (HOA) and knee osteoarthritis (KOA) as outcomes; then, we performed inverse analyses of them. We used Inverse-variance weighted (IVW) analysis as the primary analysis, and sensitivity analyses were performed. Moreover, we performed multivariate Mendelian randomization (MVMR) to estimate the independent effect of exposure on outcome after adjusting for body mass index (BMI). Summarized data for blood glucose-related traits were obtained from the MAGIC Consortium study of the glucose trait genome and for OA from the UK Biobank and arcOGEN. Summarized data for BMI were obtained from the GIANT Consortium meta-analysis of individuals of European ancestry. A two-sided p value < 0.05 in UVMR was considered suggestive of significance when
p
< 0.0167 (Bonferroni correction
p
= 0.05/3 exposures) was considered statistically significant.
Results
We found significant negative genetic causality of FG for HOA and KOA, and these associations remained significant after we adjusted for the effect of BMI [odds ratios (ORs) of 0.829 (0.687–0.999,
p
= 0.049) and 0.741 (0.570–0.964,
p
= 0.025)]. HbA1c also had an independent negative genetic causal effect on HOA after adjustment for BMI [0.665 (0.463–0.954,
p
= 0.027)]. At the same time, there was no evidence of reverse genetic causality of OA on blood glucose-related traits.
Conclusion
We further elucidated the relationship between blood glucose-related traits and OA by adjusting for the effect of BMI from a genetic causal perspective. This study provides new insights to further clarify the relationship between blood glucose levels and OA, as well as the pathogenesis, etiology and genetics of OA.
Supplementary Information
The online version contains supplementary material available at 10.1186/s13098-024-01517-3.