Accumulating observational studies have identified associations between type 1 diabetes (T1D) and polycystic ovary syndrome (PCOS). Still, the evidence about the causal effect of this association is uncertain. We performed a two-sample Mendelian randomization (MR) analysis to test for the causal association between T1D and PCOS using data from a large-scale biopsy-confirmed genome-wide association study (GWAS) in European ancestries. We innovatively divided T1D into nine subgroups to be analyzed separately, including: type1 diabetes wide definition, type1 diabetes early onset, type 1 diabetes with coma, type 1 diabetes with ketoacidosis, type 1 diabetes with neurological complications, type 1 diabetes with ophthalmic complications, type 1 diabetes with peripheral circulatory complications, type 1 diabetes with renal complications, and type 1 diabetes with other specified/multiple/unspecified complications. GWAS data for PCOS were obtained from a large-scale GWAS (10,074 cases and 103,164 controls) for primary analysis and the IEU consortium for replication and meta-analysis. Sensitivity analyses were conducted to evaluate heterogeneity and pleiotropy. Following rigorous instrument selection steps, the number of SNPs finally used for T1D nine subgroups varying from 6 to 36 was retained in MR estimation. However, we did not observe evidence of causal association between type 1 diabetes nine subgroups and PCOS using the IVW analysis, MR-Egger regression, and weighted median approaches, and all P values were > 0.05 with ORs near 1. Subsequent replicates and meta-analyses also yielded consistent results. A number of sensitivity analyses also did not reveal heterogeneity and pleiotropy, including Cochran’s Q test, MR-Egger intercept test, MR-PRESSO global test, leave-one-out analysis, and funnel plot analysis. This is the first MR study to investigate the causal relationship between type 1 diabetes and PCOS. Our findings failed to find substantial causal effect of type 1 diabetes on risk of PCOS. Further randomized controlled studies and MR studies are necessary.

中文翻译：

---

1 型糖尿病与多囊卵巢综合征风险之间因果关系的遗传证据

越来越多的观察性研究已经确定 1 型糖尿病 (T1D) 和多囊卵巢综合征 (PCOS) 之间的关联。尽管如此，有关这种关联的因果效应的证据仍不确定。我们使用来自欧洲血统的大规模活检证实的全基因组关联研究 (GWAS) 的数据进行了两个样本孟德尔随机化 (MR) 分析，以测试 T1D 和 PCOS 之间的因果关系。我们创新地将T1D分为9个亚组分别进行分析，包括：1型糖尿病广义定义、1型糖尿病早发、1型糖尿病昏迷、1型糖尿病伴酮症酸中毒、1型糖尿病伴神经并发症、1型糖尿病伴眼科并发症、伴有外周循环并发症的 1 型糖尿病、伴有肾脏并发症的 1 型糖尿病以及伴有其他特定/多种/未特定并发症的 1 型糖尿病。PCOS 的 GWAS 数据来自大规模 GWAS（10,074 例病例和 103,164 例对照）进行初步分析，并通过 IEU 联盟进行复制和荟萃分析。进行敏感性分析以评估异质性和多效性。经过严格的仪器选择步骤，最终用于 T1D 九个亚组（从 6 到 36 个不等）的 SNP 数量保留在 MR 估计中。然而，使用 IVW 分析、MR-Egger 回归和加权中位数方法，我们没有观察到 1 型糖尿病九个亚组与 PCOS 之间存在因果关系的证据，并且所有 P 值均 > 0.05，OR 接近 1。分析也得出了一致的结果。许多敏感性分析也没有揭示异质性和多效性，包括 Cochran's Q 检验、MR-Egger 截距检验、MR-PRESSO 全局检验、留一分析和漏斗图分析。这是第一项调查 1 型糖尿病与 PCOS 之间因果关系的 MR 研究。我们的研究结果未能发现 1 型糖尿病与 PCOS 风险之间存在实质性因果关系。进一步的随机对照研究和 MR 研究是必要的。