Genetic instrument selection and genetic correlation between phenotypes

F-statistics of all the IVs of SUA were above the threshold of 10, indicating that the IVs were strong instruments and thus reducing the bias of the IV estimates (Supplementary Table S2). Then, SNPs with incompatible alleles (e.g., rs2749005 and rs34555420) and palindromic SNPs whose orientation could not be determined (e.g., rs12510175, rs1851285, rs1869581, rs2252862, rs2493121, rs538737, and rs7039) were eliminated by data harmonization. Furthermore, 21 SNPs associated with the outcome variable, i.e., diabetic microvascular complications (e.g., BMI, blood lipids, FBG, homocysteine, smoking status, and hypertension status), were identified and excluded by screening in the PhenoScanner database (Supplementary Table S3). Finally, the screened SNPs were used as IVs of the SUA.

Causal effect of SUA on diabetic nephropathy

There was a causal relationship between genetically elevated SUA and diabetic nephropathy [OR = 1.32, 95%(CI) = 1.07–1.63, p = 0.008] according to the IVW model. The corrected MR-PRESSO model results [OR = 1.30, 95%(CI) = 1.07–1.58, p = 0.008] were consistent with the IVW results. The results of the MR‒Egger model [OR = 1.15, 95%(CI) = 0.76–1.75, p = 0.498] and WM model [OR = 1.33, 95%(CI) = 0.98–1.79, p = 0.064] were inconsistent with those of the IVW model. We also performed MR analysis on secondary outcomes of diabetic nephropathy [type 1 diabetes mellitus (T1DM) with renal complications and type 2 diabetes mellitus (T2DM) with renal complications]. There was no significant causal relationship between SUA and renal complications in any of the four models: the IVW model [OR = 1.06, 95%(CI) = 0.77–1.46, p = 0.711], the MR-PRESSO model [OR = 1.06, 95%(CI) = 0.77–1.46, p = 0.711], the MR‒Egger model [OR = 0.88, 95%(CI) = 0.46–1.69, p = 0.697], or the WM model [OR = 1.00, 95%(CI) = 0.62–1.62, p = 0.999]. However, we obtained the opposite results: the IVW model [OR = 1.27, 95%(CI) = 1.00-1.62, p = 0.049] and MR-PRESSO model [OR = 1.27, 95%(CI) = 1.00-1.62, p = 0.050] indicated a significant causal relationship between SUA and renal complications in patients with T2DM. The sensitivity analysis showed that no SNP had a significant effect on the causal relationship estimates in the “leave-one-out” analysis (Supplementary Fig. 1). The MR‒Egger intercept test in the pleiotropy analysis indicated that the IVs did not have any directional pleiotropy (Table 1). MR analysis found a significant causal relationship between SUA and both DN and T2DM with renal complications but no significant causal relationship between T1DM and renal complications (Fig. 3A).

Table 1 Pleiotropic and heterogeneous trials of causality between serum uric acid levels and the risk of diabetic microvascular complicationsCausal effect of SUA on diabetic retinopathy

No significant causal relationship existed between SUA and diabetic retinopathy according to the IVW [OR = 1.09, 95%(CI) = 0.94–1.26, p = 0.249], MR‒Egger [OR = 1.25, 95%(CI) = 0.93–1.68, p = 0.142], or WM models [OR = 1.21, 95%(CI) = 0.98–1.49, p = 0.081] of MR. The same conclusion was obtained from the corrected MR-PRESSO model [OR = 1.08, 95%(CI) = 0.84–1.40, p = 0.550]. MR analysis for secondary outcomes of diabetic retinopathy (T1DM with ophthalmic complications, T2DM with ophthalmic complications) was also performed. The results of the IVW model [OR = 1.00, 95%(CI) = 0.83–1.22, p = 0.963], MR-PRESSO model [OR = 1.00, 95%(CI) = 0.83–1.22, p = 0.963], MR‒Egger model [OR = 1.30, 95%(CI) = 0.87–1.93, p = 0.200], and WM model [OR = 1.30, 95%(CI) = 0.95–1.78, p = 0.102] showed no significant causal relationships between SUA and T1DM with ophthalmic complications. There was also no significant causal relationship between SUA and T2DM with ophthalmic complications according to the results of the four MR models, i.e., IVW [OR = 1.17, 95%(CI) = 0.96–1.42, p = 0.115], MR‒Egger [OR = 1.18, 95%(CI) = 0.81–1.74, p = 0.392], WM [OR = 1.07, 95%(CI) = 0.79–1.45, p = 0.665] and MR-PRESSO [OR = 1.13, 95%(CI) = 0.93–1.38, p = 0.218]. Similarly, the MR‒Egger intercept tests did not suggest any directional pleiotropy in IVs (Table 1). The details are listed in Supplementary Fig. 1. The results of MR analysis indicated that there was no significant causal relationship between SUA and diabetic retinopathy (Fig. 3B).

Fig. 3

Forest plot of the association between SUA and diabetic microvascular complications. (A) Diabetic nephropathy, (B) diabetic retinopathy, and (C) diabetic neuropathy. The dot and bar indicate the causal estimate and 95%(CI) of the association between increasing SUA and diabetic microvascular complications, respectively

Causal effect of SUA on diabetic neuropathy

The results of the IVW model [OR = 1.08, 95%(CI) = 0.84–1.40, p = 0.549], MR‒Egger [OR = 1.14, 95%(CI) = 0.68–1.90, p = 0.616], and WM models [OR = 0.93, 95%(CI) = 0.63–1.36, p = 0.703] showed no significant causal relationships between SUA and diabetic neuropathy. The same conclusion was obtained from the corrected MR-PRESSO model [OR = 1.08, 95%(CI) = 0.84–1.40, p = 0.550].

Moreover, we performed MR analysis on secondary outcomes of diabetic neuropathy (T1DM with neurological complications, T2DM with neurological complications) and showed that there was no significant causal relationship between SUA and either T1DM with neurological complications or T2DM with neurological complications. We used four models, namely, IVW [OR = 1.17, 95%(CI) = 0.79–1.72, p = 0.430], MR-PRESSO [OR = 1.17, 95%(CI) = 0.79–1.72, p = 0.431], MR‒Egger [OR = 1.69, 95%(CI) = 0.79–3.61, p = 0.181] and WM [OR = 1.13, 95%(CI) = 0.61–2.12, p = 0.694], to analyze the causal relationships between SUA and neither T1DM had neurological complications. Similarly, the relationships between SUA and T2DM with neurological complications were analyzed using the IVW model [OR = 0.91, 95%(CI) = 0.69–1.21, p = 0.532], MR‒Egger model [OR = 0.71, 95%(CI) = 0.40–1.26, p = 0.248], WM model [OR 0.74, 95%(CI) = 0.48–1.17, p = 0.197] and MR-PRESSO model [OR = 0.91, 95%(CI) = 0.69–1.21, p = 0.533]. Likewise, the MR‒Egger intercept tests suggested that there was no pleiotropy in the IVs (Table 1, Supplementary Fig. 1). The results of MR analysis indicated that there was no significant causal relationship between SUA and diabetic neuropathy (Fig. 3C).