IgG glycosylation and diabetic sequelae: examining the associations with nephropathy and neuropathy
Aims: To investigate whether baseline IgG N-glycosylation profiles are associated with incident diabetic neuropathy and nephropathy.
Methods: This pooled analysis of four prospective European cohorts included 1,924 adults in neuropathy analyses and 3,263 in nephropathy analyses. Baseline IgG N-glycosylation was measured using chromatography or mass spectrometry before type 2 diabetes diagnosis in EPIC-Potsdam and after diagnosis in DiaGene, Hoorn DCS and GenoDiabMar. Participants with the relevant complication at baseline were excluded. Neuropathy was assessed by clinical examination in DiaGene and Hoorn DCS and by standardised physician reports classified under ICD-10 EPIC-Potsdam. Nephropathy was identified using albuminuria or physician-reported renal outcomes. Multivariable Cox models were pooled using a random-effects meta-analysis, with appropriate multiple testing adjustment.
Results: During follow-up, 674 participants developed neuropathy and 639 developed nephropathy. No IgG glycan peak or derived glycosylation trait was significantly associated with incident neuropathy after multiple-testing correction. In contrast, agalactosylated, asialylated IgG-GP3 and IgG-GP4 were associated with increased nephropathy risk (HR=1.13, 95% CI 1.04–1.24 and HR=1.15, 95% CI 1.05–1.26, respectively; both FDR<0.05). Digalactosylated IgG-GP14 and sialylated IgG-GP18 were associated with reduced nephropathy risk (HR=0.86, 95% CI 0.78–0.94 and HR=0.85, 95% CI 0.77–0.94, respectively; both FDR=0.02).
Conclusions: IgG N-glycosylation profiles were prospectively associated with incident nephropathy, supporting further investigation of glycosylation-related immune mechanisms in diabetic complications. Three glycan structures showed nominal associations with incident neuropathy, although these did not withstand multiple-testing correction.
Comments. This study has a large number of incident cases of neuropathy and nephropathy and uses rigorous statistical methodology. However, monofilament and vibration testing may miss early or isolated small fibre neuropathy, while EPIC-Potsdam is not a dedicated resource for diabetes research and diabetic complications relying on physician-reported outcomes and ICD-10 coding. Indeed, differences in how diabetic peripheral neuropathy was diagnosed may have introduced misclassification. Incident diagnoses may therefore reflect recognition of established disease rather than its onset, potentially obscuring associations with IgG glycosylation. By comparison, because albuminuria is measured quantitatively against established thresholds, nephropathy status may have been more consistently categorised as positive or negative for the statistical analyses than neuropathy status. Differences in outcome classification may therefore contribute to the contrasting associations observed for nephropathy and neuropathy. There may also be genuine neuroimmunological differences between painful and insensate presentations of diabetic peripheral neuropathy. Analysing neuropathy as a single clinical outcome may therefore obscure phenotype-specific associations with IgG glycosylation. Further prospective and longitudinal studies with methods to identify small, mixed and large fibre diabetic peripheral neuropathy together with characterising somatosensory phenotypes could determine whether IgG glycosylation is associated with incident diabetic neuropathy.
Maryam Ferdousi
Reference. Martinez Escobedo C, Birukov A, Heijmans R, Kuxhaus O, Eichelmann F, Slieker RC, Memarian E, Simunic-Briski N, Rutters F, Blom MT, Riera M, Crespo M, Wuhrer M, Barrios C, van Hoek M, 't Hart LM, Sijbrands E, Lauc G, Schulze MB, Wittenbecher C. IgG N-Glycosylation and Incident Neuropathy and Nephropathy in People With Type 2 Diabetes: A Pooled Analysis of Prospective Cohort Studies. Diabetes Care. 2026 Sep 1;49(9):1714-1719. doi: 10.2337/dc26-0620. PMID: 42467059; PMCID: PMC13493340.
