Plasma N-glycans and risk of type 2 diabetes and CVD
| Type 2 diabetes . | |||
|---|---|---|---|
| Training cohort . | Validation cohort . | ||
| Plasma N-glycan . | HR per SD (95% CI) . | Plasma N-glycan . | HR per SD (95% CI) . |
| GP18−1 | 2.37 (1.96–2.86) | GP18−1 | 1.95 (1.47–2.58) |
| GP20 | 1.58 (1.33–1.88) | GP20 | 1.55 (1.22–1.97) |
| GP26 | 2.57 (2.13–3.09) | GP26 | 1.91 (1.49–2.43) |
| GP322 | 0.79 (0.64–0.98) | GP322 | 1.05 (0.74–1.48) |
| GP35−0.5 | 0.50 (0.42–0.58) | GP35−0.5 | 0.63 (0.50–0.80) |
| GP36 | 0.56 (0.48–0.66) | GP36 | 0.53 (0.42–0.66) |
| GST2D* | 4.20 (3.57–4.94) | GST2D* | 3.73 (3.00–4.63) |
| CVD | |||
| Men | Women | ||
| Plasma N-glycan | HR per SD (95% CI) | Plasma N-glycan | HR per SD (95% CI) |
| GST2D | 1.23 (1.09–1.40) | ||
| GP23 | 1.55 (1.29–1.87) | ||
| GP16 | 0.50 (0.40–0.63) | ||
| GSCVD† | 1.89 (1.57–2.27) | GP5‡ | 0.70 (0.57–0.86) |
| Type 2 diabetes . | |||
|---|---|---|---|
| Training cohort . | Validation cohort . | ||
| Plasma N-glycan . | HR per SD (95% CI) . | Plasma N-glycan . | HR per SD (95% CI) . |
| GP18−1 | 2.37 (1.96–2.86) | GP18−1 | 1.95 (1.47–2.58) |
| GP20 | 1.58 (1.33–1.88) | GP20 | 1.55 (1.22–1.97) |
| GP26 | 2.57 (2.13–3.09) | GP26 | 1.91 (1.49–2.43) |
| GP322 | 0.79 (0.64–0.98) | GP322 | 1.05 (0.74–1.48) |
| GP35−0.5 | 0.50 (0.42–0.58) | GP35−0.5 | 0.63 (0.50–0.80) |
| GP36 | 0.56 (0.48–0.66) | GP36 | 0.53 (0.42–0.66) |
| GST2D* | 4.20 (3.57–4.94) | GST2D* | 3.73 (3.00–4.63) |
| CVD | |||
| Men | Women | ||
| Plasma N-glycan | HR per SD (95% CI) | Plasma N-glycan | HR per SD (95% CI) |
| GST2D | 1.23 (1.09–1.40) | ||
| GP23 | 1.55 (1.29–1.87) | ||
| GP16 | 0.50 (0.40–0.63) | ||
| GSCVD† | 1.89 (1.57–2.27) | GP5‡ | 0.70 (0.57–0.86) |
Individual N-glycans were mutually adjusted and controlled for age.
GST2D, N-glycan score for type 2 diabetes prediction, a weighted combination of the selected N-glycans. Weights based on risk estimates in the training cohort (nsubcohort = 1,417, ncases = 495, overlap 44) and evaluated in the validation cohort (nsubcohort = 717, ncases = 248, overlap 20).
GSCVD, the CVD prediction score in men was derived as a weighted combination of GST2D, GP23, and GP16 (men: nsubcohort = 841, ncases = 269, overlap = 33).
In women, only GP5 was selected as robust risk marker, hence no score was derived (women: nsubcohort = 1,343, ncases = 149, overlap = 11).