Table 3. Generalized linear regression results.
Generalized linear regression coefficients obtained through modeling of metabolite variations as a function of gender proportion, smoking history, body-mass index (BMI), age and AMD status. F.A.: Fatty acids. Values in bold illustrate the higher contributions of AMD status for each metabolite variation, compared to confounders. Similar metabolite variations in the two cohorts are denoted by underline.
| Compound | Gender proportion | Smoking history | BMI | Age | AMD status |
|---|---|---|---|---|---|
| Coimbra cohort | |||||
| Acetate | -4.0x10-5 | 2.1x10-5 | -5.3x10-6 | 1.0x10-6 | 1.1x10-3 |
| Acetoacetate | -3.8x10-5 | -3.1x10-5 | -3.1x10-6 | 7.3x10-6 | 6.9x10-4 |
| Creatine | 1.4x10-4 | -1.8x10-5 | 4.2x10-6 | -4.8x10-6 | 1.6x10-3 |
| Dimethyl sulfone | -1.8x10-5 | 2.8x10-5 | -4.8x10-7 | 7.5x10-6 | 2.4x10-4 |
| HC = CH F.A. | 7.6x10-4 | 2.9x10-4 | 1.2x10-5 | -3.3x10-5 | 1.5x10-2 |
| Histidine | 1.3x10-7 | -6.5x10-6 | -1.3x10-6 | 3.1x10-8 | 8.1x10-4 |
| β -hydroxybutyrate | -1.3x10-5 | -4.6x10-6 | -1.3x10-6 | 2.5x10-6 | 4.6x10-4 |
| Pyruvate | -9.9x10-5 | 1.5x10-5 | 3.3x10-6 | 6.2x10-7 | 9.3x10-4 |
| Boston cohort | |||||
| Alanine | -8.0x10-5 | 2.9x10-5 | -2.0x10-5 | -5.5x10-6 | 4.2x10-3 |
| Glutamine | -4.6x10-4 | -1.2x10-4 | -8.5x10-5 | 3.2x10-5 | 7.1x10-3 |
| Histidine | -2.3x10-5 | 4.0x10-6 | -4.0x10-6 | 3.6x10-6 | 5.1x10-4 |
| HC = CH F.A. | 3.0x10-4 | -1.7x10-4 | 5.5x10-6 | -5.4x10-5 | 1.4x10-2 |