Table 1.
Clinical characteristics at baseline eye assessment in UK Biobank (2009–2013) and from the third health check phase for EPIC-Norfolk (2004–2011)
| Baseline characteristic | Mean (SD) or (%) | |
| UK Biobank N=66 326 | EPIC N=5955 | |
| Median duration of follow-up (years) | 7.7 | 9.1 |
| Age (years) | 56.8 (8.2) | 67.6 (7.6) |
| Female (%) | 55.0% | 57.1% |
| Ethnicity (%) | ||
| White | 92.0% | 99.5% |
| Black | 2.5% | 0.1% |
| Asian | 2.5% | 0.0% |
| Other | 2.5% | 0.2% |
| Unknown/did not answer | 0.6% | 0.0% |
| Smoking (%) | ||
| Never smoker | 56.7% | 49.7% |
| Occasionally | 2.6% | N/A |
| Ex-smoker | 34.0% | 44.2% |
| Current smoker | 6.1% | 4.5% |
| Prefer not to say/missing | 0.5% | 1.6% |
| BMI (kg/m²) | 27.2 (4.7) | 26.8 (4.3) |
| Systolic BP (mm Hg) | 136.8 (18.3) | 135.7 (16.6) |
| Diastolic BP (mm Hg) | 81.5 (10.0) | 78.5 (9.3) |
| Total cholesterol (mmol/L) | 5.7 (1.1) | 5.4 (1.1) |
| LDL cholesterol (mmol/L) | 3.5 (0.9) | 3.2 (1.0) |
| HDL cholesterol (mmol/L) | 1.5 (0.4) | 1.5 (0.4) |
| Triglycerides (mmol/L) | 1.7 (1.0) | 1.7 (0.9) |
| UK Biobank | EPIC | |
| Image quality* | 0.9 (0.1) | 0.9 (0.1) |
| Arteriolar width (µm) | 86.9 (7.9) | 75.1 (6.4) |
| Venular width (µm) | 103.1 (13.0) | 91.4 (10.8) |
| Arteriolar tortuosity† | 4.4 (1.6) | 4.3 (1.6) |
| Venular tortuosity† | 3.1 (1.4) | 3.2 (1.3) |
| Arteriolar vessel area (mm²) | 1.8 (0.8) | 2.0 (0.7) |
| Venular vessel area (mm²) | 2.5 (0.9) | 2.6 (0.7) |
| Arteriolar segment SD−1 (µm−1) | 0.08 (0.03) | 0.13 (0.03) |
| Venular segment SD−1 (µm−1) | 0.10 (0.03) | 0.14 (0.04) |
| Use of medications | ||
| Blood pressure lowering | 19.6% | 35.1% |
| Cholesterol lowering | 17.9% | 22.3% |
| Prefer not to report/missing | 1.2% | N/A |
| Self-reported history | ||
| Heart attack (%) | 1.9% | 3.1% |
| Stroke (%) | 1.4% | 2.0% |
| Prefer not to report/missing | 1.2% | N/A |
| Diabetes (%) | 4.9% | 4.0% |
Values are mean (SD) or (%).
Missing data were BMI n=287, blood pressure n=207 in UK Biobank only. Framingham risk score-based models for incident MI that used lipids missing data were as follows after excluding those with prevalent events (MI or diabetes) or using lipid lowering therapy, n= 6805 for total cholesterol or HDL cholesterol for UK Biobank. Other missing variables for UK Biobank were LDL cholesterol n=4628; triglycerides n=4576; and for EPIC-Norfolk—total cholesterol n=428, LDL cholesterol n=510, HDL cholesterol n=427, triglycerides n=428)
*Image quality score generated by QUARTZ, values range from 0.6 to 1.0, higher values indicate higher image quality.
†Geometric mean exponentiated SD of the log-transformed values; the 95% range for the geometric mean is from (geometric mean÷GSD2) to (geometric mean×GSD2).
BMI, body mass index; BP, blood pressure; EPIC, European Prospective Investigation into Cancer; HDL, high-density lipoprotein; LDL, low-density lipoprotein; QUARTZ, QUantitative Analysis of Retinal vessels Topology and siZe.