Figure 1
Kaplan-Meier curve analysis in the whole derivation cohort. (A) Survival probability for all patients in the derivation cohort. (B) Survival probability for all patients in the derivation cohort grouped by statins using or not. Gray line: no statins using patients. Blackline: statins using patients. (C) Survival probability for all patients in the derivation cohort grouped by gender. Gray line: Male. Black line: Female. (D) Survival probability for all patients in the derivation cohort grouped by geriatric or not. Geriatric is defined according to the clinical definition in China. Gray line: Not Geriatric, younger than 60. Black line: Geriatric, older than 60. Survival probability: the outcome modeled in the Kaplan-Meier curves are for the composite outcome defined in ‘Methods.’
Table 1
Characteristics of patients aged 30–89 in derivation and validation cohorts 2008–2017 [CODE: finalBaselineTable.R].
| CHARACTERISTICS | DERIVATION COHORT | VALIDATION COHORT | ||
|---|---|---|---|---|
| FEMALE | MALE | FEMALE | MALE | |
| N | 39,392 | 45,238 | 51,651 | 68,240 |
| Age at diagnosis, years, median (IQR) | 59 (8.6) | 57 (9.4) | 61 (9.3) | 58 (8.9) |
| Systolic blood pressure, mmHg, mean (SD) | 131 (18.7) | 130 (18.2) | 137.6 (17.7) | 138.1 (17.1) |
| Smoking status, n (%) | ||||
| no | 38,840 (98.6) | 28,904 (63.9) | 26,061 (50.5) | 28,900 (42.4) |
| ex | 80 (0.2) | 2,599 (5.7) | 14,596 (28.3) | 24,539 (36) |
| cur | 472 (1.2) | 13,735 (30.4) | 10,994 (21.3) | 14,801 (21.7) |
| TDL cholesterol, mmol/mol, mean (SD) | 5.9 (1.6) | 5.9 (1.6) | 5.4 (1.2) | 5.1 (1.3) |
| LDL cholesterol, mmol/mol, mean (SD) | 3.4 (0.9) | 3.4 (0.9) | 3 (1.1) | 2.8 (1) |
| HDL cholesterol, mmol/mol, mean (SD) | 2.4 (0.8) | 2.3 (0.8) | 1.3 (0.4) | 1.1 (0.3) |
| Glycated hemoglobin, %, mean (SD) | 6.9 (1.5) | 7.2 (1.7) | 7.9 (2) | 8.3 (2.2) |
| Urine creatinine (umol/L) | 100.9 (66.2) | 134.2 (78.6) | 71.8 (18.8) | 85.3 (21.8) |
| Albumin-to-creatinine ratio, mean (SD) | 72.7 (141.5) | 60.5 (122.2) | 3.5 (13.5) | 4.1 (14.9) |
| Estimated glomerular filtration rate mls/min/1.73 m2, mean (SD) | 68.9 (34.9) | 66 (36.5) | 81.2 (19.4) | 86.9 (18) |
| Albuminuria, n (%) | ||||
| normal | 25,460 (64.6) | 30,996 (68.5) | 42,414 (82.1) | 51,240 (75.1) |
| micro | 11,261 (28.6) | 11,919 (26.3) | 7,741 (15) | 14,553 (21.3) |
| macro | 2,671 (6.8) | 2,323 (5.1) | 1,496 (2.9) | 2,447 (3.6) |
| Retinopathy, n (%) | ||||
| 0 | 37,487 (95.2) | 43,416 (96) | 51,643 (100) | 68,225 (100) |
| 1 | 1,905 (4.8) | 1,822 (4) | 8 (0) | 15 (0) |
| rheumatoidarthritis, n (%) | ||||
| 0 | 39,087 (99.2) | 45,058 (99.6) | 50,962 (98.7) | 67,822 (99.4) |
| 1 | 305 (0.8) | 180 (0.4) | 689 (1.3) | 418 (0.6) |
| Atrial Fibrillation, n (%) | ||||
| 0 | 39,210 (99.5) | 44,938 (99.3) | 49,806 (96.4) | 65,582 (96.1) |
| 1 | 182 (0.5) | 300 (0.7) | 1,845 (3.6) | 2,658 (3.9) |
| Prescribed statins prior to diabetes diagnosis, n (%) | ||||
| 0 | 34,925 (88.7) | 39,102 (86.4) | 36,088 (69.9) | 47,204 (69.2) |
| 1 | 4,467 (11.3) | 6,136 (13.6) | 15,563 (30.1) | 21,036 (30.8) |
| Prescribed antihypertensive medications, n (%) | ||||
| 0 | 30,349 (77) | 34,151 (75.5) | 35,987 (69.7) | 48,051 (70.4) |
| 1 | 9,043 (23) | 11,087 (24.5) | 15,664 (30.3) | 20,189 (29.6) |
| 5-year CVD event rate, n (%) | 7,340 (18.63) | 9,954 (22.0) | 2,579 (5) | 3,660 (5.4) |
| 10-year CVD event rate, n (%) | 8,048 (20.43) | 10,779 (23.83) | 3,647 (7.1) | 5,116 (7.5) |
Table 2
Adjusted hazard ratios (95% CI) for cardiovascular disease for the basic and extended models in derivation cohort.
| MODEL NAME | BASIC | EXTENDED |
|---|---|---|
| Age at diagnosis | 1.027 [1.026–1.028]*** | 1.025 [1.024–1.027]*** |
| Sex: | ||
| male | 1 | 1 |
| female | 0.821 [0.797–0.846]*** | 0.86 [0.831–0.889]*** |
| Rheumatoid Arthritis | 1.63 [1.399–1.899]*** | 1.65 [1.416–1.922]*** |
| Hypertension | 2.873 [2.785–2.964]*** | 2.758 [2.672–2.847]*** |
| Prescribed statins prior to diabetes diagnosis | 2.264 [2.178–2.353]*** | 2.37 [2.279–2.464]*** |
| Prescribed antihypertensive medications | 0.635 [0.611–0.659]*** | 0.628 [0.604–0.653]*** |
| Smoking status | ||
| Never smoker | 1 | 1 |
| Ex-smoker | 1.232 [1.18–1.285]*** | |
| Current smoker | 1.273 [1.18–1.373]*** | |
| Albuminuria | ||
| Normal | 1 | |
| Microalbuminuria | 1.096 [1.06–1.134]*** | |
| Macroalbuminuria | 1.244 [1.172–1.32]*** | |
| Estimated glomerular filtration rate, mls/min/1.73 m2, | ||
| (0, 15) | 0.748 [0.631–0.888]*** | |
| （ 15, 30） | 1 | |
| (30, 60) | 1.146 [1.087–1.208]*** | |
| (60, 90) | 1.308 [1.24–1.381]*** | |
| >90 | 1.354 [1.285–1.427]*** | |
| Glycated hemoglobin, %, | 0.904 [0.894–0.914]*** | |
| LDL cholesterol, mmol/mol, | 0.828 [0.814–0.842]*** | |
| Center | 1.693 | 0.558 |
| Harrell’s C-statistic | 0.718 [0.716–0.72] | 0.727 [0.725–0.729] |
| Nagelkerke’s R2 | 0.11 | 0.123 |
| AIC | 374,308.188 | 373,182.563 |
| Internal Validation C-statistic (bootstrap) | 0.718 [0.716–0.72] | 0.727 [0.725–0.729] |
[i] *** p < 0.001.
Figure 2
Calibration plots for observed versus predicted five-year risk of CVD as estimated using the basic and the extended models in the derivation cohort. Gray dashed line reflects perfect agreement between observed and predicted risk. (A) Calibration plots using the basic model. (B) Calibration plots using the extended model.