Our study included 8,722 participants without diabetes at baseline. The mean age of study participants was 46±12 years, and 66.5% were men; 33.9% were hypertensive; 28.3% had hyperlipidemia; 5.4% were known to have prior coronary artery disease (CAD), and only a few of them had heart failure (Table1).
Table 1
Baseline characteristics of the study cohort
| Total | METs < 6 | METs 6-9 | METs 10-11 | METs ≥ 12 | p-value |
| 8722 | 589 (6.7%) | 3077 (35.3%) | 2930 (33.6%) | 2126 (24.4%) |
Age (year) | 45.9 ± 12.06 | 50.9 ± 12.37 | 48.4 ± 11.87 | 45.5 ± 11.81 | 41.3 ± 10.96 | <.0001 |
Gender (Male) | 5798 (66.5%) | 254 (43.1%) | 1410 (45.8%) | 2157 (73.6%) | 1977 (93.0%) | <.0001 |
Height (cm) | 162.2 ± 19.51 | 157.8 ± 18.52 | 159.4 ± 17.97 | 164.2 ± 17.85 | 165.0 ± 23.49 | <.0001 |
Weight (kg) | 79.7 ± 16.55 | 81.7 ± 18.08 | 81.3 ± 17.51 | 79.5 ± 16.52 | 77.2 ± 14.11 | <.0001 |
BMI (kg/m2) | 29.4 ± 6.15 | 32.3 ± 7.22 | 31.2 ± 6.58 | 28.5 ± 5.23 | 26.9 ± 5.10 | <.0001 |
Cardiovascular risk factors | | | | | | |
Hypertension | 2955 (33.9%) | 258 (43.8%) | 1154 (37.5%) | 987 (33.7%) | 556 (26.2%) | <.0001 |
Hyperlipidaemia | 2465 (28.3%) | 172 (29.2%) | 880 (28.6%) | 884 (30.2%) | 529 (24.9%) | 0.0005 |
Known CAD | 474 (5.4%) | 27 (4.6%) | 157 (5.1%) | 184 (6.3%) | 106 (5.0%) | 0.0938 |
Known CHF | 66 (0.8%) | 4 (0.7%) | 30 (1.0%) | 23 (0.8%) | 9 (0.4%) | 0.1593 |
Prior CABG | 131 (1.5%) | 10 (1.7%) | 57 (1.9%) | 43 (1.5%) | 21 (1.0%) | 0.0885 |
Prior MI | 454 (5.2%) | 30 (5.1%) | 151 (4.9%) | 172 (5.9%) | 101 (4.8%) | 0.2524 |
Prior PCI | 306 (3.5%) | 16 (2.7%) | 92 (3.0%) | 123 (4.2%) | 75 (3.5%) | 0.0538 |
Smoking | 277 (3.2%) | 34 (5.8%) | 88 (2.9%) | 107 (3.7%) | 48 (2.3%) | <.0001 |
Lung disease | 348 (4.0%) | 37 (6.3%) | 145 (4.7%) | 124 (4.2%) | 42 (2.0%) | <.0001 |
Lab results* | | | | | | |
Cholesterol | 4.6 (3.87, 5.30) | 4.4 (3.80, 5.20) | 4.5 (3.84, 5.28) | 4.5 (3.87, 5.33) | 4.6 (3.95, 5.35) | 0.0142 |
Triglyceride | 1.4 (0.99, 1.86) | 1.4 (1.04, 1.94) | 1.3 (0.98, 1.82) | 1.4 (0.98, 1.89) | 1.4 (1.02, 1.91) | 0.0128 |
HDL | 1.0 (0.88, 1.22) | 1.0 (0.87, 1.22) | 1.1 (0.90, 1.25) | 1.0 (0.87, 1.22) | 1.0 (0.88, 1.18) | <.0001 |
LDL | 2.8 (2.22, 3.49) | 2.7 (2.10, 3.32) | 2.8 (2.19, 3.43) | 2.8 (2.22, 3.51) | 2.9 (2.29, 3.62) | <.0001 |
Haemoglobin | 143 (129, 154) | 133.0 (121, 144) | 136.0 (123, 148) | 146.0 (132, 156) | 152.0 (142, 160) | <.0001 |
Haematocrit | 0.4 (0.39, 0.46) | 0.4 (0.37, 0.43) | 0.4 (0.37, 0.44) | 0.4 (0.40, 0.46) | 0.4 (0.42, 0.47) | <.0001 |
Calcium | 2.3 (2.25, 2.40) | 2.3 (2.20, 2.35) | 2.3 (2.23, 2.39) | 2.3 (2.25, 2.39) | 2.4 (2.28, 2.42) | <.0001 |
BUN | 4.6 (3.80, 5.80) | 4.5 (3.60, 6.00) | 4.5 (3.60, 5.80) | 4.7 (3.80, 5.80) | 4.8 (4.00, 5.80) | <.0001 |
Creatinine | 75.0 (64.00, 89.00) | 69.5 (60.00, 85.00) | 70.0 (60.00, 84.00) | 75.0 (65.00, 88.00) | 81.0 (72.00, 93.00) | <.0001 |
eGFR | 94.7 (80.39, 109.42) | 92.8 (77.29, 106.46) | 93.7 (78.85, 107.83) | 95.8 (81.84, 110.79) | 94.3 (81.44, 110.09) | 0.0001 |
hsCRP | 2.5 (0.93, 5.00) | 6.0 (2.78, 11.00) | 3.5 (1.24, 6.16) | 2.7 (0.85, 5.00) | 1.4 (0.78, 2.38) | 0.0001 |
Vitamin-D | 32.9 (23.00, 48.00) | 29.8 (21.00, 47.00) | 33.0 (22.20, 49.00) | 33.3 (23.15, 49.40) | 32.6 (23.90, 46.00) | 0.0001 |
Medications | | | | | | |
Beta-blockers | 710 (8.1%) | 61 (10.4%) | 268 (8.7%) | 247 (8.4%) | 134 (6.3%) | 0.0017 |
Calcium channel blockers | 503 (5.8%) | 54 (9.2%) | 202 (6.6%) | 159 (5.4%) | 88 (4.1%) | <.0001 |
ACEI | 512 (5.9%) | 40 (6.8%) | 177 (5.8%) | 191 (6.5%) | 104 (4.9%) | 0.0753 |
ARB | 988 (11.3%) | 86 (14.6%) | 386 (12.5%) | 341 (11.6%) | 175 (8.2%) | <.0001 |
ACEARB | 1288 (14.8%) | 107 (18.2%) | 489 (15.9%) | 452 (15.4%) | 240 (11.3%) | <.0001 |
Aspirin | 875 (10.0%) | 60 (10.2%) | 309 (10.0%) | 328 (11.2%) | 178 (8.4%) | 0.0123 |
Digoxin | 48 (0.6%) | 7 (1.2%) | 20 (0.6%) | 12 (0.4%) | 9 (0.4%) | 0.0847 |
PPI | 744 (8.5%) | 59 (10.0%) | 271 (8.8%) | 284 (9.7%) | 130 (6.1%) | <.0001 |
Plavix | 2057 (23.6%) | 149 (25.3%) | 718 (23.3%) | 717 (24.5%) | 473 (22.2%) | 0.2169 |
Statin | 1388 (15.9%) | 86 (14.6%) | 483 (15.7%) | 506 (17.3%) | 313 (14.7%) | 0.0674 |
Diuretic | 357 (4.1%) | 41 (7.0%) | 154 (5.0%) | 110 (3.8%) | 52 (2.4%) | <.0001 |
Stress test results | | | | | | |
Rest HR (bpm) | 81.9 ± 16.12 | 86.9 ± 18.45 | 84.8 ± 17.00 | 81.1 ± 15.05 | 77.5 ± 14.21 | <.0001 |
Peak HR (bpm) | 155.5 ± 22.13 | 130.9 ± 25.28 | 149.8 ± 21.76 | 158.5 ± 18.79 | 166.7 ± 17.53 | <.0001 |
Rest SBP (mmHg) | 129.8 ± 16.94 | 133.0 ± 19.06 | 130.7 ± 18.34 | 129.1 ± 16.04 | 128.4 ± 14.98 | <.0001 |
Peak SBP (mmHg) | 160.8 ± 25.92 | 152.7 ± 32.48 | 160.7 ± 28.64 | 162.4 ± 24.55 | 160.7 ± 21.29 | <.0001 |
Rest DBP (mmHg) | 79.2 ± 9.88 | 79.5 ± 10.07 | 78.6 ± 10.38 | 79.6 ± 9.45 | 79.5 ± 9.62 | 0.0034 |
Peak DBP (mmHg) | 82.8 ± 10.63 | 82.3 ± 10.89 | 83.4 ± 10.77 | 83.4 ± 10.76 | 81.2 ± 10.06 | <.0001 |
Peak METs | 10.1 ± 2.80 | 4.7 ± 0.95 | 8.0 ± 1.04 | 10.7 ± 0.62 | 13.6 ± 1.48 | <.0001 |
Chronotropic Incompetence | 2275 ( 26.1% ) | 391 ( 66.4% ) | 1013 ( 32.9% ) | 591 ( 20.2% ) | 280 ( 13.2% ) | <.0001 |
Duke Treadmill Score | 5.5 ± 6.12 | 1.8 ± 5.79 | 4.0 ± 5.43 | 6.2 ± 5.73 | 7.9 ± 6.56 | <.0001 |
Low | 6040 (69.3%) | 114 (19.4%) | 2064 (67.1%) | 2221 (75.8%) | 1641 (77.2%) | <.0001 |
Moderate | 2447 (28.1%) | 454 (77.1%) | 925 (30.1%) | 597 (20.4%) | 471 (22.2%) |
High | 235 (2.7%) | 21 (3.6%) | 88 (2.9% ) | 112 (3.8% ) | 14 (0.7%) |
CAD: Coronary artery disease, CHF: Congestive heart failure, CABG: Coronary artery bypass grafting, MI: Myocardial infarction, PCI: Percutaneous coronary intervention, HDL: High density lipoprotein, LDL: Low density lipoprotein, BUN: Blood urea nitrogen, eGFR: estimated Glomerular filtration rate, ACEI: Angiotensin converting enzyme inhibitors, ARB: Angiotensin Receptor blocker, PPI: Proton pump inhibitors, HR: Heart rate, SBP: Systolic blood pressure, DBP: Diastolic blood pressure, METs: Metabolic equivalent of tasks. |
Table (2) Adjusted multivariate Cox regression model and predictive value of the finding of exercise stress testing |
| Model (1) | Model (2) | Model (3) | Model (4A) | Model (4B) |
AIC | 36619.287 | 36573.79 | 36408.081 | 36253.129 | 36227.135 |
C-Statistics | 0.6221 | 0.6289 | 0.6581 | 0.6817 | 0.6839 |
IAUC | 0.6256 | 0.6352 | 0.6581 | 0.6911 | 0.6948 |
NRI | 0.3757 | 0.2451 | 0.3562 | 0.1595 | 0.129 |
IDI | 0.0123 | 0.0019 | 0.0065 | 0.0026 | 0.0026 |
| HR | 95% CI | p | HR | 95% CI | p | HR | 95% CI | p | HR | 95% CI | p | HR | 95% CI | p |
Age (years) | 1.032 | ( 1.03-1.04 ) | <.001 | 1.028 | ( 1.02-1.03 ) | <.001 | 1.028 | ( 1.02-1.03 ) | <.001 | 1.022 | ( 1.02-1.03 ) | <.001 | 1.02 | ( 1.02-1.02 ) | <.001 |
Gender Female vs Male | 1.131 | ( 1.03-1.24 ) | 0.009 | 1.176 | ( 1.07-1.29 ) | 0.001 | 1.138 | ( 1.04-1.25 ) | 0.007 | 0.892 | ( 0.81-0.99 ) | 0.027 | 0.87 | ( 0.79-0.96 ) | 0.007 |
Resting heart rate (bpm) | 1.002 | ( 1.00-1.00 ) | 0.109 | 1.003 | ( 1.00-1.01 ) | 0.011 | 1.004 | ( 1.00-1.01 ) | 0.003 | 1.002 | ( 1.00-1.01 ) | 0.109 | 1.001 | ( 1.00-1.00 ) | 0.452 |
Cardiovascular risk factors | | | | | | | | | | | | | | | |
Hypertension | | | | 1.189 | ( 1.08-1.31 ) | 0.000 | 1.138 | ( 1.03-1.26 ) | 0.015 | 1.06 | ( 0.95-1.18 ) | 0.284 | 1.059 | ( 0.95-1.18 ) | 0.293 |
Hyperlipidemia | | | | 1.196 | ( 1.08-1.32 ) | 0.000 | 1.502 | ( 1.36-1.66 ) | <.001 | 1.537 | ( 1.39-1.70 ) | <.001 | 1.539 | ( 1.39-1.71 ) | <.001 |
Smoking | | | | 1.202 | ( 0.94-1.54 ) | 0.150 | 1.241 | ( 0.97-1.59 ) | 0.090 | 1.19 | ( 0.93-1.52 ) | 0.167 | 1.178 | ( 0.92-1.51 ) | 0.193 |
Lung disease | | | | 1.06 | ( 0.77-1.46 ) | 0.720 | 0.961 | ( 0.69-1.34 ) | 0.814 | 0.94 | ( 0.67-1.31 ) | 0.716 | 0.908 | ( 0.65-1.27 ) | 0.572 |
Known CAD | | | | 1.154 | ( 0.98-1.36 ) | 0.092 | 1.149 | ( 0.97-1.37 ) | 0.117 | 1.18 | ( 0.99-1.40 ) | 0.060 | 1.189 | ( 1.00-1.41 ) | 0.049 |
Known CHF | | | | 0.588 | ( 0.36-0.97 ) | 0.039 | 0.483 | ( 0.29-0.81 ) | 0.005 | 0.431 | ( 0.26-0.72 ) | 0.001 | 0.429 | ( 0.26-0.71 ) | 0.001 |
Medications | | | | | | | | | | | | | | | |
Statins | | | | | | | 0.332 | ( 0.28-0.40 ) | <.001 | 0.338 | ( 0.28-0.41 ) | <.001 | 0.338 | ( 0.28-0.41 ) | <.001 |
PPI | | | | | | | 1.632 | ( 1.32-2.01 ) | <.001 | 1.563 | ( 1.27-1.93 ) | <.001 | 1.568 | ( 1.27-1.94 ) | <.001 |
CCB | | | | | | | 0.994 | ( 0.81-1.21 ) | 0.955 | 1.005 | ( 0.82-1.23 ) | 0.964 | 0.999 | ( 0.82-1.22 ) | 0.993 |
BB | | | | | | | 1.697 | ( 1.38-2.09 ) | <.001 | 1.615 | ( 1.31-1.99 ) | <.001 | 1.588 | ( 1.29-1.96 ) | <.001 |
ACEARB | | | | | | | 0.996 | ( 0.86-1.15 ) | 0.951 | 0.979 | ( 0.85-1.13 ) | 0.771 | 0.988 | ( 0.86-1.14 ) | 0.866 |
Diuretic | | | | | | | 1.187 | ( 0.95-1.48 ) | 0.124 | 1.182 | ( 0.95-1.47 ) | 0.138 | 1.164 | ( 0.93-1.45 ) | 0.178 |
METs (Categories) | | | | | | | | | | | | | | | |
METs 6-9 | | | | | | | | | | 0.871 | ( 0.75-1.01 ) | 0.073 | | | |
METs 10-11 | | | | | | | | | | 0.619 | ( 0.52-0.73 ) | <.001 | | | |
METs ≥ 12 | | | | | | | | | | 0.482 | ( 0.39-0.59 ) | <.001 | | | |
METs (Continuous) | | | | | | | | | | | | | 0.9 | ( 0.88-0.92 ) | <.001 |
Chronotropic incompetence | | | | | | | | | | 1.119 | ( 1.01-1.24 ) | 0.030 | 1.056 | ( 0.95-1.17 ) | 0.303 |
Duke Risk Score | | | | | | | | | | | | | | | |
High Risk | | | | | | | | | | 1.293 | ( 1.00-1.67 ) | 0.048 | 1.257 | ( 0.97-1.62 ) | 0.078 |
Moderate Risk | | | | | | | | | | 1.26 | ( 1.14-1.39 ) | <.001 | 1.21 | ( 1.10-1.33 ) | <.001 |
Participants with the highest exercise capacity (≥12METs) were younger (41±3 vs. 51±12 years, p<0.001); more often males (91% versus 43%, p<0.001); had lower mean body mass index (26±9 vs. 32±3kg/m2, p<0.001) and less likely to have hypertension or hyperlipidemia (26% vs. 44% and 25% vs. 29%, respectively, p<0.001) in comparison with the lowest exercise capacity group (METs<6). No apparent differences between CRF categories regarding prior coronary artery disease (CAD) and heart failure were observed. At peak exercise, the heart rate was higher among the highest achievers (166±2 vs. 131±25 bpm, p<0.001) as well as systolic blood pressure (161±21 vs. 153±33 mmHg, p<0.001). Moreover, The highest METs was associated with a lower risk of Duke Treadmill risk, while high Duke Treadmill risk was observed in those with lower METs achieved (for low Duke score: 77.2% vs. 19.4%) and (for high Duke score 0.7% vs. 2.7%), p<0.001.
Over a median follow-up period of 5.24 (IQR: 2.17-8.78) years, there were 2,280 (25.96%) new cases of diabetes. The unadjusted 5-year accumulative incidence rates of diabetes across categories of CRF (<6, 6<9, 10<11, and ≥12 METs) were 25.1%, 16.9%, 9.4%, and 6.2%, respectively. The cumulative rates of incidence of diabetes by the end of 15 years follow up period was as follows: 40.6%, 31.9%, 23.2%, and 17.3% for those who achieved <6, 6<9, 10<11, and ≥12 METs, respectively (Figure 1).
Examination of the crude association between categories of METs and risk for incident diabetes using a Kaplan-Meier cumulative incidence curve revealed a significant trend across categories of METs (log-rank <0.001) (Figure 2). In a multivariate Cox regression model adjusted for potential confounders, we found a 10% reduction in the risk of incident diabetes with higher METs achieved (HR, 0.90; 95% CI, 0.88–0.92; p<0.001) (model 4B) and risk for incident diabetes reduced per each METs achieved; 13% (6<9 METs), 38% (10<11), and 52% (≥12 METs) compare to lower METs (<6 METs) p<0.001. Adding the exercise capacity information (METs achieved, CI, and Duke Risk score) improves the overall predictability of the model (model 4A) over the baseline model (Table 2). We examined the predictive value of the finding of exercise stress testing, and we found superiority of METs groups in prediction incidence diabetes above the other models resulted in a significant reclassification of the study cohort and significant improvement of the area under the curve above the primary model (Table 2)
A non-linear inverse relationship was observed between baseline exercise capacity and risk of incident diabetes after adjusted for known cofounders. A gradual decline in the risk of development of diabetes was observed for every improvement of exercise capacity (METs increase) above 6 METs (Figure 3)