Table 1.
Baseline patient characteristics, by serum potassium, before and after propensity score matching
| Before matching |
After matching |
|||||
|---|---|---|---|---|---|---|
| N (%) or mean (±SD) | ≥4 mEq/L (N = 1189) | <4 mEq/L (N = 5656) | P | ≥4 mEq/L (N = 1187) | <4 mEq/L (N = 1187) | P |
| Age (years) | 63 (±11) | 64 (±11) | 0.035 | 63.5 (±11.1) | 63.4 (±11.3) | 0.859 |
| Age ≥65 years | 590 (50%) | 3001 (53%) | 0.041 | 604(51%) | 589 (50%) | 0.566 |
| Female | 363(31%) | 1318(23%) | <0.0001 | 367 (31%) | 362 (31%) | 0.859 |
| Non-white | 198 (17%) | 745 (13%) | 0.002 | 197 (17%) | 197 (17%) | 1.000 |
| Body mass index, kg/m2 | 27 (±6) | 27 (±5) | 0.645 | 27.3 (±5.5) | 27.3 (±5.6) | 0.822 |
| Duration of HF (months) | 30 (±37) | 29 (±36) | 0.333 | 28.8 (±36.0) | 29.8 (±36.3) | 0.477 |
| Primary cause of HF | ||||||
| Ischemic | 737 (62%) | 3981 (70%) | 750 (63%) | 737 (62%) | ||
| Hypertensive | 183 (15%) | 527 (9%) | <0.0001 | 153 (13%) | 183 (15%) | 0.067 |
| Idiopathic | 177 (15%) | 796 (14%) | 209 (18%) | 176 (15%) | ||
| Others | 92 (8%) | 352 (6%) | 75 (6%) | 91 (8%) | ||
| Prior myocardial infarction | 682 (57%) | 3620 (64%) | <0.0001 | 678 (57%) | 682 (58%) | 0.901 |
| Current angina pectoris | 316 (27%) | 1539 (27%) | 0.667 | 319 (30%) | 316 (27%) | 0.926 |
| Hypertension | 649 (55%) | 2604 (46%) | <0.0001 | 652 (55%) | 647 (55%) | 0.869 |
| Diabetes mellitus | 315 (27%) | 1639 (29%) | 0.090 | 302 (25%) | 315 (27%) | 0.574 |
| Chronic kidney disease | 527 (44%) | 2589 (46%) | 0.370 | 537 (45%) | 525 (44%) | 0.650 |
| Medications | ||||||
| Pre-trial digoxin use | 535 (45%) | 2346 (42%) | 0.026 | 533 (45%) | 533 (45%) | 1.000 |
| Trial use of digoxin | 601 (51%) | 2822 (50%) | 0.702 | 612 (52%) | 599 (51%) | 0.622 |
| ACE inhibitors | 1096 (92%) | 5283 (93%) | 0.129 | 1094 (92%) | 1095 (92%) | 1.000 |
| Hydralazine & nitrates | 26 (2%) | 80 (1%) | 0.053 | 26 (2%) | 25 (2%) | 1.000 |
| Diuretics | 987 (83%) | 4343 (77%) | <0.0001 | 986 (83%) | 985 (83%) | 1.000 |
| Potassium-sparing diuretics | 95 (8%) | 388 (7%) | 0.171 | 110 (9%) | 94 (8%) | 0.272 |
| Potassium supplement | 487 (41%) | 1638 (29%) | <0.0001 | 465 (39%) | 485 (41%) | 0.426 |
| Symptoms and signs of HF | ||||||
| Dyspnea at rest | 286 (24%) | 1223 (22%) | 0.070 | 292 (25%) | 286 (24%) | 0.811 |
| Dyspnea on exertion | 889 (75%) | 4264 (75%) | 0.657 | 894 (75%) | 887 (75%) | 0.776 |
| Limitation of activity | 892 (75%) | 4305 (76%) | 0.433 | 918 (77%) | 890 (75%) | 0.193 |
| Jugular venous distension | 172 (15%) | 690 (12%) | 0.034 | 183 (15%) | 171 (14%) | 0.526 |
| Third heart sound | 309 (26%) | 1312 (23%) | 0.043 | 328 (28%) | 308 (26%) | 0.379 |
| Pulmonary râles | 209 (18%) | 892 (16%) | 0.129 | 205 (17%) | 208 (18%) | 0.914 |
| Lower extremity edema | 292 (25%) | 1161 (21%) | 0.002 | 299 (25%) | 290 (24%) | 0.704 |
| NYHA functional class, % | ||||||
| Class I | 164 (14%) | 809 (14%) | 171 (14%) | 164 (14%) | ||
| Class II | 630 (53%) | 3076 (54%) | 0.125 | 624 (53%) | 629 (53%) | 0.899 |
| Class III | 362 (30%) | 1670 (30%) | 364 (31%) | 361 (30%) | ||
| Class IV | 33 (3%) | 101 (2%) | 28 (2%) | 33 (3%) | ||
| Heart rate (/minute), | 80 (±13) | 78 (±13) | <0.0001 | 80 (±13) | 80 (±13) | 0.783 |
| Blood pressure (mm Hg) | ||||||
| Systolic | 129 (±22) | 127 (±20) | 0.038 | 128 (±21) | 129 (±21) | 0.620 |
| Diastolic | 76 (±11) | 75 (±11) | 0.001 | 76 (±12) | 76 (±12) | 0.363 |
| Chest radiograph findings | ||||||
| Pulmonary congestion | 182 (15%) | 753 (13%) | 0.070 | 194 (16%) | 182 (15%) | 0.536 |
| Cardiothoracic ratio >0.5 | 776 (65%) | 3366 (60%) | <0.0001 | 758 (64%) | 774 (65%) | 0.520 |
| Serum creatinine (mg/dL) | 1.27 (±0.4) | 1.3 (±0.4) | 0.047 | 1.3 (±0.4) | 1.3 (±0.4) | 0.568 |
| Estimated glomerular filtration rate, ml/min per 1.73 m2 | 64 (±21) | 63 (±21) | 0.177 | 63 (±20) | 64 (±21) | 0.357 |
| Ejection fraction (%) | 32 (±13) | 32 (±12) | 0.346 | 33 (±13) | 32 (±13) | 0.466 |
| Ejection fraction >45% | 162 (14%) | 704 (12%) | 0.270 | 154 (13%) | 161(51%) | 0.717 |
*
A random sample of 1187 patients were selected from 5656 patients with normal potassium and were paired with 1187 patients with low serum potassium in the matched cohort. This was done to assemble a pre-match cohort of the same size (n=2374) of that of the post-match cohort (n=2374) to avoid artificial inflation of the significance of intergroup differences in baseline covariates due to larger pre-match sample size (n=6845)