Background
Methods
Design, setting, participants, data sources and statistics
Average standardized urinary output calculation
Modeling 90-day mortality
Handling missing covariate data
Ethics
Results
Population
All patients (n = 15,550) | Training set (n = 12,440) | Validation set (n = 3110) | P value | |
---|---|---|---|---|
Demographics | ||||
Male, n (%) | 10,283 (66.1) | 8221 (66.1) | 2062 (66.3) | 0.84 |
Female, n (%) | ||||
Age at ICU admission, median (IQR), years | 65.0 (53.0, 75.0) | 65.0 (53.0, 75.0) | 65.0 (53.0, 75.0) | 0.50 |
Pre-admission body weight, median (IQR), kg | 75.0 (65.0, 87.0) | 75.0 (65.0, 87.0) | 75.0 (65.0, 87.0) | 0.37 |
Retrieved from medical records | 12,362 (79.5) | 9897 (79.6) | 2465 (79.3) | 0.84 |
Imputed (multiple imputations) | 3188 (20.5) | 2543 (20.4) | 645 (20.7) | |
Baseline creatinine, median (IQR), μmol/L | 69.0 (54.0, 90.0) | 69.0 (53.0, 90.0) | 70.0 (54.0, 91.0) | 0.07 |
Comorbidities | ||||
Charlson score, median (IQR) | 4.0 (2.0, 6.0) | 4.0 (2.0, 6.0) | 4.0 (2.0, 6.0) | 0.19 |
Chronic kidney disease, n (%) | 1610 (10.4) | 1255 (10.1) | 355 (11.5) | 0.03 |
Hypertension, n (%) | 7286 (47.0) | 5804 (46.8) | 1482 (47.8) | 0.33 |
Diabetes, n (%) | 2926 (18.9) | 2307 (18.6) | 619 (20.0) | 0.09 |
Heart failure, n (%) | 3454 (22.3) | 2747 (22.2) | 707 (22.8) | 0.45 |
Chronic obstructive pulmonary disease, n (%) | 1806 (11.7) | 1455 (11.7) | 351 (11.3) | 0.54 |
Myocardial infarction, n (%) | 5430 (35.0) | 4357 (35.1) | 1073 (34.6) | 0.60 |
Chronic liver disease, n (%) | 1188 (7.7) | 957 (7.7) | 231 (7.5) | 0.65 |
Cancer, n (%) | 2900 (18.7) | 2312 (18.6) | 588 (19.0) | 0.70 |
ICU admission characteristics | ||||
Type of admission | 0.63 | |||
Medical | 6534 (42.8) | 5248 (43.0) | 1286 (42.1) | |
Surgical scheduled | 4374 (28.7) | 3487 (28.6) | 887 (29.0) | |
Surgical unscheduled | 4345 (28.5) | 3461 (28.4) | 884 (28.9) | |
Main ICU diagnosis | 0.08 | |||
Cardiovascular surgery | 3505 (23.1) | 2812 (23.1) | 693 (22.7) | |
Other heart disease | 1963 (12.9) | 1602 (13.2) | 361 (11.8) | |
Septic shock | 1762 (11.6) | 1387 (11.4) | 375 (12.3) | |
Neurologic | 1614 (10.6) | 1293 (10.6) | 321 (10.5) | |
Trauma | 1236 (8.1) | 958 (7.9) | 278 (9.1) | |
Acute respiratory insufficiency | 1313 (8.6) | 1061 (8.7) | 252 (8.3) | |
Cardiopulmonary arrest | 897 (5.9) | 731 (6.0) | 166 (5.4) | |
Other | 2908 (19.1) | 2306 (19.0) | 602 (19.8) | |
sCr at ICU admission, median (IQR), μmol/L | 83.0 (65.0, 112.0) | 83.0 (65.0, 112.0) | 83.0 (65.0, 114.0) | 0.32 |
Mechanical ventilation within 24 h of ICU admission | 9152 (58.9) | 7346 (59.1) | 1806 (58.1) | 0.33 |
Noradrenaline within 24 h of ICU admission | 10,702 (68.8) | 8547 (68.7) | 2155 (69.3) | 0.54 |
Modified SAPS II scorea, median (IQR) | 37.0 (28.0, 48.0) | 37.0 (28.0, 48.0) | 37.0 (28.0, 48.0) | 0.97 |
Relationship between UO and 90-day mortality
Proposal for a new definition of oliguria
Confirmatory analyses
Time window | Suggested oliguria threshold (ml/kg/h) | All validation patients | Validation patients with oliguria | ||
---|---|---|---|---|---|
Base model | Full model | Base model | Full model | ||
3 h | < 0.2 | 85.3 | 86.1 | 82.1 | 83.0 |
6 h | < 0.2 | 85.3 | 85.9 | 80.0 | 81.5 |
12 h | < 0.3 | 85.3 | 86.1 | 82.5 | 83.7 |
24 h | < 0.3 | 85.3 | 86.1 | 79.8 | 82.0 |