Table 1.
Characteristic | Total (N = 20,741) | Metformin users (N = 1905) | Metformin nonusers (N = 18,836) | SMD* |
---|---|---|---|---|
Age, median [IQR], years | 69 [58–77] | 70 [63–76] | 69 [58–77] | 0.24 |
Male sex | 11,697 (56) | 1193 (63) | 10,504 (56) | 0.14 |
Charlson Comorbidity Index | 0.52 | |||
0 | 6894 (33) | 305 (16) | 6589 (35) | |
1 or 2 | 8147 (39) | 737 (39) | 7410 (39) | |
3 or higher | 5700 (27) | 863 (45) | 4837 (26) | |
Diabetes mellitus | 4594 (22) | 1903 (100) | 2691 (14) | 3.45 |
Sulfonylureas | 476 (2) | 250 (13) | 226 (1) | 0.48 |
Insulin | 1473 (7) | 443 (23) | 1030 (5) | 0.52 |
Other antihyperglycemic agents | 445 (2) | 259 (14) | 186 (1) | 0.50 |
Preadmission eGFR, median [IQR], ml/min/1.73 m2 | 80 [58–95] | 77 [58–92] | 80 [58–95] | 0.07 |
≥ 60 ml/min/1.73 m2 | 12,892 (62) | 1345 (71) | 11,547 (61) | 0.48 |
45–60 ml/min/1.73 m2 | 2189 (11) | 278 (15) | 1911 (10) | |
≤ 45 ml/min/1.73 m2 | 2442 (12) | 221 (12) | 2221 (12) | |
Missing | 3218 (16) | 61 (3) | 3157 (17) | |
ICU admission type | 0.20 | |||
Medical | 9942 (48) | 1019 (53) | 8923 (47) | |
Emergency surgical | 6344 (31) | 456 (24) | 5888 (31) | |
Elective surgical | 3149 (15) | 345 (18) | 2804 (15) | |
Missing | 1306 (6) | 85 (4) | 1221 (6) | |
Time from hospital admission to ICU admission, median [IQR], h† | 5.1 [0.0–30.5] | 5.3 [0.0–29.4] | 5.1 [0.0–30.7] | 0.06 |
SAPS-II score, median [IQR] | 40 [30–52] | 42 [31–53] | 40 [30–52] | 0.09 |
Missing | 11,456 (55) | 1024 (54) | 10,432 (55) | |
Mechanical ventilation | 9305 (45) | 815 (43) | 8490 (45) | 0.05 |
Inotropes or vasopressors | 8943 (43) | 854 (45) | 8089 (43) | 0.04 |
Renal replacement therapy | 1257 (6) | 152 (8) | 1105 (6) | 0.09 |
AKI stage within 24 h | 0.47 | |||
No AKI | 10,597 (51) | 982 (52) | 9615 (51) | |
1 | 3584 (17) | 388 (20) | 3196 (17) | |
2 | 1436 (7) | 198 (10) | 1238 (7) | |
3 | 1877 (9) | 261 (14) | 1616 (9) | |
Missing | 3247 (16) | 76 (4) | 3171 (17) | |
30-day mortality | 4367 (21) | 346 (18) | 4021 (21) | 0.08 |
Data are expressed as no. (%) or median [IQR]
*As general guidance, it is suggested that effect sizes are likely to be “small” when an SMD approximates 0.2, likely to be “medium” when an SMD is 0.5, and “large” when an SMD is higher than 0.8
†In total, data are missing for 19 (0.1%) patients
SMD standardized mean difference, eGFR estimated glomerular filtration rate, ICU intensive care unit, SAPS-II Simple Acute Physiology Score II, AKI acute kidney injury