Table 3.
Subgroup effects of TXA according to diabetes status in the ATACAS trial
| Patients without diabetes | Patients with diabetes | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| TXA (n = 1512) | Placebo (n = 1514) | RR (95% CI) | P | TXA (n = 799) | Placebo (n = 806) | RR (95% CI) | P | Interaction P | |
| Outcome | |||||||||
| Total mediastinal drainage, median (IQR), mL | 820 (550-1250) | 1090 (750-1560) | — | <.001 | 730 (500-1130) | 1033 (700-1490) | — | <.001 | .41 |
| Blood transfusion up to 24 h after surgery, n (%) | 441 (29.2) | 739 (48.8) | 0.60 (0.54-0.66) | <.001 | 281 (35.2) | 391 (48.5) | 0.72 (0.64-0.82) | <.001 | .012 |
| Reoperation for bleeding or tamponade, n (%) | 22 (1.5) | 47 (3.1) | 0.47 (0.28-0.77) | .003 | 10 (1.3) | 18 (2.2) | 0.56 (0.26-1.21) | .14 | .70 |
| Total red cell transfusion, n (%), no. of units | 2 (1-3) | 2 (2-4) | — | <.001 | 2 (1-4) | 2 (2-4) | — | .074 | .31 |
| Length of stays* | HR (95% CI) | HR (95% CI) | |||||||
| ICU stay, h | 26 (21-54) | 29 (22-66) | 1.09 (1.01-1.17) | .022 | 40 (22-72) | 41 (22-71) | 1.01 (0.91-1.11) | .87 | .22 |
| Hospital stay, d | 8 (6-12) | 8 (6-14) | 1.09 (1.02-1.17) | .028 | 90 (7-14) | 9 (7-14) | 0.97 (0.88-1.07) | .54 | .074 |
These data have not been published previously. Bold indicates significant results.
HR, hazard ratio; ICU, intensive care unit; IQR, interquartile range.
*
Length-of-stay data are not normally distributed; therefore, these were analyzed as times to events using Cox regression. The hazard ratio is the risk for discharge on any given day: an HR > 1 indicates a higher rate of discharge, meaning a greater likelihood of discharge on a given day compared with placebo (ie, a shorter length of stay). The dash indicates that the risk ratio cannot be determined