Background: Decreasing the ICU length of stay (LOS) after cardiac operations may increase ICU recidivism, obviating the benefit of early discharge.
Methods: From January 1, 1994 to January 1, 1998, there were 2,388 consecutive cardiac operations, from which 2,228 patients were discharged alive from the ICU and had sufficient information to determine their incidence of ICU return, the reasons for their return, their ICU LOS (initial and secondary LOS), and mortality.
Results: A decrease occurred in the initial ICU LOS from 1994 through 1997 (medians for 1994, 1995, 1996, and 1997, respectively: 31 h, 26.4 h, 24.5 h, and 24 h; and means, respectively: 69.4 +/- 139.8, 62.8 +/- 114.1, 52.5 +/- 104.0, and 56.2 +/- 103.4 h; p = 0.048). In association with this, however, ICU recidivism increased (as percentage of discharges, respectively: 3.9%, 4.2%, 6.1%, and 8.4%; p = 0.005). Inclusive of secondary ICU LOS, the total ICU LOS hours still decreased over the 4-year period. Most notably, the incidence of readmission increased with longer initial LOS (initial LOS quartiles from shortest to longest: 3.9%, 5.2%, 4.7%, and 9.2%; p = 0.0008). Predictors of ICU recidivism included preoperatively, a history of congestive heart failure, and a lower mean left ventricular ejection fraction (52.7 +/- 19.3% vs 49.8 +/- 21.5%; p = 0.0080); and, postoperatively, an increased mean weight gain (8.5 +/- 5.6 kg vs 10.3 +/- 4.7 kg; p = 0.040) and longer mean initial ventilator time (157 +/- 299 h vs 35 +/- 107 h; p = 0.038). The most common reason for readmission was pulmonary problems.
Conclusions: Over the years studied, the initial ICU LOS after cardiac operations has decreased in association with a significant increase in ICU recidivism. Importantly, however, patients readmitted to the ICU are those with longer initial LOSs. Decreased initial stay does not account for our increased ICU recidivism, and efforts to decrease ICU recidivism can focus on the patients with poor preoperative cardiac function and longer initial ICU stays.