| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2007;83:1089-1095
© 2007 The Society of Thoracic Surgeons
a Department of Cardiovascular Anesthesia and Intensive Care, IRCCS Policlinico S. Donato, Milan, Italy
b Department of Internal Medicine, IRCCS Policlinico S. Donato, Milan, Italy
Accepted for publication October 2, 2006.
* Address correspondence to Dr Ranucci, Cardiovascular Anesthesia Department, IRCCS Policlinico S. Donato, Via Morandi 30, San Donato Milanese, 20097 Milan, Italy (Email: cardioanestesia{at}virgilio.it).
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and MethodsResultsCommentReferences |
|---|
Methods: A multivariable predictive model for early discharge from the intensive care unit was established using a stepwise forward logistic regression. The analysis was retrospectively conducted on 9120 consecutive patients undergoing cardiac surgical procedures at our institution.
Results: A total of 1874 patients were discharged early from the intensive care unit. Factors associated with early discharge were ejection fraction, lowest hematocrit on cardiopulmonary bypass, lowest temperature on cardiopulmonary bypass, and the presence of the goal-oriented strategy (odds ratio, 5.5; 95% confidence interval, 4.8 to 6.3). Factors associated with late discharge were age, preoperative serum creatinine value, unstable angina, congestive heart failure, redo operation, combined operation, and cardiopulmonary bypass duration. An extubation time of 4 hours after the arrival in the intensive care unit was associated with the peak rate of early discharge. Patients being early discharged according to the goal-oriented strategy did not demonstrate a different complication rate compared with patients treated with a standard strategy.
Conclusions: Early discharge from the intensive care unit depends on a combination of preoperative and intraoperative factors, but most of all on the presence of a goal-oriented strategy. A very early extubation is not required for an early discharge from the intensive care unit.
|
Introduction |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentReferences |
|---|
Early extubation is classically considered one of the main steps in a fast-track process [2], and different anesthesia protocols have been proposed to achieve this goal. In the 1990s, the common procedure for fast-track cardiac anesthesia included a limited intravenous opioid dose with the concomitant use of volatile anesthetics or propofol, or both [3, 4]. Subsequently, many authors proposed the use of a totally intravenous anesthesia [5–8]. Regardless of the anesthetic regimen, there is still an open debate about the optimal extubation time after cardiac operations, even if it is generally accepted that early extubation is one of the main determinants of early discharge from the ICU [1, 2, 4, 6, 9].
Many studies have addressed the success rate of a fast-track program in different conditions; however, nearly all of the studies (1) included only patients who had undergone isolated CABG operations [7, 9–12]; (2) were powered to compare different fast-track cardiac anesthesia techniques [5–7, 10]; (3) were powered to compare different postoperative analgesia managements [8, 12], and (4) with few exceptions [13], fewer than 600 patients were analyzed.
The primary end point of this study was to determine the predictors for an early discharge (the morning after the operation) from the ICU after cardiac operations, with specific respect to the application of a goal-oriented strategy. The role of extubation time in this process and the outcome profile in patients being discharged early from the ICU are secondary end points.
|
Patients and Methods |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentReferences |
|---|
Patients
A series of 9120 consecutive patients who had cardiac operations at our institution in the period May 1, 2000, through March 31, 2006, was retrospectively analyzed using the local computerized database. Data before the year 2000 were unavailable owing to the lack of a systematic data collection. All patients gave a written consent to the anonymous scientific treatment of their data. The local ethical committee waived the need for approval.
The database includes all the adult (>18 years) patients undergoing cardiac surgical procedures. Every kind of cardiac operation was included, excepted cardiac transplantation, which is not performed in our institution. The only exclusion criterion was death during the operation or within the first 24 hours after the operation (66 patients). A final series of 9054 adult patients was admitted to the statistical analysis.
Intraoperative Anesthetic Management, Monitoring, and Cardiopulmonary Bypass
Premedication included atropine sulphate (0.5 mg), promethazine (50 mg), and fentanyl (50 to 100 µg, according to the patient’s weight). Anesthesia was induced with an intravenous infusion of remifentanil at a starting dose of 0.5 µg · kg–1 · min–1 and a midazolam bolus of 0.2 mg/kg. Cisatracurium besylate (0.2 mg/kg) was administered to allow tracheal intubation. The anesthesia was maintained with a continuous infusion of remifentanil at a dose of 0.05 µg · kg–1 · min–1 to 1 µg · kg–1 · min–1 (titrated on the basis of the hemodynamic response) and midazolam at 0.1 mg · kg–1 · h–1.
All the operations were performed with cardiopulmonary bypass (CPB). CPB was conducted using closed or open circuits, standard or phosphorylcholine coated hollow-fiber oxygenators, and roller or centrifugal pumps, according to availability. Regardless of the circuit used, the priming volume was always minimized to 800 to 1000 mL.
Intraoperative standard monitoring included the usual invasive arterial blood pressure monitoring at the radial site plus a triple lumen central venous catheter for central venous pressure measurement. In patients with a reduced ejection fraction (<0.30), a pulmonary artery catheter was inserted, and in case of a valve procedure or combined operations, transesophageal echocardiographic monitoring was applied.
After CPB discontinuation and protamine administration, an intramuscular dose of 10 to 15 mg of morphine (according to the patient’s weight) was administered, and an intravenous infusion of ketorolac (30 mg) and tramadol (100 mg) was started and continued until the end of the operation. Simultaneously, the remifentanil infusion was decreased, to be stopped at the end of the operation.
At the skin closure, an elastomeric pump was connected to a peripheral venous catheter to guarantee a continuous infusion of ketorolac (0.05 mg · kg–1 · h–1) and tramadol (0.15 mg · kg–1 · h–1) throughout the first 24 hours of the postoperative course. The last cisatracurium besylate bolus was administered at the sternal closure. Patients were transferred to the ICU with the midazolam infusion at 0.05 mg · kg–1 · h–1.
Postoperative Management
All the patients were admitted to the adult postcardiac surgery ICU, where the nursing ratio is 1:2 patients. Postoperative analgesia was maintained with the elastomeric pump infusion and, if needed, with an intravenous dose of morphine (5 mg). The midazolam infusion was maintained during the first hour, considered as the observation period. Criteria for starting the extubation process were at least 2 hours from the arrival in ICU, temperature higher than 36°C, hemodynamically stable, chest tube drainage of less than 100 mL/h, and hourly urine output exceeding 0.5 mL/kg. Once this profile was established, the midazolam infusion was stopped if it was still being used.
Patients were considered eligible for extubation when they met the following criteria: responds to command, pulse oximetry oxygen saturation (SpO 2) exceeding 95% at a fraction of inspired oxygen (FIO 2) of less than 0.5; pH exceeding 7.3; PaCO 2 of less than 50 mm Hg, and adequate respiratory mechanics. Patients were considered eligible for ICU discharge when they met the following criteria: no mental confusion; SpO 2 exceeding 90% at a FIO 2 of less than 0.5; no uncontrolled arrhythmia; no active chest tube drainage; hourly urine output exceeding 0.5 mL/kg, and no vasopressors or inotropic agents of any kind.
Intensive Care Unit Discharge Strategy
From 2000 through 2003, no specific strategy for discharging patients the morning after the operation was applied.
At the beginning of 2004, a specific goal-oriented strategy, aimed to discharge patients the morning after the operation was applied. This strategy was based on a series of changes in the ward and ICU organization and protocols that are listed and compared with the standard strategy in Table 1.
|
|
Statistical Analysis
The primary outcome variable was considered the early discharge from the ICU. No distinction was made with respect to the eligible or actual day of discharge.
Preoperative and intraoperative variables were initially tested for association with early discharge using univariate analysis. Continuous variables were tested using the Student t test for unpaired data, and categoric variables were tested using a relative risk analysis with 95% confidence intervals (CI).
The variables associated with an early discharge from the ICU were included in a multivariable analysis (stepwise forward logistic regression) to identify the independent predictors of early discharge and their respective power within the model. To avoid overfitting of the model, the number of independent variables admitted to the model was limited to 1 for every 100 events; multicollinearity was controlled using a correlation matrix for the independent variables, considering a value of correlation <0.8 as the highest acceptable limit [14]. Tolerance statistics were applied to confirm the absence of multicollinearity, considering a tolerance value >0.3 for excluding multicollinearity [14]. Model performance was assessed using a receiver operating characteristic curve analysis for discrimination and the Hosmer-Lemeshow goodness-of-fit test for calibration.
Subsequent analyses applied to the association between early discharge from the ICU and mechanical ventilation duration, and to the outcome profile of early discharged patients included Pearson 
2 test, relative risk analysis, and Student t test for unpaired data.
Data in tables are reported as mean ± standard deviation of the mean or as count and percentage. A p < 0.05 was considered significant for all the statistical tests. Statistical calculations were performed using the SPSS 11.0 (SPSS Inc, Chicago, IL) statistical program.
|
Results |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentReferences |
|---|
|
All the above-mentioned variables were included into a multivariable stepwise forward logistic regression with the early discharge as the dependent variable. The multivariable model excluded factors that were significant at the univariate analysis, owing to the presence of intercorrelation between independent variables (the baseline hematocrit was intercorrelated with the lowest hematocrit on CPB; previous vascular surgery with age; chronic renal failure with serum creatinine value; COPD with age; diabetes on medication with age and serum creatinine value) or to the few number of events (cardiogenic shock).
The final multivariable model (Table 4) included the six continuous variables of age, ejection fraction, baseline hematocrit, serum creatinine level, lowest hematocrit on CPB, and lowest temperature on CPB, and the five categoric variables of unstable angina, congestive heart failure, redo operation, combined operation, and goal-oriented strategy. The presence of the goal-oriented strategy was associated with a successful early discharge from the ICU with an OR of 5.5 (95% CI, 4.8 to 6.3).
|
2 = 5.53, p = 0.7).
Extubation Time, Early Discharge From the ICU, and Outcome
In a separated post hoc analysis, we considered the extubation time in patients with or without a goal-oriented strategy. In this analysis, we considered only patients being extubated within 12 hours (2082 in the goal-oriented strategy group and 2783 in the other group) to avoid the inclusion of patients who needed prolonged intubation owing to postoperative problems. Mean extubation time (hours) was 9.6 ± 2.3 for the group without a goal-oriented strategy and 8.2 ± 2.9 for the group with a goal oriented strategy (p < 0.001), and the distribution of extubation times significantly differed between groups (Fig 1).
|
|
|
|
Comment |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentReferences |
|---|
From a practical standpoint, it is worthwhile to notice that the strategy applied does not include any changes in the intraoperative management and only minor changes in the ICU management, conversely concentrating resources and technical equipment in the ward. This enhanced level of patient care in the ward makes the ICU staff more confident in discharging patients early, after 2 hours from the extubation, and is not associated with an increased reentry rate or adverse events in the ward, even though this strategy was applied to a patient series with a more severe preoperative risk profile.
Within this context, the extubation time has a limited role. Patients treated with the goal-oriented strategy had significantly lower extubation times, but the difference in clinical terms is negligible (1.4 hours). Conversely, the success rate with respect to the extubation time is almost parallel in the groups with or without the goal-oriented strategy, being always significantly higher for each extubation time in the goal-oriented strategy group. The peak of success in both groups was obtained at an extubation time of 4 hours: before and after this timing, the success rate declines.
It seems reasonable that the reason for a lower rate of early discharge from the ICU in patients being very early extubated could be a higher rate of postextubation complications, leading the ICU staff to postpone the discharge. Our analysis confirms this hypothesis, revealing a significant higher rate of postextubation complications in patients extubated after 2 or 3 hours versus patients extubated after 4 hours.
The interpretation of this data is speculative, because our work is a retrospective analysis and we have no explanation for the higher rate of surgical revisions in patients being very early extubated. Conversely, there is a combination of events (need for inotropic support, decreased urine output, ventricular arrhythmias) that may be interpreted as a moderate low cardiac output state not present before the extubation but possibly triggered by the increased cardiac work load to sustain the spontaneous ventilation.
Most importantly, the presence of a higher rate of transient neurologic dysfunctions in these patients may explain the delayed discharge from the ICU and could be related to our intraoperative anesthetic protocol, which is based on midazolam as the hypnotic agent. This management has been already used in fast-track cardiac anesthesia [5, 15] with successful early extubation and discharge from the ICU. One study [5], however, demonstrated that propofol-based anesthesia was associated with a shorter intubation time, whereas midazolam-based anesthesia in elderly patients was associated with a higher rate of mental function impairment. Therefore, a different anesthetic management more oriented to early extubation could induce different results, and we cannot exclude that by using shorter-acting intravenous hypnotic drugs, a different optimal timing for extubation could be established.
Our study confirms the role of other covariates within a fast-track program in cardiac surgery. In a study addressing the role of preoperative risk factors in determining the success of a fast-track program, Cheng and associates [10] demonstrated that the success rate decreased as a function of preoperative risk score and age. Brucek and associates [8], in a model of ultra-fast track extubation, found that surgery duration and preoperative lung dysfunction were significant predictors of failure in extubation immediately after surgery.
In our model, both the patient risk profile and operation-related factors significantly contribute in determining the discharge time from the ICU. In addition to the already mentioned variables, we found that severe hemodilution was an independent predictor of late discharge from the ICU. Many articles have recently addressed this topic, demonstrating that severe hemodilution during CPB is associated with a number of postoperative complications, mainly related to renal dysfunction [16–19], but also to many other organ dysfunctions and death [20, 21]. We recently demonstrated that the reason for this association is a low oxygen delivery during CPB [22]; however, the role of hemodilution-triggered homologous blood transfusions must be mentioned in the determinism of hemodilution-related complications [23].
The main strength of our study relies on the number of patients and variables considered. We are, however, aware of the limitations of our approach. The study is not prospective, and the patients have not been randomized for the goal-oriented strategy. Moreover, our model is not validated in a second series of patients.
In conclusion, a successful program of early discharge from the ICU depends on the (1) preoperative risk profile of the patients, (2) type of operation, temperature, and duration of CPB, (3) application of strategies aimed to reduce hemodilution during CPB; and, the most important point, (4) on the application of a goal-oriented strategy with increased resources in the ward and a short postextubation observational period in the ICU. The extubation time is correlated to success only after the fourth postoperative hour. In our model, a very early extubation is not a guarantee of higher success rate in early discharging patients from the ICU.
|
References |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentReferences |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
