Impact of Immediate Versus Delayed Tracheal Extubation on Length of ICU Stay of Cardiac Surgical Patients

Background The question whether early extubation leads to shorter post-operative length of ICU stay has rapidly gained importance over the past few decades considering the correlation between length of hospital stay and costs. Although some investigations did not find a difference between post-operative length of stay in early and late extubated patients, most observational studies and small randomized clinical trials showed significant benefits for early extubation.

Ultra-fast track anesthesia (UFTA) aims at immediate extubation of cardiac surgical patients at the end of the operation. There are few contraindications to the adoption of early extubation protocols. Generally most cardiac surgical patients, presenting for either elective or emergent surgery, have adequate ventilatory function. If patients were not intubated and ventilated preoperatively, they are not likely to require prolonged mechanical ventilation.

Goals of the study:

The primary goal of this study was to assess the impact of adopting UFTA on the length of ICU stay of patients undergoing elective cardiac surgery. The incidence of complications during their ICU stay was recorded as secondary outcome.

Methods From February, 2011, through October, 2013, 52 consecutive patients underwent open heart surgeries and were managed by the same anesthesiologists. Adult patients (>18yrs) scheduled for primary elective cardiac operations were included in the study.

Excluded from the study were patients undergoing emergency/redo operations, those already intubated preoperatively, patients with preoperative uncontrolled diabetes (HbA1C >5.9 mg/dl), cardiogenic shock, poor left ventricular function (ejection fraction (EF) < 45%), severe pulmonary hypertension (pulmonary artery systolic pressure (PASP) > 55 mmHg), severe renal impairment (creatinine clearance (CC) < 50 ml.min-1) or regular dialysis, and patients deliberately kept intubated for hemodynamic instability, and/or concerns of postoperative bleeding.

Sample size calculation Based on two-tailed α error probability of 0.05 and β error probability of 0.2 (power of 80%), a total sample size of 52 patients equally allocated into two groups was required to detect a presumed minimum clinically significant effect size d of 0.8 in length of ICU stay. Statistical power calculations was performed using computer program G*Power 3 for Windows. (Franz Faul, Universität Kiel, Germany).

Randomization participants were randomly allocated to either the UFTA group or conventional group using computer generated random numbers and opaque sealed envelopes.

Preoperative evaluation Routine preoperative investigations included complete blood picture, bleeding profile, liver and kidney functions tests, fasting blood glucose level, chest X-ray, twelve-lead electrocardiogram (ECG), transthoracic echocardiography, and coronary angiography if indicated. Logistic EuroSCORE II (European System for Cardiac Operative Risk Evaluation) was calculated to all participants.

Patients in both groups received their usual medications, excluding oral hypoglycemic and ACEI and ARB's, up to the time of surgery. Premedication with intravenous midazolam, 1-3 mg was given to patients upon arrival to the anesthesia bay 20-30 min before admission to the operating room.

Intraoperative monitoring was achieved using standard 5-lead ECG, invasive arterial blood pressure, pulse oximeter, and nasopharyngeal temperature probe, and transesophageal echocardiography.

Conduct of anesthesia Induction of anesthesia was performed with midazolam 0.03-0.05 mg/kg, fentanyl 1-2 mcg/kg, and propofol 1-2 mg/kg. Orotracheal intubation was facilitated by atracurium 0.5 mg/kg. Anesthesia was maintained with sevoflurane titrated to an expired MAC between 1-1.5, and a continuous infusion of morphine at 10 to 20 mcg/kg/h. Additional doses of fentanyl and atracurium were given as needed.

Standard median sternotomy was performed for all but one patient, who had mitral valve replacement through lateral thoracotomy. One patient had off-pump coronary artery bypass grafting.

Post bypass fluid balance was maintained with warmed lactated/acetated Ringer's to maintain patient's nasopharyngeal temperature above 35.5°C. Hemodynamic instability (defined as systolic blood pressure less than 90 mmHg and/or mean arterial pressure less than 60 mmHg), if encountered, was corrected with intravenous bolus doses of fluids, table positioning, and/or norepinephrine boluses 4-8 mcg per dose. While closing the subcutaneous tissue, patients received intravenous paracetamol 15 mg/kg.

At the completion of surgery, for UFTA group, the inhalational anesthetic concentration was reduced gradually to 0.4 expired MAC. If patients met the preset extubation criteria (SpO2 > 95% with FiO2 < 0.6, ETCO2 < 50 mmHg, spontaneous respiratory rate < 24/min), and train of four (TOF) > 90%, residual muscle relaxation was antagonized with neostigmine and atropine 0.05 mg/kg and 0.02 mg/kg respectively. The trachea then was extubated and oxygen delivered via open face mask 10 L/min during the transfer from the operating theatre to the ICU. Patients were questioned about recall of intraoperative events 12 and 24 hours after surgery.

For patients in the conventional group, patients were transferred to the ICU intubated and sedated with propofol infusion 50 - 70 mcg/kg/min and morphine 10 - 20 mcg/kg/hr and mechanical ventilation continued.

As per the local ICU protocol and after discontinuation of sedation, patients were judged eligible for extubation if they were awake and able to respond comprehensively to simple verbal commands, provided that they were hemodynamically stable and had normal ventilatory mechanics, acid-base status, PaO2, and PaCO2 at an inspired FiO2 of 0.4.

Postoperative pain control Postoperative pain management was multimodal. Each patient received a combination of the following treatments: continuous intravenous morphine 10-20 mcg/kg/hr, intravenous morphine on demand 0.05mg/kg (if pain scored 4 or more on a 0 to 10 pain scale with lockout time of 10 min), and intravenous paracetamol 15 mg/kg every 6 hours. Pain was assessed with Visual Analogue Scale (VAS), every 6 hours for the first 24 hours of their ICU stay, by the intensivist in charge.

Data collection Preoperative data included (1) demographic data (2) preoperative diagnosis and comorbidities (3) preoperative investigations and (4) logistic Euro SCORE II.

Intraoperative hemodynamic parameters, urine output, arterial oxygen saturation (SpO2), arterial blood gas analysis (ABG), standard 5 lead ECG analysis with ST segment analysis, total operative time, cardiopulmonary bypass time, aortic cross clamp time, and total opioid consumption (fentanyl \& morphine) were continuously recorded.

Postoperative data collected included (1) time to extubation (in hours), (2) the total length of ICU stay (in hours), (3) pain score (assessed every 6 hours), (4) total opioid requirements, (5) incidence of complications, including reintubation, excessive mediastinal bleeding, need for reopening, and postoperative nausea and vomiting, (6) incidence of postoperative ischemia, detected by ECG changes (ST segment elevation or depression), creatinine kinase (CK), creatinine kinase - myocardial band (CK-MB) and Troponin I, (7) incidence of postoperative renal impairment, detected by diminished urine output < 0.5 ml/kg/hr and/or rising creatinine level, and (8) incidence of respiratory dysfunction detected by ABG changes and decreased SpO2.

Statistical Analysis Obtained data was presented as mean (SD), median (interquartile ranges), or numbers and percentages as appropriate. Nominal variables were analyzed using Chi-squared (χ²) test or Fischer exact test as appropriate. Continuous variables were analyzed using unpaired Student's t-test or univariate two-group repeated measures analysis of variance (ANOVA) with post hoc Dunnett's test as appropriate. Nominal and non-normally distributed variables were analyzed using Mann-Whitney U test. Statistical calculations were performed using Microsoft® Office Excel 2010 and SPSS 16.0 for windows. P value < 0.05 was considered statistically significant.