SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Small single-institutional studies performed prior to the introduction of organ allocation using the Model for End-Stage Liver Disease (MELD) suggest that early airway extubation of liver transplant recipients is a safe practice. We designed a multicenter study to examine adverse events associated with early extubation in patients selected for liver transplantation using MELD score. A total of 7 institutions extubated all patients meeting study criteria and reported adverse events that occurred within 72 hours following surgery. Adverse events were uncommon: occurring in only 7.7% of 391 patients studied. Most adverse events were pulmonary or surgically related. Pulmonary complications were usually minor, requiring only an increase in ambient oxygen concentration. The majority of surgical adverse events required additional surgery. Analysis of a limited set of perioperative variables suggest that blood transfusions and technical factors were associated with an increased risk of adverse events. In conclusion, while early extubation appears to be safe under specified circumstances, there are performance differences between institutions that remain to be explained. Liver Transpl 13:1557–1563, 2007. © 2007 AASLD.

There is an increasing trend for physicians to extubate patients immediately following major surgery to facilitate early discharge from or avoid admission to the intensive care unit.1 This practice is primarily driven by the need to reduce cost and improve resource utilization by eliminating unnecessary medical interventions and care.2, 3 Large and multiinstitutional randomized controlled studies confirm that extubation of patients within 8 hours after cardiac surgery reduces cost by decreasing intensive care unit and hospital length of stay.1–4 The failure of early extubation to influence morbidity and mortality in these study groups provides robust evidence for the broader adoption of early extubation in other patient populations as a cost-effective practice that does not compromise patient outcome.4

Successful early extubation has been reported in liver transplantation patients.5–8 In our previous study, investigators at the University of Colorado found that few complications occurred in patients who were extubated early using standard extubation criteria.5 However, the safety of early extubation across institutional boundaries could not be evaluated because the sample size was small, extubated patients had a very low rate of adverse outcomes and variation in clinical practice, and patient characteristics were limited. Questions remained as to whether the same frequency and type of adverse outcomes would occur in larger study cohorts and if other institutions would report a similar or different experience.

Other studies also report the use of early extubation in liver transplant recipients.7, 8 However, the criteria used to select patients for extubation differed between study sites. The resulting unique character of each study cohort prevents a direct comparison of findings and makes the summation of information into a larger database difficult. Further, most of these studies preceded the introduction of the Model for End-Stage Liver Disease (MELD) as an allocation tool. Because MELD prioritizes patients using different criteria than the earlier United Network for Organ Sharing allocation tool,9 it is uncertain if the outcomes from previous studies on early extubation can be expected in patients selected for transplantation using MELD scores.

Physicians have insufficient information about the type and frequency of adverse outcomes in transplant recipients who are extubated early. Consequently, they cannot identify candidates at greater risk of experiencing complications that necessitate a reinstitution of mechanical ventilation, an event that would deem the practice of early extubation as unsafe. Thus the purpose of this study was as follows: (1) to provide descriptive information on the type, severity, and frequency of adverse outcomes from a larger population of liver transplant recipients who shared the same extubation criteria; (2) to examine the similarities and differences in adverse outcomes between institutions; and (3) to determine if these adverse outcomes could be predicted by patient or operative characteristics.

PATIENTS AND METHODS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

General Design and Patients

We enlisted 7 institutions to prospectively collect data that described adverse outcomes in liver transplantation patients who were extubated early. Institutional review board approval was obtained at each study site. The 5 institutions in the United States were the University of Colorado, University of California-San Francisco, University of Pennsylvania, University of Washington, and Vanderbilt University. The 2 participating institutions in Europe were St James University, Leeds, UK, and the University of Pisa, Pisa, Italy. To minimize the well-recognized effect that learning new skills has on the rate and type of complications associated with early extubation,6, 10 we only enrolled institutions that had at least 3 yr of experience using early extubation following liver transplantation surgery as part of their routine clinical care.

The data collection period was limited to 2 yr to minimize changes in clinical practice that could influence the rate of adverse outcomes in the study population. No one has shown that all liver transplant recipients benefit from routine postoperative ventilation but some think that the greater severity of illness and added burden of incomplete donor organ function may place transplant recipients at risk of respiratory failure for an extended period of time. We therefore recorded all adverse events that occurred within 72 hours following surgery in contrast to the more common 24 hours used in most perioperative outcome studies.11 We reasoned that it is the frequency and nature of all perioperative adverse events that place patients at risk of respiratory failure, potentially making early extubation unsafe. Further, it is often impossible to attribute an adverse event exclusively to either specific surgical or anesthetic clinical care choices. Thus, investigators included all adverse events regardless of their relationship to extubation.

The study consortium defined early extubation as removal of the endotracheal tube immediately following surgery. No more than 1 hour elapsed between the conclusion of surgery and extubation. All members of the transplant team at each institution agreed to extubate every patient who met a single set of published criteria previously shown to be conservative in liver transplantation patients.6 Criteria were as follows: patients who followed verbal commands, a positive gag reflex, tidal volumes >6 mL/kg, respiratory rate <20, oxygen saturation >96% while breathing spontaneously (FiO2 ≤50%), normocarbia judged by end tidal carbon dioxide, reversal of neuromuscular blockade judged by peripheral nerve stimulator and clinical assessment, core body temperature between 36.5°C and 37.5°C, and hemodynamic stability with no vasoactive drug support.

To minimize excessive variability, we excluded patients with coronary artery disease, pulmonary vascular disease, retransplant operations, multiorgan transplants, and patients with Grade 4 encephalopathy. Entry criteria did not specify the use of any surgical approach and the choice of technique (excision of the intrahepatic inferior vena cava vs. piggyback) was determined by the surgical team. The study sites used a similar anesthesia protocol. All patients were induced using propofol (1.5-2.0 mg/kg) and succinylcholine (1 mg/kg) or Vecuronium (0.1 mg/kg) in a rapid sequence and subsequently paralyzed with vecuronioum or Cis-atracurium to maintain at least 1 of the train of 4 twitches. Anesthesia was maintained using either inhaled desflurane or isoflurane and fentanyl (15 μg/kg or less). No other sedative-hypnotic agents were administered. Neuromuscular relaxation was reversed using neostigmine (50 μg/kg) and glycopyrrolate (0.01-0.02 μg/kg). Following extubation, patients were transferred to a care site that had continuous monitoring and a reduced patient to nursing ratio.

Study Variables

Our analysis evaluated 11 preoperative and 6 intraoperative variables shown in Tables 1 and 2, respectively. Body mass index (kg/m2), serum creatinine, international normalized ratio, Child-Turcotte-Pugh score, MELD score, whether the patient received a living donation, and the severity of ascites and encephalopathy were determined on the day of surgery. In the dichotomous schema used to categorize ascites, well controlled was defined as less than 1 L of fluid in the peritoneum at the start of surgery while greater than 1 L of fluid was considered poorly controlled. Encephalopathy was graded from 0 to 3 using standard criteria.12

Table 1. The Distribution of Preoperative Variables for the 391 Study Subjects
 Site 1Site 2Site 3Site 4Site 5Site 6Site 7P-value
  1. Medians followed by (means ± SD) are given for continuous variables. Counts are provided for categorical variables. P-value is for difference in populations. Diagnosis is classified as (1) Alcohol, (2) Viral, (3) Cholestatic, (4) Combined, (5) Other. All patients with grade 4 encephalopathy were excluded from study.

  2. Abbreviations: INR, international normalized ratio; BMI, body mass index; Encephl, encephalopathy.

Study Subjects (N)76771671811339 
Total Transplants (N)232128112192253169184 
Age (yrs)53.5 (52.7 ± 10.4)48 (46.7 ± 14.7)52 (50.4 ± 9.5)50 (47.9 ± 12.6)50 (50.5 ± 11.4)48 (47.9 ± 10.5)53 (52.4 ± 8.6)0.0361
Male (%)795783618974690.37
White (%)6210097856776740.0001
Diagnosis 1,2,3,4,55/51 0/13/70/3 1/2/15/28 2/30/67/16 28/8/82/2/2/11/112/37/17/18/294/17/7/2/90.0001
BMI (kg/m2)26.2 (27.4 ± 5.4)30.4 (30.4 ± 8.0)24.2 (24.0 ± 3.0)24.6 (25.0 ± 4.0)25 (27.1 ± 6.3)25.7 (26.3 ± 3.8)27.1 (28.7 ± 4.9)0.0001
Child-Pugh score9 (8.7 ± 2.1)8 (8.1 ± 1.3)8 (8.2 ± 2.2)8 (8.0 ± 1.9)8 (8.4 ± 2.5)10 (10.0 ± 2.6)10 (9.7 ± 1.7)0.0001
MELD score13.5 (15.7 ± 7.7)19 (18.6 ± 7.0)11 (11.9 ± 4.9)13 (13.5 ± 4.5)20.5 (19.8 ± 6.3)22 (21.6 ± 6.6)18 (18.5 ± 4.2)0.0001
% Ascites > 1L844386605650770.0001
Encephal 0,1,2,320/29/21/64/2/1/014/39/15/318/27 17/52/12 3/130/34 30/1922/10 7/00.0001
Creatinine (mg/dL)0.9 (1.06 ± 0.41)1 (1.23 ± 0.68)0.8 (0.94 ± 0.72)0.89 (0.91 ± 0.24)1.25 (1.28 ± 0.47)1.1 (1.30 ± 0.67)1 (1.03 ± 0.37)0.0001
INR (sec)1.35 (1.57 ± 0.56)1.6 (1.71 ± 0.41)1.4 (1.48 ± .34)1.3 (1.43 ± 0.39)1.35 (1.54 ± 0.59)1.53 (1.84 ± 0.85)1.5 (1.62 ± 0.45)0.0005
Table 2. The Distribution of Intraoperative Variables in 391 Study Subjects
 Site 1Site 2Site 3Site 4Site 5Site 6Site 7P-value
  1. Medians followed by (means ± SD) are given for continuous variables. P-value is for difference in populations.

% Living Donation1200001600.0001
Surg time (hours)6.9 (7.03 ± 1.59)4.77 (4.57 ± 0.72)10 (10.01 ± 1.49)5.4 (5.57 ± 1.17)5 (5.32 ± 0.96)5.5 (5.79 ± 1.50)5.33 (5.32 ± 1.26)0.0001
RBC (units)3 (4.1 ± 4.0)0 (0 ± 0)0 (1.7 ± 2.5)2 (2.6 ± 2.5)2 (2.6 ± 2.7)5 (6.3 ± 6.0)1 (1.6 ± 1.7)0.0001
% Starts 10pm-4am16045322250.0001
% Bypass00100100100000.0001
% Caval Excision0010010007000.0001

Classification of Postoperative Adverse Outcomes

Adverse outcomes were classified using a 1 to 5 scale as outlined by the National Cancer Institute, Common Terminology Criteria for Adverse Events, version 3.0.13 Using this schema, Grade 1 is a mild adverse effect, with moderate and severe adverse effects categorized as Grades 2 and 3. Grade 4 is a life-threatening complication with death as Grade 5. If a patient experienced more than 1 complication, then all adverse events were independently categorized. All adverse outcomes were classified by a single investigator, blinded to the institution, and not involved in the collection of study site data (Table 3). A different investigator acted as a study co-coordinator and contacted participating centers to verify information about the recorded observations.

Table 3. Type and Severity of 30 Primary and 2 Secondary Adverse Outcomes in 391 Patients that Occurred Within 72 Hours of Surgery
CategoryGrade 1Grade 2Grade 3Grade 4Grade 5Complications
  1. The neurological complications of Grade 3 and 4 occurred secondary to other complications.

Pulmonary91   10 (31.2%)
Neurological1211 5 (15.6%)
Renal  4  4 (12.5%)
Hemorrhage 13  4 (12.5%)
Surgical  3  3 (9.3%)
Constitutional1 1  2 (6.4%)
Cardiac 11  2 (6.4%)
Death    22 (6.4%)
Total11 (34.4%)5 (15.6%)13 (40.6%)1 (3.1%)2 (6.3%)32

Statistical Methods

Baseline characteristics and adverse event rates were calculated for each site. For continuous data, the Kruskal-Wallis statistic was applied to test for population differences across sites. The Pearson chi-squared coefficient was applied to categorical and binary data elements with Fisher's exact test calculated for items with more than 25% cell counts less than 5. Unadjusted P values are given for each comparison. We analyzed the relationship between all adverse outcomes and patient/surgical factors by logistic regression (Table 4). We repeated the analysis after excluding all Grade 1 adverse outcomes to examine the relationship between more serious complications and patient/surgical factors (Table 4).

Table 4. Logistic Regression Results
Any Adverse OutcomeAny Adverse OutcomeSerious Adverse Outcome
ORCIORCI
BMI0.950.87-1.030.990.93-1.05
MELD1.050.97-1.141.110.99-1.24
Male1.570.89-2.751.230.23-6.54
Length of Surgery (Hours)1.050.86-1.271.200.77-1.88
Units RBC1.191.15-1.221.171.06-1.28
Night Start1.490.63-3.511.250.17-9.09
Bypass24.713.57-170.8866.8413.15-339.73
Piggyback4.790.64-36.0811.831.37-102.45
Pseudo R-squared 0..2043 0.3101
c-statistic 0.8091 0.8679

Independent variables were chosen based on their ability to predict adverse outcomes in previous studies,5 association with preoperative severity of illness, and influence on successful extubation after other types of surgery, to control for variation across study sites and to address the issue of whether differing surgical techniques might affect adverse outcomes following early extubation. Because patient characteristics differed by site and patients receiving transplantation may share some unmeasured commonalities with other patients at the same site, confidence intervals were adjusted using the Huber/White/sandwich estimator for robust calculation of variance with a cluster correction to allow for lack of independence within groups (study sites).

The explanatory power for each outcome is given as a c-statistic and as a pseudo R-squared value. The c-statistic is the area under the receiver operator curve and assesses the model's discriminant function, while the R-squared value estimates the percent of outcome variation explained by the model. The Hosmer-Lemeshow goodness-of-fit statistic, an indicator of model calibration, is also provided. All analyses were performed using Stata 8.2 (StataCorp, College Station, TX).

RESULTS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Characteristics of Study Sites

A total of 391 patients from the 7 participating centers were entered for analysis. All sites performed more than 100 liver transplants during the 2-yr study period; however, early extubation rates varied by site from 5 to 67% of cases Patient populations differed significantly by study site on all variables tested (Tables 1 and 2). The number of living donor operations reported by each institution varied from none to 16% of study subjects. The use of veno-venous bypass (VVBP) showed institutional preference, as 4 centers never used bypass while 3 centers always used bypass. Similar discrepancies between institutions were also observed for surgical technique. Length of surgery, the number of night starts, and the number of red blood cell units used during the operation also varied between institutions (Table 2).

Description of Adverse Outcomes

A total of 30 patients (7.6%) experienced a single primary adverse event. A total of 2 additional secondary complications resulted in 32 total adverse events in the study cohort (Table 3). Although 21 of the 32 events (65%) occurred within 36 hours after surgery, 11 of these were only Grade 1 in severity. Of the 10 pulmonary adverse effects, 9 were Grade 1 in severity, consisting of transient hypoxemia that required an increase in ambient oxygen delivery. The remaining patient required intubation and ventilation for persistent hypoxemia but was extubated without additional complication within 24 hours (Grade 2). There were 6 patients who required a second surgery within 36 hours (Grade 3): 3 for postoperative hemorrhage, 1 with a perforated duodenum, 1 who required repair of a biliary leak, and another to recover retained surgical instruments. One of the patients that experienced postoperative bleeding did not require relaparotomy (Grade 2). Primary neurological complications occurred in 3 patients, 1 who experienced mild agitation (Grade 1), another who had delirium (Grade 2), and a third who had a single generalized seizure (Grade 2). Fever developed in 1 patient but resolved spontaneously (Grade 1). Aside from the 6 patients who required relaparotomy and 1 patient with respiratory insufficiency, no patients required reintubation or an increase in their acuity of care.

In contrast, the 11 adverse outcomes that occurred after the first 36 hours of care were more serious, with 10 of these events being Grade 3 or more in severity. The types of adverse events were varied, with 1 case of fever caused by culture positive fungal infection (Grade 3). A total of 4 cases met criteria for acute renal failure,14 each scored as a Grade 3 complication. A total of 3 cases resolved within the 72-hour observation period while the remaining patient required dialysis. There were 2 cardiac complications, 1 new onset atrial fibrillation (Grade 2) and an episode of myocardial ischemia with a troponin leak (Grade 3). A total of 2 patients died (Grade 5) in the third postoperative day, 1 died after being readmitted to the intensive care unit for a gastrointestinal bleed following unrecognized acute portal vein thrombosis and the other was a sudden death of unknown etiology in a patient who had been discharged from the intensive care unit that day. A total of 2 patients had secondary neurological complications consisting of a change in mental status (Grade 3) following new onset of atrial fibrillation and as a complication of fungal infection (Grade 4). The patients who died required reintubation and an increased acuity of care that were unanticipated.

Adverse outcomes were not equally distributed between institutions, with 2 centers accounting for 65% of all complications within 72 hours. One center reported adverse outcomes in 53% of their extubated recipients while another reported a 12.6% complication rate. Of 4 cases of renal failure, 3 occurred at the latter institution. Each of these centers reported 1 death. The rate of adverse outcomes at the remaining 5 centers was 3.6% (range 0-4.6%).

Study Variables as Predictors of Patient Outcome

Table 4 provides results from the logistic regression models developed to evaluate the relationship between patient characteristics, intraoperative factors, surgical technique, and adverse outcome. The c-statistics indicate good discriminant ability (c-statistic >0.8) for any negative outcome and for negative outcomes greater than Grade 1. Based upon the pseudo R-squared values, a significant proportion of the variance is attributable to independent variables included in the model (approximately 20% and 30% for any adverse outcome and serious adverse outcomes, respectively), and the model fit is acceptable as indicated by the Hosmer-Lemeshow test (0.4068 for any adverse outcome and 0.7287 for serious adverse outcomes).

In the multivariate analysis, no single patient characteristic predicted an adverse outcome. Among intraoperative variables, the number of units of packed red cells transfused contributed significantly to adverse outcome (Table 4). Each unit of blood transfused conferred 15 to 20% increased risk of a complication. The use of VVBP predicted adverse outcomes and conferred a 24- and 66-fold increase in risk for all negative outcomes and those greater than Grade 1, respectively. In contrast, the piggyback technique as opposed to intrahepatic inferior vena caval excision was associated with a 10-fold increased risk of serious adverse events but had no effect on minor adverse outcomes.

DISCUSSION

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

In 391 patients who met the same conservative criteria for early extubation, adverse events occurred in only 7.7% of liver transplant recipients within 72 hours of surgery. The rate fell to 3.6% when the 2 institutions that experienced higher complication rates were excluded. Institutions that had more recorded events also had more serious ones. However, more serious events (Grade 4 and 5) occurred after 36 hours, a time when most transplantation patients are extubated using very conservative weaning protocols.15 Most adverse events were either pulmonary or surgically-related and were of Grade 3 or less in severity. Variables associated with adverse events following surgery were the number of units of blood transfused, the piggyback technique, and the use of VVBP.

Physicians elect ventilatory support following surgery to prevent primary respiratory failure or respiratory failure caused by other adverse events. Extubation and elective weaning protocols have good prognostic value to identify patients who will suffer primary respiratory failure.16, 17 Thus, differences in extubation preferences mainly stem from what risk a physician thinks a patient has of experiencing an adverse event that could cause respiratory failure. To determine the utility of immediate postoperative extubation in liver transplantation patients, physicians require evidence that the practice does not adversely affect patients. We therefore collected data on the nature and frequency of complications following extubation and examined differences in outcome between institutions.

Our patients had significantly fewer perioperative complications than commonly reported for transplant recipients who were routinely ventilated. It is possible that early extubation inadvertently selects patients who would have better outcomes regardless of how their ventilation was managed postoperatively. Better outcomes are also reported for patients who can be extubated within 1 hour following cardiac surgery,18 and immediately after major vascular19 and esophageal surgery.20 In fact, some suggest that extubation and weaning criteria could be crude markers of global physiological recovery.6 However, in contrast to patients who undergo major surgery; recovery of liver transplantation patients depends on characteristics of the recipient and the donor organ.

Shorter ventilation times are associated with better outcomes in other studies of liver transplant recipients.5 More adverse events in ventilated patients may also results from injury caused by ventilation or variation in institutional outcomes. Esophagectomy and cardiac surgery patients who were routinely ventilated had complications that were clearly mitigated by early or immediate extubation.18, 19, 21 The latter had significantly fewer pulmonary complications and better hemodynamic stability. Evidence from other studies suggests that early extubation may have prevented adverse events in some of our patients. However, we did not have a similar group of patients who were routinely ventilated at each institution for comparison. Thus, study design prevented us from directly measuring the effect of routine postoperative ventilation on outcome.

Even though our patients experienced few complications when the results were summed, the rate of adverse events reported by 2 of the participating was significantly greater. The number of units of blood transfused, piggyback technique, and VVBP were associated with more adverse outcomes in our study patients; however, the differential distribution of these variables between institutions did not completely explain the higher rate of adverse outcomes at 2 of the study sites. Physician experience did not appear to influence our findings, as there was only a small correlation (−0.1778; P = 0.0004) between the number of extubations performed at each institution and the number or severity of adverse outcomes. Variability in outcome was not limited to adverse events. We also observed significant differences in the relative number of patients extubated at each institution. Study sites extubated all patients that met standard criteria; however, individual study sites reported rates varying from 5 to 67% of patients. The variability in outcome within our own study leads us to conclude that institution-specific practices explain some of the better outcomes in our study.

Institutional differences in outcome are endemic in all aspects of medicine and our study is no exception. This study was not designed to probe the causes of interinstitutional variability in recipients extubated immediately after surgery. Because differences in patient characteristics did not explain outcome variability, we suggest that intraoperative practices that were not measured in this study need further evaluation. There are a number of important variables that we did not collect in this study that may have affected patient outcome. For example, we did not measure the correlation between donor graft function and outcome. Also, we did not evaluate the interaction between intraoperative fluid management and patient outcome. There is ample evidence that both of the latter factors affect patient outcome and thus adverse events.20, 22, 23 There were few adverse events in this study. Thus, we did not have enough observations to examine the effects of other important practices on these adverse outcomes. This would require a much larger database of observations.

We examined how a selected group of attributes interacted with the occurrence of adverse outcomes in our patients._MELD score did not have any influence on adverse outcome in the early postoperative period. This was an unexpected finding, considering that a number of studies have observed a correlation between MELD and both short- and long-term outcome.24, 25 However, our average MELD scores for each institution ranged from only 11.9 + 4.9 to 21.6 + 6.6; values are far below the threshold of 24 usually observed to be associated with adverse posttransplantation outcomes at some institutions.26 There were only 51 patients in this study who had a MELD score greater than 24. That the rate of adverse events in this group of patients was only 7.8% suggests that MELD score did not affect early extubation outcome in this study population.

The number of units of blood transfused and VVBP were associated with an increase in the number of all adverse outcomes while piggyback technique correlated with more serious complications within the first 72 hours following surgery. A correlation between the number of units of blood transfused during liver transplantation surgery, use of VVBP, and negative outcomes including death is well-described in the literature.27, 28 A similar relationship between the piggyback technique and the number of adverse events is not described, but there is a correlation between piggyback and increased blood loss at some institutions.29 Our evidence suggests that further studies should target the effect of VVBP and the piggyback technique on the number of perioperative complications, resource utilization, and consequently cost.

Even though variables such as transfusion and the use of VVBP are statistically associated with an increased rate of early postoperative adverse outcomes, it is possible that the variables themselves are not causative but rather represent markers for other practice patterns that were not measured in this study. That other institutionally-based clinical care choices could be hidden within some practices and influence adverse outcomes is suggested by a previous single-institutional study that did not find any effect of the amount of blood transfused on the rate of postoperative complications in liver transplant recipients who were extubated immediately after surgery.5 We could only partially correct for this phenomenon by using the robust clustering techniques described in the Patients and Methods section.

We have no means to completely separate potential unmeasured variation in practices associated with use of VVBP or piggyback technique from the use of the practice itself. We were able, however, to examine the likelihood that practice variation between study sites biased our results by creating models in which we dropped the indicators for surgical technique and inserted indicators for each institution. These models only explained an additional 2 to 5% of the variability and overfit the data. They do, however, offer evidence that the factors included in our model account for our results as opposed to correlated but unmeasured site-specific practices. First, the magnitude of all effects except gender remained remarkably stable (e.g., for any adverse event the odds ratio for units of blood transfused went from 1.19 to 1.17; MELD from 1.05 to 1.03). Second, no variables changed from being significant to insignificant or vice versa by conventional standards, although the confidence interval around the MELD score did increase in size.

In summary, liver transplant recipients who are extubated immediately after surgery using the standard criteria experience a low rate of adverse outcomes within 72 hours. Yet, there are considerable performance differences between institutions in the number and severity of perioperative adverse events that can cause patient compromise. Of all variables studied, only the amount of blood transfused, the piggyback technique, and the use of VVBP were associated with adverse outcomes. However, it is unclear whether these directly affect outcome or are surrogate variables for other practices not examined in this study. This study is limited by the small number of adverse outcomes observed, the small number of participating institutions, and the lack of information about other aspects of perioperative care that may affect patient outcome. In view of these limitations, our conclusion require further testing. This would require a larger dataset that adequately represents the full spectrum of perioperative practices.

REFERENCES

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES
  • 1
    Hawkes CA, Dhileepan S, Foxcroft D. Early extubation for adult cardiac surgical patients. Cochrane Database Syst Rev 2003; CD003587.
  • 2
    Cheng D. Impact of early extubation on hospital discharge on hospital discharge. J Cardiothorac Vasc Anesth 1998; 12(Suppl 2): 3544.
  • 3
    Calafiore AM, Scipioni G, Teodori G, Di Giammarco G, Di Mauro M, Canosa C, et al. Day 0 intensive care unit discharge—risk or benefit for the patient who undergoes myocardial revascularization? Eur J Cardiothorac Surg 2002; 21: 377384.
  • 4
    Cheng DC. Pro: early extubation after cardiac surgery decreases intensive care unit stay and cost. J Cardiothorac Vasc Anesth 1995; 9: 460464.
  • 5
    Mandell MS, Lezotte D, Kam I, Zamudio S. Reduced use of intensive care after liver transplantation: Influence of early extubation. Liver Transpl 2002; 8: 676681.
  • 6
    Mandell MS, Lockrem J, Kelly SD. Immediate tracheal extubation after liver transplantation: experience of two transplant centers. Anesth Analg 1997; 84: 249253.
  • 7
    Neelakanta G, Sopher M, Chan S, Pregler J, Steadman R, Braunfeld M, Csete M. Early tracheal extubation after liver transplantation. J Cardiothorac Vasc Anesth 1997; 11: 165167.
  • 8
    Biancofiore G, Romanelli AM, Bindi ML, Cosani G, Boldrini A, Barristini M, et al. Very early tracheal extubation without predetermined criteria in a liver transplant recipient population. Liver Transpl 2001; 7: 777782.
  • 9
    Brown RS, Kumar KS, Russo MW, Kinkhabwala M, Rudow DL, Harren P, et al. Model for End-Stage Liver Disease and Child-Turcotte-Pugh score as predictors of pretransplantation disease severity, posttransplantation outcome, and resource utilization in United Network for Organ Sharing status 2A patients. Liver Transpl 2002; 8: 278284.
  • 10
    Biancofiore G, Bindi ML, Romanelli AM, Boldrini A, Bisa M, Esposito M, et al. Fast track in liver transplantation: 5 years' experience. Eur J Anaesthesiol 2005; 22: 584590.
  • 11
    Hines R, Barash PG, Watous G, O'Connor T. Complications occurring in the postanesthesia care unit. Anesth Analg 1992 74: 503509.
  • 12
    Howdle PD. Clinical evaluation of liver disease. In: O'GradyJG, LakeJR, HowdlePD, eds. Comprehensive clinical hepatology, 1st ed. London: Mosby, 2000: 2:4.12:4.15.
  • 13
    National Institute of Cancer Evaluation Program, Common Terminology Criteria for Adverse Events, Version 3.0 December 12, 2003, Available at: https://webapps.ctep.nci.nih.gov/webobjs/ctc/webhelp/welcome_to_ctcae.htm; Last accessed July 30, 2006.
  • 14
    Uchino S, Kellum JA, Bellomo R, Doig G, Morimatsu H, Morgera S, et al. Acute renal failure in critically ill patients. A multinational and multicenter study. JAMA 2005; 294: 813818.
  • 15
    Carton EG, Plevak DJ, Kranner PW, Rettke SR, Geiger HJ, Coursin DB. Perioperative care of the liver transplant patient. Part 2. Anesth Analg 1994; 78: 382399.
  • 16
    Chavez A, Cruz R, Zaritsky A. Spontaneous breathing trial predicts successful extubation in infants and children. Pediatr Crit Care Med 2006; 7: 324328.
  • 17
    Martinez A, Seymour C, Nam M. Minute ventilation recovery time, a predictor of extubation outcome. Chest 2203; 123: 12141221
  • 18
    Moon MC, Abdoh A, Hamilton GA, Lindsay WG, Duke PC, Pascoe EA, Del Rizzo DF. Safety and efficacy of fast track in patients undergoing coronary artery bypass surgery. J Card Surg 2001; 16: 319326.
  • 19
    Cohen J, Loewinger J, Hutin K, Sulkes J, Zelikovski A, Singer P. The safety of immediate extubation after abdominal aortic surgery: A prospective, randomized trial. Anesth Analg 2001; 93: 15461549.
  • 20
    Lanuti M, de Delva PE, Maher A, Wright DC, Gaissert HA, Wain JC, et al. Feasibility and outcomes of an early extubation policy after esophagectomy. Ann Thorac Surg 2006; 82: 20372041.
  • 21
    Bardstrup B, Tonnensen H, Beier-Holgersen R, Hjortso E, Ording H, Lindorff-Larsen K, et al. Effects of intravenous restriction on postoperative complications: comparison of two perioperative fluid regimes: a randomized assessor blinded multicenter trial. Ann Surg 2003; 238: 641648.
  • 22
    Feng S, Goodrich NP, Bragg-Gresham JL, Dykstra DM, Punch JD, DebRoy MA, et al. Characteristics associated with liver graft failure: the concept of a donor risk index. Am J Transplant 2006; 6: 783790.
  • 23
    Rossaint R, Slama K, Jaeger M, Konrad M, Pappert D, Bechstein W, et al. Fluid restriction and early extubation for successful liver transplantation. Transpl Proc 1990; 22: 15331534.
  • 24
    Cholongitas E, Marelli L, Shusang V, Senzolo M, Rolles K, Patch D, Burroughs AK. A systematic review of the performance of the Model for End-Stage Liver Disease (MELD) in the setting of liver transplantation. Liver Transpl 2006; 12: 10491061.
  • 25
    Xia VW, Braunfeld M, Neelakanta G, Hu K-Q, Nourmand H, Levin P, et al. Preoperative characteristics and intraoperative transfusion and vasopressor requirements in patients with low vs. high MELD scores. Liver Transpl 2006; 12: 614620.
  • 26
    Desai NM, Mange KC, Crawford MD, Abt PL, Frank AM, Markmann JW, et al. Predicting outcome after liver transplantation; utility of the model for end-stage liver disease and a newly derived discrimination function. Transplantation 2004; 77: 99106.
  • 27
    Fan ST, Lo CM, Liu CL, Yong BH, Wong J. Determinants of hospital mortality of adult recipients of right lobe liver transplantation. Ann Surg 2003; 238: 864870.
  • 28
    Ramos E, Dalmau A, Sabate A, Lama C, Llado L, Figueras J, Jaurrieta E. Intraoperative red blood cell transfusion in liver transplantation: influence on patient outcome, prediction of requirements, and measures to reduce them. Liver Transpl 2003; 9: 13201327.
  • 29
    Parilla P, Sanchez Bueno F, Figueras J, Jaurrieta E, Mir J, Margarit C, et al. Analysis of the complications of the piggyback technique in 1,112 liver transplants. Transplantation 1999; 67: 12141217.