|
|
||||||||
a Department of Anesthesiology, University of Virginia Health Sciences Center, University of Virginia Health System, Charlottesville, Virginia
b Department of Public Health Sciences, Division of Biostatistics and Epidemiology, University of Virginia Health Sciences Center, University of Virginia Health System, Charlottesville, Virginia
c Department of Surgery, University of Virginia Health Sciences Center, University of Virginia Health System, Charlottesville, Virginia
d Department of Anesthesiology and Intensive Care, University Hospital Muenster, Muenster, Germany
e Department of Community Medicine and Injury Control Research Center, West Virginia University, Morgantown, West Virginia
Accepted for publication June 13, 2011.
* Address correspondence to Dr Blank, Department of Anesthesiology, University of Virginia Health System, Box 800710, Charlottesville, VA 22908 (Email: rsb8p{at}virginia.edu).
| Abstract |
|---|
|
|
|---|
Methods: We conducted a retrospective cohort study; charts of 129 patients who underwent elective pneumonectomy at the University of Virginia were reviewed. Logistic regression was used to estimate the effect of anesthetic factors on the odds of at least one respiratory complication. Linear regression models were fit to assess the impact of these outcomes on length of stay (LOS).
Results: The incidence of respiratory complications in this cohort was 21%. In univariate analysis total nonblood fluids (p = 0.001), and the blood products packed red blood cells (p < 0.001), plasma (p < 0.001), and platelets (p = 0.044) were significantly associated with respiratory complications. In a multivariable logistic regression analysis, single unit transfusion of any blood product (packed red blood cells, plasma, or platelets) was identified as a major risk factor for respiratory complications after controlling for covariates (odds ratio = 1.47, 95% confidence interval 1.06 to 2.05). Respiratory failure and complications were closely related to LOS, increasing the LOS by a factor of 4.7 (95% confidence interval 3.51 to 6.18) and 3.5 (95% confidence interval 2.69 to 4.41), respectively.
Conclusions: Blood product transfusion affects respiratory function and is an independent risk factor for respiratory complications after pneumonectomy.
| Introduction |
|---|
|
|
|---|
| Patients and Methods |
|---|
|
|
|---|
Charts were reviewed for all 129 patients meeting the inclusion criteria, who underwent pneumonectomy at the University of Virginia Health System from January 1997 through May 2008. Inclusion criteria were age greater than or equal to 18 years and elective pneumonectomy. Data of interest were divided into exposure and outcome variables and were abstracted on two different forms at different times or by different investigators to reduce the likelihood of bias. The primary exposures of interest were anesthetic factors, patient characteristics, comorbidities, pulmonary function tests, and surgical factors. The primary outcome was a composite outcome of all respiratory complications. Possible respiratory complications included respiratory failure (defined as the need for initial mechanical ventilation more than 48 hours postoperatively or the need for reintubation due to respiratory insufficiency), acute lung injury (ALI), acute respiratory distress syndrome (ARDS), atelectasis requiring bronchoscopy, pneumonia, and bronchopleural fistula. Acute lung injury and ARDS were defined according to the criteria set forth in the 1994 American European Consensus Conference [5]. Atelectasis was diagnosed in patients with radiographic evidence of atelectasis and clinical evidence of an elevated alveolar-arterial oxygen gradient requiring bronchoscopy. The diagnosis of pneumonia was made when radiographic evidence of lung consolidation and positive sputum cultures were both present. Atrial dysrhythmia was diagnosed when any atrial dysrhythmia including atrial fibrillation, atrial flutter, and supraventricular tachycardia required medical or electrical therapy. Secondary outcomes included hospital length of stay (LOS) and discharge status (mortality).
Clinical Management
Criteria for consideration of pneumonectomy surgery included a Zubrod performance status of 2 or less, adequate cardiopulmonary reserve, and the absence of an active infection. For patients with known lung cancer, a pneumonectomy was not performed on patients with multistation positive N2 nodal disease, patients with known or suspected M1 disease, or any patient on relatively high-dose immunosuppressive medications. Except where contraindicated, patients received a thoracic epidural catheter preoperatively, typically placed in the interspaces between T6 and T9. One hundred and twenty-three patients (95%) received epidural catheters for postoperative pain management. Contraindications to epidural analgesia included patient refusal, coagulopathy, and antiplatelet therapy. Patients received a balanced general anesthetic, typically with sevoflurane or isoflurane in oxygen, supplemented with fentanyl, morphine, or hydromorphone. After induction and neuromuscular blockade, the trachea was intubated with an appropriately sized double lumen endotracheal tube and the position verified with fiberoptic bronchoscopy. Arterial catheters were placed for continuous arterial pressure monitoring and central venous catheters were placed in selected cases. One-lung ventilation was initiated after patient positioning; tidal volumes were selected based on ideal body weight and ventilator pressures, typically in the volume control mode. Efforts were made to limit inspiratory pressures to below 30 cm water; protective ventilation strategies and the minimization of inspired oxygen fraction were routinely used during much of the study period. During the study period, intraoperative fluid management was moderately restrictive and was titrated at the discretion of the attending anesthesiologist based on standard hemodynamic criteria. Hemodynamic goals included the maintenance of arterial blood pressure within 25% of the patient's preoperative average blood pressure. Vasopressors including phenylephrine were used if needed to maintain arterial blood pressure. The incidence of intraoperative hypotension and use of vasopressors were abstracted and analyzed in this study.
Postoperatively, most patients were weaned from mechanical ventilatory support and extubated in the operating room. Epidural catheters were dosed with local anesthetic, typically 0.2% ropivacaine, prior to emergence and an infusion of local anesthetic (0.125% bupivacaine and hydromorphone 6 to 9 mcg/mL) was initiated. Transport was to an intensive care unit specializing in the care of patients undergoing cardiothoracic surgery where every patient spent the first postoperative night. Patients were not fed for at least 24 hours or until a bedside swallow study could be passed. Moderate fluid restriction was maintained in the intensive care unit. Maintenance fluids (typically lactated Ringer solution) were infused at 1 mL · kg–1 · hour–1 until oral fluids were permitted, usually on the first postoperative day. Evidence of pulmonary edema as evidenced by chest radiography, physical examination, or increased oxygen requirements prompted a decrease in intravenous fluid administration and diuretic therapy. Patients were mobilized out of bed within 12 hours unless mechanical ventilation was required. All nonventilated patients received incentive spirometry. All patients received scheduled prophylactic antibiotic therapy for 24 hours. Transfer to an adjacent step-down unit typically occurred on the following day. Length of stay data were abstracted and analyzed.
Statistical Analysis
Frequencies with percentages and medians with interquartile ranges were tabulated for categoric and continuous data, respectively. The
2 and Mann-Whitney U tests were conducted to test for differences between patients with and without the primary outcomes. When expected cell counts were 5 or less the Fisher exact test was used. Multivariable logistic regression models were fit using exact methods when indicated to quantify the association between anesthetic factors such as fluid and blood product administration, covariates, and respiratory complications. We prespecified potential risk predictors based on the results of univariate analysis and previously published risk factors. We further conducted variable clustering analysis in determining risk predictors based on an appropriate similarity matrix of the candidate predictors. Fluid exposures were categorized as follows: all nonblood product fluids were summed as total fluids; patients receiving any packed red blood cells, plasma, and platelets were coded as having received blood products. Pneumonectomy type was defined as either standard or nonstandard (including carinal, completion, extrapleural, or intrapericardial). Surgical indications for pneumonectomy were grouped as either benign or malignant. In addition to these factors, we also adjusted for American Society of Anesthesiologists (ASA) status (ASA 3 or 4 versus ASA 2) and duration of anesthesia in the multivariable model. Internal model validation was determined by bootstrap model validation, a method to assess how accurately our model would predict outcomes for a new sample of data. A bootstrapped corrected C-index or area under the receiver operator characteristic curve was utilized as a measure of overall predictive discrimination, which is defined in this study as the ability to separate those patients who developed respiratory complications after the surgery from those who did not. An area under the receiver operator characteristic curve area of 0.5 indicates no discrimination while an area of 1.0 indicates perfect discrimination. The LOS data were positively skewed and therefore transformed to the natural logarithm scale to normalize the outcome. Linear regression models were fit to assess whether there was an association between respiratory failure and respiratory complications, respectively, and LOS. These models were adjusted for the same set of prespecified factors described in the multivariable logistic model. We likewise assessed the associations between total fluids, blood products, and LOS. All analyses were performed in SAS version 9.1 (SAS Institute, Inc, Cary, NC, [2002-2003]) or R, A Language and Environment for Statistical Computing, version 2.9 (R Foundation for Statistical Computing, Vienna, Austria [2011]).
| Results |
|---|
|
|
|---|
|
|
|
In an effort to ascertain whether significant changes in clinical practice, exposures, and outcomes have occurred during the study period we analyzed these variables as a function of study year. Most exposure and outcome variables did not significantly vary as a function of time during the study. These include respiratory failure, respiratory complications, LOS, blood loss, fluid administration, blood product transfusion, surgical indication, duration of anesthesia, and ASA status. As expected, maximum tidal volume during one-lung ventilation declined over the study period (p = 0.004).
Adjusting for patient characteristics such as age, gender, body mass index, and ASA status, LOS in patients who developed respiratory failure and respiratory complications was increased by a factor of 4.7 (95% CI, 3.51 to 6.18) and 3.5 (95% CI, 2.69 to 4.41), respectively (Fig 1). Adjusting for the same set of covariates described in the multivariable respiratory complications model, blood product administration (1.16-fold increase per unit blood product; 95% CI, 1.09 to 1.25), pneumonectomy type (1.42-fold increase [standard versus other]; 95% CI, 1.04 to 1.95), and duration of anesthesia (1.16-fold increase per 100 minutes; 95% CI, 1.03 to 1.31) were significant risk factors for prolonged LOS in the multivariable model.
|
| Comment |
|---|
|
|
|---|
In the present study, intraoperative infusion of parenteral fluids (total fluids) was related to postoperative respiratory complications in univariate analysis but was not identified as an independent risk factor in multivariable analysis. Discrepancies between studies may reflect the inherent difficulty in identifying and controlling for confounding in retrospective studies of perioperative management. The administration of parenteral fluids during pneumonectomy is likely to accompany numerous other factors which may adversely affect outcomes after pulmonary resection including surgical indication [3, 10] and surgical complexity [11, 12], operative time, or anesthetic duration [4, 13], pneumonectomy type [10], intraoperative complications [14] including hemorrhage [15], patient age [1, 3, 10, 12, 16], comorbidities [1, 4, 10–12, 16–19], and ASA status [13, 18, 20]. Thus it is unclear whether the effect of parenteral fluids in this context is directly responsible for adverse respiratory outcomes or rather whether it simply accompanies other variables more likely to be directly injurious.
Several studies have shown that intraoperative transfusion of blood products is related to postoperative morbidity [1, 12, 14–16, 21, 22] and mortality [3, 11, 12, 16, 23] in patients undergoing pulmonary resection surgery. However, a specific association between transfusion and major morbidity after pneumonectomy (multivariable analysis) has been previously reported in one study [21], but not in others examining the role of this exposure [1, 3, 20]. In this study, multivariable analysis controlling for potential confounders demonstrated that the transfusion of one unit of any blood product was associated with a significant increase in the risk of postoperative respiratory complications. Again, discrepancies between studies may result from the degree to which confounding has been controlled in multivariable analyses. For example, pneumonectomy type and duration have been shown to predict blood product transfusion [21] and may confound the relationship between these variables. Preoperative hemoglobin concentration likewise has been associated with complications after pneumonectomy [1, 3, 20], but it is not clear whether this is an independent effect.
Patients who developed respiratory complications in the present study had a lower preoperative hemoglobin concentration compared with the group that did not subsequently experience this complication. Thus, intraoperative transfusion of packed red blood cells may be a surrogate for preoperative anemia, the disease states causing anemia, or for a technically difficult or complicated surgery. It appears likely from these studies that anemia poses two related risks in surgical patients; the risk inherent to anemia or the disease states causing anemia and the risks imposed from an increased likelihood of blood product transfusion. In the present study, however, the association between blood product transfusion and respiratory complications was independent of surgical indication, pneumonectomy type, and anesthetic duration, and was not altered by controlling for preoperative hemoglobin concentration or preoperative anemia, suggesting that blood product transfusion is an independent risk factor for adverse respiratory events.
The present study is limited by its retrospective nature, limitations in the quantity and nature of data that could be accurately abstracted from patient charts, the total number of patients studied, and potential changes in surgical, anesthetic, and postoperative care that may have occurred over the study period. Specifically, the institution of protective ventilation strategies, which became the clinical standard of practice for one-lung ventilation during the study period, may have had clinical impact in reducing lung injury [24, 25] in these patients. Most exposures and all reported outcomes, however, did not vary significantly during study period.
These findings support the concept that blood product administration affects respiratory function and the development of respiratory complications after pneumonectomy. The identification of ideal transfusion triggers, use of strict transfusion strategies, and the optimization of preoperative hemoglobin concentration may represent suitable targets for study in prospective trials.
| Acknowledgments |
|---|
|
|
|---|
| Footnotes |
|---|
|
|
|---|
| References |
|---|
|
|
|---|
Related Article
This article has been cited by other articles:
![]() |
R. G. Evans and B. Naidu Does a conservative fluid management strategy in the perioperative management of lung resection patients reduce the risk of acute lung injury? Interact CardioVasc Thorac Surg, September 1, 2012; 15(3): 498 - 504. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. Licker Invited Commentary Ann. Thorac. Surg., October 1, 2011; 92(4): 1194 - 1194. [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 |