Impact of Off-Pump Coronary Artery Bypass Graft Surgery on Postoperative Pulmonary Complications in Patients With Chronic Lung Disease

Transcription

1 ADULT CARDIAC ORIGINAL ARTICLES: ADULT CARDIAC ADULT CARDIAC SURGERY: The Annals of Thoracic Surgery CME Program is located online at To take the CME activity related to this article, you must have either an STS member or an individual non-member subscription to the journal. Impact of Off-Pump Coronary Artery Bypass Graft Surgery on Postoperative Pulmonary Complications in Patients With Chronic Lung Disease Faraz Kerendi, MD, Michael E. Halkos, MD, John D. Puskas, MD, Omar M. Lattouf, MD, Patrick Kilgo, MS, Robert A. Guyton, MD, and Vinod H. Thourani, MD Cardiothoracic and Vascular Surgeons, Austin, Texas; Division of Cardiothoracic Surgery, Emory University School of Medicine, Atlanta; and Rollins School of Public Health, Emory University, Atlanta, Georgia Background. Off-pump coronary artery bypass graft surgery (OPCAB) has proven to be beneficial in many high-risk subgroups. This study aims to determine whether OPCAB lowers the incidence of pulmonary complications among patients with chronic lung disease (CLD) when compared with on-pump coronary artery bypass graft surgery (ONCAB). Methods. From 2002 to 2007, 7,060 patients underwent isolated coronary artery bypass graft surgery in an academic center. Patients were classified according to surgery type (ONCAB or OPCAB) and presence or absence of CLD. A propensity score was produced to estimate each patient s likelihood of being assigned to OPCAB on the basis of 39 preoperative risk factors. Multiple logistic regression models and adjusted odds ratios with 95% confidence intervals were used to evaluate the effect of surgery type, CLD, and their interaction on pulmonaryrelated complications and mortality. Results. Among OPCAB patients, 15.3% (720 of 4,693) had CLD compared with 11.2% (264 of 2,367) for ONCAB. Off-pump coronary artery bypass graft surgery was performed in 73.2% of CLD patients compared with 66.5% in those without CLD (p < ). Chronic lung disease was associated with a greater incidence of prolonged ventilation, reintubation, pneumonia, intensive care unit hours, and non home discharge. After propensity score adjustment, OPCAB was associated with a significantly reduced incidence of prolonged ventilation, pneumonia, intensive care unit stay, and mortality. No significant interactions existed between surgery type and CLD status, suggesting that OPCAB was equally beneficial to patients with and without CLD. Conclusions. In this series, patients with CLD were more likely to undergo OPCAB. Patients with CLD are at significantly greater risk of pulmonary-related complications than patients without CLD. Off-pump coronary artery bypass graft surgery reduced the incidence of pulmonary complications and mortality in all patients. Importantly, this benefit was seen similarly for patients with and without CLD. (Ann Thorac Surg 2011;91:8 15) 2011 by The Society of Thoracic Surgeons As techniques in off-pump coronary artery bypass graft surgery (OPCAB) have progressed, evidence has emerged that high-risk patients with multiple associated comorbidities may benefit from avoiding the deleterious effects of cardiopulmonary bypass [1, 2]. The advantages of OPCAB have been demonstrated in a multitude of retrospective studies as well as several randomized controlled trials comparing OPCAB with conventional on-pump coronary artery bypass graft surgery (ONCAB) [3, 4]. Many of these studies have shown that OPCAB may be advantageous in certain high-risk subgroups, including patients with left ventricular dysfunction [5, 6], renal failure [7, 8], and diabetes [9], and elderly patients [10]. Patients with chronic lung disease (CLD) represent one such high-risk subgroup, which represents a large percentage of cardiac surgery patients. Postoperative complications in these patients may range from relatively Accepted for publication Aug 3, Address correspondence to Dr Kerendi, Cardiothoracic and Vascular Surgeons, PA, 1010 W 40th St, Austin, TX 78756; fkerendi@ ctvstexas.com. Dr Puskas discloses that he has financial relationships with Medtronic and Maquet by The Society of Thoracic Surgeons /$36.00 Published by Elsevier Inc doi: /j.athoracsur

2 Ann Thorac Surg KERENDI ET AL 2011;91:8 15 OPCAB IN PATIENTS WITH LUNG DISEASE benign to life threatening (ie, respiratory failure requiring tracheostomy, severe pneumonia, or acute respiratory distress syndrome). Despite this, there has been a paucity of data regarding any potential benefit of OPCAB in patients with CLD undergoing surgical coronary revascularization. Although several studies have been designed to evaluate the benefit of OPCAB on postoperative pulmonary function, the results have been contradictory. In a prospective, randomized controlled trial comparing OPCAB with ONCAB at Emory, Staton and associates [11] reported an improvement in gas exchange and extubation times for OPCAB patients. Other similar studies, however, have failed to confirm these beneficial findings [12, 13]. We propose that there may be a disparity in outcomes in that OPCAB may be of greater benefit in patients with underlying CLD than in patients with comparatively normal lung function. Therefore, the goals of this study are to assess postoperative pulmonary complications in patients with CLD undergoing coronary artery bypass graft surgery (CABG) and to compare outcomes of OPCAB compared with ONCAB in patients with a history of CLD. Material and Methods The Society of Thoracic Surgeons (STS) Adult Cardiac Database was queried for all patients who underwent primary isolated CABG at the Emory University Hospitals between January 2002 and April This time frame was chosen to coincide with the entire period during which CLD was categorized as mild, moderate, or severe in the institutional STS database. The study cohort consisted of 7,060 consecutive patients, including urgent and emergent patients. Medical records from this retrospective, single-center cohort study included demographic data, preexisting comorbidities, surgeon identity, operative strategy, and clinical outcomes. The study was approved by the Emory University Institutional Review Board in compliance with the regulations of the Health Insurance Portability and Accountability Act and the Declaration of Helsinki, with a waiver for individual patient consent. Surgical Technique Each patient underwent OPCAB or ONCAB, performed at the discretion of faculty surgeons. Off-pump coronary artery bypass graft surgery was performed with one of several commercially available cardiac positioning and coronary artery stabilizing devices, using techniques that have been previously described. On-pump coronary artery bypass graft surgery was performed with standard techniques, using roller head pumps, membrane oxygenators, cardiotomy suction, arterial filters, cold antegrade and retrograde blood cardioplegia, and moderate systemic hypothermia (32 to 34 C). Patients who were converted intraoperatively from OPCAB to ONCAB or from ONCAB to OPCAB were entered into the database and analyzed according to the operation they ultimately received. Intraoperative conversion began to be recorded as part of the institutional database in 2004, when a data field for conversion was introduced into the STS national adult cardiac database. During the study period, 126 patients (1.8%) were converted from OPCAB to ONCAB. No patients were recorded as having converted from ONCAB to OPCAB. Of the 126 converted patients, 5 of them died (4.0%), which is not statistically higher than the nonconverted group (n 127 deaths, 1.8%; p 0.07). Of the 5 converted deaths, 3 of them were in patients with no CLD and 2 of them were in patients with mild CLD. Long-Term Follow-Up Long-term follow-up was derived from the Social Security Death Index (SSDI) database. The SSDI is a publicuse national database of death records extracted from the US Social Security Administration s Death Master File Extract. Persons who have died since 1963 who had a social security number and whose death has been reported to the SSA will be listed in the SSDI. Thus, for each patient who died before the cutoff date of March 31, 2007, a mortality date was provided, allowing construction of Kaplan-Meier long-term survival curves. Cause of death was not considered; this study seeks to describe all-cause mortality between CLD and surgery types. Demographics and Preoperative Data Before analysis, preoperative risk factors for the outcomes of interest were identified and harvested from the STS database (Table 1). Standard STS definitions of each risk factor and outcome were used ( Specifically, mild CLD was defined as a forced expiratory volume in 1 second of 60% to 75% of predicted or treatment with oral bronchodilator or chronic inhaler therapy. Moderate CLD was defined as a forced expiratory volume in 1 second of 50% to 59% of predicted or chronic steroid therapy aimed at lung disease. Severe CLD was defined as a forced expiratory volume in 1 second of less than 50% or room air partial pressure of oxygen less than 60 mm Hg or room air partial pressure of carbon dioxide greater than 50 mm Hg. The institutional database was populated by trained personnel devoted exclusively to this task; consequently, missing data were scarce. Data were 100% complete for the critical risk factors of interest (surgery type and CLD) as well as for each major postoperative hospital outcome. Data were missing for the following variables: body mass index (n 8; 0.1%), body surface area (n 8; 0.1%), height (n 8; 0.1%), ejection fraction (n 284; 4.0%), STS predicted risk of mortality (n 2; 0.1%), last creatinine level (n 40; 0.6%), and weight (n 3; 0.1%). Data Management and Statistical Analysis All data for consecutive patients were entered into a computerized cardiac surgical database, using the data fields and definitions of the STS National Adult Cardiac Database. Checks for data quality are used both at the institutional level and before final entry into the STS National Adult Cardiac Database. A multiple imputation algorithm was used to impute values that reflect the uncertainty surrounding the missing data. This was done 9 ADULT CARDIAC

3 ADULT CARDIAC 10 KERENDI ET AL Ann Thorac Surg OPCAB IN PATIENTS WITH LUNG DISEASE 2011;91:8 15 Table 1. Preoperative Risk Factors and Demographics Used for Propensity Scoring a Risk Factor ONCAB (N 2,103) No CLD OPCAB (N 3,973) p Value ONCAB (N 264) CLD OPCAB (N 720) p Value Age (y) 61.5 (10.4) 62.8 (11.4)) (10.2) 65.6 (10.5) Female sex 503 (23.9) 1,155 (29.1) (36.0) 274 (38.1) 0.55 Body mass index ((kg/m 2 )/100) 0.3 (0.12) 0.3 (0.17) (0.07) 0.30 (0.19) 0.24 Cerebrovascular disease 280 (13.3) 593 (14.9) (38.3) 210 (29.2) Previous stroke 149 (7.1) 335 (8.4) (22.7) 127 (17.6) 0.07 Diabetes 813 (38.7) 1,451 (36.5) (40.9) 276 (38.3) 0.46 Chronic kidney disease 104 (5.0) 337 (5.7) (6.4) 67 (9.3) 0.15 Dialysis-dependent renal failure 44 (2.1) 106 (2.7) (1.5) 25 (3.5) 0.11 Congestive heart failure 401 (19.1) 647 (16.3) (31.1) 230 (31.9) 0.79 Ejection fraction b Left main disease ( 50% stenosis) 582 (27.7) 952 (24.0) (33.0) 184 (25.6) 0.02 Previous myocardial infarction 1,179 (56.1) 1,941 (48.9) (65.5) 422 (58.6) Number of diseased vessels b Hypertension 1,708 (81.2) 3,243 (81.6) (84.9) 609 (84.6) 0.92 Peripheral vascular disease 247 (11.8) 526 (13.2) (29.2) 195 (27.1) 0.52 Current smoker 547 (26.0) 1,053 (26.5) (40.9) 294 (40.8) 0.98 Previous smoker 1,411 (67.1) 2,674 (67.3) (82.6) 619 (86.0) 0.19 Surgical status Elective 1,434 (68.2) 2,885 (72.6) (66.7) 472 (65.6) Urgent 462 (22.0) 969 (24.4) (25.8) 229 (31.8)... Emergent 192 (9.1) 119 (3.0) (7.2) 19 (2.6)... Salvage 15 (0.7) 0 (0)... (0.4) 0 (0)... Previous cardiovascular intervention 709 (33.7) 1,277 (32.1) (48.1) 300 (41.7) 0.07 Previous CABG 163 (7.8) 82 (2.1) (7.2) 16 (2.2) Previous PCI 332 (15.8) 640 (16.1) (18.2) 115 (16.0) 0.41 Intraaortic balloon pump 258 (12.3) 146 (3.7) (10.6) 25 (3.5) STS predicted risk of mortality (0.046) (0.03) (0.06) (0.06) 0.31 a Year of surgery, surgeon identity, and eight indicators of missing variables were also included in the model. b Mean standard deviation. CABG coronary artery bypass graft; CLD chronic lung disease; ONCAB on-pump coronary artery bypass graft surgery; OPCAB off-pump coronary artery bypass graft surgery; PCI percutaneous coronary intervention; STS Society of Thoracic Surgeons. to avoid selection bias that can occur by deleting cases with missing variables of interest. Ten data sets were imputed, and parameter estimates of the ten data sets were combined using methods originally described by Schaffer [14]. Data were assumed to be missing at random. Patients were classified according to CLD and the surgery type (OPCAB or ONCAB) they received. To reduce the effect of selection bias, propensity scores, described by Blackstone [15] and D Agostino [16], were calculated for each patient on the basis of 39 risk factors available preoperatively. For the propensity score calculation, a multiple logistic regression model was used nonparsimoniously to model OPCAB (yes or no) as a function of all 39 risk factors (Table 1). The resulting conditional probability of a patient receiving OPCAB is the propensity score. The goal of the propensity score adjustment is to postrandomize or balance the groups with respect to their preoperative risk factors so that nonconfounded comparisons of group effects can be estimated in an unbiased fashion. To statistically evaluate the main effects of CLD and surgery type, three different multivariable logistic regression model types were constructed for each outcome of interest. Adjusted odds ratios associated with OPCAB and CLD, along with 95% confidence intervals, were computed for each adverse outcome and each model type. All logistic models were adjusted with the propensity score. Primary outcomes of interest included prolonged ventilation ( 48 hours), total ventilator hours, need for reintubation, pneumonia, acute respiratory distress syndrome, total intensive care unit (ICU) stay (in hours), non home discharge, and 30-day mortality. First, a model consisting of CLD as a dichotomous variable (presence or absence), surgery type (OPCAB or ONCAB), and their interaction was fit. If the interaction was statistically insignificant, then that term was removed from the model and the main effects alone were evaluated. This model was primarily designed to evaluate the effects of CLD and surgery type in unison and to determine whether OPCAB lessened or worsened outcomes in the presence of CLD. Second, a model consist-

4 Ann Thorac Surg KERENDI ET AL 2011;91:8 15 OPCAB IN PATIENTS WITH LUNG DISEASE Table 2. Unadjusted Outcomes of Patients With or Without Chronic Lung Disease According to Surgery Type Outcome ONCAB (N 2,103) No CLD OPCAB (N 3,973) p Value ONCAB (N 264) CLD 11 OPCAB (N 720) p Value ADULT CARDIAC Prolonged ventilation ( 48 h) 209 (10.0) 187 (4.9) (17.2) 73 (10.4) Postoperative ventilator time (h) a Reintubation 78 (3.7) 116 (2.9) (9.1) 55 (7.6) 0.46 Pneumonia 104 (5.0) 125 (3.2) (9.5) 57 (7.9) 0.43 ARDS 12 (0.6) 6 (0.2) (0.4) 5 (0.7) 0.57 Sepsis 42 (2.0) 28 (0.7) (3.0) 17 (2.4) 0.55 Permanent stroke 41 (2.0) 39 (1.0) (1.1) 8 (1.1) 0.97 Reexploration for 42 (2.0) 96 (2.4) (1.5) 24 (3.3) 0.13 bleeding/tamponade Acute renal failure requiring 90 (4.3) 91 (2.3) (5.3) 35 (4.9) 0.78 dialysis Atrial fibrillation 409 (19.5) 645 (16.2) (25.0) 165 (22.9) 0.49 Home discharge (survivors) 1,934 (94.3) 3,731 (94.8) (83.8) 625 (89.0) In-hospital mortality 52 (2.5) 36 (0.9) (2.7) 18 (2.5) day mortality 58 (2.8) 45 (1.1) (3.0) 21 (2.9) 0.93 a Mean standard deviation. ARDS acute respiratory distress syndrome; CLD chronic lung disease; ONCAB on-pump coronary artery bypass graft surgery; OPCAB off-pump coronary artery bypass graft surgery. ing of CLD severity (none, mild, moderate, severe) and surgery type was constructed. This model compared the severity levels of CLD to patients with no CLD, adjusted for the propensity score and surgery type. Third, a series of four stratified models comparing surgery type within each CLD severity level (none, mild, moderate, severe) was fit to determine whether the effect of OPCAB differs across CLD severity. For continuous outcomes, OPCAB and CLD main effects were compared with general linear models, adjusted for the propensity score. Long-term survival comparisons were made using Cox proportional hazards regression (adjusted) models and Kaplan-Meier product-limit estimates (unadjusted). Kaplan-Meier curves were generated that provide survival estimates at postoperative times. Differences between CLD level Kaplan-Meier estimates were determined by log-rank tests. These estimates include operative deaths. Adjusted long-term survival comparisons were made by using Cox proportional hazards regression to model the instantaneous hazard of death as a function of CLD, surgery type, and their interaction, adjusted for the propensity score and patient age. Hazard ratios were generated for each model term along with 95% confidence intervals. The data were managed and analyzed using SAS version 9.1 (SAS Institute Inc, Cary, NC) and STATA version 9.0 (StataCorp, College Station, TX). Unadjusted comparisons were performed with 2 tests and twosample Student s t tests for categorical and continuous predictors, respectively. All statistical tests were twosided using an alpha level of 0.05 for significance. No adjustments for multiple tests were made. Results Preoperative Risk Factors The OPCAB group consisted of 4,693 patients (66.5%), and the ONCAB group consisted of 2,367 patients (33.5%). A total of 647 patients (9.2%) were characterized as mild CLD, 148 (2.1%) as moderate CLD, and 189 (2.7%) as severe CLD. Preoperative demographics and risk factors are listed in Table 1 and compared for patients with and without CLD according to surgery type (OPCAB versus ONCAB). Patients with CLD were more likely to undergo OPCAB compared to those without CLD (73.2% versus 66.5%; p ). Among patients with CLD, those undergoing OPCAB were older, but had a lower incidence of cerebrovascular disease, left main disease, previous myocardial infarction, previous CABG, and need for intraaortic balloon pump. Among those without CLD, OPCAB patients were older, more likely to be female, and had a higher ejection fraction. They also had a lower incidence of left main disease, previous myocardial infarction, previous CABG, need for intraaortic balloon pump, and higher STS predicted risk of mortality. All other risk factors were similar between groups (Table 1). Unadjusted Outcomes Comparison of unadjusted outcomes for patients without CLD revealed that OPCAB was favorable, with significantly fewer postoperative complications (including pneumonia, sepsis, acute respiratory distress syndrome, stroke, acute renal failure, and atrial fibrillation) as well as decreased in-hospital and 30-day mortality. Among patients with CLD, those who underwent OPCAB had a

5 ADULT CARDIAC 12 KERENDI ET AL Ann Thorac Surg OPCAB IN PATIENTS WITH LUNG DISEASE 2011;91:8 15 Table 3. Multiple Logistic Regression Analysis Comparing Outcomes on Basis of Presence or Absence of Chronic Lung Disease and Surgery Type and Their Interaction Outcome OPCAB AOR Surgery Type p Value CLD AOR CLD p Value Interaction p Value Prolonged ventilation 48 h 0.52 (0.40, 0.68) (1.33, 2.69) Reintubation 0.77 (0.53, 1.12) (1.61, 4.18) Pneumonia 0.66 (0.47, 0.93) (1.29, 3.22) ARDS 0.41 (0.12, 1.44) (0.09, 5.26) Mean ICU stay (h) 31.8 vs vs Home discharge (survivors) 1.07 (0.80, 1.45) (0.22, 0.48) day mortality 0.54 (0.31, 0.93) (0.50, 2.45) AOR adjusted odds ratio; ARDS acute respiratory distress syndrome; CI confidence interval; CLD chronic lung disease; ICU intensive care unit; OPCAB off-pump coronary artery bypass graft surgery. shorter postoperative ventilator time and were more likely to be discharged to home (Table 2). Multiple Logistic Regression Analysis Separate multiple logistic regression models were created to evaluate the effect of surgery type (OPCAB versus ONCAB), presence or absence of CLD, as well as their interaction after adjusting for 39 perioperative risk factors (Table 3). No significant interactions existed between surgery type and CLD status, indicating that although OPCAB was equally beneficial to patients with and without CLD, there was no added disproportionate benefit in those with CLD. Chronic lung disease was associated with significantly longer postoperative ventilator and ICU times (p 0.001), and a greater incidence of reintubation (p 0.001), pneumonia (p 0.002), and non home discharge (p 0.001). Off-pump coronary artery bypass graft surgery significantly reduced postoperative ventilator and ICU times (p for both), as well as the incidence of pneumonia (p 0.018) and 30-day mortality (p 0.028). Outcomes were also examined based on the severity of CLD mild, moderate, or severe. When compared with those without CLD, patients with mild, moderate, or severe CLD all had a greater incidence of prolonged ventilator requirement, reintubation, pneumonia, and non home discharge. In addition, those with severe CLD had a significantly greater 30-day mortality compared with those without CLD (Table 4). A comparison of OPCAB versus ONCAB within each category of CLD severity, however, failed to achieve statistical significance for any end point (Table 5), further suggesting that OPCAB does not interact with COPD severity levels. Survival Analysis One-, 3-, and 5-year survival rates are summarized in Table 6. The adjusted Cox proportional hazards regression model revealed no OPCAB effect (hazard ratio 0.93; p 0.49) on survival. However, compared with patients without CLD, mild CLD patients had significantly shorter survival (hazard ratio 1.47; p 0.002) as did moderate (hazard ratio 2.67; p 0.001) and severe (hazard ratio 4.47; p 0.001) CLD patients. Kaplan- Meier survival estimates significantly differed according to CLD severity level (p 0.001; Fig 1), although surgery type had no significant effect on long-term survival (p 0.06; Fig 2). Comment The results of this single-institution, retrospective cohort study corroborate previous findings that CLD is a significant risk factor for postoperative pulmonary complications and mortality after CABG [17 19]. Specifically, patients with CLD had longer postoperative ventilator and ICU times and were at greater risk for reintubation, Table 4. Multiple Logistic Regression Analysis Comparing Outcomes on Basis of Severity of Chronic Lung Disease (Mild, Moderate, Severe) Outcome Mild CLD vs None AOR Moderate CLD vs None AOR Severe CLD vs None AOR Prolonged ventilation 48 h 1.66 (1.26, 2.20) a 2.71 (1.70, 4.34) a 3.36 (2.25, 5.01) a Reintubation 1.97 (1.39, 2.81) a 3.78 (2.20, 6.48) a 4.58 (2.91, 7.21) a Pneumonia 2.20 (1.60, 3.02) a 3.46 (2.05, 5.85) a 2.36 (1.36, 4.07) a Non-home discharge 1.92 (1.45, 2.56) a 3.71 (2.38, 5.80) a 3.58 (2.38, 5.39) a ARDS 2.22 (0.75, 6.60) 2.48 (0.33, 18.7) 2.21 (0.29, 16.8) 30-day mortality 1.59 (0.90, 2.82) 1.52 (0.47, 4.86) 3.65 (1.73, 7.70) a a Significantly different from no chronic lung disease at p AOR adjusted odds ratio; ARDS acute respiratory distress syndrome; CI confidence interval; CLD chronic lung disease.

6 Ann Thorac Surg KERENDI ET AL 2011;91:8 15 OPCAB IN PATIENTS WITH LUNG DISEASE Table 5. Multiple Logistic Regression Analysis Comparing Outcomes of Off-Pump Coronary Artery Bypass Graft Surgery Versus On-Pump Coronary Artery Bypass Graft Surgery on Basis of Severity of Chronic Lung Disease (Mild, Moderate, Severe) Outcome OPCAB vs ONCAB Mild CLD AOR (N 647) OPCAB vs ONCAB Moderate CLD AOR (N 148) 13 OPCAB vs ONCAB Severe CLD AOR (N 189) ADULT CARDIAC Prolonged ventilation 48 h 0.65 (0.32, 1.31) 1.38 (0.39, 4.94) 0.42 (0.15, 1.15) Reintubation 0.78 (0.32, 1.90) 3.53 (0.62, 20.0) 0.53 (0.16, 1.79) Pneumonia 0.72 (0.33, 1.61) 0.80 (0.20, 3.28) 0.87 (0.20, 3.75) Non-home discharge 0.83 (0.43, 1.60) 0.83 (0.27, 2.51) 1.02 (0.36, 2.90) ARDS 1.66 (0.09, 31.2) Not estimable Not estimable 30-day mortality 0.40 (0.10, 1.62) Not estimable Not estimable AOR adjusted odds ratio; ARDS acute respiratory distress syndrome; CI confidence interval; CLD chronic lung disease; ONCAB on-pump coronary artery bypass graft surgery; OPCAB off-pump coronary artery bypass graft surgery. pneumonia, and non home discharge. We also found that those with severe CLD had significantly higher 30-day mortality when compared with patients without CLD and that the diagnosis of CLD is associated with decreased long-term survival. The primary objective of this study was to determine whether OPCAB would reduce the incidence of postoperative pulmonary complications, particularly in those with a preoperative diagnosis of CLD. Indeed, the riskadjusted occurrences of several important outcomes were reduced in patients undergoing OPCAB, including postoperative ventilator and ICU times, pneumonia, and 30-day mortality. Although these benefits were seen for all patients, including those with and without CLD, we did not show any additional disproportionate advantage of OPCAB for patients with CLD. Postoperative pulmonary complications account for a significant increase in resource utilization and hospital costs in cardiac surgery as these patients have longer ICU and hospital stays and are frequently transferred to rehabilitation facilities for further recovery. Various measures have been proposed to reduce the risk of pulmonary complications, such as preoperative inspiratory muscle training [20], smoking cessation, and the use of systemic corticosteroids and antibiotics in selected cases [18]. Likewise, routine postoperative use of incentive spirometry and inhaled bronchodilators, as well as adequate pain control may contribute to improved outcomes. In addition to these measures, the results of the present study indicate that avoidance of cardiopulmonary bypass may be beneficial. The precise mechanism by which OPCAB contributes to improved pulmonary outcomes is unknown, but is likely to be multifactorial. It has been proposed that the systemic inflammatory state associated with cardiopulmonary bypass results in increased lung vascular permeability [21] and impairment of gas exchange, particularly in patients with chronic underlying lung disease. In addition to directly avoiding these deleterious effects, OPCAB has been shown to result in reduced blood transfusion requirements [4], which would indirectly prevent further transfusion-related lung injury. Less blood loss and increased hemodynamic stability in OPCAB patients may also translate into earlier extubation, thereby avoiding ventilator-related barotrauma and the potential infectious risks associated with prolonged mechanical ventilation. Beyond the utilization of cardiopulmonary bypass, there are additional surgical factors that may contribute to pulmonary complications, which have not been accounted for in this study. It has been postulated that sternotomy and sternal spreading may lead to some Table 6. Kaplan Meier Product-Limit Survival Estimates for Each level of Surgery Type and Chronic Lung Disease Variable 1-Year Survival 3-Year Survival 5-Year Survival No CLD Mild CLD Moderate CLD Severe CLD All OPCAB Patients All ONCAB Patients CLD chronic lung disease; ONCAB on-pump coronary artery bypass graft surgery; OPCAB off-pump coronary artery bypass graft surgery. Fig 1. Survival by chronic lung disease severity. Survival is decreased in patients with mild, moderate (mod), and severe chronic lung disease compared with those without chronic lung disease.

7 ADULT CARDIAC 14 KERENDI ET AL Ann Thorac Surg OPCAB IN PATIENTS WITH LUNG DISEASE 2011;91:8 15 Fig 2. Survival by surgery type: -pump coronary artery bypass graft surgery (OPCAB) versus on-pump coronary artery bypass graft surgery (ONCAB). No significant difference is seen. degree of pulmonary dysfunction by altering chest wall mechanics. Off-pump coronary artery bypass graft surgery is not beneficial in this regard, and in fact may be disadvantageous, as a greater degree of sternal spreading is required, particularly when grafting the lateral wall. Also, when harvesting the internal mammary artery, pleural preservation may have a protective effect on the lungs. Although we do not have data regarding pleural preservation, it has not been our standard practice to do so. In addition, the use of both internal mammary arteries may further confound the issue by violating both pleural spaces. Indeed, patients in the CLD group were less likely to have one (91.9% versus 93.6%) or both (5.6% versus 9.4%; p ) internal mammary arteries used. This study has several additional limitations that must be addressed. Although we have used propensity scoring and multivariable regression analyses to adjust for preoperative risk factors, our data are retrospective and carry the deficiencies inherent to a nonrandomized trial. However, the large number of patients included in this investigation would be difficult to enroll in a prospective study. Despite including 5 years of data and more than 7,000 patients, there may have been important differences in the CLD group that we were unable to demonstrate owing to a lack of power, particularly when analyzing groups based on the severity of CLD. Finally, the classifications of CLD used in this study were based on definitions from the STS Adult Cardiac Surgery Database and not on any uniform objective preoperative assessment of pulmonary function. Nevertheless, our results indicate that the beneficial effects of OPCAB were seen for all patients, not just those with a diagnosis of CLD. In summary, patients with CLD are at significantly greater risk of postoperative pulmonary complications and mortality after CABG, and have decreased long-term survival. Although OPCAB did not reduce mortality in the subgroup of patients with CLD, it was associated with a significant reduction of pulmonary-related complications, including prolonged ventilation, pneumonia, ICU stay, non home discharge, and 30-day mortality in all patients. References 1. Yilmaz M, Saba D, Karal I, et al. Postoperative outcomes after off-pump coronary artery bypass grafting in EuroSCORE lowand high-risk women. Heart Surg Forum 2007;10: Al-Ruzzeh S, Ambler G, Asimakopoulos G, et al. Off-Pump Coronary Artery Bypass (OPCAB) surgery reduces riskstratified morbidity and mortality: a United Kingdom Multi- Center Comparative Analysis of Early Clinical Outcome. Circulation 2003;108(Suppl 1):II Mack MJ, Pfister A, Bachand D, et al. Comparison of coronary bypass surgery with and without cardiopulmonary bypass in patients with multivessel disease. J Thorac Cardiovasc Surg 2004;127: Puskas JD, Williams WH, Duke PG, et al. Off-pump coronary artery bypass grafting provides complete revascularization with reduced myocardial injury, transfusion requirements, and length of stay: a prospective randomized comparison of two hundred unselected patients undergoing off-pump versus conventional coronary artery bypass grafting. J Thorac Cardiovasc Surg 2003;125: Darwazah AK, Abu Sham a RA, Hussein E, Hawari MH, Ismail H. Myocardial revascularization in patients with low ejection fraction or 35%: effect of pump technique on early morbidity and mortality. J Card Surg 2006;21: Sharoni E, Song HK, Peterson RJ, Guyton RA, Puskas JD. Off pump coronary artery bypass surgery for significant left ventricular dysfunction: safety, feasibility, and trends in methodology over time an early experience. Heart 2006;92: Sajja LR, Mannam G, Chakravarthi RM, et al. Coronary artery bypass grafting with or without cardiopulmonary bypass in patients with preoperative non-dialysis dependent renal insufficiency: a randomized study. J Thorac Cardiovasc Surg 2007;133: Dewey TM, Herbert MA, Prince SL, et al. Does coronary artery bypass graft surgery improve survival among patients with end-stage renal disease? Ann Thorac Surg 2006;81: Puskas JD, Sharoni E, Petersen R, et al. Angiographic graft patency and clinical outcomes among diabetic patients after off-pump versus conventional coronary artery bypass grafting: results of a prospective randomized trial. Heart Surg Forum 2003;6(Suppl 1):S Demaria RG, Carrier M, Fortier S, et al. Reduced mortality and strokes with off-pump coronary artery bypass grafting surgery in octogenarians. Circulation 2002;106(12 Suppl 1):I Staton GW, Williams WH, Mahoney EM, et al. Pulmonary outcomes of off-pump vs on-pump coronary artery bypass surgery in a randomized trial. Chest 2005;127: Cox CM, Ascione R, Cohen AM, Davies IM, Ryder IG, Angelini GD. Effect of cardiopulmonary bypass on pulmonary gas exchange: a prospective randomized study. Ann Thorac Surg 2000;69: Montes FR, Maldonado JD, Paez S, Ariza F. Off-pump versus on-pump coronary artery bypass surgery and postoperative pulmonary dysfunction. J Cardiothorac Vasc Anesth 2004;18: Schafer JL. Analysis of Incomplete Multivariate Data. Boca Raton, FL: Chapman & Hall/CRC; Blackstone EH. Comparing apples and oranges. J Thorac Cardiovasc Surg 2002;123: D Agostino RB. Tutorial in biostatistics: propensity score methods for bias reduction in the comparison of a treatment to a non-randomized control group. Stat Med 1998;17: Manganas H, Lacasse Y, Bourgeois S, Perron J, Dagenais F, Maltais F. Postoperative outcome after coronary artery bypass grafting in chronic obstructive pulmonary disease. Can Respir J 2007;14:19 24.

8 Ann Thorac Surg KERENDI ET AL 2011;91:8 15 OPCAB IN PATIENTS WITH LUNG DISEASE 18. Smetana GW. Preoperative pulmonary evaluation. N Engl J Med 1999;340: Rosenfeld R, Smith JM, Woods SE, Engel AM. Predictors and outcomes of extended intensive care unit length of stay in patients undergoing coronary artery bypass graft surgery. J Card Surg 2006;21: Hulzebos EH, Helders PJ, Favié NJ, De Bie RA, Brutel de la Riviere A, Van Meeteren NL. Preoperative intensive inspiratory muscle training to prevent postoperative pulmonary complications in high-risk patients undergoing CABG surgery: a randomized clinical trial. JAMA 2006;296: Messent M, Sinclair DG, Quinlan GJ, Mumby SE, Gutteridge JM, Evans TW. Pulmonary vascular permeability after cardiopulmonary bypass and its relationship to oxidative stress. Crit Care Med 1997;25: ADULT CARDIAC INVITED COMMENTARY Pulmonary complications are a major source of morbidity and mortality after cardiac surgery and have a significantly prolonged intensive care unit (ICU) and hospital stay [1]. Risk factors for pulmonary complications include advanced age, reduced preoperative pulmonary function, current smoking, postoperative pain, poor cough, effusions, abdominal distention, left internal mammary artery harvest with pleural entry, and most importantly poor cardiac function. The use of cardiopulmonary bypass (CPB) is believed to reduce postoperative lung function by apnea and atelectasis during the bypass run, by an increase in interstitial water, and by inflammatory mediators and microemboli damage to the alveolar membrane. In a large cohort study, the use of cardiopulmonary bypass was an independent risk factor for postoperative intubation beyond 48 hours [1]. In a small prospective, randomized study of on-pump coronary artery bypass (ONCAB) and off-pump coronary artery bypass (OPCAB), it was found that the OPCAB group had improved gas exchange and had earlier extubation when compared with ONCAB [2]. Therefore, should we try to avoid CPB, especially in those patients with reduced pulmonary function when performing bypass surgery? The large retrospective study by Kerendi and colleagues [3] tries to analyze the difference between OP- CAB and ONCAB vis a vis pulmonary complications in patients classified preoperatively as having no, mild, moderate, or severe lung disease. They are to be commended for demonstrating the relationship between the risk of pulmonary complications, early mortality, and long-term mortality to the severity of preoperative lung disease. This is important information and may help in our preoperative decision-making process. They also applied robust statistical applications to achieve their conclusion that OPCAB reduced the incidence of pulmonary complications in all patients as compared with ONCAB; although true, this statement requires clarification and should be nuanced. The greatest impact of OPCAB in reducing prolonged ventilation, ventilator time, pneumonia, ARDS, sepsis, in hospital mortality, and 30-day mortality was among those patients without preoperative lung disease. In those with lung disease, only a reduction in prolonged ventilation and ventilation times was noted. When comparing OPCAB with ONCAB within each category of lung disease, there was no significant difference in primary endpoints among groups. One would have believed that if OPCAB reduces pulmonary complications in a group without lung disease, presumably by avoiding cardiopulmonary bypass, then this effect should be even more pronounced and concordant in a group of patients with lung disease. Unfortunately, this does not seem to be the case. It is known that a sternotomy, in addition to causing pain and discomfort, will provoke a reduction in lung volumes and impair rib-cage expansion and coordination [4, 5]. During OPCAB, the sternum may be aggressively retracted and stretched. It is believed that OPCAB decreases lung compliance due to increase lung fluid and rotation of the heart during circumflex grafting [3]. Is it possible that these mechanisms, and others yet identified, are as important in the development of pulmonary complications as the use of cardiopulmonary bypass among patients with pre-existing lung disease? Perhaps the noted reduction in pulmonary complications among those patients without lung disease undergoing OPCAB is more of a reflection of different anesthetic and ICU management and faster extubation rather than avoidance of cardiopulmonary bypass. At the present time, I believe that we are unable to conclude that OPCAB should be performed preferentially among patients with preoperative lung disease. Kevin Lachapelle, MD Cardiac Surgery McGill University Health Center 687 Pine Ave W, Ste S8.30 Montreal, QC, Canada H3A 1A1 kevin.lachapelle@muhc.mcgill.ca References 1. Reddy LC, Grayson AD, Griffiths EM, Pullan MD, Rashid A. Logistic risk model for prolonged ventilation after adult cardiac surgery. Ann Thorac Surg 2007;84: Staton GW, Williams WH, Mahoney EM, et al. Pulmonary outcomes of off-pump vs on-pump coronary artery bypass surgery in a randomized trial. Chest 2005;123: Kerendi F, Halkos ME, Puskas JD, et al. Impact of off-pump coronary artery bypass graft surgery on postoperative pulmonary complications in patients with chronic lung disease. Ann Thorac Surg 2011;91: Locke TJ, Griffiths TL, Mould H, Gibson GJ. Rib cage mechanics after median sternotomy. Thorax 1990;45: Braun SR, Birnbaum ML, Choprs PS. Pre and postoperative pulmonary function abnormalities in coronary artery revascularization surgery. Chest 1978;73: by The Society of Thoracic Surgeons /$36.00 Published by Elsevier Inc doi: /j.athoracsur