|Home | About | Journals | Submit | Contact Us | Français|
We aim to describe our experience in coronary artery bypass grafting (CABG) with or without cardiopulmonary bypass by comparing intraoperative and postoperative outcomes.
From January 1993 to June 2013, 3097 patients underwent consecutive emergency and scheduled CABG surgery. A total of 1770 patients underwent on-pump CABG (ONCABG) and 1327 off-pump CABG (OPCABG). A propensity score matching was performed to identify appropriate matched-pair patients; univariable and multivariable logistic regression analyses were performed to assess significant predictors of hospital and 30-day morbidity and mortality composite end-points. Morbidity composite end-point was defined as any renal, pulmonary, cardiovascular and neurological complication that occurred during hospital stay. We collected all-cause mortality data during the study period.
We identified 1004 patients in each group. There were no significant differences in thirty day mortality, 2.8 vs 3.8%, in OPCABG and ONCABG, respectively (P = 0.21). Cardiovascular, neurological, respiratory and renal complications were more frequent in the ONCABG group: 13.9 vs 8.7% (P < 0.001), 3.9 vs 2.2% (P = 0.03), 13.5 vs 7.5% (P < 0.001), 7.1 vs 5.3% (P = 0.095), respectively. The long-term all-cause mortality rate was 12.3 vs 12.9% in the OPCABG versus ONCABG group (P = 0.42), respectively. In both uni- and multivariable analysis preoperative renal failure, chronic obstructive pulmonary disease and ONCABG were independent predictors of mortality and morbidity composite end-points.
OPCABG is associated with less postoperative morbimortality and shorter hospital and intensive care unit length of stay. ONCABG resulted as an independent predictor of morbidity and mortality composite end-point. No statistically significant differences were observed in long-term all-cause mortality between groups.
Coronary artery disease is the leading cause of cardiovascular morbidity and mortality in the general population . In some severely-ill patients medical therapy confers a poor survival advantage relative to surgical revascularization. The goal of all surgical revascularization procedures is to perform a technical procedure with minimal morbidity and mortality, enhanced long-term survival and maximal freedom from recurrent symptoms and need for further revascularization procedures. On-pump coronary artery bypass graft (ONCABG) continues to be the gold standard to achieve these goals. Despite improved techniques and surgical experience, part of the morbidity related to CABG is caused by the cardiopulmonary bypass (CPB) , prompting off-pump coronary artery bypass graft (OPCABG). However, there are robust data in the literature suggesting that OPCABG is not superior to ONCABG regarding early outcomes and might be associated with inferior early and late graft patency rates and likely compromised long-term survival [3–5]. On the other hand, two recent large, international and randomized trials have shown no difference in 30-day and 1-year clinical outcomes between on- and off-pump surgery when performed by experienced surgeons [6, 7]. Complete off-pump procedures in the hands of highly trained teams appear to be associated with a reduced risk of early morbidity, such as stroke, wound and respiratory infections, as well as fewer transfusions and shorter hospital stay [8–10].
We aim to describe our experience in ONCABG versus OPCABG surgery comparing intraoperative and postoperative outcomes at 30-day and long-term all-cause mortality after 20 years of surgical experience in a single institution in which 45% of coronary revascularizations were performed under OPCABG.
This is a retrospective, observational cohort study of prospectively collected data from consecutive patients who underwent isolated CABG at the Consorcio Hospital General of Valencia between January 1993 and June 2013. The data collection was extracted from our database (PalexData®, Barcelona, Spain). Long term survival information was provided by the National Registry of Health Care System. The resulting base sample contained detailed clinical information on 3097 patients, of whom 1327 underwent OPCABG and 1770 underwent ONCABG. In order to reduce the effects of treatment selection bias and potential confounding factors, we used a propensity score matching analysis by building a binary logistic regression model with the main preoperative risk variables and comorbidities, which resulted in 1004 patients who underwent OPCABG matched with an equal number of patients who underwent ONCABG. Surgical procedure selection was at the discretion of the operating surgeon based on patients’ comorbidities perceived as high risk for CPB and personal preferences. The procedures were performed by more than one surgeon and all of them have more than 10 years of experience and have been trained in OPCABG. Patients with concomitant cardiac procedures, such as valve replacement or repair, carotid endarterectomy or CABG reoperations were excluded. Long-term follow-up for all-cause mortality was obtained from Local and National Health Care System records.
Anaesthetic and surgical techniques were standardized for all patients. Patients undergoing ONCABG, CPB was instituted with the use of ascending aortic cannulation and a venous cannulation of the right atrium. CPB was carried out with non-pulsatile flow (in most cases), alpha-stat pH blood gas management, haematic cardioplegia with topical hypothermia, moderate systemic hypothermia (32–34°C) and pump flow rates to achieve a mean arterial pressure of 60–80 mmHg. The membrane oxygenator was primed with 1000 ml of crystalloid solution, 500 ml of Gelafundine® (B.Braun Melsungen AG, Melsugen, Germany), 250 ml of 20% mannitol and 5000 UI of heparin. Myocardial protection was achieved with intermittent hyperkalaemic antegrade and retrograde warm blood cardioplegia. For OPCABG, adjustable tissue stabilizer was used to expose the coronaries and provide stabilization that enhances visibility at the anastomotic site. During surgery heparin at 300 IU/kg for the ONCABG and 150 IU/kg for the OPCABG was administered. Activated clotting time was maintained over 480 s for ONCABG and over 300 s for OPCABG. The effect of heparin was reversed with protamine sulphate at a ratio of 1 : 1 after the distal anastomoses was finished in the OPCABG and after coming off-pump in ONCABG. In both techniques proximal anastomoses were performed with a partial occluding clamp in most patients. Postoperative management of all patients followed standard practice guidelines from surgery to discharge. This included admission to the intensive care unit (ICU) from the operating room, with subsequent transfer to a ward. In the ICU, patients were managed according to the unit protocol implementing fast-track extubation.
Continuous data were expressed as mean ± SD, and categorical data were expressed as percentages. The Kolmogorov–Smirnov test was used to check for normality of data in the two groups before further analysis. Differences between OPCABG and ONCABG were compared with a χ2 statistical test for categorical variables and T or Wilcoxon rank sum tests, as appropriate.
A propensity score matching was performed to identify appropriate matched-pair of patients (1 : 1) between groups by building a binary logistic regression model with the main preoperative risk variables and comorbidities defined as above. We performed a one-to-one matched analysis without replacement on the basis of the estimated propensity score of each patient. Table Table11 presents all baseline characteristics of covariates under consideration for models. After propensity score matching, a general linear regression model with the pairs as random effects was used to test for differences between treatment groups. The goodness of fit of the method was evaluated by Hosmer–Lemeshow test (P = 0.31). The discriminative power of the model was evaluated by a receiver operating characteristic curve, contrasting the probability of belonging to the group of OPCABG according to logistic regression against the real location of each patient in each group and resulted in an area under the curve of 0.75 [95% confidence interval (CI) 0.70–0.89].
Stepwise univariable and multivariable logistic regression analysis was performed to assess independent predictors of the 30-day morbidity composite end-point (renal, pulmonary, cardiovascular and neurological complications) and the morbimortality composite end-point (mortality at 30 days and morbidity). A significance level of 0.2 was used to enter in the multivariable logistic regression analysis. Results are reported as percentages and odds ratio and 95% CIs. Overall survival was estimated by using the Kaplan–Meier method and was expressed as percentage. P-values are two-sided for the overall population, a value of 0.05 was considered statistically significant. Statistical analyses were performed with SPSS statistical software (version 18; IBM, corporation, Armonk, NY, USA).
Baseline characteristics of CABG patients are reported in Table Table1.1. A total of 1327 patients underwent OPCABG and 1770 underwent ONCABG. Compared with the ONCABG group, patients in the OPCABG group were older and had a statistically significant higher prevalence of hypertension, peripheral artery disease, history of previous myocardial infarction and previous interventional percutaneous revascularization (P < 0.05).
Of the total CAGB procedures performed, 42.8% were carried out using OPCAB during the study period.
After propensity score analysis, 1004 pairs of patients were matched. In the matched cohorts, no significant differences between the two groups for any covariate were observed (Table (Table1).1). The average grafts performed was higher in the on-pump group, 3.50 ± 0.96 vs 2.87 ± 0.99 (P < 0.001). No significant differences in anterior territory revascularization (left anterior descending and diagonal arteries) were found, 98 vs 97.6% (P = 0.54) in ONCABG versus OPCABG, respectively. The lateral and postero-inferior wall were significantly less revascularized in the OPCABG versus ONCABG: 64.5 vs 79.2% (P < 0.001) and 60.6 vs 76.3% (P < 0.001), respectively. A statistically significant difference was found when we compared the number of grafts in the OPCABG group if surgery was performed during the initial 10 years of surgical experience versus the last 10 years (2.7 ± 1 vs 3.1 ± 0.9, respectively, P < 0.001).
There were no significant differences in mortality between groups, 2.8 vs 3.8% in OPCABG versus ONCABG, respectively (P = 0.21). Concerning clinical outcomes: cardiovascular 13.9 vs 8.7% (P < 0.001), neurological 3.9 vs 2.2% (P = 0.03), respiratory 13.5 vs 7.5% (P < 0.001) and renal complications 7.1 vs 5.3% (P = 0.95) were more frequent in the ONCABG group (Table (Table22).
ICU and hospital length of stay (LOS) were longer in the ONCABG group: 4.1 ± 2.6 vs 3.4 ± 2.3 days (P < 0.001) and 9.7 ± 5.8 vs 7.8 ± 4.1 days (P < 0.001), respectively.
Causes of mortality are reported in Table Table33.
The mean follow-up period was 7.4 ± 4.3 years for OPCABG and 9.4 ± 4.2 years for ONCABG (P < 0.001). The long-term record for all-cause mortality was 12.3 vs 12.9% in the OPCABG versus ONCABG group, respectively (P = 0.42) (Fig. (Fig.11).
In both uni- and multivariable analysis, preoperative renal failure [odds ratio (OR) 2.5 (1.4–4.68)] chronic obstructive pulmonary disease [OR 1.58 (1.2–2.4)] and ONCABG [OR 1.9 (1.5–2.4)] were independent predictors of morbidity and morbimortality composite end-points (Tables (Tables44–7).
CABG can be performed with or without CPB. However, the preferable technique is unclear and the results are controversial.
Our study, based on a large cohort of patients, shows that OPCABG is a safe technique for patients with ischaemic heart disease and is associated with a non-inferior morbidity and mortality and shorter length of ICU and hospital stay. However, the number of grafts performed is higher in the ONCABG and the lateral and postero-inferior walls are significantly less revascularized in the OPCABG group. ONCABG was found to be an independent predictor of morbidity composite end-point and mortality and morbidity composite end-point.
Despite the continuous evolution, CPB has potential harm as a result of physiological processes and can result in tissue damage and organ dysfunction. As the interest for OPCABG has been gradually increasing, several concerns have been raised and authors are still trying to answer if OPCABG reduces hospital morbidity and mortality, if OPCABG has the same outcomes especially regarding security, long-term survival and quality of grafts and if it is possible to perform the full revascularization using OPCABG .
Murzi et al.  studied after propensity score matching 584 pairs of patients undergoing OPCABG and ONCABG. They observed lower in-hospital mortality (0.5 vs 2.9%, P= 0.01), lower incidence of stroke (0 vs 0.9%, P = 0.02), postoperative renal dysfunction (4.9 vs 10%, P = 0.001), pulmonary complications (10.2 vs 16.6%, P = 0.002) and fewer infectious complications (3.5 vs 6.2%, P = 0.03) in OPCABG compared with ONCABG. They also observed less atrial fibrillation (20.6 vs 26.8%, P = 0.01) and less inotropic support (38 vs 46.5%, P = 0.03) in the off-pump patients.
Liu et al.  in a recent propensity score analysis of 54 matched-pair patients demonstrated lower incidence of postoperative myocardial infarction and similar mortality and major complications in both groups. Our results are consistent with data reported [13, 14] and, in our propensity matched groups, we also observed less cardiovascular, neurological and pulmonary complications. We also demonstrated less infectious and renal complications, albeit without statistical significance probably due to sample size and low number of events. The haemodynamic instability that may occur while performing revascularization on a beating heart is proposed as the main reason for not using the OPCABG technique. As technology has evolved with new stabilization tools and the experience of the surgical team and haemodynamic management have been improved, this complication might have been reduced . It may explain the reduced cardiovascular, pulmonary and neurological complications reported in our data after 20 years of experience. Consistent with our data, CPB has been reported to be an independent predictor of in-hospital mortality .
Furthermore, an randomized controlled trial (RCT) of patients over 75 years old randomized to ONCABG and OPCABG did not find any significant difference with respect to a composite outcome of death, stroke, myocardial infarction or new renal replacement therapy at 30 days and at 12 months. However, repeat revascularization occurred more frequently in the OPCABG . Similarly, meta-analysis of RCT and observational studies did not find significant improvements in early mortality and morbidity in the off-pump group  and in a 1-year follow-up .
Although many studies showed that early mortality of OPCAB is comparable with the one obtained with the use of CPB, the results related to the medium and long term are even more controversial. By assessing the follow-up between 1 and 3 years of 401 patients who participated in two RCTs (BHACAS I and II) Angeli et al.  observed that survival free of cardiac events, including death, was similar in surgeries with and without CPB. A similar result was observed in the MASS III  study after 5 years of follow-up, with a trend towards greater survival in patients undergoing surgery without CPB, which reached significance in the fifth year of follow-up, but with no significant difference in the seventh year.
Some the most important concerns related to long-term outcome are the early graft patency and the quality of anastomoses in the beating heart with the off-pump technique. A recent meta-analysis by Moller et al. , which included 86 trials with a total of 10 716 participants, showed that OPCABG increased all-cause mortality compared with ONCABG (3.7 vs 3.1%, P = 0.04) and resulted in fewer distal anastomoses (MD −0.28; P < 0.0001). In addition, they did not demonstrate any significant benefit regarding myocardial infarction, stroke, renal insufficiency or coronary reintervention in the off-pump group. Similarly, Hattler et al.  and Takagi et al.  showed a significantly lower patency rate for grafts and less effective revascularization in the OPCABG group. At 1 year after surgery, patients with less revascularization had higher adverse event rates .
In our study, we did not observe significantly higher 30-day mortality in the OPCABG group but we found fewer distal anastomoses performed and an inferior revascularization rate in the inferior and posterior wall. However, we observed an increase in the mean number of grafts performed over the years, which also could explain our good global outcomes with the OPCABG technique. Indeed, our short- and long-term all-cause mortality was not different between groups despite the fact that the number of grafts initially was lower in OPCABG. It is noteworthy that revascularization of anterior territory was not different between groups.
In the latest meta-analysis by Takagi and Umemoto  which included 22 studies, the number of grafts per patient and the index of completeness of revascularization was significantly greater for the ONCABG group. Previously, Takagi et al. had demonstrated a statistically significant 37% reduction in follow-up mortality with complete revascularization compared with incomplete revascularization (heart rate, 0.63; 95% CI, 0.53–0.75; P < 0.00001) . Furthermore, they demonstrated a statistically significant 7% increase in long-term (more than 5 years of follow-up) all-cause mortality with OPCABG. However, the pooled analysis of five randomized trials demonstrated a statistically non-significant 14% increase in mortality with OPCABG. In contrast, we had a mean follow-up of 7.4 years and we did not observe differences in all-cause mortality.
Proponents of OPCAB have suggested that earlier studies which data have been included in the meta-analysis did not use modern stabilizers, heart-positioning devices or intracoronary shunts and the surgical and anaesthesiologist's experience was inadequate. In large randomized trials involving OPCABG surgeons who had at least 2 years of experience and who had performed at least 100 procedures, there was no difference in death, mortality or major adverse cardiac events between OPCABG and ONCABG surgery [6, 23].
In our hospital, an explanation for good outcomes might be due to the fact that OPCABG is a routine procedure, in which all professionals are consistently involved. Moreover, specific protocols for intraoperative anticoagulation therapy, postoperative antiplatelet therapy, perioperative anesthetic management and availability of immediate cardiopulmonary bypass if needed have been developed. Cardiac surgeons are experienced in OPCABG and outcomes have improved substantially from the beginning of OPCABG procedures to date.
This is a retrospective observational analysis of our institutional database and has the limitations of the available collected information and the bias inherent to a retrospective study. Although patient selection bias is sought to be solved with a propensity score matching, the fact that the decision to perform a procedure is based sometimes in the perceived high comorbidity for CPB might introduce bias difficult to be controlled in this study. However, some of our surgeons performed routinely all the procedures under OPCABG. The learning curve and its impact on selection bias have not been assessed in the study. The heterogeneous enrollment over time in both groups is also a limitation due to the fact that changes in standard of care could also have had an impact on our outcomes, which we could not correct for with propensity score matching.
We used the number of grafts to analyse revascularization but not the definition of complete revascularization specifically for every patient or any other completeness of revascularization index. This feature could have introduced a selection bias. Moreover, the trends towards more grafts over time in the OPCABG could have been a source of bias. The impact on morbidity and mortality is unclear although we do not observe differences over time.
Even though the multivariable logistic regression identified the use of CPB as an independent predictor of morbidity and morbimortality, we included major and non fatal cardiovascular, neurological, renal and pulmonary morbidity to increase sensitivity to the effect of CBP.
Our 30-days follow-up for complications and mortality is clearly a limitation due to the fact that any fatal event immediately after this period might have been related to the surgery and we could have missed this event.
Although long-term follow-up was achieved for all patients in our series and deaths were comprehensively recorded, we could not define the precise cause of death for every patient. Differences in cardiovascular death between groups could not be determined. Although no statistically significant differences were observed in long-term all-cause mortality between groups, differences in overall survival between both can still not be excluded.
We did not collect the data of the reconversion rate on an emergency basis, which has also been shown to increase morbidity and mortality.
We did not collect the postoperative bleeding and the transfusion rate data, which might have been proven to be predictors of worse outcomes and could be different in both techniques.
In conclusion, our study indicates that both coronary revascularization techniques are safe options and OPCABG. We could not show statistically significant differences in outcomes between both of them. However, OPCABG is associated with less postoperative morbimortality and shorter hospital and ICU LOS. OPCABG may be an alternative when there are contraindications for cannulation of the aorta or for CPB. Long-term follow-up showed no differences in all-cause mortality despite the reduced number of grafts performed in the OPCABG during the first years of study.
It would be important for cardiac surgical societies to develop guidelines for the practice of OPCAB and establish criteria for surgeon training, patient selection, institutional requirements, standardization of operative techniques and postoperative management to guarantee short- and long-term outcomes for OPCABG comparable with ONCAB, which remains the gold standard for coronary surgical revascularization.
Conflict of interest: none declared.