rimmed mean ventilation times showed 9.35 hours in the off-pump group versus 10.6 hours in the on-pump group, representing a modest reduction in ventilation time. Similarly, trimmed mean lengths of ICU stay were 1.2 days in the off-pump group versus 1.4 days in the onpump group (p = 0.62).
Total hospital stay remained shorter for off-pump patients with a mean length of 8.5 days versus 10.5 days for on-pump patients (p = 0.33). Trim mean of hospital length of stay was 6.9 days in the off-pump group versus 8.4 days for the on-pump group (p = 0.33) (Figure 2).
There was one death in the off-pump group, leading to an in hospital mortality of 2.7%. This patient had under-
Figure 1. Trim mean of ventilation times in hours for offpump versus on-pump coronary artery bypass grafting.
Figure 2. Comparison of off-pump and on-pump group applying Trim Mean for patients’ length of stay on intensive care unit (ICU), high dependency unit (HDU), ward and hospital stay.
gone emergency conversion from OPCAB to on-pump CAB. In the on-pump group no patient died during hospital stay.
Chest tube drainage was lower for OPCAB (484 milliliter) compared to the on pump surgery (744 milliliter), and this difference was statistically significant (p = 0.04). Accordingly, mean discharge hemoglobin (Hb) in the off-pump group was considerably higher than that in the on-pump group (10.6 g/dl versus 10.0 g/dl, p < 0.01). Blood transfusion was required in 14% of the off-pump patients and 45% of the on-pump patients, with a mean of 0.7 and 1.3 red blood cell units, respectively (p = 0.22).
Our data showed almost identical frequency of atrial fibrillation between groups with 13.51% in off-pump surgery and 13.27% in CABG surgery (p = 0.97). Rates of intra-aortic balloon pump insertion (0.8% versus 0.0%, p = 0.56) and rates for bleeding requiring re-opening (3.54% versus 2.70%, p = 0.64) were both slightly higher in off-pump cases. However there were no sternal wound infections in the off-pump group and two in the on-pump group (1.7%, p = 0.42).
No significant differences were observed for rates of acute renal failure and gastrointestinal complications. Surprisingly the number of pneumothoraces requiring chest drain was significantly higher in the OPCAB group (10.81% versus 2.65%, p = 0.04) while pleural effusion rates were similar (2.70% versus 2.65%, p = 0.31) (Figure 3).
Cerebro-vascular accident as measure for neurological outcome was recorded during hospital stay. There was one CVA in the off-pump group (2.7%) and none in the on-pump group (p = 0.11).
We present the results of our early experience in the offpump technique for patients requiring first time, isolated CABG and compared them with a control group of patients receiving conventional on-pump CABG over the same time frame.
Operation time was only slightly shorter in the offpump group and was not showing clear trend towards shorter (or longer) times throughout the reported time frame. With off-pump cardiac surgery there is a potential for shorter operating time due to circumventing cardiopulmonary bypass. In our study taking specialised equipment onto the table, making connections to external sources and ensuring that this equipment worked properly was time consuming. Therefore this effect is possibly a reflection of our early experience. In retrospect, this definitely is a part of the procedure where time efficiencies could be made. Use of a dedicated theatre team in our early experience could also have made this process more efficient. In the long term, we hope to see much
Figure 3. Complications following off-pump and on-pump heart surgery of 150 analysed patients.
reduced operating times.
Reduction in ventilation time, length of ICU stay and hospital length of stay were important and encouraging findings in this analysis.
Theatre protocol for patients undergoing bypass surgery without use of heart-lung machine was modified slightly with regards to Heparin dosages. Close communication between surgeon and anaesthetic team was mandatory. We observed a reduced need for fluids (data not shown). The protocol for CAB patients of our institution with one day of ICU stay followed by one day in HDU was kept up due to logistic limitations and therefore we believe that there is scope to reduce these times further. It is noteworthy that total hospital stay was significantly shorter in the off-pump group, presumably reflecting much quicker recovery in this group.
Overall mortality and morbidity differed non-significantly in our groups. However, requirement for prolonged mechanical ventilation > 24 h and the need for IABP were significantly higher in conventional CABG group. The largest randomized controlled trial comparing off-pump and on-pump surgery failed to show a difference in mortality or major complications at one month following surgery in low risk patients  . In contrast, a mortality benefit has been observed in large observational studies. The difference is possibly because most RCTs recruit low risk patients whereas the large observational series include patients with significant comorbidities that potentially stand to benefit more from the offpump approach  . The one patient who died in the off-pump group had undergone emergency conversion to on-pump CABG, namely urgently instituted CPB after commencement of anastomoses. This is nearly consistent with the 2.2% off-pump to on-pump surgery conversion rate reported in the Society of Thoracic Surgeons data base study . Emergency conversion to on-pump surgery can have deleterious effects on patient outcomes [17-19] , and is associated with increased perioperative and 3-year mortality, morbidity, and enhanced economic costs (summarized in ). Factors leading to all conversions (elective and emergency) are predominantly hemodynamic compromise (64.1%), as it was the case in our study, followed by anatomical considerations (problematic target sizes) (10.5%), electrophysiological disturbances (9.9%), graft occlusion/ malfunction (7.0%), and other or unknown causes (8.4%) . Strategies to prevent emergency conversion, including elective conversion as another approach with similar outcomes to onpump surgery (ONCAB) or successful OPCAB cases, are being discussed elsewhere  and need emphasized. Our patient did not fall in the very early learning period; the surgeon had operated more than twenty patients OPCAB at that time.
In our study we found both, significantly reduced drainage from chest tubes after surgical procedure and higher haemoglobin levels on discharge from hospital in the off-pump group. Higher transfusion requirements contribute to increased morbidity and mortality risk after cardiac surgery. Operative mortality was significantly in patients who were given blood with (or without re-operation than in patients who did not receive blood products (8.7% blood given and re-operation for bleeding vs. 6.6% no blood given and re-operation for bleeding, 2.1% blood given and no re-operation for bleeding vs. 0.099% no blood used and no re-reoperation). The authors of this propensity matched, prospective cohortinvestigation came to the conclusion that a better attentiveness to surgical techniques for control and decrease bleeding is likely to lower morbidity and mortality .
Postoperative atrial fibrillation (POAF) bears the potential for prolonged hospital length of stay and neurological and renal complications . New onset atrial fibrillation after isolated coronary artery bypass surgery independently predicts mortality (hazard ratio, 1.2)  . In our study POAF was similar in both groups. In contrast to our findings, a recent five-year follow up study demonstrated a significantly higher incidence of AF in off-pump compared to on-pump cases ( 35% versus 4%)  . This rate appears to be a lot higher than in both of our study groups. Another randomized study suggested a reduced risk for atrial fibrillation .
Although there were no clear differences in our analysis concerning neurological outcome, it is widely suggested that this can be improved in OPCAB and other minimized techniques in surgical revascularization [24- 26]. Brizzio et al. recently reported a risk of stroke in off-pump compared to on-pump patients as low as 1.0% versus 2.4% (p < 0.01) . The overall stroke rate was 1.7%, which is similar to previous reports. Interestingly, the authors demonstrated, on analysis, that after controlling for certain preoperative risk factors such as history of stroke, age more than 70 years, left ventricular ejection fraction < 35%) and the Society of Thoracic Surgeons (STS) mortality risk score, pump status was significantly and independently associated with the mortality outcome after a stroke, with off-pump patients experiencing an 84% (first time and re-operation), and 80% (only first time operation), retrospectively, reduced risk of death after stroke . In another study that entailed subgroup analysis of outcomes from the SYNTAX trial, the event rate of stroke one year after surgery tended to be lower with the off-pump no-touch technique when compared with the conventional surgical arm (0.8% vs. 2.2%) and similar to the percutaneous interventional arm of the SYNTAX trial (0.6%) . Furthermore, better cognitive outcomes in off-pump compared to on-pump coronary revascularization have been observed at three months after surgery although this difference disappears after 1 year and remains absent at 5 years (OCTOPUS Study) . Notably, the stroke rate after five years was indistinct with 1.4% in the off-pump group and 3.6% in on-pump group (p = 0.28) .
The significantly higher rate of pneumothoraces in OPCAB cases was surprising and cannot be explained. The only difference in the two groups is that in OPCAB both pleurae were opened in all cases in contrast to on pump with right pleura opened in only a few cases. Opening both pleurae in off-pump surgery helps to improve hemodynamics and gives optimal positioning of the heart . Cautious intra and postoperative handling as well as more statistical analysis will be necessary to shed light on this issue in the near future.
A significant advantage of off-pump surgery is that it uniquely permits the use of an aortic no-touch technique [28,29] by avoiding aortic manipulation with use of either pedicled or composite arterial grafts, with or without anastomotic devices. The aortic no-touch technique was employed in several of the patients included in our study. In SYNTAX trial, the off-pump CABG no-touch technique was compared with conventional CABG or PCI. It showed superior results in terms of repeat revascularization (1.3% OPCAB No-Touch with total arterial Y-graft, 5.9% CABG SYNTAX, 13.5% Taxus SYNTAX), symptomatic graft occlusion (0.8% vs. 3.4% vs. 3.3%), and also in the percentage of major adverse cardiovascular or cerebrovascular events (5.3% vs. 12.4% vs. 17.8%)  .
Optimisation of hemodynamics, stabilization and visualization allow for optimal condition to construct anastomoses. The stabilizers used were up to date and significant learning curves reported in the literature were reported for periods of major technique development, as introduction of deep traction sutures and incorporation of modern stabilizers  . We used intraluminal shunts in OPCAB cases and consider these highly advantaging for exposure of intima from both native and conduit vessel, thereby avoiding inadvertent inclusion of the back wall of the native site in the suture line.
The incorporation of OPCAB technique into established surgical practice is done differently throughout the institutions. S. Hoff outlined methods to improve adoption of OPCAB as the following appropriate patient selection, individualized grafting strategy, peer to peer training of the entire team, graded clinical experience (on-pump beating heart, cannulated OPCAB, wet pump, dry pump, and minimally invasive/ advanced OPCAB) . These points were accomplished to satisfactory degree.
Patient selection is generally considered crucial for safe adoption of the off-pump technique. OPCAB-experienced surgeons suggest exclusion of patients with depressed left ventricular function, left main disease, and three-vessel disease in a surgeon’s early experience  . Yet no indications and contraindications were defined by any authorized institution for training and routine OPCAB surgery. One patient in the rather early experience was high risk. During the study period the surgeon established an overall rate of OPCAB (24.67%). It is our feeling that the learning curve can be negotiated with a fairly high initial frequency of ONCAB. However, a reasonable future option of technique incorporation is the formalized training in surgical adoption of the OPCAB technique, prevention and management of emergency conversion as well as leading of a multidisciplinary team . Providing structured training in OPCAB is an effort being made by The American Association for Thoracic Surgery (AATS) and The Society of Thoracic Surgeons (STS). The regarding joint committee (Committee on New Technology Assessment) developed a detailed program, that includes didactic sessions, live animal and cadaver training, observational visits to the institutions of surgeons who are experienced in off-pump surgery and visits by those surgeons (preceptors) to the trainees’ home institutions . Training to surgical residents we consider paramount in establishing the OPCAB technique in an institution. Eventually additional monitoring using control or cumulative sum (CUSUM) charts may be of great value in the process of introduction of OPCAB and accreditation. These provide a case by case monitoring and, hence, allow for tight quality control, if boundary definitions are put carefully .
Our study contains several limitations. Most of all it was a non-randomized retrospective study from which selection bias and diminished comparability result. Although patients were not randomized or matched in any formal manner, we consider that the groups were sufficiently comparable for the purposes of this study. Due to the rough measurement parameter (cerebrovascular accident) neurological outcome was not evaluated in detail, as can be done by cognitive performance tests. Due to the character of the study we did not examine graft patency which is a hotly debated outcome measure in comparison between OPCAB and on-pump CAB.
The goal of this study was to audit surgical outcomes during the early experience with the off pump technique. We have shown that experienced surgeon could incorporate the off-pump technique safely into surgical practice with relatively little extra training using careful patient selection and up to date materials. The experience will allow us to improve upon these results and bring the benefits of off-pump coronary surgery to the population of patients in need for this option. This article was intended to encourage surgeons yet indistinctive about the technique to start incorporating OPCAB into ones surgical technique. A quick learning curve can be achieved due to careful patient selection, high frequency of OPCAB, use of modern surgical technique and devices, training at teaching centers and a dedicated theatre team.
C. N. thanks Prof. Dr. Ing. Günther Kundt, Institute for Biostatistics and Informatics in Medicine and Ageing Research, University of Rostock, Medical Faculty for his advices.