Ann Thorac Surg 2002;73:1866-1873
© 2002 The Society of Thoracic Surgeons
Original article: cardiovascular
Evaluation of the effectiveness of off-pump coronary artery bypass grafting in high-risk patients: an observational study
Martin H. Chamberlain, FRCSa,
Raimondo Ascione, MDa,
Barnaby C. Reeves, DPhilb,
Gianni D. Angelini, FRCS*a
a Bristol Heart Institute, University of Bristol, Bristol UK
b Health Services Research Unit, London School of Hygiene and Tropical Medicine, London, United Kingdom
Accepted for publication February 18, 2002.
* Address reprint requests to Dr Angelini, Bristol Heart Institute, University of Bristol, Bristol Royal Infirmary, Bristol BS2 8HW, UK
e-mail: g.d.angelini{at}bristol.ac.uk
 |
Abstract
|
|---|
Background. Coronary artery bypass grafting in high-risk patients carries substantial morbidity. We compared the effectiveness of off-pump revascularization with that of conventional coronary artery bypass grafting using cardiopulmonary bypass and cardioplegic arrest in consecutive high-risk patients.
Methods. From April 1996 to December 2000, clinical data for consecutive patients undergoing coronary artery revascularization were prospectively entered into a database. Data were extracted for all patients considered to be high risk, defined as the presence of one or more of ten adverse prognostic factors. Hospital mortality and early morbidity were compared between two groups of patients, the on-pump and off-pump groups.
Results. The study group comprised 1,570 consecutive high-risk patients, 332 (21.1%) of whom underwent an off-pump operation. Patients in the on-pump group had fewer high-risk factors and lower Parsonnet scores and were less likely to be 75 years of age or older, to have peripheral vascular disease or hypercholesterolemia, or to have sustained a previous transient ischemic attack. However, they were more likely to be assigned to a higher Canadian Cardiovascular Society class and had more extensive coronary artery disease and were more likely to have unstable angina, to require urgent or emergency operations, and to receive more grafts than those undergoing off-pump procedures. Unadjusted odds ratios for intensive care unit or high-dependency unit stay, total length of stay, blood loss of more than 1,000 mL, postoperative hemoglobin and transfusion requirement all showed a highly significant benefit for the off-pump group (p 
0.005; odds ratios, 0.33 to 0.65). After adjustment for prognostic variables, odds ratios remained essentially unaltered (adjusted odds ratio estimates 0.36 to p < 0.05) except for blood loss of more than 1,000 mL (adjusted odds ratio estimate, 0.82; p = 0.22). Sensitivity analyses confirmed the robustness of these findings.
Conclusions. Off-pump coronary artery bypass grafting is safe, effective, and associated with reduced morbidity in high-risk patients compared with conventional coronary artery revascularization.
|
Introduction
|
|---|
Despite improvements in surgical, anesthesia, and cardiopulmonary bypass techniques, coronary artery bypass grafting in high-risk patients is still associated with substantial morbidity [1–4]. In a recent randomized controlled trial, we [5, 6] demonstrated that performing revascularization on the beating heart without cardiopulmonary bypass reduces morbidity in patients having elective procedures. However, little is known about the effectiveness of off-pump coronary artery bypass grafting in high-risk patients. Although the effects of off-pump surgical procedures have also been investigated in a large database study that was more inclusive than our randomized controlled trial [7], this study did not specifically examine the benefits for high-risk patients as a separate group and was not able to control for diverse prognostic factors because compre-hensive prognostic data were not available in the database.
As the results of the randomized controlled trial and other studies on subsystem organ dysfunction [5, 8–13] became apparent, surgeons at our institution with the expertise to perform off-pump operations were reluctant to take part in a second randomized trial including high-risk patients. Therefore, they started to perform off-pump operations on high-risk patients on the assumption that such patients had even more to gain from a less invasive technique than those at less risk of morbidity. The use of a high-quality prospective database to document all patients undergoing coronary artery revascularization in our institution created the opportunity to evaluate the effectiveness of off-pump versus on-pump coronary artery bypass grafting in high-risk patients using a cohort study design.
|
Material and methods
|
|---|
Patient selection and data collection
A standard set of perioperative data is collected prospectively for every patient undergoing coronary artery revascularization at our institution. This report analyzes data (to distinguish from mortality rate of 2.4%, reprinted later for all high-risk patients) for patients having coronary artery bypass grafting between April 1996 and December 2000 (hospital mortality rate for all patients undergoing coronary artery bypass grafting, 1.1%). The dataset includes five different sections to be filled in consecutively by the anesthetist, the surgeon, the intensive care unit (ICU), the high-dependency unit (HDU), and the ward nurses. Data are entered into a database (Patient Analysis & Tracking Systems: Dendrite Clinical Systems, London, UK).
Any patient with one or more of the following characteristics was considered to be at high risk for early mortality or morbidity: age of 75 years or more; ejection fraction lower than 0.30; recent myocardial infarction (<1 month); congestive cardiac failure (current or past); previous stroke; serum creatinine level of 150 µmol/L or higher; respiratory impairment (current chronic obstructive airways disease or asthma); peripheral vascular disease; redo bypass procedure; and intraoperative endarterectomy. Data were extracted from the database for consecutive high-risk patients who had undergone coronary artery revascularization during the study period.
Allocation to the on-pump or off-pump group was based on the preference and the expertise of the surgeons carrying out the operations and, during the early part of data collection when experience was still being gained with off-pump techniques, on coronary anatomy and number of grafts required. The decision about which method to use was made after the chest had been opened and the anatomy of the coronary vessels could be explored.
Anesthesia and operative technique
Anesthesia technique and heparin sodium management were standardized and have previously been reported [10]. In the on-pump group, cardiopulmonary bypass was instituted using ascending aortic cannulation and two-stage venous cannulation of the right atrium. Myocardial protection was achieved with intermittent antegrade hyperkalemic warm blood cardioplegia [10]. In the off-pump group, we used two previously described methods of exposure and stabilization of the heart [14, 15]. Postoperative clinical management was carried out according to a standard unit protocol as previously reported [10].
Management of discharge
We aim to discharge patients who have had coronary artery bypass grafting on the fifth postoperative day. The suitability of patients to be discharged either home or to another hospital for further convalescence is made by an independent physician according to our unit protocol. The decision to discharge is based on a satisfactory routine checkup on day 4 consisting of clinical examination, full blood cell count, urea and electrolyte levels, electrocardiogram, and chest roentgenogram. If the patient is medically unfit on day 5, hospitalization is prolonged, and further investigations may be performed depending on the clinical status.
Clinical data collection, monitoring, and definitions
Data characterizing perioperative clinical outcome were entered prospectively into the Patient Analysis & Tracking System. Hospital mortality included all deaths within 30 days of operation irrespective of where the death occurred and all deaths in the hospital after 30 days among patients who had not been discharged after the index operation. Perioperative myocardial infarction, ST segment changes, pacing, arrhythmias, and inotropic requirement were recorded and defined as previously reported [10]. Pulmonary complication included chest infection, ventilation failure, reintubation, and tracheostomy [10]. Postoperative blood loss was defined as total chest tube drainage [16]. Neurological complication included permanent and transient strokes [17]. Renal complication included acute renal failure, defined as the requirement of hemodialysis or an elevated creatinine level (>200 µmol/L). Infective complication included septicemia and sternal and leg wound infections, defined by positive culture and requirement of antibiotic therapy [10].
Statistical analysis
Conversions from an off-pump to on-pump operation were analyzed in the off-pump group, ie, by intention-to-treat. The analyses were carried out in stages. First, several known prognostic factors were compared between the on-pump and off-pump groups. Any prognostic factor that showed evidence of an association with type of operation (p < 0.20) was considered to be a potential confounding factor. Second, outcomes for the on-pump and off-pump groups were compared using multiple logistic regression (STATA version 7.0) both before and after adjusting for possible confounding factors. All regression models included consultant teams as a fixed effect, thereby taking account of clustering of patients within teams and allowing the outcome frequency to vary by team. Finally, two sensitivity analyses were conducted to test the robustness of the findings: one excluding salvage and emergency operations and one investigating the effect of calendar time (before and after April 1, 1999) and the interaction between calendar time and type of operation.
Some continuous outcomes were dichotomized because the distributions were severely skewed, eg, total length of stay (7 days versus >7 days), blood loss (1,000 mL versus >1,000 mL), and postoperative hemoglobin level (<10 g/dL versus 
10 g/dL). Adjusted models included all high-risk or prognostic factors that were observed to be associated with type of operation (p < 0.2). Exact p values and confidence intervals are presented throughout. The large number of prognostic variables and outcomes of interest resulted in a large number of statistical comparisons. No correction was made for multiple comparisons, but our interpretation of the findings takes into account the consistency of the findings and their magnitude as well as their significance.
|
Results
|
|---|
A total of 1,570 patients with one or more high-risk characteristics underwent coronary artery bypass grafting; 1,238 (78.9%) had a conventional on-pump procedure and 332 (21.1%), an off-pump procedure. Six consultant teams carried out the operations. One consultant team did not perform any off-pump operations during the study period. The proportion of off-pump operations done by the other five teams ranged from 5% to 77%. The proportion of off-pump operations increased steadily during the study period, as did the average number of grafts in off-pump operations (Fig 1). There were three conversions from off-pump to on-pump operations.
The distributions of a number of prognostic variables in the on-pump and off-pump groups are shown in Table 1.
Several of the prognostic characteristics were distributed unevenly between groups (p < 0.1), although neither group included a higher proportion of patients at highest risk or with greater comorbidity prior to operation. Patients in the on-pump group had, on average, fewer high-risk factors and a lower Parsonnet score and were less likely to be 75 years of age or older, to have peripheral vascular disease or hypercholesterolemia, or to have had a previous transient ischemic attack. However, on average, they were more likely to be assigned to a higher Canadian Cardiovascular Society class; had more extensive coronary heart disease; and were more likely to have unstable angina, to require an urgent or emergency operation, and to receive more grafts than those in the off-pump group.
The distributions of outcomes between groups are presented in Table 2,
and the unadjusted and adjusted effect sizes for off-pump versus on-pump revascularization are shown in Table 3. The unadjusted odds ratios (ORs) for length of stay in the ICU or HDU, total length of stay, blood loss of more than 1,000 mL, pulmonary complication, absence of sinus rhythm, postoperative hemoglobin level of less than 10 g/dL, and transfusion requirement (red blood cells, platelets, or fresh frozen plasma) all showed a highly significant benefit for the off-pump group (p 0.005), with ORs ranging from 0.33 to 0.65. The effect size estimates remained essentially unaltered after adjustment for prognostic variables (point estimates of ORs ranging from 0.36 to 0.68; p < 0.05) except the estimate for blood loss of more than 1,000 mL (adjusted OR, 0.82; p = 0.22). The risk of a postoperative hemoglobin level of less than 10 g/dL was significantly lower in the off-pump group both before and after adjustment (after adjustment: OR, 0.50; 95% confidence interval, 0.34 to 0.72; p < 0.001), even though patients in this group were less likely to have received a blood transfusion.
Effect sizes for other adverse outcomes were also stable after adjustment for potential confounding factors. Almost all OR estimates were less than unity, a finding suggesting the benefit of off-pump operation, although the majority did not reach significance. Adjusted ORs often suggested that the benefits of off-pump revascularization are large (ORs were < 0.7 for an additional nine outcomes; see Table 3), but confidence intervals for these estimates included unity because of the rarity of many of the complications (see Table 2).
The decision to include all high-risk patients during the study period was taken so that the study would be representative of our entire experience with off-pump coronary artery bypass grafting in an unselected way, thus increasing the generalizability of our findings. However, from the point of view of scientific rigor, there were some disadvantages to this decision. For example, emergency and salvage operations were included and were not distributed evenly between the two groups, and there was a possibility of changes in the effectiveness of the off-pump method during the course of the study as a result of evolution of either the technique or the skills of the surgeons. Therefore, additional sensitivity analyses were carried out to test the robustness of the findings.
In the sensitivity analyses, OR estimates for off-pump operation for the different outcomes did not alter substantially after emergency and salvage operations were excluded (n = 67; data are available on request). Off-pump operations were a relatively small proportion of all operations before April 1, 1999, (noff = 94/995, 9%) compared with after April 1, 1999 (noff = 238/575, 41%). Both calendar time and interaction of calendar time with type of operation were significant predictors of outcome for absence of sinus rhythm, blood loss of more than 1,000 mL, transfusion of platelets, and a combined ICU and HDU stay greater than 2 days. For all of these outcomes, the risk of an adverse outcome in the on-pump group was greater before April 1, 1999 than after that date; and the relative benefit of off-pump revascularization compared to on-pump revascularization was greater before April 1, 1999 than after that date. In these analyses, the effect of off-pump operation decreased substantially after April 1, 1999, but the effect of off-pump coronary artery bypass grafting before April 1, 1999, was estimated to be even stronger than when calendar time was not considered (data are available on request). In addition, there was an effect of calendar time, but no interaction, for perioperative myocardial infarction, pulmonary complication, postoperative hemoglobin level of <10 g/dL or lower, transfusion of red blood cells, transfusion of fresh frozen plasma, ICU stay greater than 1 day, and discharge to another hospital rather than home. For all of these outcomes except postoperative hemoglobin level, the risk of an adverse outcome was greater before April 1, 1999, than afterward. For postoperative hemoglobin level, the risk of a low level was greater in the later period, a finding that is consistent with the lower risk of transfusion during the same period. Including calendar time in the models did not substantially alter the OR estimates for off-pump operation for these outcomes (data are available on request).
|
Comment
|
|---|
This study suggests that off-pump coronary artery bypass grafting is safe and effective in high-risk patients. These findings complement the results from our randomized controlled trial of on-pump versus off-pump revascularization in elective "low-risk" patients [5, 6] and suggest that for a surgeon with expertise using the technique, the off-pump method should be preferred for the majority of patients.
The inclusiveness of the study is an important feature contributing to this conclusion. The study describes the complete experience using off-pump revascularization in our unit over about 4.5 years. We deliberately analyzed the entire cohort of consecutive patients meeting our definition of high risk, including emergency and salvage operations and operations carried out by trainees and by surgeons who had little experience with off-pump surgical procedures as well as surgeons who had performed many off-pump operations. The sensitivity analyses confirmed that the findings were largely unaltered when these factors were taken into consideration.
Because of the novel nature of the technique, it is also inevitable that the technique evolved during the period of data collection, eg, as a result of the introduction of a stabilizer [10]. These developments might lead one to expect that the effectiveness of the off-pump operation would have increased with time. However, where an interaction between calendar time and surgical technique was found, the benefit of off-pump revascularization was always greater in the earlier part of the study, despite the frequency of adverse outcomes being higher, on average, during that same period. This finding may be due to the fact that the off-pump technique was initially used for technically easier procedures.
The largest benefits were observed for the variables transfusion, length of stay in the ICU and the HDU, and total length of stay. It might be argued that without the blinding of the patients and their carers, these outcomes could be biased by knowledge of the type of operation. We believe that bias is an unlikely explanation for the results. First, none of the health care staff were aware during the period of the study that the comparison was going to be made; the data were collected prospectively as part of routine care for all patients undergoing a cardiac surgical-procedure, but the study objective and the analysis plan were not established until the end of the study period. Second, as already described, standard unit protocols are used to make decisions about transfusion, as it is unethical to expose a patient to the risks of transfusion unnecessarily. These protocols were applied carefully throughout the period of the study and would have minimized the opportunity for bias. Third, there are similar unit protocols for use of ICU and HDU beds and for discharging patients home or to another hospital; these resources are limited, and there are strong pressures to implement the guidelines to maintain the work load of the unit.
The findings suggest that off-pump coronary artery bypass grafting is beneficial for other outcomes, but, even with the relatively large sample size, the study did not have sufficient power to detect as significant clinically important benefits for outcomes that occurred infrequently. A larger cohort of patients, preferably with a higher proportion of patients undergoing off-pump operations, is required to obtain more precise estimates of these potential benefits.
Without random allocation (surgeons who had the expertise to carry out off-pump revascularization were not prepared to randomize patients), we cannot exclude the possibility that the findings arise from confounding factors. However, this is an unlikely explanation for the benefits observed from off-pump surgical procedures for several reasons. First, the method for allocating patients to the on-pump or off-pump groups (see Material and Methods) does not appear to have created two groups with substantially different risks of adverse outcomes, although the off-pump group required, on average, fewer grafts. Second, the effects of off-pump operations were consistently beneficial across a wide range of outcomes. Third, the effects were large for some outcomes, such as transfusion, ICU and HDU stay, and total length of stay. Finally, consideration of potential confounding by a large number of prognostic factors did not alter the findings.
Bias could also have arisen from missing data had there been a tendency for data to be missing selectively for the highest-risk patients in the off-pump group. It can be seen from Tables 1 and 2 that there were few missing data (<5%) for all prognostic variables and for all outcomes, with the exception of blood loss, transfusion requirement, and postoperative hemoglobin level. Data for these variables were not collected from April 1996 to April 1997, which accounts for the majority of the missing data. If anything, there was a tendency for on-pump patients with missing data for these variables to have greater comorbidity than off-pump patients, as judged by comparing their other outcomes (data are available on request). Bias of this kind, if real, would have acted to dilute rather than exaggerate the effect of off-pump revascularization for transfusion outcomes, postoperative hemoglobin level, and blood loss.
Delaying the decision about the type of procedure (on- pump versus off-pump) to use until the anatomy of the coronary vessels could be explored minimized the need to convert off-pump operations to on-pump operations, and only three conversions occurred. Carrying out intention-to-treat analyses avoided bias that would otherwise have arisen if conversions to on-pump revascularization had occurred in the off-pump patients most at risk [18].
The study could be criticized for assessing multiple aspects of early postoperative outcome, which means that many statistical tests had to be carried out. However, there is no single primary outcome other than death that surgeons consider representative of early outcome, and hospital mortality is now such a rare event (only 2.4% in this entire cohort of high-risk patients) that it has become an insensitive measure of early outcome. Nor is there an established method for combining the different aspects of outcome that are considered to reflect poorer outcome, eg, need of inotropic support or transfusion. Nevertheless, the direction of the effects of off-pump coronary artery bypass grafting was consistent for almost every outcome assessed, and the findings overall were consistent with those of our randomized controlled trial in a sample of patients who were not at high risk [5,6]. We argue that this is strong evidence for concluding that off-pump revascularization is, at least, beneficial even if the precise magnitude of these benefits requires further investigation for some outcomes. Given that the off-pump technique is also less expensive than the on-pump method and has equivalent midterm outcomes [6], it appears justifiable to promote more widespread adoption of the technique.
In conclusion, this study suggests that off-pump coronary artery bypass grafting in high-risk patients is safe and is associated with reduced morbidity when compared with conventional on-pump coronary artery revascularization.
|
References
|
|---|
-
Cremer J., Martin M., Redl H., et al. Systemic inflammatory response syndrome after cardiac operations. Ann Thorac Surg 1996;61:1714-1720.[Abstract/Free Full Text]
-
Buffolo E., Silva de Andrade J.C., Rodrigues Branco J.N., Teles C.A., Aguiar L.F., Gomes W.J. Coronary artery bypass grafting without cardiopulmonary bypass. Ann Thorac Surg 1996;61:63-66.[Abstract/Free Full Text]
-
Lui B., Belboul A., Larsson S., Roberts D. Factors influencing haemostasis and blood transfusion in cardiac surgery. Perfusion 1996;11:131-143.[Medline]
-
Holloway D.S., Summaria L., Sandesara J., Vagher J.P., Alexander J.C., Caprini J.A. Decreased platelet number and function and increased fibrinolysis contribute to postoperative bleeding in cardiopulmonary bypass patients. Thromb Haemost 1988;59:62-67.[Medline]
-
Ascione R., Lloyd C.T., Underwood M.J., Lotto A.A., Pitsis A.A., Angelini G.D. Economic outcome of off-pump coronary artery bypass surgery: a prospective randomized study. Ann Thorac Surg 1999;68:2237-2242.[Abstract/Free Full Text]
-
Angelini GD, Taylor FC, Reeves BC, Ascione R. Beating heart against cardioplegic arrest studies (BHACAS 1 and 2): a pooled analysis of two randomised controlled trials. Lancet (in press).
-
Cleveland J.C., Jr, Shroyer A.L.W., Chen A.Y., Peterson E., Grover F.L. Off-pump coronary artery bypass grafting decreases risk-adjusted mortality and morbidity. Ann Thorac Surg 2001;72:1282-1289.[Abstract/Free Full Text]
-
Ascione R., Lloyd C.T., Underwood M.J., Lotto A.A., Pitsis A.A., Angelini G.D. Inflammatory response after coronary revascularization with or without cardiopulmonary bypass. Ann Thorac Surg 2000;69:1198-1204.[Abstract/Free Full Text]
-
Bouchard D., Cartier R. Off-pump revascularization of multivessel coronary artery disease has a decreased myocardial infarction rate. Eur J Cardio-thorac Surg 1998;14:S20-S24.[Abstract/Free Full Text]
-
Ascione R., Caputo M., Calori G., Lloyd C.T., Underwood M.J., Angelini G.D. Predictors of atrial fibrillation after conventional and beating heart coronary surgery. A prospective, randomized study. Circulation 2000;102:1530-1535.[Abstract/Free Full Text]
-
Ascione R., Lloyd C.T., Underwood M.J., Gomes W.J., Angelini G.D. On-pump versus off-pump coronary revascularization: evaluation of renal function. Ann Thorac Surg 1999;68:493-498.[Abstract/Free Full Text]
-
Arom K.V., Flavin T.F., Emery R.W., Kshettry V.R., Janey P.A., Petersen R.J. Safety and efficacy of off-pump coronary artery bypass grafting. Ann Thorac Surg 2000;69:704-710.[Abstract/Free Full Text]
-
Moshkovitz Y., Lusky A., Mohr R. Coronary artery bypass without cardiopulmonary bypass: analysis of short-term and mid-term outcome in 220 patients. J Thorac Cardiovasc Surg 1995;110:979-987.[Abstract/Free Full Text]
-
Watters M.P., Ascione R., Ryder I.G., Ciulli F., Pitsis A.A., Angelini G.D. Haemodynamic changes during beating heart coronary surgery with the ‘Bristol Technique’. Eur J Cardio-thorac Surg 2001;19:34-40.[Abstract/Free Full Text]
-
Lucchetti V., Angelini G.D. An inexpensive method of heart stabilization during coronary artery operations without cardiopulmonary bypass. Ann Thorac Surg 1998;65:1477-1478.[Abstract/Free Full Text]
-
Ascione R., Williams S., Lloyd C.T., Sundaramoorthi T., Pitsis A.A., Angelini G.D. Reduced postoperative blood loss and transfusion requirement after beating-heart coronary operations: a prospective randomized study. J Thorac Cardiovasc Surg 2001;121:689-696.[Abstract/Free Full Text]
-
Lloyd C.T., Ascione R., Underwood M.J., Gardner F., Black A., Angelini G.D. Serum S-100 protein release and neuropsychologic outcome during coronary revascularization on the beating heart: a prospective randomized study. J Thorac Cardiovasc Surg 2000;119:148-154.[Abstract/Free Full Text]
-
Begg C., Cho M., Eastwood S., et al. Improving the quality of reporting of randomized controlled trials. The CONSORT statement. J Am Med Assoc 1996;276:637-639.[Medline]
Top
Abstract
Introduction
) and on-pump (
) high-risk operations by year and (B) average number of grafts per patient in each group by year. Data for 1996 include April through December only.