How Safe Is It to Train Residents to Perform
Off-Pump Coronary Artery Bypass Surgery?
CARDIOVASCULAR
George Asimakopoulos, FRCS, PhD, A. Paul Karagounis, FRCS, MPhil,
Oswaldo Valencia, MD, David Rose, MD, Gunaratnam Niranjan, FRCS,
and Venkatachalan Chandrasekaran, FRCS
Department of Cardiothoracic Surgery, St Georges Hospital, London, United Kingdom
Background. The technique of off-pump coronary ar-
tery bypass graft (OPCABG) surgery differs considerably
from on-pump CABG. This study investigates the impact
of surgical training on clinical outcome in patients un-
dergoing OPCABG.
Methods. All 251 OPCABG cases performed by one
service over an 18-month period were analyzed. The 83
operations (33%) performed by two trainees under super-
vision were compared with the 168 operations (67%)
performed by an experienced consultant surgeon. Patient
and disease characteristics, intraoperative and postoper-
ative data, morbidity and mortality were analyzed using
univariate and multivariate analysis. Data were extracted
from a prospective database.
Results. Patients operated on by the consultant were
more likely to have had unstable angina (p 0.003, odds
ratio [OR] 3.5), impaired left ventricular function
(ejection fraction <0.3; p 0.005, OR 2.4), or previous
C ardiopulmonary bypass (CPB) has been employed
for the vast majority of coronary artery bypass
grafting (CABG) operations since the emergence of car-
diac surgery. Cardiopulmonary bypass is, however, asso-
ciated with complications such as coagulation abnormal-
ities and dysfunction of major organs [1]. Despite
significant reduction in in-hospital mortality rates over
the past decades [2, 3], pulmonary and renal dysfunction
and also strokes continue to affect patients after CABG [4,
5]. Off-pump coronary artery bypass graft (OPCABG)
surgery has gained popularity over the last decade and is
currently being performed by many surgeons worldwide
[6]. There is a growing body of literature suggesting that
OPCABG is a safe alternative to on-CPB coronary artery
bypass (on-CABG) and it may be associated with reduc-
tion in postoperative morbidity and mortality [7].
Performing OPCABG is undoubtedly demanding tech-
nically as the surgeon is faced with a beating heart and
not a bloodless field that contrasts with conditions in
on-CABG. The potential clinical advantages of OPCABG
Accepted for publication July 18, 2005.
Presented at the Poster Session of the Forty-first Annual Meeting of The
Society of Thoracic Surgeons, Tampa, FL, Jan 24 26, 2005.
Address correspondence to Dr Chandrasekaran, Department of Cardio-
thoracic Surgery, St. Georges Hospital, Blackshaw Rd, London SW17
0QT, United Kingdom; e-mail: v.chandra@fsmail.net.
2006 by The Society of Thoracic Surgeons
Published by Elsevier Inc
cardiac surgery (p 0.03). They were more likely to
receive three or more grafts (p 0.017, OR 2.0).
Operative mortality was 2.4% (consultant) and 0% (train-
ees; p 0.31). Postoperative morbidity, such as reopera-
tion for bleeding (consultant 3% versus trainees 1.2%),
stroke (0.6% versus 1.2%), and hemofiltration (3.6% ver-
sus 0%) was similar between the two patient groups. Stay
in the intensive care unit was not significantly different
in the two groups.
Conclusions. In our experience, trainee surgeons are
less likely to operate on patients with unstable angina or
cardiac dysfunction. Operative morbidity and mortality
are, however, similar in patients operated on by either
an experienced consultant surgeon or trainees. We
believe OPCABG can be taught safely to trainees
under supervision.
(Ann Thorac Surg 2006;81:568 72)
2006 by The Society of Thoracic Surgeons
and the advances made in surgical technology have
made this procedure an essential part of a cardiothoracic
training program. The aim of this study is to review
clinical data of patients undergoing OPCABG and com-
pare clinical outcomes after procedures performed by a
consultant surgeon or a trainee as the primary operator
over an 18-month period.
Material and Methods
Patients and Collection of Data
This study involved retrospective data analysis of pa-
tients undergoing consecutive isolated CABG under the
care of one consultant surgeon (V.C.) at one institution,
between July 2002 and December 2003. In all, 280 patients
underwent CABG during the study period, of whom 251
patients underwent OPCABG. The remaining 29 patients
underwent on-CABG as a part of a randomized trial and
were, therefore excluded from this analysis. There were
no intraoperative conversions from OPCABG to on-
CABG during the study period.
The clinical data were collected prospectively in line
with the appended Minimum Dataset (MDS) defined by
The Society of Cardiothoracic Surgeons. The current
MDS, and its associated definitions, is compatible with all
existing initiatives in the United Kingdom such as the UK
0003-4975/06/$32.00
doi:10.1016/j.athoracsur.2005.07.054
Ann Thorac Surg
2006;81:568 72
Heart Valve Registry, the Central Cardiac Audit Data-
base, and the British Cardiac Intervention Society data-
base. The definitions and data fields are also compatible
with evolving European initiatives, The Society of Tho-
racic Surgeons, the American College of Cardiology, and
the Healthcare Financing Administration in the United
States [2]. Local validation of the collected data is per-
formed regularly, and external validation is being per-
formed by the Society on a 3- to 5-yearly cycle.
Anesthetic and Operative Techniques
The anesthetic and operative techniques used for
OPCABG at our institution have been described in a
separate study report [8]. Briefly, anesthetic premedica-
tion included morphine (10 mg) and hyoscine (0.3 mg)
administered intramuscularly 2 hours before the opera-
tion. Anesthesia was introduced with midazolam (100 to
200 g/kg), fentanyl (150 to 200 g), and pancuronium (50
to 100 g/kg), and sustained with propofol (5 to 10
mg kg1 hour1). Anticoagulation was achieved using
150 U/kg of heparin. The activated clotting time was
maintained above 300 seconds. Heparin was reversed
completely with protamine sulphate at the end of the
procedure. The operation was performed through me-
dian sternotomy using the Guidant Acrobat SUV Vac-
uum Stabilizer System (OM-9000; Santa Clara, Califor-
nia). The exposure of the target vessel was facilitated by
the use of swabs in the transverse sinus (deep pericardial
stay sutures were not used), allowing the right side of the
pericardium down and proximal vessel occlusion by a
soft bulldog clamp. All distal anastomoses were con-
structed with 8-0 polypropylene and the proximal onto
the aorta with 6-0 polypropylene. Mean arterial blood
pressure was maintained between 50 and 70 mm Hg
during the procedure by maintaining optimal preload,
repositioning the heart, and selective use of vasoconstric-
tors, such as metaraminol and norepinephrin.
Training Method
The two trainees participating in this study had surgical
experience equivalent to year 3 and 4 of the UK National
Training Program. Both had received previous training in
on-pump CABG involving more than 100 cases each. In
our institution, approximately one third of coronary sur-
gery is done off pump. The training consultant performs
85.5% of revascularization operations off pump. The
trainees during the study period performed exclusively
off-pump CABG. The patients operated upon by the
trainees were selected by assessing their suitability for
training taking into account the urgency of the operation,
their co-morbidities, the quality of the coronary arteries,
and the number of grafts required. Patients with poor left
ventricular function and unstable hemodynamics were
not used for training purposes. Training in OPCABG
focused, therefore, on teaching the technique of handling
the heart, positioning the stabilizer, and performing the
anastomoses at the front of the heart initially. Further-
more, the training progressed to gradually increasing
levels of complexity and responsibility according to the
surgical abilities of the trainee. More specifically, arterial
ASIMAKOPOULOS ET AL 569
TRAINING IN OFF-PUMP BYPASS SURGERY
and venous conduits were used as indicated and sequen-
tial and Y-graft anastomoses were included in the teach-
ing. All the training took place under the supervision of
the same consultant surgeon, who assisted in the major-
ity of cases and allowed the trainees to perform the
CARDIOVASCULAR
operations on their own at the final stages of the training
period with the trainer being in the vicinity of the
operating room. The sequence of grafting consisted of
the left anterior artery grafted first if it was critically
diseased, followed by the vessels of the left side of the
heart and the vessels of the posterior wall.
Statistical Methods
Contingency tables for categorical data were compared
between the two groups using Pearsons 2 test or Fish-
ers exact test. Distributions of continuous data were
observed and t tests or Mann-Whitney U tests were
performed accordingly for comparisons of groups. When
no normal distribution of data was detected, variables
were presented as median with 25th and 75th percentiles
(interquartile range). The European System for Cardiac
Operative Risk Evaluation (EuroSCORE) was used to
analyze patients according to predicted risk for operative
mortality [9]. Odds ratios with 95% confidence intervals
were calculated for all compared variables. Logistic re-
gression was used to correlate preoperative patient and
disease characteristics with postoperative outcomes. All p
values less than 0.05 indicate a statistically significant
difference between patient groups. All analyses were
carried out using the statistical software SPSS 11.0 (SPSS,
Chicago, Illinois).
Results
Of the 251 OPCABG patients included in this study, 168
(66.9%) underwent OPCABG by the consultant and 83
(33.1%) by a trainee as the primary surgeon. There were
no intraoperative conversions from OPCABG to on-
CABG during the study period, whereas the total con-
version rate for the trainer is approximately 0.8%. More
specifically, in year 2002, of 138 OPCABs there were 2
conversions to on-CABG; in year 2003, there were no
conversions of 156 OPCABs; and in year 2004, there were
2 conversions of 183 OPCABs.
The demographic and preoperative clinical character-
istics of the two patient groups are presented in Table 1.
The two groups characteristics were similar with the
exception of age (older in the trainee group, p 0.034),
angina Canadian Cardiovascular Society class (higher in
the consultant group, p 0.001) and left ventricular
ejection fraction (lower in the consultant group, p
0.005). All nine reoperations included in the study, were
performed by the consultant (p 0.03). The predicted
mortality risk, as calculated by EuroSCORE, was identi-
cal in the two patient groups.
Table 2 shows intraoperative and postoperative char-
acteristics of the patients. Trainees operated on signifi-
cantly fewer patients requiring three or more grafts (p
0.017), but the mean number of grafts per patient was 3.5
for the consultant group and 3.25 for the trainee group,
 570 ASIMAKOPOULOS ET AL Ann Thorac Surg
TRAINING IN OFF-PUMP BYPASS SURGERY 2006;81:568 72

Table 1. Demographic and Preoperative Clinical Characteristics of Patients Operated on by Consultant or Trainee
Consultant Trainee
p Value
Characteristics n % n % (Odds Ratio, CI)
CARDIOVASCULAR

Sex
Female 34 20.2 23 27.7
Male 134 79.8 60 72.3
Total 168 100 83 100 0.20
Age, median in years 67 71
Interquantile range (5973) (6174) 0.034
Body mass index, median 27 27
Interquantile range (2429) (2430) 0.77
EuroSCORE, median 4 4
Interquantile range (26) (24) 0.11
CCS class I or II 113 67.3 73 88.0
CCS class III or IV 55 32.7 10 12.0 0.001 (3.5, 1.77.4)
Previous myocardial infection 77 45.8 34 41.0 0.54
Previous PCI 2 1.2 1 1.2 1
Diabetes mellitus 38 22.6 15 18.1 0.51
Previous stroke 16 9.5 10 12.0 0.52
1- or 2-Vessel disease 35 20.8 24 28.9
3-Vessel disease 133 79.2 59 71.1 0.16
LVEF 0.5 109 64.9 68 81.9
0.49 59 35.1 15 18.1 0.005 (2.4, 1.34.6)
Preoperative IABP 2 1.2 0 0 1
Elective procedure 106 63.1 61 73.5 0.10
Reoperation 9 5.4 0 0 0.03

CCS Canadian Cardiovascular Society; CI confidence interval; EuroSCORE European System for Cardiac Operative Risk Evaluation;
IABP intra-aortic balloon pump; LVEF left ventricular ejection fraction; PCI percutaneous coronary intervention.

which is not statistically significantly different. This is more, the trainees performed 27.5% of the grafts to the
explained by the consultant having operated on higher circumflex and the posterior descending arteries of the
number of patients requiring more than 4 grafts. Further- heart (the consultant performed 309 grafts, whereas the

Table 2. Intraoperative and Postoperative Clinical Characteristics of Patients Operated on by Consultant or Trainee
Consultant Trainee
p Value
Characteristics n % n % (Odds Ratio, CI)

1 or 2 Grafts 37 22.0 30 36.1

3 Grafts 131 78.0 53 63.9 0.017 (2.0, 1.13.5)
LIMA used 163 97.0 82 98.8 0.23
Arterial grafts only 61 36.3 31 37.3 0.69
Postoperative IABP 4 2.4 0 0 0.31
Resternotomy for bleeding 5 3.0 1 1.2 0.67
Mediastinal infection 1 0.6 1 1.2 0.67
Postoperative laparotomy 2 1.2 0 0 1
Tracheostomy 2 1.2 0 0 1
Stroke 1 0.6 1 1.2 0.55
Hemofiltration 6 3.6 0 0 0.18
ICU stay, median in hours 24 22
Interquantile range (2045) (1930) 0.016
Postoperative stay, median in days 6 6
Interquantile range (38) (48) 0.95
30-Day death 4 2.4 0 0 0.31

CI confidence interval; IABP intra-aortic balloon pump; ICU intensive care unit; LIMA left internal mammary artery.
Ann Thorac Surg
2006;81:568 72
trainees performed 117 in total). The patients in the trainee
group stayed on the intensive care unit for a slightly
shorter period (median, 24 versus 22 hours; p 0.016).
Rates of serious complications remained lower in both
groups. There were four 30-day deaths in the consultant
patient group (2.4%), whereas no patients died in the
trainee group. Two of the four deaths were emergencies
involving patients transferred from the catheterization
laboratory after failed percutaneous coronary interven-
tions. The total mortality of the firm for CABG in the
study period was therefore 1.6% (4 of 251).
Comment
In our institution, approximately 31% of coronary surgery
is done off pump. The training surgeon and his or her
team including the trainees during the period of 2002 to
2004 performed 477 OPCABGs (85.5%) and 81 on-CABGs.
The results of this study demonstrate that OPCABG can
be safely performed by cardiac surgical trainees under
supervision. Mortality and morbidity can be very low.
More specifically, 30-day mortality among the 83 patients
who underwent surgery by a trainee was zero, and the
only major complication involved one resternotomy for
bleeding, one mediastinal infection requiring sternal
rewiring, and one stroke. An intra-aortic balloon pump
was placed postoperatively only in 4 patients in the
consultant patient group. Mortality in the consultant
patient group was 2.4% (4 patients), and morbidity was
slightly but not significantly higher than in the trainee
patient group. The reason for this finding is that emer-
gent operations, as well as patients with particularly
impaired left ventricular function, were invariably oper-
ated on by the consultant. In the trainee group, the
median for stay in the intensive care unit was 22 hours,
with the longest stay being 136 hours.
It is commonly believed that operations allocated to
trainees involve patients with low predicted operative
risk. In our study, patients in the trainee group were
significantly older but had, overall, lower angina class
and higher left ventricular ejection fraction. Operative
risk scoring, as assessed by the EuroSCORE, demon-
strated median scores of four in both groups, with a 75th
percentile of four in the trainee group versus six in the
consultant group. The EuroSCORE is an additive risk
prediction system, similar in concept to the Parsonnet
score, and was constructed using data from 128 European
cardiac surgical centers [9]. It was developed using
mainly on-CABG patients and has undergone extensive
external validation with satisfactory discrimination and
calibration [10]. The EuroSCORE also showed satisfac-
tory prediction properties for OPCABG patients [11]. Our
study suggests that overall average predicted risk does
not differ between patients operated on by trainees as
compared with consultants, although particularly high
risk operations are performed by a consultant surgeon. It
is possible that patients with particularly diffuse coronary
artery disease and small coronary vessels were more
likely to undergo operation by the consultant, although it
would be difficult to quantify this statement.
ASIMAKOPOULOS ET AL 571
TRAINING IN OFF-PUMP BYPASS SURGERY
Small sample size is the main limitation of our study.
Low rates of outcome measures, such as mortality and
morbidity, precluded our regression analysis from de-
tecting correlation between patient characteristics and
outcome. Low complication rates were, nevertheless,
CARDIOVASCULAR
encouraging. A nonsignificant trend for cases performed
by trainees to develop fewer postoperative complications
reinforces the notion that trainees can be taught to
operate safely off-CPB. Analysis of higher numbers of
OPCABG cases performed by a variety of trainees should
be performed in future studies to reconfirm the main
findings of this trial. Long-term follow-up of patients will
also be of interest.
The effect of training on clinical outcome after cardiac
surgery has been the subject of previous publications.
Some authors attempted to compare outcome between
consultant and trainee cases directly and also by account-
ing for possible treatment selection bias. Oo and associ-
ates [12] adjusted for case mix by using propensity score
and concluded that when trainees operate on lower risk
cases, they produce better clinical outcome than their
trainers. In their study, CABG was performed off-CPB
only in 26.3% of cases. and trainees operated on merely
9.9% of the total patient population. Mean EuroSCORE
was 2.9 for trainees and 3.5 for consultants. In a trial
conducted in a UK center, hospital costs were not in-
creased when trainees performed the surgery [13]. When
trainees operated on higher risk on-CABG cases than the
trainers, in a separate study, they achieved equally sat-
isfactory results [14]. The same group also showed that it
is safe to train junior surgeons in the performance of
complex cardiac procedures, such as mitral valve repair
[15]. Trainees performed 79 mitral valve repairs with a
hospital mortality of 3.8% versus 4.3% for staff surgeons.
Similarly, propensity score-matched cases showed no
significantly different complication rates in the two pa-
tient groups. These findings are significant, as few trainee
surgeons have the opportunity to gain substantial expe-
rience under supervision in mitral valve repair.
The need to provide junior surgeons with adequate
exposure to cardiac surgical techniques without CPB has
been recognized in the literature over the last few years.
The experience of a single resident in OPCABG training
was presented in an early report [16]. In this publication,
the trainee tended to use the OPCABG technique for
cases that required, on average, a smaller number of
grafts than on-CABG cases performed during the same
training period. Only 56% of these patients received a
circumflex artery anastomosis. The same group con-
ducted a postal survey assessing the exposure of the
North American cardiac surgical residents to OPCABG.
Response at the time (1999) suggested that the majority of
residents would not reach proficiency in performing
OPCABG during their residency [17]. It is likely that
training in the technique has become more sufficient by
the year 2005.
In the United Kingdom, experience of training in
OPCABG has been previously reported by a small num-
ber of cardiac surgical departments. Jenkins and cowork-
ers [18] reported findings of five patients who underwent
 572 ASIMAKOPOULOS ET AL
TRAINING IN OFF-PUMP BYPASS SURGERY
OPCABG by a trainee and consented to early postoper-
ative coronary angiography. This revealed 17 satisfactory
(100%) distal anastomoses. The concept and outcome of
training in OPCABG in a single UK cardiac surgical
center has been presented comprehensively in a series of
CARDIOVASCULAR
publications, all largely concerning the same patient
population [19 22]. The last report of the series summa-
rizes outcome on 990 CABG cases performed by trainees,
of which 474 were OPCABGs. These operations were
compared with those performed by an experienced con-
sultant surgeon. Predicted and observed risk of death or
serious complications was similar for trainee and consul-
tant operations. The authors conclude that it is safe to
teach OPCABG to trainee surgeons, and that continuous
performance monitoring for trainees is possible and
desirable.
In summary, the results of this study suggest that
trainees can be taught to perform OPCABG safely under
supervision with low rate of complications, including in
patients with triple-vessel disease and grafting to the
posterior wall of the heart. These findings are in agree-
ment with previous literature reports. We believe the
trainees should be given sufficient exposure to operate on
patients undergoing OPCABG after adequate exposure
to on-CABG surgery. Patients should be reassured that
safety is not compromised by the presence of a trainee as
the primary operator in their surgery.
References
1. Asimakopoulos G. Mechanisms of the systemic inflamma-
tory response syndrome. Perfusion 1999;14:269 77.
2. Keogh B, Kinsman R. National adult cardiac data base report
1990-2000. The Society of Cardiothoracic Surgeons of Great
Britain and Ireland, 2000 (www.scts.org).
3. Ferguson TB Jr, Hammill BG, Peterson ED, DeLong ER,
Grover FL, for the STS National Database Committee. A
decade of changerisk profiles and outcomes for isolated
coronary artery bypass grafting procedures, 1990-1999: a
report from the STS National Database Committee and the
Duke Clinical Research Institute. Society of Thoracic Sur-
geons. Ann Thorac Surg 2002;73:480 90.
4. Weerasinghe A, Hornick P, Smith P, Taylor K, Ratnatunga C.
Coronary artery bypass grafting in non-dialysis-dependent
mild-to-moderate renal dysfunction. J Thorac Cardiovasc
Surg 2001;121:10839.
5. Asimakopoulos G, Smith PL, Ratnatunga CP, Taylor KM.
Pulmonary dysfunction and adult respiratory distress syn-
drome after cardiopulmonary bypass. Ann Thorac Surg
1999;68:110715.
Ann Thorac Surg
2006;81:568 72
6. Yacoub M. Off-pump coronary bypass surgery in search of
an identity. Circulation 2001;104:17435.
7. Gobran SR, Goldman S, Ferdinand F, et al. Outcomes after
usage of a quality initiative program for off-pump coronary
artery bypass surgery: a comparison with on-pump surgery.
Ann Thorac Surg 2004;78:201521.
8. Asimakopoulos G, Karagounis AP, Valencia O, et al. Renal
function after cardiac surgery off- versus on-cardiopulmonary
bypass: analysis using the Cockroft-Gault formula for estimat-
ing creatinine clearance. Ann Thorac Surg 2005;79:2024 31.
9. Nashef S, Roques F, Michael P, Gauducheau E, Lemeshow S,
Salamon R. European System for Cardiac Operative Risk
Evaluation (EuroSCORE). Eur J Cardiothorac Surg 1999;16:
9 13.
10. Asimakopoulos G, Al-Ruzzeh S, Ambler G, et al. Existing risk
stratification systems are useful for patient consultation but
should not be used to compare surgical performance between
institutions. Eur J Cardiothorac Surg 2003;23:935 42.
11. Al-Ruzzeh S, Asimakopoulos G, Ambler G, et al. Validation
of four different risk stratification systems in patients off-
pump coronary artery bypass surgery: A UK multi-centre
analysis of 2223 patients. Heart 2003;89:4325.
12. Oo AY, Grayson AD, Rashid A. Effect of training on out-
comes following coronary artery bypass graft surgery. Eur
J Cardiothorac Surg 2004;25:591 6.
13. Goodwin AT, Birdi I, Ramesh GJ, et al. Effect of surgical
training on outcome and hospital costs in coronary surgery.
Heart 2001;85:454 7.
14. Baskett RJF, Buth KJ, Legare J-F, et al. Is it safe to train
residents to perform cardiac surgery? Ann Thorac Surg
2002;74:10439.
15. Baskett RJF, Kalavrouziotis D, Buth KJ, Hirsch GM, Sullivan
JAP. Training residents in mitral valve surgery. Ann Thorac
Surg 2004;78:1236 40.
16. Karamanoukian HL, Panos AL, Bergsland J, Salerno TA.
Perspectives of a cardiac surgery resident in-training on
off-pump coronary bypass operations. Ann Thorac Surg
2000;69:42 6.
17. Ricci M, Karamanoukian HL, DAncona G, et al. Survey of
resident training in beating heart operations. Ann Thorac
Surg 2000;70:479 82.
18. Jenkins D, Al-Ruzzeh S, Khan S, et al. Multivessel off-pump
coronary artery bypass grafting can be taught to trainee
surgeons. J Card Surg 2003;18:419 28.
19. Caputo M, Chamberlain MH, Ozalp F, et al. Off-pump
coronary operations can be safely taught to cardiothoracic
trainees. Ann Thorac Surg 2001;71:12159.
20. Caputo M, Bryan AJ, Capoun R, et al. The evolution of
training in off-pump coronary surgery in a single institution.
Ann Thorac Surg 2002;74(Suppl):14037.
21. Ascione R, Reeves BC, Pano M, Angelini GD. Trainees
operating in high-risk patients without cardiopulmonary
bypass: a high-risk strategy? Ann Thorac Surg 2004;78:26 33.
22. Caputo M, Reeves BC, Rogers CA, Ascione R, Angelini GD.
Monitoring the performance of residents during training in
off-pump coronary surgery. J Thorac Cardiovasc Surg 2004;
128:90715.