RESULTS

Unmatched analysis
Baseline patient characteristics are presented in Table 1. The overall mean age was 66.6 ± 8.5 years and 53 patients (15%) were female. Patients in the redo-OPCAB group were on average 2 years older (67.8 ± 8.1 vs 65.7 ± 8.6 years, P = 0.02) and had twice the amount of peripheral vascular disease (27% vs 16%, P = 0.02). The average number of diseased coronary vessels was lower in the redo-OPCAB (2.3 ± 0.8 vs 2.6 ± 0.6, P < 0.01) reflecting the number of bypass grafts performed (2.3 ± 0.8 vs 2.6 ± 0.6, P < 0.01). Completeness of revascularization was similar at 76% vs 73% redo-OPCAB vs redo-CABG respectively (P = 0.9). Postoperative outcomes are shown in Table 2. The incidence of the composite end-point was higher in the redo-CABG group (25% vs 16%, P = 0.06). In-hospital mortality was 2.8 times higher in the redo-CABG group (6 vs 2%, P = 0.09). The other variables of the composite included: postoperative AKI (16 vs 13%, P = 0.43), stroke (1 vs 1%, P = 0.78) and severe LCO needing IABP (10 vs 3%, P =0.01), all redo-CABG vs redo-OPCAB respectively. Peak of postoperative SCrea was 134.7 ± 85.1 µmol/l and 126.04 ± 62.3 µmol/l for the redo-CABG and redo-OPCAB respectively (P = 0.30). Reopening for bleeding was higher in redo-CABG group (3 vs 1%, P =0.19), while hospital stay was 0.6 days longer (8.8 ± 6.9 vs 8.2 ± 6.2 days, P = 0.07) in the same group.
Table 1 Pre- and intraoperative characteristics
Unmatched analysis
Overall (n = 349) Redo-OPCAB (n = 143) Redo-CABG (n = 206) P-value
Age (years) 66.6 (8.5) 67.8 (8.1) 65.7 (8.6) 0.02
Female gender (%) 53 (15) 21 (15) 32 (15.5) 0.82
BMI (kg/m2) 27.2 (4.2) 27.9 (4.4) 26.9 (4.7) 0.07
Reduced LVEF <50% 117 (33) 46 (32) 71 (34) 0.65
Diabetes (%) 78 (22) 31 (22) 47 (23) 0.80
Hypertension (%) 252 (72) 110 (77) 142 (69) 0.08
CKD (%) 6 (2) 2 (1) 4 (2) 0.70
Preoperative creatinine (µmol/l) 103.1 (29) 101.8 (26.3) 103.9 (30.8) 0.52
eGFR (ml/min/1.73 m2) 64.4 (16.8) 64.7 (16.4) 64.2 (17.1) 0.76
Previous CVA (%) 40 (11) 18 (13) 22 (11) 0.56
PVD (%) 72 (21) 38 (27) 34 (16) 0.02
COPD (%) 53 (15) 25 (17) 28 (14) 0.31
Smoking history (%) 266 (76) 110 (77) 156 (76) 0.76
EuroSCORE 7 (2.7) 6.9 (2.5) 7.2 (2.8) 0.49
NYHA Class 3 or 4 (%) 149 (43) 53 (37) 96 (47) 0.10
CCS Class 3 or 4 (%) 228 (65) 89 (62) 139 (67) 0.37
Urgent surgery (%) 138 (39) 57 (40) 81 (39) 0.91
Previous surgical procedure (%) 0.03
 CABG 273 (78) 100 (71) 173 (83)
 Valve 43 (12) 23 (16) 20 (10)
 CABG + valve ± other 33 (9) 18 (13) 15 (7)
Number of diseased vessels 2.4 (0.7) 2.3 (0.8) 2.6 (0.6) <0.01
Use of nitrates (%) 53 (15) 19 (13) 34 (16) 0.41
Number of grafts 2.5 (0.7) 2.3 (0.8) 2.6 (0.6) <0.01
Use of IMA (%) 175 (50) 81 (57) 94 (46) 0.06
Use of radial artery (%) 67 (19) 32 (22) 35 (17) 0.26
Data are reported as mean and SD for continuous variables and as total count and percentages for categorical variables.
BMI: body mass index; LVEF: left ventricular ejection fraction; CKD: chronic kidney disease; eGFR: estimated glomerular filtration rate; CVA: cerebrovascular accident; PVD: peripheral vascular disease; COPD: chronic obstructive pulmonary disease; NYHA: New York Heart Association; CCS: Canadian Cardiovasular Society; CABG: coronary artery bypass grafting; CAD: coronary artery disease; IMA: internal mammary artery.
Table 2 Operative outcomes in the unmatched population
Unmatched analysis
Overall (n = 349) Redo-OPCAB (n = 143) Redo-CABG (n = 206) P-value
In-hospital mortality (%) 15 (4) 3 (2) 12 (6) 0.09
AKI (%) 50 (14) 18 (13) 32 (16) 0.43
CVA (%) 3 (1) 1 (1) 2 (1) 0.78
Use of IABP (%) 25 (7) 4 (3) 21 (10) 0.01
Composite outcome (%) 73 (21) 23 (16) 50 (25) 0.06
Reoperation for bleeding (%) 8 (2) 1 (1) 7 (3) 0.19
Completeness of revascularization (%) 257 (74) 108 (76) 149 (73) 0.92
Hospital stay (days) 8.6 (6.6) 8.2 (6.2) 8.8 (6.9) 0.07
Data are reported as mean and SD for continuous variables and as total count and percentages for categorical variables.
AKI: acute kidney injury (defined as peak of postoperative creatinine >% higher than preoperative value); CVA: cerebrovascular accident; IABP: intra-aortic balloon pump (either intraoperative or postoperative).
The mean follow-up time was 8.88 ± 5.1 years for the entire population. Figure 2 shows the Kaplan–Meier survival curves for the 2 unmatched groups. The long-term survival was similar between groups (log-rank test: P = 0.78). Late survival at 1-, 5- and 10-year was similar between groups at 95 vs 91.3%, 86.7 vs 83.5%, and 67.2 vs 68.4% for redo-OPCAB vs redo-CABG respectively. A multivariate logistic regression model indicated body mass index [odds ratio (OR_ 0.93, 95% confidence interval (CI) 0.86–0.99, P = 0.04) and EuroSCORE (OR 1.18, 95% CI 1.06–1.31, P = 0.01) as independent predictors of the composite outcome.
Figure 2  Kaplan–Meier survival curves for redo-OPCAB and redo-CABG for the entire cohort.

Propensity score matching analyses A and B
Distribution of baseline characteristics for both matched analyses A and B was similar between groups (Table 3).
Table 3 Pre- and intraoperative characteristics of patients in matched analysis
Characteristic Matched Analysis A Matched Analysis B
Redo-OPCAB (n = 111) Redo-CABG (n = 111) P-value SMD* Redo-OPCAB (n = 88) Redo-CABG (n = 88) P-value SMD*
Age (years) 67.8 (8.6) 67.9 (7.2) 0.53 0.01 67.3 (7.7) 67.7 (7.5) 0.74 0.05
Female gender (%) 15 (13) 12 (11) 0.54 0.08 11 (13) 9 (11) 0.65 0.07
BMI (kg/m2) 27.4 (3.7) 27.1 (4) 0.49 0.09 27.8 (3.9) 27.4 (4.3) 0.65 0.08
Reduced LVEF <50% 40 (36) 38 (34) 0.78 0.04 29 (34) 30 (36) 0.86 0.03
Diabetes (%) 26 (23) 29 (26) 0.63 0.06 19 (23) 23 (27) 0.44 0.11
Hypertension (%) 83 (75) 81 (73) 0.76 0.04 60 (71) 62 (74) 0.70 0.05
CKD (%) 1 (1) 1 (1) 1.00 <0.01 0 (0) 1 (1) NA 0.16
Preop creatinine (µmol/l) 100.1 (25.4) 102 (29) 0.76 0.07 96.7 (20) 98.6 (24) 0.82 0.08
eGFR (ml/min/1.73 m2) 66 (16) 65.8 (18) 0.84 0.01 67.9 (16) 67.7 (17.7) 0.83 0.01
Previous CVA (%) 15 (13) 14 (13) 0.84 0.03 9 (11) 9 (11) 1.00 <0.01
PVD (%) 23 (21) 25 (22) 0.73 0.04 13 (15) 15 (18) 0.66 0.06
COPD (%) 20 (18) 17 (15) 0.59 0.07 10 (12) 8 (9) 0.59 0.08
Smoking history (%) 82 (73) 81 (73) 0.88 0.02 65 (78) 60 (71) 0.35 0.14
EuroSCORE 7 (2.7) 6.9 (2.7) 0.83 0.05 7.1 (2.5) 7.2 (2.8) 0.71 0.06
NYHA Class 3/4 (%) 43 (39) 44 (40) 0.89 0.02 35 (42) 33 (39) 0.76 0.05
CCS Class 3/4 (%) 67 (60) 73 (66) 0.43 0.11 57 (68) 55 (65) 0.73 0.05
Urgent surgery (%) 44 (40) 48 (44) 0.57 0.07 33 (39) 28 (33) 0.41 0.12
Previous type of surgery (%) 0.91 0.06
 CABG 84 (76) 83 (75) NA NA
 Valve 16 (14) 18 (16) NA NA
 CABG + valve ± other 11 (10) 10 (9) NA NA
Number of diseased vessels 2.4 (0.7) 2.4 (0.7) 0.79 0.03 2.3 (0.8) 2.4 (0.7) 0.29 0.13
Use of nitrates (%) 18 (16) 17 (15) 0.84 0.03 11 (13) 10 (12) 0.78 0.04
No of grafts 2.1 (0.8) 2.4 (0.8) <0.01 2 (0.8) 2.2 (0.8) 0.03
Use of IMA (%) 62 (56) 52 (47) 0.21 44 (51) 29 (34) 0.02
Use of radial artery (%) 24 (22) 19 (17) 0.39 20 (24) 16 (19) 0.47
Data are reported as mean and SD for continuous variables and as total count and percentages for categorical variable.
SMD: standardized mean difference (reported for the variables included in the propensity score matching process); BMI: body mass index; LVEF: Left ventricular ejection fraction; CKD: chronic kidney disease; CVA: cerebrovascular accident; PVD: peripheral vascular disease; COPD: chronic obstructive pulmonary disease; NYHA: New York Heart Association; CCS: Canadian Cardiovascular Society.
a SMD (reported for the variables included in the propensity score-matching process).

Matched Analysis A
In propensity score-Matched Analysis A the number of patients was 111 in each group, including those who had undergone any type of primary surgery (CABG only, valve only and CABG plus valve with/without other). Postoperative outcomes are shown in Table 4. The rate of the composite end-point was 21% in the redo-CABG and 17% in the redo-OPCAB groups (P = 0.25). This included in-hospital mortality (4 vs 2%, P = 0.25), AKI (14% vs 12%, P = 0.56), stroke (1% vs 1%, P = 1) and severe LCO requiring IABP (10 vs 3%, P = 0.03), all redo-CABG vs redo-OPCAB. Reopening for bleeding was 4 times higher in the redo-CABG group (4% vs 1%, P = 0.17) and the length of stay was slightly longer in the redo-CABG group (8.7 ± 5.6 vs 8.1 ± 5.6 days, P = 0.18). Peak of postoperative SCrea was 127.9 ± 59.8 µmol/l and 123.3 ± 60.5 µmol/l for redo-CABG and redo-OPCAB, respectively (P = 0.34). Figure 3A shows the Kaplan–Meier survival curves. The long-term survival was similar between groups (log-rank test: P = 0.5). Late survival at 1, 5 and 10 years was similar between groups at 94.6 vs 91%, 83.2 vs 79.8% and 65.1 vs 60.8% for redo-OPCAB vs redo-CABG, respectively. The number of graft was higher in the redo-CABG group (2.4 ± 0.8 vs 2.1 ± 0.8, P <0.01). After stepwise selection process, the multivariable logistic regression model included EuroSCORE (OR = 1.25, 95% CI: 0.99–1.59, P = 0.06) and preoperative creatinine (OR = 1.02, 95% CI 0.99–1.04, P = 0.15) as predictors of composite outcome.
Table 4 Operative outcomes in matched analysis
Matched Analysis A Matched Analysis B
Characteristic Redo-OPCAB (n = 111) Redo-CABG (n = 111) P-value Redo-OPCAB (n = 84) Redo-CABG (n = 84) P-value
In-hospital mortality (%) 2 (2) 5 (4) 0.25 0 (0.0) 4 (5) 0.13
AKI (%) 13 (12) 16 (14) 0.56 5 (6) 9 (11) 0.24
CVA (%) 1 (1) 1 (1) 1.00 0 (0) 0 (0) NA
Usage of IABP (%) 3 (3) 10 (10) 0.03 2 (2) 9 (11) 0.02
Composite outcome (%) 17 (15) 23 (21) 0.25 7 (8) 16 (19) 0.04
Reoperation for bleeding (%) 1 (1) 4 (4) 0.17 0 (0) 1 (1) 1.00
Completeness of revascularization (%) 77 (69) 91 (82) 0.02 57 (68) 60 (71) 0.60
Hospital stay 8.1 (6.2) 8.67 (5.6) 0.15 8 (5.4) 8.5 (5) 0.17
AKI: acute kidney injury (defined as peak of postoperative creatinine >% higher than preoperative value); CVA: cerebrovascular accident; IABP: intra-aortic balloon pump (either intraoperatively or postoperatively only).
Figure 3  Kaplan–Meier survival curves for redo-OPCAB and redo-CABG in the propensity score-Matched Analysis A and Matched Analysis B.

Matched Analysis B
In propensity score-Matched Analysis B, the number of patients was 84 in each group, including those undergone isolated CABG only as the primary operation. Postoperative outcomes are shown in Table 4. The rate of the composite end-point was reduced in the redo-OPCAB group (19 vs 8%, P = 0.04) and distributed as follows: in-hospital mortality (5 vs 0%, P = 0.13), AKI (11 vs 6%, P = 0.24), stroke (0% vs 0.0%, P = NA) and severe LCO requiring IABP (11 vs 2%, P = 0.02), all redo-CABG vs redo-OPCAB, respectively. Reopening for bleeding was 1% vs 0.0% (P = 1), completeness of revascularization was 71 vs 68% (P = 0.6) and length of stay was 8.5 ± 5.4 vs 8.5 ± 5 days (P = 0.17), all redo-CABG vs redo-OPCAB, respectively. Peak of postoperative SCrea was 135.3 ± 100.6 µmol/l and 110 ± 45 µmol/l for redo-CABG and redo-OPCAB, respectively (P = 0.01). Figure 3B shows the Kaplan–Meier survival for the 2 groups. The long-term survival was similar between groups (log-rank test: P = 0.61). Late survival at 1, 5, and 10 years was 97.6 vs 95.2%, 87.5 vs 88.9% and 71.6 vs 71.7% for redo-OPCAB vs redo-CABG, respectively. The multivariable logistic regression model identified redo-CABG (OR = 3.83, 95% CI: 1.14–12.8, P = 0.03) and urgent surgery (OR = 9.8, 95% CI 0.91–104.2, P = 0.06) as predictors for composite outcome.