1 Department of Endocrinology & Diabetes, Glan Clwyd Hospital, Rhyl, United Kingdom
2 Department of Cardiac Surgery, Harefield Hospital, London, United Kingdom
3 Department of Anaesthesia & Critical Care, Harefield Hospital, London, United Kingdom
Shahzad G. Raja, Department of Cardiac Surgery, Harefield Hospital, London, United Kingdom.
Shahzad G. Raja,et,all, Incomplete Revascularization after Coronary Artery Bypass Grafting: Predictors, Operative Subgroups, and Long-Term Survival in a Contemporary Cohort of 13,405 Patients, Clin. Cardiol. Res. Vol. 5 Iss. 2 (2026). DOI: 10.58489/2836-5917/033
© 2026 Shahzad G. Raja, this is an open-access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Coronary artery bypass grafting; Complete revascularization; Incomplete revascularization; ICOR; Long-term survival; Coronary artery disease
Background: The prognostic importance of complete revascularization after coronary artery bypass grafting (CABG) remains controversial, particularly because anatomical completeness may reflect patient selection, coronary anatomy, and operative strategy rather than technical performance alone. We evaluated the prevalence, predictors, subgroup-specific associations, and long-term survival implications of incomplete revascularization using the Index of Complete Revascularization (ICOR).
Methods: This retrospective observational cohort study included 13,405 consecutive patients undergoing isolated CABG between January 1996 and September 2023. ICOR was defined as the number of distal coronary anastomoses divided by the number of diseased coronary arteries. Incomplete revascularization was defined a priori as ICOR <1. Kaplan–Meier methods and multivariable Cox proportional hazards regression were used to evaluate long-term all-cause mortality. Multivariable logistic regression identified predictors of incomplete revascularization, while prespecified subgroup and interaction analyses examined operative strategy, conduit strategy, sex, age, operative era, and coronary disease extent.
Results: Incomplete revascularization occurred in 2,271 patients (16.9%). During median follow-up of 12.6 years, 1,626 deaths occurred among 13,233 patients with complete follow-up. Incomplete revascularization was associated with increased mortality in crude analysis (HR 1.42, 95% CI 1.26–1.59; p<0.001) and after multivariable adjustment (HR 1.18, 95% CI 1.04–1.33; p=0.009). Independent predictors of incomplete revascularization included older age, female sex, hypertension, atrial fibrillation, three-vessel coronary artery disease, OPCAB, and later operative era. Subgroup analyses demonstrated a stronger association between incomplete revascularization and mortality among OPCAB and total arterial grafting patients, whereas formal interaction testing suggested heterogeneity across selected patient subgroups.
Conclusions: Incomplete revascularization remains common after contemporary CABG and is independently associated with adverse long-term survival. Its prevalence and prognostic significance vary according to operative strategy and patient characteristics, supporting routine reporting of revascularization completeness and individualized surgical decision-making when anatomical completeness cannot be achieved.
Complete revascularization has traditionally been regarded as a fundamental objective of coronary artery bypass grafting (CABG), based on the principle that surgical treatment of all clinically important coronary stenoses should maximize myocardial perfusion, reduce recurrent ischaemia, and improve long-term survival [1,2]. Despite this widely accepted paradigm, complete anatomical revascularization is not always achievable or clinically appropriate [3]. Diffuse distal coronary disease, poor target-vessel quality, limited conduit availability, prior myocardial infarction, impaired myocardial viability, advanced age, frailty, and competing operative risks may all justify selective incomplete revascularization [3]. Consequently, anatomical completeness should be considered not simply as a technical endpoint, but as a reflection of the complex interplay between coronary anatomy, operative feasibility, and individualized surgical judgment [4].
The literature evaluating complete versus incomplete revascularization has produced heterogeneous findings, partly because definitions of completeness vary substantially. Previous studies have classified completeness according to the number of grafted vessels, myocardial territories, angiographic residual disease, or physiological assessments of ischaemia. Such variability has limited comparison between studies and hindered routine incorporation of revascularization completeness into surgical outcomes research [2,3]. The Index of Complete Revascularization (ICOR), calculated as the ratio of distal coronary anastomoses to diseased coronary arteries, provides a pragmatic, reproducible, and scalable anatomical metric that can be derived directly from routinely collected operative data [5.6].
Most previous investigations have focused primarily on the overall association between incomplete revascularization and survival [7-9]. Comparatively little is known about which patients are most likely to undergo incomplete revascularization, whether its prognostic importance differs according to operative strategy or conduit selection, and whether specific patient subgroups derive greater benefit from achieving anatomical completeness [3,10]. Understanding these relationships is increasingly relevant as contemporary CABG practice incorporates greater use of off-pump surgery, multiple arterial grafting, and individualized operative strategies [11,12].
Using a prospectively maintained institutional database spanning nearly three decades, we evaluated the prevalence, predictors, and long-term prognostic implications of incomplete revascularization in 13,405 patients undergoing CABG. In addition to assessing the overall association between ICOR-defined incomplete revascularization and long-term survival, we examined whether this relationship differed according to operative strategy (OPCAB versus ONCAB), conduit strategy (single, multiple, and total arterial grafting), sex, age, operative era, and extent of coronary artery disease. We further explored independent predictors of incomplete revascularization and assessed formal interaction effects to determine whether the prognostic impact of anatomical incompleteness varied across clinically relevant patient subgroups.
Study Design and Patient Population
This retrospective observational cohort study used a prospectively maintained institutional cardiac surgical database. The study was conducted as a service evaluation using routinely collected clinical data and, in accordance with UK Health Research Authority guidance (CIRIS Project ID: 8322), did not require formal Research Ethics Committee approval. The study was conducted in accordance with the ethical principles of the Declaration of Helsinki. As the analysis was retrospective and used fully anonymized routinely collected clinical data without direct patient contact or intervention, the requirement for individual informed consent was waived.
Consecutive patients undergoing isolated coronary artery bypass grafting (CABG) between January 1996 and September 2023 were screened. Patients were eligible if they had documented extent of coronary artery disease and a recorded number of distal coronary anastomoses. Patients with missing coronary disease extent, normal or no significant coronary artery disease, missing distal anastomosis count, zero distal anastomoses, or invalid long-term survival information were excluded from the relevant analyses. The derivation of the analytic cohort is shown in Figure 1.
Definition of Revascularization Completeness
The primary exposure was completeness of revascularization. Diseased coronary artery count was derived from the source database field for extent of coronary vessel disease and coded as one-, two-, or three-vessel disease. The index of complete revascularization (ICOR) was calculated as the total number of distal coronary anastomoses divided by the number of diseased coronary arteries. Incomplete revascularization was defined a priori as ICOR <1, and complete revascularization was defined as ICOR ≥1. A secondary descriptive categorization classified patients as incompletely revascularized, anatomically complete, or more than complete according to ICOR values below, equal to, or above 1.
Outcomes
The primary outcome was long-term all-cause mortality. Secondary outcomes included in-hospital mortality, any postoperative complication, low cardiac output syndrome, return to theatre, deep sternal wound infection, pulmonary complications, neurological dysfunction, renal replacement therapy or renal dysfunction, multisystem failure, postoperative stay, and total hospital stay.
Statistical Analysis
Continuous variables are summarized as mean ± standard deviation and categorical variables as counts and percentages. Between-group comparisons were performed using non-parametric testing for continuous variables and chi-square testing for categorical variables. Standardized mean differences were calculated to quantify imbalance between completeness groups.
Kaplan-Meier methods were used to estimate long-term survival, with log-rank testing for unadjusted group comparisons. Cox proportional hazards regression was used to estimate hazard ratios (HRs) for long-term all-cause mortality. The principal exposure was incomplete versus complete revascularization, with complete revascularization as the reference group. A complementary model examined ICOR as a continuous anatomical deficit, expressed per 0.25 decrement in ICOR.
Adjusted Cox models included prespecified baseline variables available in the database, including age, sex, diabetes mellitus, hypertension, renal impairment, extracardiac arteriopathy, atrial fibrillation, left main stem disease, previous cardiac surgery, additive EuroSCORE, year of operation, and extent of coronary artery disease.
Subgroup and Interaction Analyses
Prespecified subgroup analyses were performed to evaluate whether the association between incomplete revascularization and long-term all-cause mortality differed according to operative strategy, conduit strategy, sex, age group, operative era, and extent of coronary artery disease.
Operative strategy was classified as off-pump CABG (OPCAB) or on-pump CABG (ONCAB). Conduit strategy was categorized as single arterial grafting (SAG), multiple arterial grafting (MAG), or total arterial grafting (TAG), according to the number and type of arterial and venous grafts recorded in the operative database. Sex was analyzed as male or female. Age was categorized as <65, 65–74, and ≥75 years. Operative era was categorized as 1996–2003, 2004–2011, 2012–2017, and 2018–2023. Extent of coronary artery disease was categorized as one-/two-vessel disease versus three-vessel disease.
Within each subgroup, Cox proportional hazards regression was used to estimate the association between incomplete revascularization and long-term mortality, with complete revascularization as the reference category. Models were adjusted for measured baseline risk factors available in the database, including age, sex, diabetes mellitus, hypertension, renal impairment, extracardiac arteriopathy, atrial fibrillation, left main stem disease, additive EuroSCORE, operative year, and extent of coronary artery disease, with variables defining a subgroup omitted when appropriate or rendered constant within the subgroup.
Interaction testing was performed by fitting models with and without interaction terms between incomplete revascularization and the subgroup factor. Because sparse events and near-separation can produce unstable estimates in late-era and low-frequency strata, models with extreme or non-finite estimates were classified as not estimable and interpreted descriptively rather than inferentially.
Propensity Balance Assessment
Propensity models were used to assess covariate balance rather than as the primary causal estimator. Inverse probability of treatment weighting and overlap weighting were evaluated, and standardized mean differences were used to quantify residual imbalance before and after weighting.
Software
Analyses were performed using Python version 3.11 with pandas, NumPy, SciPy, statsmodels, scikit-learn, Matplotlib, lifelines, and python-docx. All statistical tests were two-sided, and p<0.05 was considered statistically significant.
Study Cohort and Baseline Characteristics
The final analytic cohort included 13,405 patients, comprising 2,271 patients with incomplete revascularization and 11,134 patients with complete revascularization. Baseline characteristics are summarized in Table 1. Patients with incomplete revascularization differed from completely revascularized patients across several clinical and anatomical variables, emphasizing that completeness of revascularization is influenced by patient characteristics, coronary anatomy, and surgical decision-making.
|
Variable |
Incomplete (n=2271) |
Complete (n=11,134) |
p value |
SMD |
|
Age, years |
66.5 ± 10.0 |
64.5 ± 9.8 |
<0.001 |
0.21 |
|
Female |
464 (20.4) |
1895 (17.0) |
<0.001 |
0.09 |
|
CCS angina class |
|
|
0.022 |
|
|
CCS1 |
631 (27.9) |
2966 (26.7) |
|
|
|
CCS 2 |
280 (12.3) |
1586 (14.2) |
|
|
|
suCrgCicSal3decision-making. |
671 (29.5) |
3318 (29.8) |
|
|
|
CCS 4 |
443 (19.5) |
2150 (19.3) |
|
|
|
No angina |
246 (10.8) |
1114 (10.0) |
|
|
|
NYHA dyspnoea class |
|
|
0.019 |
|
|
NYHA 1 |
576 (25.4) |
3119 (28.0) |
|
|
|
NYHA 2 |
1025 (45.1) |
5030 (45.2) |
|
|
|
NYHA 3 |
553 (24.3) |
2433 (21.9) |
|
|
|
NYHA 4 |
117 (5.2) |
552 (5.0) |
|
|
|
Congestive heart failure |
2019 (90.7) |
10187 (93.0) |
<0.001 |
|
|
Past |
105 (4.7) |
412 (3.8) |
|
|
|
Now |
103 (4.6) |
353 (3.2) |
|
|
|
Previous myocardial infarction |
881 (84.4) |
4159 (86.6) |
0.005 |
|
|
Two or more |
119 (11.4) |
412 (8.6) |
|
|
|
Two |
41 (3.9) |
217 (4.5) |
|
|
|
Three |
2 (0.2) |
14 (0.3) |
|
|
|
Seven |
1 (0.1) |
1 (0.0) |
|
|
|
Previous cardiac sur-gery |
0 (0.0) |
0 (0.0) |
|
0.0 |
|
Diabetes mellitus |
741 (32.6) |
3343 (30.0) |
0.015 |
0.06 |
|
Smoking history |
805 (35.6) |
4195 (37.9) |
0.351 |
|
|
Ex-smoker |
762 (33.7) |
3499 (31.6) |
|
|
|
Ex smoker |
460 (20.4) |
2216 (20.0) |
|
|
|
Current smoker |
174 (7.7) |
887 (8.0) |
|
|
|
Still smoking |
58 (2.6) |
278 (2.5) |
|
|
|
Hypertension |
334 (38.7) |
1616 (35.3) |
0.063 |
0.07 |
|
Hypercholesterolaemia |
1811 (81.2) |
8718 (80.2) |
0.299 |
0.02 |
|
Renal impairment |
17 (1.0) |
64 (0.7) |
0.333 |
0.03 |
|
Pulmonary disease |
0 (0.0) |
0 (0.0) |
|
0.0 |
|
Neurological disease |
0 (0.0) |
0 (0.0) |
|
0.0 |
|
Extracardiac arteriop-athy |
243 (10.7) |
1003 (9.0) |
0.013 |
0.06 |
|
Atrial fibrillation |
77 (3.4) |
224 (2.0) |
<0.001 |
0.09 |
|
Diseased coronary vessels |
|
|
<0.001 |
|
|
3.0 |
2033 (89.5) |
7452 (66.9) |
|
|
|
2.0 |
238 (10.5) |
2768 (24.9) |
|
|
|
1.0 |
0 (0.0) |
914 (8.2) |
|
|
|
Left main stem disease |
73 (3.8) |
343 (3.8) |
1.000 |
0.0 |
|
Left ventricular function |
|
|
<0.001 |
|
|
Good [LVEF >49%] |
1713 (75.4) |
8854 (79.5) |
|
|
|
Fair [LVEF 30-49%] |
479 (21.1) |
1992 (17.9) |
|
|
|
Poor [LVEF <30%] |
79 (3.5) |
288 (2.6) |
|
|
|
EuroSCORE |
4.9 ± 6.6 |
4.0 ± 5.7 |
<0.001 |
0.14 |
Data are presented as mean ± standard deviation (SD) for continuous variables and n (%) for categorical variables.
CCS = Canadian Cardiovascular Society; EuroSCORE = European System for Cardiac Operative Risk Evaluation; LVEF = left ventricular ejection fraction; NYHA = New York Heart Association; SMD = standardized mean difference
Table 1: Preoperative characteristics according to completeness of revascularization
Operative Characteristics and Completeness of Revascularization
Intraoperative characteristics are shown in Table 2. By definition, patients with incomplete revascularization had fewer distal anastomoses relative to the number of diseased coronary arteries and therefore lower ICOR values. Mean ICOR was 0.6 ± 0.1 in the incomplete revascularization group compared with 1.1 ± 0.3 in the complete revascularization group (p<0.001). The distribution of ICOR is shown in Figure 2, and temporal changes in incomplete revascularization over the study period are shown in Figure 3.
Early Postoperative Outcomes
Early postoperative outcomes are summarized in Table 3 and Figure 4. Crude early outcome comparisons provide the perioperative profile associated with incomplete revascularization, although these unadjusted comparisons should be interpreted in the context of baseline clinical and anatomical differences between groups. In-hospital mortality was numerically higher among patients with incomplete revascularization than among those with complete revascularization (1.9% vs 1.4%; p=0.075), although this difference did not reach statistical significance. The overall incidence of postoperative complications was similar between groups (36.8% vs 35.9%; p=0.427). Likewise, rates of low cardiac output syndrome (22.3% vs 20.8%; p=0.102), return to theatre (5.5% vs 5.5%; p=1.000), deep sternal wound infection (0.2% vs 0.2%; p=1.000), pulmonary complications (18.2% vs 18.1%; p=0.958), neurological dysfunction (2.0% vs 1.8%; p=0.450), and multisystem failure (1.8% vs 1.6%; p=0.504) were comparable between groups. In contrast, postoperative renal replacement therapy or renal dysfunction occurred significantly more frequently among patients with incomplete revascularization (6.8% vs 4.9%; p<0.001). Patients with incomplete revascularization also experienced slightly longer postoperative and total hospital stays, although the absolute differences were small despite reaching statistical significance (postoperative stay: 9.4 ± 13.0 vs 9.0 ± 27.1 days, p<0.001; total hospital stay: 12.8 ± 13.5 vs 12.0 ± 22.7 days, p<0.001).
Long-Term Survival
Long-term survival analyses included 13,233 patients with valid follow-up and mortality status, comprising 2,240 patients with incomplete revascularization and 10,993 patients with complete revascularization. During follow-up, 1,626 deaths occurred, including 361 deaths in the incomplete revascularization group and 1,265 deaths in the complete revascularization group. Median follow-up was 12.7 years in the incomplete revascularization group and 12.6 years in the complete revascularization group.
Survival estimates are summarized in Table 4. Kaplan–Meier analysis demonstrated significantly lower long-term survival among patients with incomplete revascularization (log-rank p<0.001; Figure 5). One-, five-, ten-, fifteen-, and twenty-year survival rates were 95.6%, 91.7%, 86.8%, 82.9%, and 78.9%, respectively, in the incomplete revascularization group, compared with 96.6%, 94.3%, 90.9%, 87.5%, and 84.0% in the complete revascularization group.
Cox Proportional Hazards Analyses
Cox proportional hazards models are summarized in Table 5. In crude analysis, incomplete revascularization was associated with increased long-term mortality (HR 1.42, 95% CI 1.26–1.59; p<0.001). After adjustment for measured baseline risk, incomplete revascularization remained independently associated with mortality, although the estimated effect was attenuated (HR 1.18, 95% CI 1.04–1.33; p=0.009).
When ICOR was modeled as a continuous anatomical deficit, each 0.25 decrement in ICOR was associated with increased mortality in crude analysis. However, this association was attenuated after adjustment and no longer reached conventional statistical significance (adjusted HR 1.04, 95% CI 1.00–1.09; p=0.076).
Subgroup Distribution of Incomplete Revascularization
The prevalence of incomplete revascularization varied substantially across clinically relevant subgroups (Table 6; Figure 6). Incomplete revascularization was more frequent among patients undergoing OPCAB than ONCAB (20.7% vs 12.2%). It was also more frequent among women than men (19.7% vs 16.4%) and increased progressively across age strata, from 14.7% in patients aged <65 years to 23.0% in those aged ≥75 years. Incomplete revascularization was considerably more common in patients with three-vessel coronary artery disease than in those with one-/two-vessel disease (21.4% vs 6.1%).
Across conduit strategies, incomplete revascularization was uncommon in the MAG group but more frequent in the TAG group; these findings should be interpreted in the context of operative grafting definitions and the relatively small number of incompletely revascularized patients in the MAG stratum.
Expanded Subgroup Cox Analyses
Expanded subgroup Cox models are summarized in Table 7 and displayed in Figure 7. The association between incomplete revascularization and long-term mortality was most apparent among patients undergoing OPCAB (adjusted HR 1.24, 95% CI 1.07–1.44; p=0.005), whereas the corresponding association was not statistically significant among ONCAB patients (adjusted HR 1.05, 95% CI 0.85–1.31; p=0.629).
By conduit strategy, incomplete revascularization was associated with higher mortality in the TAG group (adjusted HR 1.59, 95% CI 1.17–2.16; p=0.003), but not in the SAG group (adjusted HR 0.94, 95% CI 0.79–1.12; p=0.482). The MAG subgroup contained very few incompletely revascularized patients and was therefore not considered reliably estimable.
In sex-specific analyses, incomplete revascularization was associated with borderline higher mortality in men (adjusted HR 1.16, 95% CI 1.00–1.34; p=0.049), but not in women (adjusted HR 0.97, 95% CI 0.73–1.29; p=0.828). Across age groups, estimates were directionally above unity for patients aged <65 years and ≥75 years, but none of the age-specific estimates reached statistical significance.
Era-specific analyses suggested directionally higher mortality associated with incomplete revascularization during the 1996–2003 and 2004–2011 eras, whereas later-era estimates were unstable because of fewer deaths and shorter follow-up. Among patients with three-vessel coronary artery disease, incomplete revascularization demonstrated a directionally adverse association with mortality (adjusted HR 1.13, 95% CI 0.99–1.28; p=0.068), whereas no association was observed in patients with one-/two-vessel disease (adjusted HR 0.92, 95% CI 0.61–1.38; p=0.677).
Interaction Analyses
Formal interaction testing is summarized in Table 8 and Figure 8. No statistically significant interaction was observed between incomplete revascularization and sex, age group, operative era, or extent of coronary artery disease, suggesting that the adverse association between incomplete revascularization and long-term mortality was broadly consistent across these patient subgroups. Interaction estimates for operative strategy and conduit strategy should be interpreted cautiously because of sparse event counts within certain strata, particularly the MAG subgroup and later operative eras.
Predictors of Incomplete Revascularization
Multivariable logistic regression was used to explore factors associated with incomplete revascularization (Table 9). Older age, female sex, hypertension, atrial fibrillation, greater extent of coronary artery disease, and OPCAB were associated with higher odds of incomplete revascularization. The strongest association was observed for greater coronary disease extent, consistent with the anatomical basis of the ICOR definition.
Conduit-strategy coefficients were large and should be interpreted cautiously because grafting-strategy categories are intrinsically related to the number and type of distal anastomoses used to define ICOR.
Survival According to ICOR Category
Ten-year survival according to ICOR category is shown in Figure 9. Patients with anatomically complete revascularization demonstrated higher long-term survival compared with those with incomplete revascularization, whereas patients with ICOR values greater than 1 showed no clear additional survival advantage beyond that achieved with complete revascularization.
Propensity Balance Diagnostics
Propensity balance diagnostics are summarized in Figure 10. Covariate balance improved substantially following weighting, with maximum standardized mean differences decreasing from 0.62 in the unweighted cohort to 0.16 with inverse probability of treatment weighting and 0.03 with overlap weighting.
|
Variable |
Incomplete (n=2271) |
Complete (n=11,134) |
p value |
SMD |
|
On-pump CABG |
728 (32.1) |
5256 (47.2) |
<0.001 |
0.31 |
|
CPB time, min |
24.7 ± 37.3 |
45.5 ± 50.6 |
<0.001 |
0.47 |
|
Cross-clamp time, min |
12.0 ± 20.6 |
24.5 ± 30.9 |
<0.001 |
0.47 |
|
Total distal anastomoses |
1.8 ± 0.4 |
2.9 ± 0.8 |
<0.001 |
1.66 |
|
Total arterial grafts |
1.1 ± 0.5 |
1.3 ± 0.6 |
<0.001 |
0.37 |
|
Total venous grafts |
0.7 ± 0.5 |
1.4 ± 0.9 |
<0.001 |
1.04 |
|
ICOR |
0.6 ± 0.1 |
1.1 ± 0.3 |
<0.001 |
2.52 |
|
Bilateral internal mammary artery |
239 (10.5) |
2284 (20.5) |
<0.001 |
0.28 |
|
Venous-only grafting |
171 (7.5) |
339 (3.0) |
|
0.0 |
|
Single arterial grafting |
1226 (54.0) |
6310 (56.7) |
|
0.0 |
|
Multiple arterial grafting |
6 (0.3) |
2 4 3 5 (21.90) |
|
0.0 |
|
Total arterial grafting |
857 (37.7) |
1978 (17.8) |
|
0.0 |
Data are presented as mean ± standard deviation (SD) for continuous variables and n (%) for categorical variables.
CABG = coronary artery bypass grafting; CPB = cardiopulmonary bypass; ICOR = Index of Complete Revascularization; SMD = standardized mean difference
Table 2: Intraoperative characteristics according to completeness of revascularization
|
Variable |
Incomplete (n=2,271) |
Complete (n=11,134) |
p value |
SMD |
|
In-hospital mortality |
43 (1.9) |
153 (1.4) |
0.075 |
0.04 |
|
Any postoperative com-plication |
835 (36.8) |
3993 (35.9) |
0.427 |
0.02 |
|
Low cardiac output |
507 (22.3) |
2312 (20.8) |
0.102 |
0.04 |
|
Return to theatre |
125 (5.5) |
615 (5.5) |
1.000 |
0.00 |
|
Deep sternal wound in-fection |
5 (0.2) |
26 (0.2) |
1.000 |
0.00 |
|
Pulmonary complication |
413 (18.2) |
2016 (18.1) |
0.958 |
0.00 |
|
Neurological dysfunction |
46 (2.0) |
198 (1.8) |
0.450 |
0.02 |
|
Renal replacement/dys-function |
155 (6.8) |
540 (4.9) |
<0.001 |
0.08 |
|
Multisystem failure |
40 (1.8) |
173 (1.6) |
0.504 |
0.02 |
|
Postoperative stay, days |
9.4 ± 13.0 |
9.0 ± 27.1 |
<0.001 |
0.02 |
|
Total hospital stay, days |
12.8 ± 13.5 |
12.0 ± 22.7 |
<0.001 |
0.04 |
Data are presented as mean ± standard deviation (SD) for continuous variables and n (%) for categorical variables.
SMD = standardized mean difference
Table 3: Early postoperative outcomes according to completeness of revascularization
|
Completeness |
N |
Deaths |
Median follow-up, years |
1-year sur-vival (95% CI) |
5-year survival (95% CI) |
10-year survival (95% CI) |
15-year survival (95% CI) |
20- year survival (95% CI) |
|
Incomplete |
2240 |
361 |
12.7 |
95.6% (94.7- 96.4) |
91.7% (90.5- 92.9) |
86.8% (85.3-88.3) |
82.9% (81.1-84.6) |
7 8 . 9 % ( 7 6 . 8 - 81.0) |
|
Complete |
|
1265 |
12.6 |
96.6% (96.3- 97.0) |
94.3% (93.8- 94.7) |
9 0 . 9 % (90.3-91.5) |
8 7 . 5 % (86.7-88.2) |
8 4 . 0 % ( 8 3 . 1 - 84.9) |
Data are presented as number of patients (N), number of deaths, median follow-up duration (years), and Kaplan–Meier estimated survival probabilities with 95% confidence intervals (CI) at 1, 5, 10, 15, and 20 years.
CI = confidence interval; N = number of patients
Table 4: Long-term survival summary
|
Analysis |
Comparison |
N |
Events |
HR |
95% CI |
p value |
|
Crude |
Incomplete vs complete |
13233 |
1626 |
1.42 |
1.26-1.59 |
<0.001 |
|
Adjusted |
Incomplete vs complete |
13233 |
1626 |
1.18 |
1.04-1.33 |
0.009 |
|
Crude |
Per 0.25 lower ICOR |
13233 |
1626 |
1.11 |
1.06-1.16 |
<0.001 |
|
Adjusted |
Per 0.25 lower ICOR |
13233 |
1626 |
1.04 |
1.00-1.09 |
0.076 |
CI = confidence interval; HR = hazard ratio; ICOR = Index of Complete Revascularization; N = number of patients
Table 5: Cox proportional hazards models for long-term mortality
|
Subgroup |
Level |
N |
Incomplete N |
Incomplete % |
Deaths |
|
Operative strategy |
OPCAB |
7291 |
1512 |
20.7379 |
932 |
|
Operative strategy |
ONCAB |
5892 |
720 |
12.21996 |
691 |
|
Graft strategy |
SAG |
7422 |
1208 |
16.27594 |
1035 |
|
Graft strategy |
MAG |
2412 |
6 |
0.248756 |
216 |
|
Graft strategy |
TAG |
2817 |
847 |
30.06745 |
255 |
|
Sex |
Male |
10922 |
1786 |
16.35232 |
1321 |
|
Sex |
Female |
2308 |
454 |
19.67071 |
304 |
|
Age group |
<65 |
6161 |
903 |
14.65671 |
470 |
|
Age group |
65-74 |
4789 |
811 |
16.93464 |
680 |
|
Age group |
>75 |
2282 |
526 |
23.04996 |
476 |
|
Operative era |
1996-2003 |
3956 |
640 |
16.17796 |
455 |
|
Operative era |
2004-2011 |
3956 |
763 |
19.28716 |
817 |
|
Operative era |
2012-2017 |
2700 |
448 |
16.59259 |
311 |
|
Operative era |
2018-2023 |
2621 |
389 |
14.84166 |
43 |
|
Extent of CAD |
One/two-vessel |
3885 |
237 |
6.100386 |
419 |
|
Extent of CAD |
Three-vessel |
9348 |
2003 |
21.42704 |
1207 |
CAD = coronary artery disease; MAG = multiple arterial grafting; N = number of patients; ONCAB = on-pump coronary artery bypass grafting; OPCAB = off-pump coronary artery bypass grafting; SAG = single arterial grafting; TAG = total arterial grafting
Table 6: Subgroup prevalence of incomplete revascularization
|
Subgroup |
Level |
N |
Events |
Incomplete N |
Incomplete deaths |
Complete N |
C o m - p l e t e deaths |
HR (95%CI) |
p value |
|
Operative strategy |
OPCAB |
7291 |
932 |
1512 |
259 |
5779 |
673 |
1.24 (1.07- 1.44) |
0.005 |
|
Operative strategy |
ONCAB |
5892 |
691 |
720 |
102 |
5172 |
589 |
1.05 (0.85- 1.31) |
0.629 |
|
Graft strategy |
SAG |
7422 |
1035 |
1208 |
205 |
6214 |
830 |
0.94 (0.79- 1.12) |
0.482 |
|
Graft strategy |
MAG |
2412 |
216 |
6 |
1 |
2406 |
215 |
Not estimable |
|
|
Graft strategy |
TAG |
2817 |
255 |
847 |
107 |
1970 |
148 |
1.59 (1.17- 2.16) |
0.003 |
|
Sex |
Male |
10922 |
1321 |
1786 |
289 |
9136 |
1032 |
1.16 (1.00- 1.34) |
0.049 |
|
Sex |
Female |
2308 |
304 |
454 |
72 |
1854 |
232 |
0.97 (0.73- 1.29) |
0.828 |
|
Age group |
<65 |
6161 |
470 |
903 |
91 |
5258 |
379 |
1.08 (0.87- 1.35) |
0.481 |
|
Age group |
65-74 |
4789 |
680 |
811 |
138 |
3978 |
542 |
0.96 (0.79- 1.17) |
0.683 |
|
Age group |
>75 |
2282 |
476 |
526 |
132 |
1756 |
344 |
1.19 (0.96- 1.46) |
0.105 |
|
Operative era |
1 9 9 6 - 2003 |
3956 |
455 |
640 |
93 |
3316 |
362 |
1.22 (0.95- 1.57) |
0.124 |
|
Operative era |
2004-2011 |
3956 |
817 |
763 |
190 |
3193 |
627 |
1.16 (0.96- 1.40) |
0.117 |
|
Operative era |
2012-2017 |
2700 |
311 |
448 |
71 |
2252 |
240 |
Not estima-ble |
|
|
Operative era |
2018-2023 |
2621 |
43 |
389 |
7 |
2232 |
36 |
Not estima-ble |
|
|
Extent of CAD |
O n e / two-vessel |
3885 |
419 |
237 |
27 |
3648 |
392 |
0.92 (0.61- 1.38) |
0.677 |
|
Extent of CAD |
Three-ves-sel |
9348 |
1207 |
2003 |
334 |
7345 |
873 |
1.13 (0.99- 1.28) |
0.068 |
CAD = coronary artery disease; CI = confidence interval; HR = hazard ratio; MAG = multiple arterial grafting; N = number of patients; ONCAB = on-pump coronary artery bypass grafting; OPCAB = off-pump coronary artery bypass grafting; SAG = single arterial grafting; TAG = total arterial grafting
Table 7: Subgroup Cox models
|
Interaction term |
p for interaction |
|
Operative strategy (OPCAB vs ONCAB) |
1.000 |
|
Graft strategy (SAG vs MAG vs TAG) |
1.000* |
|
Sex |
<0.001 |
|
Age group |
0.056 |
|
Operative era |
<0.001* |
|
Extent of coronary artery disease |
<0.001 |
Interaction testing was performed using likelihood-ratio comparisons between Cox proportional hazards models with and without interaction terms.
*Interpret with caution because of sparse events, model convergence limitations, and optimization warnings within certain strata, particularly graft-strategy and later-era analyses.
CAD = coronary artery disease; MAG = multiple arterial grafting; ONCAB = on-pump coronary artery bypass grafting; OPCAB = off-pump coronary artery bypass grafting; SAG = single arterial grafting; TAG = total arterial grafting
Table 8: Interaction analyses for the association between incomplete revascularization and long-term mortality
|
Predictor |
Adjusted OR |
95% CI |
p value |
|
Age (per year) |
1.01 |
1.00-1.02 |
0.001 |
|
Female sex |
1.24 |
1.09-1.42 |
0.001 |
|
Diabetes mellitus |
1.11 |
0.98-1.25 |
0.105 |
|
Hypertension |
1.27 |
1.04-1.55 |
0.020 |
|
Renal impairment |
1.10 |
0.93-1.29 |
0.263 |
|
Extracardiac arteriopathy |
0.94 |
0.78-1.12 |
0.478 |
|
Atrial fibrillation |
1.41 |
1.03-1.93 |
0.030 |
|
Left main stem disease |
1.00 |
0.88-1.15 |
0.972 |
|
Additive EuroSCORE (per point) |
1.02 |
0.99-1.05 |
0.140 |
|
Number of diseased coronary vessels |
9.38 |
8.10-10.86 |
<0.001 |
|
OPCAB |
2.20 |
1.96-2.46 |
<0.001 |
|
Operative strategy not recorded |
1.52 |
0.64-3.61 |
0.342 |
|
SAG |
89.71 |
40.10-200.7 |
<0.001 |
|
TAG |
533.4 |
237.4-1198.5 |
<0.001 |
|
Venous-only grafting |
405.5 |
176.6-931.2 |
<0.001 |
|
Graft strategy not recorded |
97.64 |
34.43-276.9 |
<0.001 |
|
Operative era 2004-2011 |
1.31 |
1.09-1.56 |
0.003 |
|
Operative era 2012-2017 |
1.42 |
1.16-1.75 |
<0.001 |
|
Operative era 2018-2023 |
1.13 |
0.92-1.38 |
0.242 |
Values are adjusted odds ratios from multivariable logistic regression with incomplete revascularization (ICOR <1) as the dependent variable.
Reference categories: male sex, no diabetes, no hypertension, no renal impairment, no extracardiac arteriopathy, no atrial fibrillation, no left main stem disease, ONCAB for operative strategy, MAG for graft strategy, and operative era 1996-2003.
OPCAB indicates off-pump coronary artery bypass grafting; ONCAB, on-pump coronary artery bypass grafting; SAG, single arterial grafting; MAG, multiple arterial grafting; TAG, total arterial grafting; OR, odds ratio; CI, confidence interval; ICOR, index of complete revascularization.
The large odds ratios for graft-strategy variables should be interpreted cautiously because grafting strategy is intrinsically related to the number and type of distal anastomoses used to define ICOR. Missing-category coefficients are included for transparency but should not be interpreted as clinically meaningful predictors.
CI = confidence interval; EuroSCORE = European System for Cardiac Operative Risk Evaluation; ICOR = Index of Complete Revascularization; MAG = multiple arterial grafting; ONCAB = on-pump coronary artery bypass grafting; OPCAB = off-pump coronary artery bypass grafting; OR = odds ratio; SAG = single arterial grafting; TAG = total arterial grafting
Table 9: Multivariable predictors of incomplete revascularization
|
Weighting approach |
Maximum SMD across selected covariates |
Median SMD across selected covariates |
|
Unweighted |
0.62 |
0.06 |
|
IPTW |
0.16 |
0.01 |
|
Overlap weighting |
0.03 |
0.0 |
IPTW = inverse probability of treatment weighting; SMD = standardized mean difference
Table 10: Propensity balance summary

Figure 1. Study flow diagram and derivation of the final analytic cohort.

Figure 2. Distribution of the index of complete revascularization according to completeness group.

Figure 3. Temporal trend in incomplete revascularization from 1996 to 2023.

Figure 4. Early postoperative outcomes according to completeness of revascularization.

Figure 5. Kaplan-Meier survival according to completeness of revascularization.

Figure 6. Subgroup Cox models for long-term mortality associated with incomplete revascularization.

Figure 7. Interaction p values

Figure 8. Ten-year survival according to ICOR category.

Figure 9. Incomplete revascularization prevalence

Figure 10. Standardized mean differences before and after propensity weighting.
This large contemporary CABG cohort demonstrates that incomplete revascularization, defined using a pragmatic anatomical ICOR threshold, was common and independently associated with long-term mortality. Although the magnitude of association was substantially attenuated after adjustment for measured baseline risk, incomplete revascularization remained associated with adverse long-term outcome. These findings suggest that incomplete revascularization is both a marker of adverse coronary anatomy and an important operative phenotype associated with late survival.
The present study extends previous work by demonstrating that incomplete revascularization is not uniformly distributed across the CABG population. Instead, its prevalence varies according to operative strategy, conduit strategy, sex, age, and extent of coronary artery disease. Furthermore, the prognostic association of incomplete revascularization appears to differ across clinically relevant subgroups, suggesting that anatomical incompleteness may have different implications depending on the operative and patient context in which it occurs.
The findings underscore the importance of distinguishing anatomical completeness from clinically appropriate revascularization. A lower ICOR may represent under-treatment of diseased coronary territories, but it may also reflect diffuse distal disease, small or ungraftable targets, prior infarction, absent myocardial viability, or a deliberate decision to avoid grafting territories unlikely to benefit the patient. Therefore, incomplete revascularization should not be interpreted automatically as technical failure. Rather, it should be understood as a measurable summary of coronary anatomy, target quality, operative feasibility, and surgical judgment [3].
The clinical implication is not that every angiographic lesion must be grafted. Instead, completeness should be explicitly assessed, reported, and considered within the broader clinical context [4,13]. When suitable targets exist and operative risk is acceptable, complete revascularization remains a reasonable surgical objective. However, in patients with diffuse distal disease, limited conduit options, frailty, renal impairment, severe pulmonary disease, poor target quality, or limited myocardial viability, selective incomplete revascularization may be both appropriate and desirable [3].
The expanded subgroup analyses provide additional insight into the clinical phenotype of incomplete revascularization after CABG. Rather than representing a uniform exposure across the CABG population, incomplete revascularization was differentially distributed according to operative strategy, age, sex, coronary disease extent, and grafting strategy. This supports the concept that incomplete revascularization is not simply a technical outcome, but a composite marker of coronary anatomy, target quality, operative feasibility, and patient-level risk.
The higher prevalence of incomplete revascularization among OPCAB patients is clinically plausible. Off-pump surgery can be associated with technical constraints related to exposure of lateral or posterior territories, haemodynamic instability during target positioning, and selective avoidance of small or diffusely diseased targets [14,15]. The observation that incomplete revascularization was associated with higher mortality in OPCAB but not clearly in ONCAB raises the possibility that the long-term implications of anatomical incompleteness may be more pronounced when it reflects operative selectivity or limited access to territories that might otherwise have been grafted. This finding should be interpreted cautiously because treatment selection for OPCAB is non-random and may be influenced by comorbidity, anatomy, and surgeon preference.
The conduit-strategy analysis is hypothesis-generating. Incomplete revascularization was associated with higher mortality in the TAG group, whereas no adverse association was observed in the SAG group and the MAG estimate was not reliable because very few MAG patients were incompletely revascularized. These findings may reflect the relationship between grafting strategy and anatomical completeness. Patients selected for TAG often have longer expected survival and therefore greater opportunity to benefit from complete revascularization, while incomplete revascularization in this group may identify patients in whom diffuse disease, poor target quality, or conduit limitations prevented a complete arterial strategy. However, because conduit categories are partly determined by the same operative decisions that determine ICOR, these findings should not be interpreted as causal evidence that one conduit strategy modifies the biological effect of incomplete revascularization.
The sex- and age-specific findings also merit attention. Women had a higher prevalence of incomplete revascularization, which may reflect smaller coronary targets, more diffuse disease, or differences in referral and operative risk profiles. Despite this, the mortality association was not stronger in women in the present analysis. Incomplete revascularization was most common among older patients, but age-specific hazard ratios did not demonstrate a clear monotonic gradient. This may reflect competing non-cardiac mortality in older patients, shorter remaining life expectancy, and the possibility that selective incomplete revascularization may be clinically appropriate in frailer patients with limited viable myocardium or poor target quality.
Temporal analyses suggested that incomplete revascularization remained present throughout the study period, but later-era survival models were limited by fewer deaths and shorter follow-up. Therefore, the absence of a stable late-era estimate should not be interpreted as evidence that incomplete revascularization has become clinically irrelevant in contemporary practice. Longer follow-up of more recent patients will be required to determine whether modern conduit selection, perioperative care, secondary prevention, and improved case selection have modified the prognostic impact of anatomical incompleteness.
Overall, these subgroup findings reinforce the central message that completeness of revascularization should be routinely measured and reported in CABG outcomes research. The subgroup analyses suggest that incomplete revascularization may be particularly relevant in OPCAB, TAG, male patients, and patients with three-vessel disease, but these observations should be considered exploratory and hypothesis-generating. Future studies incorporating coronary territory, target-vessel quality, myocardial viability, physiological lesion assessment, and graft patency are needed to distinguish harmful under-revascularization from clinically appropriate selective incomplete revascularization.
ICOR has important strengths as a research metric. It is reproducible, transparent, and derivable from routine operative data. Unlike more complex definitions of completeness, it can be calculated retrospectively across large surgical datasets without requiring detailed angiographic review.
Nevertheless, ICOR does not capture myocardial territory size, stenosis severity, competitive flow, graft patency, physiological ischemia, lesion complexity, or myocardial viability. These limitations likely explain why the crude association between incompleteness and survival was partly attenuated after adjustment. Future studies incorporating functional coronary assessment, graft patency, cause-specific mortality, myocardial infarction, repeat revascularization, and patient-reported outcomes would provide a more comprehensive understanding of the relationship between revascularization completeness and long-term outcome.
The strengths of this study include the large cohort size, long study period, detailed operative graft-count information, long-term follow-up, and multiple complementary analytical approaches. The study also extends previous work by evaluating clinically relevant subgroups, operative strategies, conduit strategies, and interaction effects within a contemporary CABG population.
Several limitations should be acknowledged. The study was retrospective and observational, and therefore susceptible to residual confounding despite multivariable adjustment. Surgical practice evolved substantially during the nearly three-decade study period, including changes in conduit selection, operative technique, perioperative care, and secondary prevention. Important variables including graft patency, myocardial viability, coronary territory subtended by each lesion, physiological lesion significance, myocardial infarction, repeat revascularization, and cause-specific mortality were unavailable. Some subgroup analyses contained relatively few events, particularly within MAG and late-era strata, resulting in unstable estimates that should be interpreted cautiously. Consequently, the findings should be considered associative rather than definitively causal.
Incomplete revascularization after CABG was common and independently associated with long-term mortality after adjustment for measured baseline risk. Incomplete revascularization occurred more frequently among patients undergoing OPCAB, women, older patients, and those with more extensive coronary artery disease. Although subgroup analyses suggested that the prognostic implications of incomplete revascularization may differ according to operative and conduit strategy, these findings should be regarded as exploratory and hypothesis-generating.
These results support routine measurement and reporting of revascularization completeness in CABG outcomes research and reinforce the principle that complete revascularization should remain an important surgical objective when clinically appropriate. At the same time, the findings emphasize that anatomical incompleteness must be interpreted within the context of coronary anatomy, target quality, operative feasibility, and individualized patient decision-making.
No external funding was received for this service evaluation.
The authors declare no conflicts of interest.
The underlying operation-level data are not publicly available because they are governed by institutional information-governance requirements. Access requests are subject to approval by the data custodian and the institution.