Studies examining outcomes after coronary artery bypass grafting (CABG) by gender and/or race have shown conflicting results. It remains to be determined if, or how, gender and race are independent risk factors for CABG operative mortality. Using all consecutive patients who underwent isolated CABG at Baylor University Medical Center in Dallas, Texas, from January 2004 to October 2011, the risk-adjusted associations between gender and race, respectively, and operative mortality were estimated using a generalized propensity approach, accounting for recognized Society of Thoracic Surgeons risk factors for mortality. Women were nearly 2 times more likely to die during or within 30 days of the operation than men (odds ratio 1.96, 95% confidence interval 1.44 to 2.66, p <0.0001), while no significant mortality differences were observed among races. In conclusion, these findings suggest that women face a significantly greater risk for operative death that should be taken into account during the treatment decision-making process but that race is not associated with CABG mortality and so should not be among the factors considered.
Studies examining outcomes after coronary artery bypass grafting (CABG) by gender and race have shown conflicting results. This conflict is reflected in the clinical tools available to surgeons advising potential candidates for CABG; for example, the operative risk score models include gender, but the current CABG guidelines make no differentiation by either gender or race. Additional evidence addressing whether, and how, female gender and race affect patient CABG operative mortality is needed for the continued refinement of these tools. In this study, we analyzed a large cohort of consecutive isolated CABG patients to determine if, and to what extent, there are differences in operative mortality between men and women and among patients of different races and ethnicities.
Methods
This retrospective cohort study included all consecutive patients who underwent isolated CABG at Baylor University Medical Center in Dallas, Texas, from January 2004 to October 2011. Patients with previous valve surgery, preoperative endocarditis, and/or ventricular assist devices were excluded. The final study cohort included 8,154 consecutive patients. Data routinely collected by Baylor University Medical Center for the Society of Thoracic Surgeons (STS) Adult Cardiac Surgery Database were used to address the study research questions. The clinical and nonclinical variables included in this data set are listed in Table 1 . Race and ethnicity were categorized as white, black, hispanic, or other/unknown, and operative mortality was calculated using the STS definition: the patient died during the operation, during the hospital stay, or within 30 days of discharge.
| Characteristic | Men (n=6068, 74.4%) | Women (n=2086, 25.6%) | P-value ∗ |
|---|---|---|---|
| Age (years) | 63.7 ± 10.3 † | 65.7 ± 10.8 † | <0.003 |
| Body mass index (kg/m 2) | 29.5 ± 5.7 † | 30.2± 7.2 † | 0.979 |
| White | 80.0% | 72.4% | |
| Black | 5.7% | 12.9% | |
| Hispanic | 7.1% | 7.8% | |
| Other/Unknown | 7.3% | 6.9% | |
| Diabetes mellitus | 37.5% | 47.4% | <0.003 |
| Renal failure | 9.0% | 10.0% | >0.99 |
| Creatinine (mg/dL) | 1.3 ± 1.1 † | 1.1 ± 1.0 † | <0.003 |
| Chronic lung disease | 13.7% | 18.1% | <0.003 |
| Systemic hypertension | 80.5% | 85.0% | <0.003 |
| Peripheral vascular disease | 13.6% | 17.5% | <0.003 |
| Cerebrovascular disease | 11.6% | 17.1% | <0.003 |
| Time from last myocardial infarction to surgery (hours) | <0.99 | ||
| None | 57.3% | 58.0% | |
| ≤6 | 0.9% | 1.0% | |
| >6 but <24 | 1.3% | 1.7% | |
| ≥24 | 40.5% | 39.4% | |
| Tobacco Use | <0.003 | ||
| Never | 57.0% | 63.2% | |
| Previous | 16.2% | 10.3% | |
| Current | 26.9% | 26.5% | |
| Heart failure | 15.9% | 20.8% | <0.003 |
| Previous percutaneous coronary intervention | 27.7% | 25.1% | 0.464 |
| Previous coronary bypass | 5.9% | 4.9% | >0.99 |
| Preoperative angina pectoris | 67.5% | 71.9% | 0.004 |
| Preoperative atrial fibrillation | 17.2% | 14.9% | 0.317 |
| Preoperative ejection fraction (%) | 48.1 ± 13.9 † | 51.4 ± 14.0 † | <0.003 |
| Preoperative left main narrowing | 29.0% | 30.0% | <0.99 |
| Operation | <0.003 | ||
| Elective | 52.2% | 46.2% | |
| Non-Elective | 47.8% | 53.8% | |
| Off-pump | 26.3% | 29.1% | |
| On-pump | 73.7% | 70.9% | |
| Preoperative intra-aortic balloon pump | 15.6% | 16.1% | >0.99 |
| Unadjusted operative mortality | 2.09% | 4.75% | <.0001 |
∗ p-values using Pearson χ 2 and Bonferroni correction.
Means, standard deviations, and percentages were calculated to describe the study cohort. Differences in demographic and clinical details were tested with Wilcoxon (for continuous factors) or chi-square (for categorical factors) tests. Bonferroni correction was used to account for multiplicity.
Two propensity-adjusted (by recognized STS risk factors for mortality ; Tables 1 and 2 ) logistic regression models were used to assess the association between operative mortality and gender and between operative mortality and race. All continuous variables were modeled with restricted cubic splines with 5 knots to estimate the propensity score and to fit the propensity score in the final logistic models, avoiding the problems associated with categorization. Multiple imputation by Markov-chain Monte Carlo simulation was used to account for missing data (ejection fraction: 8.0% missing; serum creatinine level: 0.7% missing; surgical status: 0.1% missing; and body mass index: 0.1% missing). Effect modification between (1) CABG type (off-pump or on-pump), elective versus urgent or emergent CABG, and gender; and (2) CABG type, elective versus urgent or emergent CABG, and race was tested.
| Characteristic | Race | ||||
|---|---|---|---|---|---|
| White (n=6365, 78.1%) | Black (n=612, 7.5%) | Hispanic (n=593, 7.3%) | Other (n=584, 7.2%) | P-value ∗ | |
| Age (years) | 64.7 ± 10.4 † | 62.2 ± 10.3 † | 61.0 ± 10.6 † | 63.3 ± 10.5 † | <0.003 |
| Body mass index (kg/m 2) | 29.8 ± 6.1 † | 30.2 ± 6.3 † | 29.6 ± 5.8 † | 28.2± 5.7 † | <0.003 |
| Male | 76.3% | 56.1% | 72.5% | 75.5% | |
| Female | 23.7% | 44.0% | 27.5% | 24.5% | |
| Diabetes mellitus | 37.1% | 46.6% | 61.2% | 44.2% | <0.003 |
| Renal failure | 7.8% | 16.8% | 17.2% | 9.8% | <0.003 |
| Creatinine (mg/dL) | 1.1 ± 0.8 † | 1.7 ± 1.9 † | 1.5 ± 1.7 † | 1.2 ± 1.1 † | <0.003 |
| Chronic lung disease | 15.7% | 13.6% | 8.9% | 12.8% | <0.003 |
| Systemic hypertension | 81.0% | 91.5% | 83.0% | 77.4% | <0.003 |
| Peripheral vascular disease | 15.1% | 13.7% | 13.3% | 11.0% | 0.759 |
| Cerebrovascular disease | 13.2% | 15.0% | 11.1% | 10.5% | >0.99 |
| Time from last myocardial infarction to surgery (hours) | <0.003 | ||||
| None | 59.1% | 49.2% | 51.9% | 54.1% | |
| ≤6 | 0.9% | 1.0% | 1.0% | 0.7% | |
| >6 but <24 | 1.5% | 0.8% | 1.4% | 1.4% | |
| ≥24 | 38.5% | 49.0% | 45.7% | 43.8% | |
| Tobacco Use | <0.003 | ||||
| Never | 58.0% | 52.6% | 65.4% | 64.6% | |
| Previous | 15.3% | 13.6% | 10.5% | 12.7% | |
| Current | 26.7% | 33.8% | 24.1% | 22.8% | |
| Heart failure | 15.8% | 25.5% | 22.6% | 17.1% | <0.003 |
| Previous percutaneous coronary intervention | 28.2% | 26.8% | 17.9% | 24.0% | <0.003 |
| Previous coronary bypass | 6.0% | 5.2% | 2.5% | 5.7% | 0.143 |
| Preoperative angina pectoris | 69.0% | 68.1% | 66.6% | 67.0% | >0.99 |
| Preoperative atrial fibrillation | 17.7% | 12.8% | 10.8% | 14.2% | <0.003 |
| Preoperative ejection fraction (%) | 49.3 ± 13.8 ∗ | 47.2 ± 15.7 ∗ | 47.0 ± 14.3 ∗ | 49.5 ± 14.0 ∗ | <0.003 |
| Preoperative left main narrowing | 29.4% | 28.3% | 28.8% | 29.3% | >0.99 |
| Operation | <0.003 | ||||
| Elective | 52.5% | 48.6% | 43.2% | 40.7% | |
| Non-Elective | 47.6% | 51.4% | 56.9% | 59.3% | |
| Off-pump | 27.9% | 23.9% | 19.4% | 28.8% | |
| On-pump | 72.1% | 76.1% | 80.6% | 71.2% | |
| Preoperative intra-aortic balloon pump | 15.2% | 18.3% | 16.2% | 18.7% | 0.777 |
| Unadjusted operative mortality | 2.66% | 3.27% | 2.53% | 3.77% | >0.99 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
