Abstract
Background:
The long-term effects of postoperative complications following coronary artery bypass grafting (CABG) are unknown.
Methods:
Medicare-linked records from the Society of Thoracic Surgeons Adult Cardiac Surgery Database were queried for isolated CABG records from 2007 through 2012. Unadjusted and adjusted associations between individual postoperative complications and both mortality and all-cause rehospitalization were evaluated to 7 years using Cox proportional hazards models and cumulative incidence functions. Due to non-proportional hazards, associations are presented as early (0 to 90 days) and late (90 days to 7 years).
Results:
Of the 294,533 isolated CABG patients who had records linked to Medicare for long-term follow-up, (median age, 73yrs; 30% female), 120,721 (41%) experienced at least one of the complications of interest, including new onset atrial fibrillation (30.0%), prolonged ventilation (12.3%), renal failure (4.5%), reoperation (3.5%), stroke (1.9%), and sternal wound infection (0.4%). Each of the six postoperative complications was associated with a significantly increased risk of mortality and rehospitalization to 7 years, despite adjustment for baseline characteristics and the presence of multiple complications. Although the predominant effect of postoperative complications was observed in the first 90 days, the increased risk-adjusted hazard for death and rehospitalization continued through 7 years.
Conclusions:
Postoperative complications are associated with an increased risk of both early and late mortality and all-cause rehospitalization, particularly during the “value” window within 90 days of CABG. These findings underscore the need to develop avoidance strategies as well as cost-adjustment methods for each of these complications.
Classifications: Surgery, complications, Outcomes, CABG
Introduction:
Since its inception in the 1960s, coronary artery bypass grafting (CABG) has become a valuable treatment strategy in the management of coronary artery disease (1). Over this time, outcomes have continued to improve, with an operative mortality of only 1.7% reported in recent analyses (2, 3). However, the morbidity associated with this procedure continues to impact postoperative quality of life and the use of health care resources (3).
In July 2016, the Centers for Medicare and Medicaid Services (CMS) proposed new bundled payment models for acute myocardial infarction (MI) and CABG, which would bundle payments for care during the index hospitalization out to ninety days post-discharge in an attempt improve the quality of care provided (4). Similarly, existing risk models can be used to assess the short-term risk of postoperative morbidity and mortality following CABG thus allowing health care providers and payers to risk stratify patients prior to surgery (5, 6). Unfortunately, they do not address the long-term impact of these postoperative complications; an important consideration for both the patient as well as the clinician. The Society of Thoracic Surgeons Adult Cardiac Surgery Database (STS ACSD) was recently linked to Medicare records to provide long-term outcomes for clinical research purposes (7). In this study, we sought to use this linkage to investigate the effect of post-operative complications on long-term clinical outcomes.
Methods:
Patient Population
The STS ACSD, which contains clinical records from over 90% of adult cardiac surgery centers in the U.S., was queried for Medicare-linked records (Medicare data use agreement # 17758) of patients age 65 years or older who underwent an isolated CABG procedure between January 2007 and December 2012 (8). Patients who had an STS-defined postoperative complication (as captured on the STS data collection form) other than a complication of interest were excluded.
Variables
The primary exposure of interest was an in-hospital postoperative complication including new onset atrial fibrillation or flutter, reoperation, stroke (confirmed neurological deficit that was unresolved at 24 hours), renal failure (creatinine 3× baseline or ≥ 4.0, or new dialysis requirement), mediastinitis or deep sternal infection (opening of wound or operative reexploration in the setting of a positive culture if culture obtained and extended treatment of antibiotics), or prolonged ventilator support (>24 hours). The two primary outcomes of interest were long-term mortality and all-cause hospital readmission. These outcomes were chosen as mortality historically has been reported only within the short term, but complications following surgery can lead to an increased risk of mortality within the later period as well. Readmission is a surrogate for continued impact on quality of life. Outcomes were adjusted for age, sex, race/ethnicity, body mass index (BMI), diabetes, hypertension, dyslipidemia, preoperative dialysis, chronic lung disease (none versus mild or moderate versus severe), cerebrovascular disease, previous percutaneous coronary intervention, previous cardiac surgery, previous MI, New York Heart Association (NYHA) classification of heart failure, ejection fraction <55%, operative status (elective versus urgent versus emergent/salvage), cardiogenic shock, and preoperative intra-aortic balloon pump. The STS predicted risk of mortality (PROM) is determined based on pre-established STS models for isolated CABG (5).
Statistical Analysis
Patients were stratified based on the occurrence of one of the complications of interest. Baseline characteristics, operative variables, and postoperative outcomes were compared. Continuous variables were assessed using the Kruskal-Wallis test while categorical variables were assessed using the Chi-squared test.
The Kaplan-Meier method was used to evaluate long-term survival. The Kalbfleisch and Prentice estimator was used to evaluate all-cause readmission for patients with and without post-operative complications while accounting for the competing risk of death (9). Patient follow-up for mortality was administratively censored at the end of the study period (January 1, 2014). Follow-up for all-cause readmission was censored at the end of the study period or at the time of health maintenance organization enrollment. Cox proportional hazard models were then used to determine the effect of each postoperative complication. These models were performed unadjusted, adjusted for covariates, and adjusted for covariates and the presence of other postoperative complications. All models were tested for proportional hazards. Due to violations of the proportional hazards assumption, and a differing association of the complication with the outcomes in the early versus later postoperative periods, the Cox models were split into two time periods; within 90 days of the operation, and beyond 90 days postoperatively (conditional on surviving to 90 days). To examine whether the results differed by year, a sensitivity analysis was performed by including year of surgery as a covariate in the multivariable models examining the association between perioperative complications and mortality or readmission.
To investigate whether the association of post-operative complications with outcomes differed depending on the patient’s age or STS PROM, interaction terms for age and PROM with complications were included in the fully adjusted models. The categorization of STS PROM into three groups (low risk: <4%, medium risk: 4–8%, and high-risk: >8%) was based on previously published categorizations (10). Models were stratified by the effect-modifier, and the HR for the postoperative complication was presented for each level of stratum. All proportional hazard models accounted for hospital center effect using Efron approximation. This analysis was generated using SAS software, Version 9.4 (Cary, NC, USA). Institutional Review Board approval was obtained for this study prior to data analysis.
This project was funded by institutional funds and support from the Society of Thoracic Surgeons. Dr. Jawitz received funding provided by NIH grant 5T32HL069749. The authors are solely responsible for the design and conduct of this study, all study analyses, the drafting and editing of the paper, and its final contents.
Results
Of 504,565 patients over the age of 65 that were identified, 186,293 (36.9%) patients could not be linked to Medicare claims. Compared to patients that were not linked, linked patients were more likely White/Caucasian (88% vs 81%), were less likely to have a history of diabetes (40% vs 42%), and were more likely to undergo an elective CABG (44% vs 40%) (Table S1). The distribution of STS-PROM was similar between linked and unlinked cohorts (p=0.72). An additional 191 patients were excluded due to missing complication data, and 23,548 patients were excluded who had another complication besides one of the complications of interest, leaving 294,533 patients from 1,117 centers for analysis (Figure 1). A total of 120,721 (41.0%) patients experienced a postoperative complication of interest, the most common of which was atrial fibrillation (n=88,464, 30.0%, Table 1). More than one complication occurred in 26,703 (22.1%) patients. Patients who experienced a postoperative complication were more likely to be on dialysis preop (2.7% vs 1.6%, p<0.001), have severe chronic lung disease (5.7% vs 3.4%, p<0.001), and have a history of cerebrovascular disease (20.3% vs 16.8%, p<0.001, Table 2). They were also more likely to have had a previous MI (49.2% vs 41.3%, p<0.001), be NYHA class IV (30.9% vs 21.9%, p<0.001), and have an ejection fraction <55% (49.1% vs 42.7%, p<0.001). Patients who experienced a complication were significantly more likely to present with cardiogenic shock (3.4% vs 0.8%, p<0.001) and have a preoperative IABP (10.9% vs 4.9%, p<0.001). Lastly, they had a longer length of hospital stay (median 7 vs 5 days, p<0.001), were less likely to be discharged home (64.1% vs 81.0%, p<0.001), and were more likely to be discharged to an extended care facility or transitional care unit (28.2% vs 15.2%, p<0.001).
Figure 1:
Consort diagram.
Table 1:
Incidence of select postoperative complications among patients undergoing coronary artery bypass grafting.
| Complication | Frequency (Percentage) |
|---|---|
| No Complication | 173,812 (59.0%) |
| Atrial Fibrillation | 88,464 (30.0%) |
| Prolonged Ventilatory Support | 36,322 (12.3%) |
| Renal Failure | 13,359 (4.5%) |
| Reoperation | 10,280 (3.5%) |
| Stroke | 5,494 (1.9%) |
| Mediastinitis or Deep Sternal Wound Infection | 1,106 (0.4%) |
Table 2:
Preoperative and operative characteristics of patients who experienced a postoperative complication of interest as compared to patients who did not experience a postoperative complication of interest.
| Variable | Overall | Any Complication | No Complication | P |
|---|---|---|---|---|
| Age (y) | 73 (68, 78) | 74 (69, 78) | 72 (68, 77) | <0.001 |
| Female Sex | 88,450 (30.0%) | 35,831 (29.7%) | 52,619 (30.3%) | <0.001 |
| Race/Ethnicity | <0.001 | |||
| White/Caucasian | 259,017 (87.9%) | 106,881 (88.5%) | 152,136 (87.5%) | |
| Black/African-American | 13,789 (4.7%) | 5,450 (4.5%) | 8,339 (4.8%) | |
| Hispanic | 10,182 (3.5%) | 3,700 (3.1%) | 6,482 (3.7%) | |
| Asian | 4,559 (1.5%) | 1,856 (1.5%) | 2,703 (1.6%) | |
| Native American | 945 (0.3%) | 338 (0.3%) | 607 (0.3%) | |
| Other | 4,959 (1.7%) | 2,070 (1.7%) | 2,889 (1.7%) | |
| Body Mass Index (kg/m2) | <0.001 | |||
| <18.5 | 2,369 (0.8%) | 1,100 (0.9%) | 1,269 (0.7%) | |
| 18.5 - <25 | 65,420 (22.3%) | 26,928 (22.4%) | 38,492 (22.2%) | |
| 25 - <30 | 119,931 (40.8%) | 47,143 (39.2%) | 72,788 (42.0%) | |
| 30 - < 35 | 69,812 (23.8%) | 28,677 (23.8%) | 41,135 (23.7%) | |
| 35 - < 40 | 24,974 (8.5%) | 11,093 (9.2%) | 13,881 (8.0%) | |
| >40 | 11,339 (3.9%) | 5,489 (4.6%) | 5,850 (3.4%) | |
| Diabetes | 117,021 (39.7%) | 48,688 (40.3%) | 68,333 (39.3%) | <0.001 |
| Hypertension | 258,191 (87.7%) | 106,994 (88.6%) | 151,197 (87.0%) | <0.001 |
| Dyslipidemia | 248,095 (84.2%) | 101,171 (83.8%) | 146,924 (84.5%) | <0.001 |
| Dialysis | 6,052 (2.1%) | 3,220 (2.7%) | 2,832 (1.6%) | <0.001 |
| Chronic Lung Disease | <0.001 | |||
| None | 224,119 (76.1%) | 88,324 (73.2%) | 135,795 (78.1%) | |
| Mild | 37,903 (12.9%) | 16,391 (13.6%) | 21,512 (12.4%) | |
| Moderate | 19,205 (6.5%) | 8,938 (7.4%) | 10,267 (5.9%) | |
| Severe | 12,799 (4.3%) | 6,903 (5.7%) | 5,896 (3.4%) | |
| Active Smoker | 74,387 (25.3%) | 30,840 (25.5%) | 43,547 (25.1%) | <0.001 |
| Cerebrovascular Disease | 53,806 (18.3%) | 24,538 (20.3%) | 29,268 (16.8%) | <0.001 |
| Previous PCI | 221,520 (75.2%) | 89,584 (74.2%) | 131,936 (75.9%) | <0.001 |
| Previous Cardiac Surgery | 87,924 (29.9%) | 36,989 (30.6%) | 50,935 (29.3%) | <0.001 |
| Previous Myocardial Infarction | 131,248 (44.6%) | 59,393 (49.2%) | 71,855 (41.3%) | <0.001 |
| NYHA Class | <0.001 | |||
| I | 3,117 (6.3%) | 1,333 (5.3%) | 1,784 (7.4%) | |
| II | 12,860 (26.1%) | 5,817 (23.2%) | 7,043 (29.1%) | |
| III | 19,357 (39.3%) | 9,701 (38.7%) | 9,656 (39.9%) | |
| IV | 13,040 (26.5%) | 7,731 (30.9%) | 5,309 (21.9%) | |
| Ejection Fraction <55% | 128,821 (45.3%) | 57,137 (49.1%) | 71,684 (42.7%) | <0.001 |
| Emergent Status | <0.001 | |||
| Elective | 131,246 (44.6%) | 49,127 (40.7%) | 82,119 (47.2%) | |
| Urgent | 150,701 (51.2%) | 63,883 (52.9%) | 86,818 (49.9%) | |
| Emergent/Salvage | 12,508 (4.3%) | 7,671 (6.4%) | 4,837 (2.8%) | |
| Cardiogenic Shock | 5,436 (1.8%) | 4,096 (3.4%) | 1,340 (0.8%) | <0.001 |
| Preoperative IABP | 21,713 (7.4%) | 13,191 (10.9%) | 8,522 (4.9%) | <0.001 |
| Predicted Risk of Mortality | <0.001 | |||
| <4% | 243,885 (82.8%) | 92,008 (76.2%) | 151,877 (87.4%) | |
| 4–8% | 32,640 (11.1%) | 16,955 (14.0%) | 15,685 (9.0%) | |
| >8% | 18,008 (6.1%) | 11,758 (9.7%) | 6,250 (3.6%) | |
| Three-Vessel Disease | 225,828 (76.7%) | 94,787 (78.5%) | 131,041 (75.4%) | <0.001 |
| Use of IMA artery | <0.001 | |||
| None | 23,609 (8.0%) | 10,767 (8.9%) | 12,842 (7.4%) | |
| Left | 261,716 (88.9%) | 106,243 (88.0%) | 155,473 (89.4%) | |
| Right | 2,203 (0.7%) | 923 (0.8%) | 1,280 (0.7%) | |
| Both | 6,850 (2.3%) | 2,739 (2.3%) | 4,111 (2.4%) | |
| Operation Time (Hours) | 4.9 (4.1, 5.8) | 5.0 (4.2, 6.0) | 4.8 (4.0, 5.7) | <0.001 |
| CPB Utilization | 235,580 (80.0%) | 98,594 (81.7%) | 136,986 (78.8%) | <0.001 |
| CPB Time (Min) | 89 (69, 113) | 92 (71, 117) | 87 (68, 110) | <0.001 |
| Discharge Location | <0.001 | |||
| Home | 213,669 (74.2%) | 73,553 (64.1%) | 140,116 (81.0%) | |
| Extended Care/TCU | 58,586 (20.4%) | 32,349 (28.2%) | 26,237 (15.2%) | |
| Other Hospital | 2,386 (0.8%) | 1,964 (1.7%) | 422 (0.2%) | |
| Nursing Home | 11,792 (4.1%) | 6,020 (5.2%) | 5,772 (3.3%) | |
| Hospice | 389 (0.1%) | 307 (0.3%) | 82 (0.1%) | |
| Other | 1,071 (0.4%) | 641 (0.6%) | 430 (0.3%) |
Categorical variables are presented as frequency (percentage) while continuous variables are presented as median (interquartile range). PCI – Percutaneous Coronary Intervention. NYHA – New York Heart Association. IABP – Intraaortic Balloon Pump. CPB – Cardiopulmonary Bypass. TCU – Transitional Care Unit
Analysis of readmission
On unadjusted analysis of all-cause readmission, each complication was again associated with a significantly increased risk of readmission, both through 90 days and beyond (Table 3 and Figure 2). Atrial fibrillation had the weakest association with readmission through the first 90 days (unadjusted HR 1.39, 95% CI 1.36–1.41) while mediastinitis or deep sternal infection (unadjusted HR 4.91, 95% CI 4.50–5.36) had the strongest association. Beyond 90 days, atrial fibrillation (unadjusted HR 1.16, 95% CI 1.14–1.17) and reoperation (unadjusted HR 1.18, 95% CI 1.14–1.23) had the weakest association with all-cause readmission while stroke (unadjusted HR 1.65, 95% CI 1.45–1.74) and renal failure (unadjusted HR 1.64, 95% CI 1.58–1.70) were the most strongly associated with readmission. After inserting year of surgery as a covariate in a sensitivity analysis, hazard ratios remained essentially unchanged demonstrating no appreciable era effect within the study period.
Table 3:
Unadjusted and adjusted association between postoperative complications and long-term mortality and all-cause readmission following coronary artery bypass grafting.
| Within 90 Days of Procedure | Beyond 90 Days Postoperatively | |||||
|---|---|---|---|---|---|---|
| Unadjusted HR (95% CI) | Adjusted HR1 (95% CI) | Adjusted HR2 (95% CI) | Unadjusted HR (95% CI) | Adjusted HR1 (95% CI) | Adjusted HR2 (95% CI) | |
| Mortality | ||||||
| Atrial Fibrillation | 4.48 (4.20–4.74) | 3.71 (3.50–3.93) | 1.59 (1.48–1.71) | 1.44 (1.41–1.48) | 1.32 (1.29–1.35) | 1.19 (1.16–1.22) |
| Prolonged Ventilatory Support | 18.45 (17.49–19.47) | 11.58 (10.90–12.30) | 6.53 (6.08–7.01) | 2.65 (2.57–2.72) | 1.79 (1.73–1.84) | 1.58 (1.52–1.64) |
| Renal Failure | 27.23 (25.66–28.89) | 17.50 (16.28–18.80) | 5.64 (4.97–6.38) | 2.85 (2.72–2.98) | 1.80 (1.72–1.88) | 1.44 (1.34–1.54) |
| Reoperation | 13.74 (12.80–14.75) | 8.91 (8.22–9.64) | 5.39 (4.68–6.21) | 1.76 (1.67–1.85) | 1.38 (1.32–1.46) | 1.14 (1.04–1.24) |
| Stroke | 21.54 (20.03–23.16) | 13.16 (12.09–14.33) | 5.43 (4.62–6.39) | 2.72 (2.56–2.90) | 2.13 (1.99–2.28) | 1.41 (1.26–1.59) |
| Mediastinitis or Deep Sternal Infection | 13.61 (11.75–15.77) | 6.80 (5.62–8.21) | 3.59 (2.11–6.11) | 2.86 (2.53–3.23) | 2.17 (1.91–2.48) | 1.66 (1.28–2.16) |
| All-Cause Readmission | ||||||
| Atrial Fibrillation | 1.39 (1.36–1.41) | 1.33 (1.31–1.35) | 1.24 (1.21–1.26) | 1.16 (1.14–1.17) | 1.12 (1.10–1.14) | 1.09 (1.07–1.11) |
| Prolonged Ventilatory Support | 2.09 (2.04–2.14) | 1.58 (1.54–1.62) | 1.43 (1.38–1.47) | 1.53 (1.49–1.56) | 1.26 (1.23–1.29) | 1.21 (1.17–1.25) |
| Renal Failure | 2.29 (2.21–2.37) | 1.68 (1.62–1.75) | 1.46 (1.36–1.56) | 1.64 (1.58–1.70) | 1.28 (1.23–1.32) | 1.19 (1.13–1.27) |
| Reoperation | 1.63 (1.56–1.69) | 1.43 (1.37–1.48) | 1.19 (1.11–1.27) | 1.18 (1.14–1.23) | 1.11 (1.07–1.16) | 1.04 (0.98–1.10) |
| Stroke | 1.99 (1.89–2.09) | 1.60 (1.52–1.68) | 1.36 (1.26–1.48) | 1.65 (1.56–1.74) | 1.42 (1.34–1.50) | 1.31 (1.21–1.42) |
| Mediastinitis or Deep Sternal Infection | 4.91 (4.50–5.36) | 3.79 (3.45–4.16) | 6.22 (5.36–7.23) | 1.56 (1.32–1.84) | 1.38 (1.16–1.63) | 1.31 (0.96–1.79) |
Adjusted Hazard Ratio taking into account only preoperative covariates.
Adjusted Hazard Ratio taking into account whether the patient suffered from additional postoperative complications.
Figure 2:
Freedom from all-cause readmission among patients following coronary artery bypass grafting by the occurrence of postoperative complications. Patients with multiple complications are counted in each associated cohort.
When investigating the interaction of age with the risk of readmission, stroke, renal failure, mediastinitis or deep sternal wound infection, and prolonged ventilator support were all found to have a statistically significant interaction with age through 90 days. Complications among older patients were associated with a lower risk of readmission than among the younger cohort, however none of the complications had a significant interaction with age beyond 90 days. With regards to the STS PROM, all complications had a significant interaction with STS PROM except for mediastinitis or a deep sternal wound infection through 90 days. Only atrial fibrillation and prolonged ventilator support had a significant interaction with STS PROM after 90 days. Again both before and after 90 days, patients with a lower STS PROM had a higher risk of readmission associated with complications than those patients at higher risk before surgery.
Unadjusted analysis of mortality
The median follow-up for mortality was 3.7 years while the median follow-up for readmission was 1.3 years. Each of the complications of interest was associated with an increased incidence of long-term mortality on unadjusted analysis (Table 3 and Figure 3). Atrial fibrillation had the weakest unadjusted association with mortality both within 90 days of the procedure (unadjusted HR 4.48, 95% CI 4.20–4.74) as well as beyond 90 days postoperatively (unadjusted HR 1.44, 95% CI 1.41–1.48). New onset renal failure had the strongest association with unadjusted mortality within 90 days of the procedure (unadjusted HR 27.2, 95% CI 25.7–28.9), while mediastinitis or a deep sternal wound infection had the strongest association with unadjusted mortality beyond 90 days postoperatively (unadjusted HR 2.86, 95% CI 2.53–3.23).
Figure 3:
Unadjusted long-term survival of patients following coronary artery bypass grafting by the occurrence of postoperative complications. Patients with multiple complications are counted in each associated cohort. Time 0 is defined as the date of surgery.
Adjusted analysis of mortality
On adjusted analysis, whether taking into account other complications or not, each complication of interest remained significantly associated with increased long-term mortality (Table 3). A significant interaction was found with age and reoperation, renal failure, and prolonged ventilatory support in the first 90 days, and reoperation, mediastinitis or deep sternal wound infection, and prolonged ventilator support beyond 90 days (Table S2). In all instances of a significant interaction, older patients were less affected by the complication than the younger cohort. There was also a significant interaction between the STS PROM and each complication of interest except for mediastinitis or deep sternal wound infection beyond 90 days. A lower STS PROM was associated with a greater magnitude in the effect of most complications on survival both through 90 days and beyond (Table S3). The most common indication for reoperation was bleeding/tamponade (n=7,592) (Table S4).
Discussion
Despite steady improvements in outcomes, CABG is associated with substantial morbidity which negatively affects long-term quality of life (2). In this study we have demonstrated that atrial fibrillation, reoperation, stroke, renal failure, mediastinitis or a deep sternal wound infection, and prolonged ventilator support are all associated with significantly increased long-term mortality and all-cause readmission, even after adjustment. Although these associations are strongest in the first ninety days, they continue well-beyond.
Notably, there is substantial variation in the effect associated with each of the post-operative complications. Despite affecting 30% of patients, atrial fibrillation has the mildest association with long-term mortality and all-cause readmission of the studied complications. Nonetheless, it is still associated with poor long-term outcomes and therefore prevention and appropriate treatment is still essential. This is not the first study to demonstrate an association between atrial fibrillation and long-term mortality, however it was previously unclear whether atrial fibrillation is a cause of mortality or simply an associated symptom of another complication (11, 12). Patients with postoperative atrial fibrillation do have higher incidences of respiratory failure, stroke, and infections indicating it may often be secondary to other physiologic stressors (12–14). This is supported by our finding that once associated complications were accounted for in the model, the adjusted hazard ratio for mortality within the first 90 days associated with atrial fibrillation fell from 3.71 to 1.59. Furthermore, the risk of mortality and all-cause readmission associated with isolated postoperative atrial fibrillation (Table S3) was not significantly different that the risk associated with postoperative atrial fibrillation overall (including other concomitant complications).
Stroke and renal failure were very strongly associated with 90-day mortality, a finding also supported by others (15–18). John and colleagues analyzed 19,224 patients from a multicenter regional database and found that while only 1.4% of patients suffered from a stroke following CABG, the hospital mortality among these patients was 24.8% as compared to only 2.0% for the remaining population (15). Rydén and colleagues found that among a cohort of 27,929 patients in Sweden, renal failure was associated with significantly increased mortality, especially among patients with stage 3 renal failure, of which more than 20% died within the first few months of surgery (17).
Although it is difficult to predict which patients may be at risk for postoperative complications, it is important to note that many of the complications associated with increased long-term mortality and readmission have processes in place to reduce individual risk. For instance, the use of beta-adrenergic receptor blockers especially among patients who were on them preoperatively has been shown to drastically reduce rates of atrial fibrillation and associated hospital length of stay after CABG (19,20). Other studies have demonstrated that patient tobacco use is associated with an increased risk of sternal infections, and therefore smoking cessation may reduce the risk of mediastinitis (21). Nasopharyngeal and oropharyngeal decontamination with chlorhexidine has been shown to reduce the incidence of post-operative pneumonia and deep surgical site infections (22).
Bundled payment models theoretically stimulate healthcare providers to reduce overall costs while improving quality (23). With the rollout of the CMS bundled payment models for acute MI and CABG, there is increased financial incentive to reduce complications, in particular during the first 90 days post-procedure. Specifically, readmissions are associated with substantial cost, with roughly $17.4 billion dollars spent on Medicare unplanned readmissions alone in 2004 (24). Every post-operative complication investigated in this study was associated with significant increases in readmissions within 90 days, especially mediastinitis or sternal wound infection which even after adjustment was associated with a six-fold increase in the risk of readmission within the first 90 days. Efforts to reduce surgical site infections will therefore be paramount in the future medical reimbursement landscape. We have also demonstrated that there are significant longer-term implications of perioperative complications beyond the first 90 days. To reduce the incentive for hospitals to avoid higher risk patients, risk-adjusted bundled payment amounts should be considered.
There are limitations to this study which must be considered. First, as our study involved Medicare-linked records, only patients 65 years of age and older were included. Although we attempted to determine the effect of increasing age on the association of complications and long-term outcomes through interaction terms, it was not possible to determine how these complications affect patients less than 65 years of age. Also, our finding on interaction term analysis that older patients were less affected by complications than younger patients is likely the result of bias. Since older patients are more likely to die or be readmitted to the hospital independent of surgical complications, this likely diminished the effect size of the association seen between complications and mortality or readmission. Alternatively, the magnitude of the physiologic impact leading to the complications of interest may have varied between younger and older patients. Second, we were unable to link ~37% of the overall cohort with Medicare claims data. While there were some differences in baseline characteristics between the linked and unlinked cohorts (Table S1), they had similar PROM distributions and therefore, their inclusion would be unlikely to significantly change results. Further, the approximately 63% successful Medicare linkage rate of this analysis is comparable to several previously published studies using data from the STS ACSD in the same era (25–27). Third, complications often occur together, and it is difficult to assess the impact of a single complication on outcomes. Adjusting for multiple complications, the association of specific complications and outcomes was reduced demonstrating that the effect of multiple complications may be additive. Nonetheless, the significant associations after adjustment demonstrate that each complication is still individually important. Fourth, as the data on survival and freedom from hospital readmission for beyond 90 days was dependent on surviving to 90 days, it is possible that certain associations are not as strong as would be expected due to early patient drop-out. Last, longer-term follow-up beyond that attained in this study could provide further insight into how these complications impact long-term mortality and quality of life.
In conclusion, we have demonstrated long-term effects of several common complications of coronary artery bypass grafting. Although these complications have differing effects on survival and all-cause readmission, they have each been shown to have an independent, negative impact on long-term recovery following coronary artery bypass grafting. Data-proven methods for reducing the incidence and long-term clinical consequences of these complications should be implemented across institutions. Furthermore, with the rollout of bundled payments for CABG across many institutions, it is becoming increasingly important to understand the long-term implications of complications following this procedure.
Supplementary Material
Acknowledgements
The Society of Thoracic Surgeons National Database provided the data for this research and the analysis was funded by the STS Access and Publications Research Program.
Funding: This project was funded by institutional funds and support from the Society of Thoracic Surgeons. Dr. Jawitz received funding provided by NIH T-32 grant 5T32HL069749.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Conflicts of Interest: None of the authors have any conflicts of interest to declare.
References
- 1.Head SJ, Kieser TM, Falk V, Huysmans HA, Kappetein AP. Coronary artery bypass grafting: Part 1--the evolution over the first 50 years. European heart journal 2013;34(37):2862–2872. [DOI] [PubMed] [Google Scholar]
- 2.ElBardissi AW, Aranki SF, Sheng S, O’Brien SM, Greenberg CC, Gammie JS. Trends in isolated coronary artery bypass grafting: An analysis of the society of thoracic surgeons adult cardiac surgery database. The Journal of thoracic and cardiovascular surgery 2012;143(2):273–281. [DOI] [PubMed] [Google Scholar]
- 3.D’Agostino RS, Jacobs JP, Badhwar V et al. The society of thoracic surgeons adult cardiac surgery database: 2016 update on outcomes and quality. The Annals of thoracic surgery 2016;101(1):24–32. [DOI] [PubMed] [Google Scholar]
- 4.“Notice of proposed rulemaking for bundled payment models for high-quality, coordinated cardiac and hip fracture care” https://www.Cms.Gov/newsroom/mediareleasedatabase/fact-sheets/2016-fact-sheets-items/2016-07-25.Html. Accessed on September 19, 2019.
- 5.Shahian DM, O’Brien SM, Filardo G et al. The society of thoracic surgeons 2008 cardiac surgery risk models: Part 1--coronary artery bypass grafting surgery. The Annals of thoracic surgery 2009;88(1 Suppl):S2–22. [DOI] [PubMed] [Google Scholar]
- 6.Society of thoracic surgeons adult cardiac surgery risk calculator. Http://riskcalc.Sts.Org/stswebriskcalc/#/calculate. Accessed on October 11, 2019.
- 7.Jacobs JP, Edwards FH, Shahian DM et al. Successful linking of the society of thoracic surgeons adult cardiac surgery database to centers for medicare and medicaid services medicare data. The Annals of thoracic surgery 2010;90(4):1150–1157. [DOI] [PubMed] [Google Scholar]
- 8.Katzenstein HM, Bowman LC, Brodeur GM et al. Prognostic significance of age, mycn oncogene amplification, tumor cell ploidy, and histology in 110 infants with stage d(s) neuroblastoma: The pediatric oncology group experience--a pediatric oncology group study. Journal of clinical oncology : official journal of the American Society of Clinical Oncology 1998;16(6):2007–2017. [DOI] [PubMed] [Google Scholar]
- 9.Kalbfleisch JD, Prentice RL. The statistical analysis of failure time data. John Wiley & Sons; 2011. [Google Scholar]
- 10.Vassileva CM, Aranki S, Brennan JM, Kaneko T, He M, Gammie JS, Suri RM, Thourani VH, Hazelrigg S, McCarthy P. Evaluation of The Society of Thoracic Surgeons Online Risk Calculator for Assessment of Risk in Patients Presenting for Aortic Valve Replacement After Prior Coronary Artery Bypass Graft: An Analysis Using the STS Adult Cardiac Surgery Database. Annals of Thoracic Surgery 2015;100(6): 2109–2115. [DOI] [PubMed] [Google Scholar]
- 11.Villareal RP, Hariharan R, Liu BC et al. Postoperative atrial fibrillation and mortality after coronary artery bypass surgery. Journal of the American College of Cardiology 2004;43(5):742–748. [DOI] [PubMed] [Google Scholar]
- 12.Mariscalco G, Klersy C, Zanobini M et al. Atrial fibrillation after isolated coronary surgery affects late survival. Circulation 2008;118(16):1612–1618. [DOI] [PubMed] [Google Scholar]
- 13.Olshansky B. Management of atrial fibrillation after coronary artery bypass graft. The American journal of cardiology 1996;78(8a):27–34. [DOI] [PubMed] [Google Scholar]
- 14.Aranki SF, Shaw DP, Adams DH et al. Predictors of atrial fibrillation after coronary artery surgery. Current trends and impact on hospital resources. Circulation 1996;94(3):390–397. [DOI] [PubMed] [Google Scholar]
- 15.John R, Choudhri AF, Weinberg AD et al. Multicenter review of preoperative risk factors for stroke after coronary artery bypass grafting. The Annals of thoracic surgery 2000;69(1):30–35; discussion 35–36. [DOI] [PubMed] [Google Scholar]
- 16.Dacey LJ, Likosky DS, Leavitt BJ et al. Perioperative stroke and long-term survival after coronary bypass graft surgery. The Annals of thoracic surgery 2005;79(2):532–536; discussion 537. [DOI] [PubMed] [Google Scholar]
- 17.Ryden L, Ahnve S, Bell M et al. Acute kidney injury after coronary artery bypass grafting and long-term risk of myocardial infarction and death. International journal of cardiology 2014;172(1):190–195. [DOI] [PubMed] [Google Scholar]
- 18.Loef BG, Epema AH, Smilde TD et al. Immediate postoperative renal function deterioration in cardiac surgical patients predicts in-hospital mortality and long-term survival. Journal of the American Society of Nephrology : JASN 2005;16(1):195–200. [DOI] [PubMed] [Google Scholar]
- 19.Crystal E, Connolly SJ, Sleik K, Ginger TJ, Yusuf S. Interventions on prevention of postoperative atrial fibrillation in patients undergoing heart surgery: A meta-analysis. Circulation 2002;106(1):75–80. [DOI] [PubMed] [Google Scholar]
- 20.Ali IM, Sanalla AA, Clark V. Beta-blocker effects on postoperative atrial fibrillation. European journal of cardio-thoracic surgery : official journal of the European Association for Cardio-thoracic Surgery 1997;11(6):1154–1157. [DOI] [PubMed] [Google Scholar]
- 21.Abboud CS, Wey SB, Baltar VT. Risk factors for mediastinitis after cardiac surgery. The Annals of thoracic surgery 2004;77(2):676–683. [DOI] [PubMed] [Google Scholar]
- 22.Segers P, Speekenbrink RG, Ubbink DT, van Ogtrop ML, de Mol BA. Prevention of nosocomial infection in cardiac surgery by decontamination of the nasopharynx and oropharynx with chlorhexidine gluconate: A randomized controlled trial. JAMA : the journal of the American Medical Association 2006;296(20):2460–2466. [DOI] [PubMed] [Google Scholar]
- 23.Shih T, Chen LM, Nallamothu BK. Will bundled payments change health care? Examining the evidence thus far in cardiovascular care. Circulation 2015;131(24):2151–2158. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Jencks SF, Williams MV, Coleman EA. Rehospitalizations among patients in the medicare fee-for-service program. The New England journal of medicine 2009;360(14):1418–1428. [DOI] [PubMed] [Google Scholar]
- 25.Tavris DR, Brennan JM, Sedrakyan A et al. Long-term outcomes after transmyocardial revascularization. The Annals of Thoracic Surgery 2012;94(5):1500–1508. [DOI] [PubMed] [Google Scholar]
- 26.Badhwar V, Peterson ED, Jacobs JP et al. Longitudinal outcome of isolated mitral repair in older patients: results from 14,604 procedures performed from 1991 to 2007. The Annals of Thoracic Surgery 2012;94(6):1870–1877. [DOI] [PubMed] [Google Scholar]
- 27.Williams JB, Peterson ED, Brennan JM. Association between endoscopic vs open vein-graft harvesting and mortality, wound complications, and cardiovascular events in patients undergoing CABG surgery. JAMA 2012;308(5):475–484. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.