This cohort study examines the association of intraoperative transesophageal echocardiography use, 6 patient-level factors, and surgical volume of a hospital with survival outcomes of coronary artery bypass graft (CABG).
Key Points
Question
Among patients undergoing isolated coronary artery bypass graft (CABG), which subgroups derive the greatest survival benefit from intraoperative transesophageal echocardiography (TEE), and does this benefit vary by hospital surgical volume?
Findings
In this cohort study of 1.26 million patients who underwent isolated CABG, intraoperative TEE was associated with survival benefits at low and medium surgical volume hospitals, particularly among patients with greater than 50% left-main coronary stenosis, 3 or more diseased vessels, or preoperative inotrope use. No survival benefit was observed at high surgical volume hospitals.
Meaning
These findings support a more individualized approach to TEE use during isolated CABG surgery and provide a rationale for future randomized evaluation.
Abstract
Importance
The routine use of intraoperative transesophageal echocardiography (TEE) during coronary artery bypass graft (CABG) surgery remains controversial. Its benefit across different patient populations is unclear.
Objectives
To identify patient subgroups with the greatest or least likelihood to benefit from intraoperative TEE during CABG, stratified by hospital surgical volume.
Design, Setting, and Participants
This 2-stage, matched retrospective cohort study applied target trial emulation methodologies to the Society of Thoracic Surgeons (STS) Adult Cardiac Surgery Database (ACSD) to quantify the conditional treatment effect of intraoperative TEE among subpopulations undergoing isolated CABG at low, medium, and high surgical volume hospitals. The study cohort consisted of patients aged 18 years or older who underwent isolated CABG surgery between July 1, 2014, and June 30, 2022. Data analysis was conducted from August 8, 2023, to December 15, 2024.
Exposure
Receipt of an intraoperative TEE during CABG surgery.
Main Outcomes and Measures
The primary outcome was mortality within 30 days of surgery. Statistical analyses included multivariable logistic regression and multiple TEE vs without TEE matched comparisons stratified by surgical volume and patient subpopulations.
Results
Of 1 266 055 patients who underwent isolated CABG, 963 976 (76.1%) were male, and the mean (SD) age was 65.7 (10.0) years. Among these patients, 61.8% received TEE and 39.0% did not receive TEE. Intraoperative TEE use (vs without TEE) was associated with a significant survival benefit among patients treated at hospitals with low surgical volume (2.47% vs 2.94%; odds ratio [OR], 0.83 [95% CI, 0.78-0.89], P < .001) and medium surgical volume (2.09% vs 2.34%; OR, 0.89 [95% CI, 0.85-0.93], P < .001) but not high surgical volume (1.72% vs 1.77%; OR, 0.97 [95% CI, 0.91-1.03], P = .48). Among patients who underwent isolated CABG at low and medium surgical volume hospitals, TEE provided the greatest survival benefit to subpopulations with greater than 50% (vs ≤50%) left-main coronary stenosis, 3 or more (vs <3) diseased coronaries, and (3) a preoperative inotropic requirement.
Conclusions and Relevance
In isolated CABG, intraoperative TEE was associated with survival benefit at low- and medium-volume hospitals, particularly in patients with complex coronary disease or hemodynamic instability, but not at high-volume hospitals. These results highlight persistent equipoise and the need for randomized evaluation.
Introduction
Transesophageal echocardiography (TEE) is used during cardiac surgery to guide intraoperative decisions and is associated with improved outcomes after valve and aortic procedures.1,2,3 However, its benefit in isolated coronary artery bypass graft (CABG) remains uncertain, particularly for low-risk patients. While retrospective studies suggest TEE may improve outcomes in sicker patients who undergo CABG,4,5,6 the overall treatment effect is marginal5,7 and negligible in healthier subgroups.5,7
Beyond clinical equipoise, hospital-level resource limitations and staffing shortages affect TEE use during CABG.8,9 At TEE resource–limited hospitals, a precision medicine refinement for TEE allocation could improve outcomes by targeting patients most likely to benefit from TEE.
To advance prior observational work,4,5,6,7 this study aimed to identify patient subgroups with the greatest or least likelihood to benefit from intraoperative TEE during CABG, stratified by hospital surgical volume. We hypothesized that the association between TEE use and survival would vary by patient characteristics and hospital surgical volume. To accomplish this goal rigorously, we leveraged data from the Society of Thoracic Surgeons (STS) Adult Cardiac Surgery Database (ACSD) in conducting a 2-stage analysis with a target trial emulation framework10,11,12,13 to uncover heterogeneity in TEE’s survival benefit across subgroups.
Methods
Data Source and Study Population
The STS ACSD contains more than 8 million surgical records from more than 95% of US cardiac surgical centers.14 We queried data from versions 2.81, 2.90, and 4.20.2 of the database. All data management and statistical analyses were performed in accordance with the STS National Database Participant User Files Research Program data use agreement. We followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for cohort studies.15 The Advarra Institutional Review Board deemed this study exempt from review and waived the informed consent requirement because deidentified data were used.
The study cohort included all adults (aged ≥18 years) who underwent isolated CABG from July 1, 2014, to June 30, 2022. Exclusions were missing TEE data, concomitant non-CABG cardiac surgery (eg, valve, structural, and aortic), missing hospital indicator, missing diseased vessel count, and emergent salvage or unknown operative status. We elected to exclude observations with missing data in these covariates because of the size of our dataset.
Exposure and Outcome
The exposure was receipt of an intraoperative TEE during CABG surgery. Used in prior studies,3,5,9 intraoperative TEE is consistent across all versions of the STS ACSD used in this analysis.
The outcome was operative mortality. Operative mortality was defined as meeting either one of the following criteria: all-cause death occurring during the index hospitalization (including >30 days) or all-cause death after discharge but fewer than 30 days after the index surgery.16
Covariates for Matching
A comprehensive collection of independent covariates was used for matching to embed observational data into a target randomized trial (eAppendix 1 in Supplement 1). These covariates included demographics (age, sex, and race), admission status, year of surgery, preoperative comorbidities, hemodynamic parameters, laboratory values, STS Predicted Risk of Mortality (PROM) score, and hospital surgical volume (eAppendix 2 in Supplement 1). Race (categorized as American Indian or Alaska Native, Asian, Black, White, and other [including all responses other than the categories shown]) was included according to STS ACSD definitions (eFigure 1 and eTable 1 in Supplement 1).
All variables used in this analysis are standardized in the STS ACSD and are intermittently validated at individual centers by the STS Research and Analytic Center. Missing data rates among all variables included in the analyses did not exceed 0.001%, and single imputation (using median for continuous variables and mode for categorical variables) was used to address the minor residual missingness for included variables.
Potential Covariates Modifying the Treatment Effect of TEE
Six clinically relevant covariates were included in the logistic regression to estimate the conditional average treatment effect (CATE) of TEE in isolated CABG: (1) ejection fraction, (2) congestive heart failure (CHF; defined as having a New York Heart Association class I-IV16), (3) preoperative serum creatinine, (4) greater than 50% left-main coronary stenosis, (5) 3 or more diseased coronaries, and (6) inotrope use within 48 hours of surgery. These variables were selected based on a combination of clinical judgment, prior evidence of effect modification,6 prior utilization, and electronic medical record data availability.
Surgical Volume
Because CABG survival improves at higher surgical volume hospitals,17,18 we stratified analyses by surgical volume to reduce hospital-level confounding between TEE use and survival. Based on prior literature, surgical volume was classified as low with fewer than 100, as medium with 100 to 250, and as high with more than 250 annual isolated CABG cases using established volume thresholds.17,18,19,20,21
Study Design
Using target trial emulation methodologies,10,11,12,13 consistent with prior work,13,22,23 we conducted a 2-stage analysis to estimate the CATE of intraoperative TEE in isolated CABG. In stage 1, we identified subgroups who may derive the greatest and least benefit from TEE based on the 6 preselected covariates (eAppendix 4 in Supplement 1). In stage 2, we used matched cohort studies to compare operative mortality with TEE vs without TEE among the 5 TEE score categories derived in stage 1 (Figure 1; eAppendix 5 in Supplement 1).
Figure 1. Study Flow Diagram.
Stage 1 shows derivation of the transesophageal echocardiography (TEE) score, calculated as the conditional average treatment effect (CATE) of intraoperative TEE on mortality based on 6 factors (ejection fraction, congestive heart failure, preoperative serum creatinine, >50% left-main coronary stenosis, ≥3 diseased coronaries, and preoperative inotrope use). Stage 2 shows validation of the TEE score (range: 1-5, with 1 indicating most benefit and 5 indicating least benefit from TEE) via matched cohort analysis within each surgical volume stratum and score quintile. CABG indicates coronary artery bypass graft; ID, identification.
Stage 1: Identification of Subpopulations Who Benefit From TEE, and Derivation of the TEE Score
Estimating the CATE for intraoperative TEE use during isolated CABG surgery involved several steps. First, within each surgical volume strata (low, medium, and high), on a random 25% subset of data, we emulated randomization using statistical matching24 to ensure baseline covariate balance among patients who underwent isolated CABG with TEE vs without TEE. Second, we used logistic regression to model operative mortality with TEE vs without TEE conditional on the 6 key covariates (ejection fraction, CHF, preoperative serum creatinine, >50% left-main coronary stenosis, ≥3 diseased coronaries, and preoperative inotrope use). The CATE for intraoperative TEE was calculated as the difference in probability of operative mortality with TEE minus without TEE. The cohort was then divided into equal quintiles and assigned a numeric value—termed TEE score—ranging from 1 (indicating greatest TEE benefit) to 5 (indicating least TEE benefit).
Stage 2: Validation of the TEE Score via Multiple Matched Comparisons
To test whether the 5 TEE score categories would correctly estimate the clinical benefit of TEE (vs without TEE) after isolated CABG, we conducted a series of matched cohort studies using the reserved 75% subset of data. Within each surgical volume category (low, medium, and high), we compared survival with TEE vs without TEE for each TEE score subgroup.
Statistical Matching
To ensure baseline covariate balance between patients who underwent isolated CABG with TEE vs without TEE, 1:1 pair matching was performed. Consistent with prior comparative effectiveness studies,3,6,25,26 both stages 1 and 2 used exact matching27 and fine balancing28 on selected variables, along with propensity score–balanced matching for all remaining covariates.24 Statistical matching used the Mahalanobis distance with a propensity score caliper and balanced the estimated propensity score distribution, as implemented using the match2C package24 in R (R Project for Statistical Computing). Details are provided in eAppendix 3 in Supplement 1.
Statistical Analysis
Matching quality in both the training (stage 1) and testing (stage 2) phases was assessed using standardized mean differences (SMDs), with SMD less than 0.10 indicating acceptable balance.29,30 Baseline covariate distributions overall and by TEE score were summarized using descriptive statistics. Operative mortality was analyzed using the McNemar test.31 Within each TEE score subgroup, both the risk difference (RD) and odds ratio (OR) were reported.
To assess the robustness of our findings, we conducted 2 sensitivity analyses. First, we calculated E-values32 to quantify the minimum strength of association that an unmeasured confounder would need to have with both TEE use and the outcome in order to fully explain the observed association. Second, we conducted a negative control outcome analysis using new-onset postoperative atrial fibrillation (an event not plausibly affected by intraoperative TEE) to evaluate whether residual unmeasured confounding was adequately controlled.
In supplemental analyses, we also compared rates of coronary reintervention, postoperative stroke, and postoperative chest exploration for bleeding between the with TEE and without TEE groups. All hypothesis testing was 2-sided, and statistical significance was set at P < .05. Data preprocessing was performed in Stata, version 18.0 (StataCorp LLC). Matching and statistical analyses were conducted using the dplyr, exact2x2, and match2C33 packages in R, version 4.2.2 (R Project for Statistical Computing). Code is available in a GitHub repository (link provided in eAppendix 13 in Supplement 1). Data analysis was conducted from August 8, 2023, to December 15, 2024.
Results
Following exclusions (Figure 1), our study cohort consisted of 1 266 055 patients who underwent isolated CABG surgery. Of these patients, 963 976 (76.1%) were male and 302 079 (23.9%) were female, with a mean (SD) age of 65.7 (10.0) years, and 0.7% identified as American Indian or Alaska Native, 3.6% as Asian, 7.5% as Black, 84.3% as White, and 4.1% as individuals of other race.
Among the 1 255 055 isolated CABG procedures, 61.8% received TEE and 39.0% did not. Patients with TEE were demographically and hemodynamically similar to patients without TEE, including mean (SD) ejection fraction (51.87% [12.35%] vs 52.98% [11.61%]) and pulmonary artery systolic pressure (30.89 [6.46] vs 30.95 [6.99] mm Hg) laboratory values, prior procedures (both 1.6%), and mean (SD) PROM score (both 0.02% [0.03%]; lower score indicating lower risk and higher score indicating higher risk of mortality) (Table 1). Distribution across hospitals by surgical volume was comparable across low (19.8%), medium (45.1%), and high (35.1%) volumes (Table 1). Patients without TEE compared with patients with TEE were more often elective cases (39.5% vs 36.8%), were less often transfer admissions (22.7% vs 25.8%), and had lower CHF rates (17.7% vs 21.1%) (Table 1).
Table 1. Unadjusted Baseline Characteristics of the Isolated Coronary Artery Bypass Graft Cohort .
| Covariatea | Patients, No. (%) | ||
|---|---|---|---|
| Overall | Without TEE | With TEE | |
| No. | 1 266 055 | 489 895 | 776 160 |
| Age, mean (SD), y | 65.7 (10.0) | 65.7 (10.1) | 65.6 (10.0) |
| Sex | |||
| Female | 302 079 (23.9) | 119 864 (24.5) | 182 215 (23.5) |
| Male | 963 976 (76.1) | 370 031 (75.5) | 593 945 (76.5) |
| Raceb | |||
| American Indian or Alaska Native | 8484 (0.7) | 3547 (0.7) | 4937 (0.6) |
| Asian | 45 668 (3.6) | 14 140 (2.9) | 31 528 (4.1) |
| Black | 94 324 (7.5) | 36 395 (7.4) | 57 929 (7.5) |
| White | 1 067 823 (84.3) | 419 472 (85.6) | 648 351 (83.5) |
| Otherc | 52 312 (4.1) | 17 397 (3.6) | 34 915 (4.5) |
| Admit source | |||
| Elective | 591 748 (46.7) | 237 428 (48.5) | 354 320 (45.7) |
| Emergency | 335 980 (26.5) | 131 432 (26.8) | 204 548 (26.4) |
| Transfer | 311 583 (24.6) | 111 275 (22.7) | 200 308 (25.8) |
| Arrhythmia | 185 899 (14.7) | 67 462 (13.8) | 118 437 (15.3) |
| CHF | 250 227 (19.8) | 86 571 (17.7) | 163 706 (21.1) |
| CVD | 278 608 (22.0) | 104 953 (21.4) | 173 655 (22.4) |
| Laboratory values | |||
| PA systolic pressure, mean (SD), mm Hg | 30.93 (6.79) | 30.89 (6.46) | 30.95 (6.99) |
| Ejection fraction, mean (SD) | 52.30 (12.08) | 52.98 (11.61) | 51.87 (12.35) |
| Previous CABG | 20 192 (1.6) | 7875 (1.6) | 12 317 (1.6) |
| Hemoglobin, mean (SD), g/dL | 13.25 (1.98) | 13.28 (1.97) | 13.23 (1.99) |
| INR, mean (SD) | 1.05 (0.24) | 1.05 (0.24) | 1.06 (0.23) |
| Creatinine, mean (SD), mg/dL | 1.20 (1.12) | 1.18 (1.05) | 1.22 (1.16) |
| >50% Left-main coronary stenosis | 405 296 (32.0) | 154 524 (31.5) | 250 772 (32.3) |
| ≥3 Diseased coronaries | 974 761 (77.0) | 371 480 (75.8) | 603 281 (77.7) |
| Inotrope use within 48 h of surgery | 14 673 (1.2) | 5238 (1.1) | 9435 (1.2) |
| Operative status | |||
| Elective | 478 614 (37.8) | 193 351 (39.5) | 285 263 (36.8) |
| Emergent | 46 519 (3.7) | 18 094 (3.7) | 28 425 (3.7) |
| Urgent | 740 922 (58.5) | 278 450 (56.8) | 462 472 (59.6) |
| STS PROM score, mean (SD), %d | 0.02 (0.03) | 0.02 (0.03) | 0.02 (0.03) |
| Hospital surgical volume | |||
| Low | 250 490 (19.8) | 99 179 (20.2) | 151 311 (19.5) |
| Medium | 570 582 (45.1) | 218 053 (44.5) | 352 529 (45.4) |
| High | 444 983 (35.1) | 172 663 (35.2) | 272 320 (35.1) |
Abbreviations: CABG, coronary artery bypass graft; CHF, congestive heart failure; CVD, cerebrovascular disease; INR, international normalized ratio; PA, pulmonary artery; PROM, Predicted Risk of Mortality; STS, Society of Thoracic Surgeons; TEE, transesophageal echocardiography.
SI conversion factors: To convert creatinine to micromoles per liter, multiply by 88.4; hemoglobin level to gram per liter, multiply by 10.0.
Covariates represent a selection. The full, baseline covariate distributions may be viewed in eTable 1 in Supplement 1.
Race data were obtained from STS Adult Cardiac Surgery Database.
Other included all responses other than the categories shown.
PROM: lower score indicating lower risk and higher score indicating higher risk of mortality.
Matching Assessment
In stage 1, prior TEE score development postmatch SMDs for all 40 covariates across the 3 surgical volume groups were less than 0.01, well below the accepted SMD threshold (<0.10).29,30 Covariate balance for the low, medium, and high surgical volume hospitals is shown in eAppendix 7 and eTables 3 and 4; eAppendix 8 and eTables 5 and 6; and eAppendix 9 and eTables 7 and 8 in Supplement 1. In stage 2, each of the 5 TEE score categories also achieved SMD less than 0.01 for all covariates.
Heterogeneity of Intraoperative TEE Effect by Surgical Volume
Using the postmatch training data, TEE was associated with a significant survival benefit at low surgical volume hospitals (RD, −0.419% [95% CI, −0.742% to −0.097%], P = .01; OR, 0.86 [95% CI, 0.77-0.97], P = .01) and a statistically significant but marginally clinically significant survival benefit at medium surgical volume hospitals (RD, −0.236% [95% CI, −0.410% to −0.061%], P = .008; OR, 0.90 [95% CI, 0.83-0.97], P = .009). No survival benefit was observed with TEE use at high surgical volume hospitals (RD, −0.112% [95% CI, −0.284% to 0.059%], P = .20; OR, 0.93 [95% CI, 0.84-1.04], P = .21) (eAppendix 6 and eTable 2 in Supplement 1). These findings support developing separate TEE scores by surgical volume category.
Using the reserved postmatch testing data, these findings were confirmed. TEE (vs without TEE) was associated with a significant survival benefit at low surgical volume hospitals (2.47% vs 2.94%; RD, −0.470% [95% CI, −0.631% to −0.308%], P < .001; OR, 0.83 [95% CI, 0.78-0.89], P < .001). TEE was associated with a statistically significant but marginally clinically significant survival benefit at medium surgical volume hospitals (2.09% vs 2.34%; RD, −0.245% [95% CI, −0.344% to −0.146%], P < .001; OR, 0.89 [95% CI, 0.85-0.93], P < .001). No survival benefit was observed with TEE use at high surgical volume hospitals (1.72% vs 1.77%; RD, −0.00% [95% CI, −0.0001% to 0.0001%], P = .45; OR, 0.97 [95% CI, 0.91-1.03], P = .48) (Table 2).
Table 2. Individualized Treatment Effects of Transesophageal Echocardiography in Coronary Artery Bypass Graft Procedures at Low, Medium, and High Surgical Volume Hospitals.
| TEE scorea | TEE group | No. of patients | Death | RD (95% CI), % | P valueb | OR (95% CI) | P valueb | |
|---|---|---|---|---|---|---|---|---|
| No. | Mean rate (SE), % | |||||||
| Low surgical volume | ||||||||
| Pooled | Without | 74 320 | 2183 | 2.94 (0.06) | −0.470 (−0.631 to −0.308) | <.001 | 1 [Reference] | <.001 |
| Pooled | With | 74 320 | 1834 | 2.47 (0.06) | 0.83 (0.78 to 0.89) | |||
| 1 | Without | 13 695 | 691 | 5.05 (0.19) | −0.701 (−1.191 to −0.211) | .009 | 1 [Reference] | .01 |
| 1 | With | 13 695 | 595 | 4.34 (0.17) | 0.85 (0.76 to 0.95) | |||
| 2 | Without | 14 217 | 391 | 2.75 (0.14) | −0.591 (−0.945 to −0.236) | .003 | 1 [Reference] | .003 |
| 2 | With | 14 217 | 307 | 2.16 (0.12) | 0.78 (0.66 to 0.94) | |||
| 3 | Without | 15 541 | 316 | 2.05 (0.11) | −0.418 (−0.715 to −0.121) | .009 | 1 [Reference] | .01 |
| 3 | With | 15 541 | 253 | 1.63 (0.10) | 0.79 (0.67 to 0.94 | |||
| 4 | Without | 16 477 | 277 | 1.68 (0.10) | −0.334 (−0.595 to −0.072) | .01 | 1 [Reference] | .02 |
| 4 | With | 16 477 | 222 | 1.35 (0.09) | 0.80 (0.66 to 0.94) | |||
| 5 | Without | 14 390 | 506 | 3.52 (0.15) | −0.314 (−0.745 to 0.064) | .10 | 1 [Reference] | .11 |
| 5 | With | 14 390 | 457 | 3.18 (0.15) | 0.90 (0.78 to 1.02) | |||
| Medium surgical volume | ||||||||
| Pooled | Without | 163 721 | 3827 | 2.34 (0.04) | −0.245 (−0.344 to −0.146) | <.001 | 1 [Reference] | <.001 |
| Pooled | With | 163 721 | 3426 | 2.09 (0.04) | 0.89 (0.85 to 0.93) | |||
| 1 | Without | 32 633 | 1093 | 3.35 (0.10) | −0.472 (−0.732 to −0.212) | <.001 | 1 [Reference] | .001 |
| 1 | With | 32 633 | 939 | 2.88 (0.09) | 0.85 (0.77 to 0.93) | |||
| 2 | Without | 32 842 | 804 | 2.45 (0.09) | −0.298 (−0.525 to −0.072) | .02 | 1 [Reference] | .02 |
| 2 | With | 32 842 | 706 | 2.15 (0.08) | 0.87 (0.79 to 0.97) | |||
| 3 | Without | 32 099 | 441 | 1.37 (0.07) | −0.206 (−0.377 to −0.034) | .03 | 1 [Reference] | .03 |
| 3 | With | 32 099 | 375 | 1.17 (0.06) | 0.85 (0.73 to 0.98) | |||
| 4 | Without | 33 287 | 709 | 2.13 (0.08) | −0.117 (−0.329 to 0.095) | .28 | 1 [Reference] | .29 |
| 4 | With | 33 287 | 670 | 2.01 (0.08) | 0.94 (0.84 to 1.05) | |||
| 5 | Without | 32 860 | 780 | 2.37 (0.08) | −0.134 (−0.358 to 0.091) | .28 | 1 [Reference] | .29 |
| 5 | With | 32 860 | 736 | 2.24 (0.08) | 0.94 (0.85 to 1.04) | |||
| High surgical volume | ||||||||
| Pooled | Without | 128 215 | 2265 | 1.77 (0.04) | 0 (−0.0001 to 0.0001) | .45 | 1 [Reference] | .48 |
| Pooled | With | 128 215 | 2203 | 1.72 (0.04) | 0.97 (0.91 to 1.03) | |||
| 1 | Without | 25 296 | 714 | 2.82 (0.10) | 0.229 (−0.058 to 0.516) | .35 | 1 [Reference] | .36 |
| 1 | With | 25 296 | 772 | 3.05 (0.11) | 1.09 (0.98 to 1.21) | |||
| 2 | Without | 26 372 | 300 | 1.14 (0.07) | −0.053 (−0.229 to 0.123) | .56 | 1 [Reference] | .58 |
| 2 | With | 26 372 | 286 | 1.08 (0.06) | 0.95 (0.90 to 1.13) | |||
| 3 | Without | 25 049 | 326 | 1.30 (0.07) | −0.088 (−0.282 to 0.107) | .45 | 1 [Reference] | .48 |
| 3 | With | 25 049 | 304 | 1.21 (0.07) | 0.93 (0.79 to 1.09) | |||
| 4 | Without | 25 473 | 349 | 1.37 (0.07) | −0.185 (−0.379 to −0.010) | .35 | 1 [Reference] | .36 |
| 4 | With | 25 473 | 302 | 1.19 (0.07) | 0.86 (0.74 to 1.01) | |||
| 5 | Without | 26 025 | 576 | 2.21 (0.09) | −0.142 (−0.387 to 0.102) | .45 | 1 [Reference] | .48 |
| 5 | With | 26 025 | 539 | 2.07 (0.09) | 0.93 (0.82 to 1.05) | |||
Abbreviations: OR, odds ratio; RD, risk difference; TEE, transesophageal echocardiography.
TEE score range: 1 to 5, with 1 indicating most benefit and 5 indicating least benefit from TEE.
Benjamini-Hochberg correction calculation was applied to P values for both the RDs and ORs to control for a false discovery rate.
Low Surgical Volume Hospitals: Individualized TEE Treatment Effects
Using the postmatch testing data, among 148 640 patients who underwent CABG (74 320 with TEE vs 74 320 without TEE) at low surgical volume hospitals, TEE was associated with a statistically and clinically significant survival benefit for TEE scores 1 to 3 (RD, −0.701% [95% CI, −1.191% to −0.211%] to −0.418% [95% CI, −0.715% to −0.121%]; OR, 0.78 [95% CI, 0.66-0.94] to 0.85 [95% CI, 0.76-0.95]; P < .01) (Table 2). TEE score 4 showed a statistically significant but marginal clinical benefit (RD, −0.334% [95% CI, −0.595 to −0.072]; OR, 0.80 [95% CI, 0.66-0.94]; P ≤ .01). No statistically significant benefit was observed for TEE score 5 (RD, −0.314% [95% CI, −0.745% to 0.064%]; OR, 0.90 [95% CI, 0.78-1.02]; P ≥ .10) (Table 2, Figure 2).
Figure 2. Risk Difference for Death With vs Without Transesophageal Echocardiography (TEE) in Coronary Artery Bypass Graft (CABG) by Surgical Volume and TEE Score.
Error bars represent 95% CIs. TEE score range: 1 to 5, with 1 indicating most benefit and 5 indicating least benefit from TEE.
Low Surgical Volume Hospitals: Patient Characteristics Associated With Estimated Individualized TEE Benefit
At low surgical volume hospitals vs overall, patients who benefitted most from TEE (score 1) had a lower (<55%) ejection fraction (86.8% vs 41.5%), higher creatinine level (2.02 vs 1.18 mg/dL), greater than 50% left-main stenosis (50.8% vs 33.1%), 3 or more diseased coronaries (95.6% vs 76.9%), and greater preoperative inotrope use (4.7% vs 1.2%) (Table 3; eAppendix 10 and eTable 9 in Supplement 1). Patients who benefitted least from TEE (score 5) had higher CHF prevalence (73.0% vs 15.1%) compared with the overall population (Table 3; eTable 9 in Supplement 1). Notably, patients in both TEE scores 1 and 5 groups had similarly elevated PROM score (>2.5%), compared with lower-risk patients with TEE scores 2 to 3 (<1.6%), underscoring that TEE benefit varied independent of baseline surgical risk (Table 3; eTable 9 and eAppendix 14 in Supplement 1).
Table 3. Characteristics Associated With Individualized Treatment Effects of Transesophageal Echocardiography at Low and Medium Surgical Volume Hospitals.
| Characteristic | Overall, No. (%) | TEE score, No. (%)a | ||||
|---|---|---|---|---|---|---|
| 1 | 5 | |||||
| Without TEE | With TEE | Without TEE | With TEE | Without TEE | With TEE | |
| Low surgical volume | ||||||
| No. of patients | 74 320 | 74 320 | 13 695 | 13 695 | 14 390 | 14 390 |
| EF, mean (SD) | 52.54 (11.82) | 52.50 (11.91) | 44.73 (10.69) | 44.65 (10.62) | 49.17 (14.39) | 48.97 (14.53) |
| EF <55% | 30 732 (41.4) | 30 879 (41.5) | 11 770 (85.9) | 11 886 (86.8) | 6737 (46.8) | 6846 (47.6) |
| CHF | 11 312 (15.2) | 11 211 (15.1) | 521 (3.8) | 487 (3.6) | 10 519 (73.1) | 10 508 (73.0) |
| Creatinine, mean (SD), mg/dL | 1.19 (1.09) | 1.18 (1.05) | 2.07 (2.23) | 2.02 (2.15) | 1.02 (0.52) | 1.02 (0.53) |
| >50% Left-main coronary stenosis | 24 574 (33.1) | 24 611 (33.1) | 6912 (50.5) | 6961 (50.8) | 3818 (26.5) | 3855 (26.8) |
| ≥3 Diseased coronaries | 56 751 (76.4) | 57 135 (76.9) | 13 088 (95.6) | 13 090 (95.6) | 8192 (56.9) | 8259 (57.4) |
| Inotrope use within 48 h of surgery | 925 (1.2) | 893 (1.2) | 671 (4.9) | 650 (4.7) | 241 (1.7) | 237 (1.6) |
| STS PROM score, mean (SD), %b | 1.85 (2.91) | 1.84 (2.88) | 3.00 (4.19) | 2.95 (4.06) | 2.42 (3.92) | 2.44 (3.99) |
| Death rate, mean (SE), % | 2.94 (0.06) | 2.47 (0.06) | 5.05 (0.19) | 4.34 (0.17) | 3.52 (0.15) | 3.18 (0.15) |
| E-value | 1.51 | 1.51 | 1.28 | 1.28 | 1.00 | 1.00 |
| Medium surgical volume | ||||||
| No. of patients | 163 721 | 163 721 | 32 633 | 32 633 | 32 860 | 32 860 |
| EF, mean (SD) | 53.00 (11.68) | 52.99 (11.73) | 53.15 (11.35) | 53.17 (11.42) | 47.34 (12.82) | 47.35 (12.96) |
| EF <55% | 64 822 (39.6) | 65 426 (40.0) | 13 071 (40.1) | 13 319 (40.8) | 21 422 (65.2) | 21 425 (65.2) |
| CHF | 28 540 (17.4) | 28 324 (17.3) | 696 (2.1) | 657 (2.0) | 11 419 (34.8) | 11 352 (34.5) |
| Creatinine, mean (SD), mg/dL | 1.17 (1.02) | 1.16 (0.97) | 1.30 (1.32) | 1.28 (1.29) | 1.32 (1.34) | 1.29 (1.24) |
| >50% Left-main coronary stenosis | 51 914 (31.7) | 51 737 (31.6) | 31 149 (95.5) | 31 289 (95.9) | 2531 (7.7) | 2386 (7.3) |
| ≥3 Diseased coronaries | 124 996 (76.3) | 125 551 (76.7) | 32 215 (98.7) | 32 322 (99.0) | 9086 (27.7) | 9176 (27.9) |
| Inotrope use within 48 h of surgery | 1804 (1.1) | 1645 (1.0) | 1804 (5.5) | 1645 (5.0) | 0 | 0 |
| STS PROM score, mean (SD), %b | 1.83 (3.02) | 1.83 (2.98) | 2.42 (4.32) | 2.41 (4.31) | 1.96 (3.04) | 1.95 (3.02) |
| Death rate, mean (SE), % | 2.34 (0.04) | 2.09 (0.04) | 3.35 (0.10) | 2.88 (0.09) | 2.37 (0.08) | 2.24 (0.08) |
| E-value | 1.35 | 1.35 | 1.36 | 1.36 | 1.00 | 1.00 |
Abbreviations: CHF, congestive heart failure; EF, ejection fraction; PROM, Predicted Risk of Mortality; STS, Society of Thoracic Surgeons; TEE, transesophageal echocardiography.
SI conversion factgor: To convert creatinine to µmol/L, multiply by 88.4.
TEE score range: 1 to 5, with 1 indicating most benefit and 5 indicating least benefit from TEE. The complete table with results from all 5 TEE score subgroups is available in eTable 9 (low volume) and eTable 10 (medium volume) in Supplement 1.
PROM: a lower score indicates lower risk and a higher score indicates higher risk of mortality.
Medium Surgical Volume Hospitals: Individualized TEE Treatment Effects
Using the postmatch testing data, among 347 442 patients who underwent CABG (163 721 with TEE vs 163 721 without TEE) at medium surgical volume hospitals, TEE score 1 was associated with a statistically and clinically significant survival benefit (RD, −0.472% [95% CI, −0.732% to −0.212%]; OR, 0.85 [95% CI, 0.77-0.93]; P < .001). Among those with TEE scores 2 and 3, we observed a statistically significant but marginal clinical benefit (RD, −0.298% [95% CI, −0.525% to −0.072%] and −0.206% [95% CI, −0.377% to −0.034%]; OR, 0.87 [95% CI, 0.79-0.97] and 0.85 [95% CI, 0.73-0.98]; P = .03). No statistically significant benefit was observed for TEE scores 4 or 5 (RD, −0.117% [95% CI, −0.329% to 0.095%] and −0.134% [95% CI, −0.358% to 0.091%]; OR, 0.94 [95% CI, 0.84-1.05] and 0.94 [95% CI, 0.85-1.04]; P > .29) (Table 2, Figure 2).
Medium Surgical Volume Hospitals: Patient Characteristics Associated With Estimated Individualized TEE Benefit
At medium surgical volume hospitals vs overall, patients who benefitted most from TEE (score 1) had more (>50%) left-main stenosis (95.9% vs 31.6%), 3 or more diseased coronaries (99.0% vs 76.7%), and higher preoperative inotrope use (5.0% vs 1.0%) (Table 3; eTable 10 in Supplement 1). All TEE score subgroups had similar PROM score (1.1% to 2.4%), reinforcing that TEE benefit varied independent of baseline risk (Table 3; eTable 10 and eAppendix 14 in Supplement 1).
High Surgical Volume Hospitals: Individualized TEE Treatment Effects
Using the postmatch testing data, among 256 430 patients who underwent CABG (128 215 with TEE vs 128 215 without TEE) at high surgical volume hospitals, TEE was not associated with a statistically significant survival benefit overall or within any TEE score subgroup (Table 2, Figure 2).
Sensitivity and Negative Control Outcome Analyses
The sensitivity analysis showed that the primary finding—intraoperative TEE use was associated with lower mortality rates in low and medium surgical volume hospitals—was robust to a moderate level of unmeasured confounding (E-value = 1.51 for low surgical volume hospitals, and 1.35 for medium surgical volume hospitals) (Table 3). Within each surgical volume strata, the association was most robust in the TEE score 1 subgroup (Table 3).
Postoperative atrial fibrillation is an outcome that should not plausibly be associated with intraoperative TEE. A negative control outcome analysis showed that, across all hospital volume strata, TEE use was associated with slightly higher rates of postoperative atrial fibrillation in pooled analyses. The pooled ORs were as follows: 1.04 (95% CI, 1.01-1.06) for low surgical volume hospitals, 1.05 (95% CI, 1.03-1.07) for medium surgical volume hospitals, and 1.06 (95% CI, 1.04-1.08) for high surgical volume hospitals. Absolute risk differences were small (<1.1%) and not consistent across TEE score subgroups (eAppendix 11 in Supplement 1). Additionally, these results are opposite to the increased survival association observed with TEE in the low and medium surgical volume hospitals from the primary analysis. This opposite-direction association for a negative control outcome strengthens the interpretation that the observed mortality benefit associated with TEE is unlikely to be explained by a systematic bias favoring TEE.
Supplemental Clinical Outcomes
We also examined several other postoperative complications as supplemental clinical outcomes (eAppendix 11 in Supplement 1). These included coronary reintervention, postoperative stroke, and postoperative chest exploration for bleeding. These are complications for which a causal association with TEE is at least plausible. Across all hospital volume strata, we found no statistically significant differences in the rates of these outcomes between the with TEE and without TEE groups.
For coronary reintervention, pooled ORs were as follows: 1.13 (95% CI, 0.96-1.34) for low surgical volume hospitals, 0.98 (95% CI, 0.87-1.09) for medium surgical volume hospitals, and 1.12 (95% CI, 0.97-1.29) for high surgical volume hospitals. For new postoperative stroke, the pooled ORs were as follows: 1.00 (95% CI, 0.92-1.09) for low surgical volume hospitals, 1.02 (95% CI, 0.96-1.09) for medium surgical volume hospitals, and 1.00 (95% CI, 0.94-1.08) for high surgical volume hospitals. For postoperative chest exploration, the pooled ORs were as follows: 1.00 (95% CI, 0.92-1.08) for low surgical volume hospitals, 0.95 (95% CI, 0.90-1.00) for medium surgical volume hospitals, and 1.04 (95% CI, 0.97-1.10) for high surgical volume hospitals. Absolute risk differences for all 3 outcomes were small (≤0.41% for coronary reintervention, ≤1.38% for stroke, and ≤1.87% for chest exploration) and varied in direction by TEE score subgroup. These null findings may be attributable, at least in part, to the low incidence of these complications in our study population, which limited statistical power to detect small differences.
Discussion
In this cohort study of 1 266 055 patients who underwent isolated CABG with or without intraoperative TEE, we used target trial methodologies to develop and validate individualized TEE treatment decision rules. Two key findings emerged. First, of patients undergoing isolated CABG at low (<100 cases/y) and medium (100-250 cases/y) surgical volume hospitals, our 2-stage analysis uncovered an association between intraoperative TEE and improved survival among patients (1) with greater than 50% (vs ≤50%) left-main stenosis, (2) with 3 or more (vs <3) diseased coronaries, and (3) taking inotropes (vs no inotropes) within 48 hours of surgery. Second, in patients who underwent isolated CABG at high (≥250 cases/y) surgical volume hospitals, we observed no association between intraoperative TEE (vs without TEE) and increased survival, either overall or among any subpopulations.
Current American College of Cardiology/American Heart Association guidelines assign intraoperative TEE a class IIb recommendation for isolated CABG, citing unknown usefulness.34 In the absence of randomized clinical trial evidence, this classification reflects uncertainty about the clinical benefit of TEE during isolated CABG surgery.5,6,7 Our findings contribute to a growing body of work4,5,6,7 that compares clinical outcomes after isolated CABG with vs without intraoperative TEE. At low and medium surgical volume centers, the results broadly align with previous observational studies showing that TEE benefits sicker patients but not healthier ones.4,5,6,7 More specifically, our TEE score analysis identified patients with 3 or more diseased vessels (vs <3) to be more likely to benefit from TEE, a finding consistent with a 2024 analysis using private administrative claims data.6
The survival benefit to TEE among subgroups with greater than 50% left-main disease, 3 or more diseased coronaries, or preoperative inotrope use persisted even among patients with similar PROM scores, suggesting heterogeneity in TEE effectiveness even within lower-risk groups. This finding builds on the study by Metkus et al,5 which found no benefit of TEE among patients with low estimated mortality (<4%) by demonstrating that even within similar overall risk categories, certain clinical phenotypes appear more responsive to TEE. While PROM remains a valuable tool for benchmarking surgical outcomes, our findings suggest that PROM scores may be insufficient as the sole basis for identifying patients most likely to benefit from intraoperative TEE. These results underscore the need for more granular, physiology-informed approaches to patient selection—particularly as institutions seek to optimize the use of limited perioperative imaging resources.
Unexpectedly, we found no survival benefit from TEE among patients with CHF. This result contrasts with a 2024 claims-based study showing increased survival in this group with the use of TEE.6 This discrepancy may reflect differences in CHF classification using International Statistical Classification of Diseases and Related Health Problems, Tenth Revision codes in claims data6 compared with New York Heart Association class in this study. Alternatively, this result may indicate that CHF is not actually a modifier of TEE benefit. TEE is particularly valuable for detecting acute, reversible hemodynamic deterioration, such as reduced ejection fraction due to hypotension and decreased coronary perfusion, which may not be as relevant in patients with chronic, fixed ventricular dysfunction due to CHF.35,36,37
Another surprising finding from the current study was the absence of TEE benefit at high surgical volume centers. This result contrasts with prior work linking TEE to improved outcomes in higher-risk patients.4,5,6 A likely explanation is that high surgical volume hospitals already achieve low mortality (1.72% with TEE vs 1.77% without TEE), potentially offsetting any added value from TEE via factors, such as surgical expertise, intensive care unit resources, and improved staffing. It is also possible that within the high surgical volume centers, institutional practices are already standardized and highly protocolized, thereby diminishing any incremental value of additional intraoperative monitoring with TEE. Although high-volume hospitals vary in size and resources, the absence of any detectable benefit across this stratum suggests that the utility of TEE may plateau once a certain threshold of institutional capability to improve outcomes is reached. In contrast, the more pronounced benefit in the low and medium surgical volume hospitals may reflect greater variability in intraoperative management and outcomes, creating an opportunity for TEE to meaningfully play a role in care decisions.
An alternative explanation may be residual unobserved confounding at the clinician level. One critical unmeasured confounder may be the cardiac anesthesiologist. While not directly measured in this study or previous studies due to the lack of anesthesiologist identifiers,4,5,6,7 cardiac anesthesiologists (who often staff high surgical volume centers) may affect postoperative outcomes more than the TEE itself. Specifically, cardiac anesthesiologist involvement may encompass a bundle of co-occurring practices such as goal-directed hemodynamic management, enhanced communication with the surgical team, use of inotropic or vasopressor titration protocols, real-time biventricular function assessments, and proactive optimization of volume status and afterload modification in the context of cardiopulmonary bypass and myocardial protection strategies. Thus, in these settings, TEE may offer only limited incremental benefit because outcomes at high surgical volume centers are already optimized through alternative mechanisms. In contrast, at the low and medium surgical volume centers (where previous research has shown that cardiac anesthesiologist staffing is associated with high likelihood of TEE use during CABG8), TEE may serve as a marker for broader cardiac anesthesiologist involvement. Consequently, the observed clinical benefit associated with TEE use at the low surgical volume centers may, in part, reflect the impact of anesthesiologist expertise along with the TEE.
Our findings have practical implications. At lower surgical volume centers, key patient-level factors (eg, >50% left-main coronary stenosis, ≥3 diseased coronaries, and inotrope use) are routinely captured in the electronic medical record. These patient-level parameters could be used as a clinical decision support to guide TEE use or cardiac anesthesiologist staffing, particularly in TEE resource–limited settings. For instance, if a hospital only has the capacity to use intraoperative TEE in 25% of isolated CABGs, targeted TEE allocation to patients with TEE-responsive conditions may improve outcomes. To support this allocation, we developed a clinical software tool that calculates a TEE benefit score based on preoperative characteristics and hospital surgical volume (eAppendix 12 in Supplement 1).
While our clinical decision support tool was developed to guide targeted TEE use in higher-risk patients, our findings also highlight an opportunity to reduce routine TEE use in cases with lower-risk for CABG. Specifically, the findings raise important considerations for adopting a more judicious, risk-stratified approach to intraoperative TEE, particularly among patients with a low likelihood of postoperative complications. Avoiding routine TEE in these populations may offer several advantages. First, from an operational perspective, it may improve operating room efficiency by eliminating the time required for probe placement and image acquisition and interpretation. Second, while life-threatening complications are rare,38,39 structural heart literature has reported subclinical or endoscopically evident esophageal injuries in up to 86% of patients,40 suggesting a need for caution even in the absence of overt harm. Third, selective TEE use may reduce resource demands, including equipment wear, sterilization time, and staffing, which is particularly relevant for hospitals facing capacity or personnel constraints. Taken together, these considerations support a pragmatic care model in which TEE is not used by default in low-risk isolated CABG but remains readily available for intraoperative hemodynamic changes or unanticipated events. This risk-stratified, on-demand approach aligns with TEE practices in noncardiac surgery and may help optimize the balance between clinical benefit, procedural risk, and resource utilization in cardiac surgical care.
Fourth, the lack of observed survival benefit from intraoperative TEE among patients who underwent isolated CABG at high surgical volume centers (including patients in higher-risk subgroups) suggests a potential opportunity to further explore the comparative effectiveness of TEE in this setting. While the logistics and cost of a large randomized clinical trial would be considerable, our findings highlight a subgroup of patients in whom the marginal benefit of TEE appears limited, making randomized evaluation ethically and clinically justifiable.
Limitations
This study must be interpreted with awareness of its limitations. First, the observational, nonrandomized design of this study cannot confirm a causal association between TEE and increased survival because of the inability to eliminate residual confounding between patients who underwent surgery with TEE vs without TEE. Given the granularity of patient-level factors captured in the STS ACSD and the rigorous matching techniques used in this study, the most likely source of residual unobserved confounding is clinician level (as opposed to patient level). Examples of clinician-level confounding could be surgeon preference for (or against) intraoperative TEE or the availability (or unavailability) of a cardiac anesthesiologist with the certification to perform an intraoperative TEE. Because the STS ACSD does not include detailed clinician- or hospital-level variables, we stratified hospitals by surgical volume, which is a well-established proxy for institutional capability. We then matched patients within each volume stratum to mitigate confounding related to unmeasured institutional practices.
Second, while we purposefully conducted our analyses within each of the 3 surgical volume categories, we were unable to adjust for a wide variety of hospital-level factors that could confound the association between TEE and increased survival, such as postoperative or intraoperative care process differences between hospitals. Third, the STS ACSD does not include sufficient detail to classify CABG procedures as guideline-indicated or nonindicated with respect to percutaneous coronary intervention candidacy or SYNTAX scoring. However, the matched analyses balanced coronary anatomy across the with TEE and without TEE comparison groups, and we found that patients with left-main or multivessel disease appeared to derive greater benefit from TEE regardless of surgical volume. Fourth, while the STS ACSD does not include a variable for return to cardiopulmonary bypass across all data versions, we conducted a supplementary analysis of unplanned coronary reintervention as a proxy for certain TEE-detected intraoperative findings. Although not a perfect surrogate, this outcome provides insight into potential downstream consequences of TEE use during CABG.
Fifth, we were unable to compare gastroesophageal complication rates between patients who underwent isolated CABG with TEE vs without TEE. While life-threatening complications directly attributable to intraoperative TEE are rare (<0.01%),38,39 interventional cardiology literature has reported a gastroesophageal injury rate that is much higher (22% to 40%)40,41 than has been previously reported in the cardiac surgical literature (0.1% to 0.4%).38,39 Consequently, a key question for future investigation is whether the benefits of intraoperative TEE as an invasive hemodynamic monitor outweigh the risks of gastroesophageal injury in patients who underwent isolated CABG surgery.
Conclusions
Among patients who underwent isolated CABG at low and medium surgical volume hospitals, the survival benefit of intraoperative TEE was greatest in those with greater than 50% left-main coronary stenosis, 3 or more diseased vessels, or preoperative inotropic support. In contrast, no survival benefit was observed at high surgical volume hospitals, either overall or within high-risk subgroups. These findings support a more individualized approach to TEE use and underscore the need for a randomized clinical trial to define TEE’s role in isolated CABG surgery.
eAppendix 1. List of Matching Covariates
eAppendix 2. Descriptive Analysis
eFigure 1. Density Plot of the Proportion of CABG Surgeries Using TEE by Hospital Surgical Volume
eTable 1. Covariate Distribution Among Training and Testing Data, Stratified by Receipt of TEE
eAppendix 3. Details on Statistical Matching Methodology
eAppendix 4. Methods for Identification of Subgroups That Benefit From TEE (Stage One)
eAppendix 5. Methods for Validating TEE Score (Stage Two)
eAppendix 6. Preliminary Results
eTable 2. 30-Day Mortality in the Training Data Set Stratified by Hospital Volume and TEE Receipt
eAppendix 7. Low Surgical Volume Hospitals
eTable 3. Low Volume Covariate Distribution Among Post-Match Testing Data
eTable 4. Low Volume Hospitals Training Data Selected Covariate Distribution by TEE Score
eAppendix 8. Medium Surgical Volume Hospitals
eTable 5. Medium Volume Covariate Distribution Among Post-Match Testing Data
eTable 6. Medium Volume Hospitals Training Data Selected Covariate Distribution by TEE Score
eAppendix 9. High Surgical Volume Hospitals
eTable 7. High Surgical Volume Hospitals Covariate Distribution Among Post-Match Testing Data
eTable 8. High Surgical Volume Hospitals Training Data Selected Covariate Distribution by TEE Score
eAppendix 10. Testing Data Results
eTable 9. Characteristics Associated With Predicted Individualized TEE Treatment Effects at Low Surgical Volume Hospitals
eTable 10. Characteristics Associated With Predicted Individualized TEE Treatment Effects at Medium Surgical Volume Hospitals
eAppendix 11. Additional Outcome Analyses
eAppendix 12. Risk Score Calculator
eAppendix 13. Code Documentation
eReferences
eAppendix 14. Two-stage Target Trial Matched Analysis
Data Sharing Statement
References
- 1.Nicoara A, Skubas N, Ad N, et al. Guidelines for the use of transesophageal echocardiography to assist with surgical decision-making in the operating room: a surgery-based approach: from the American Society of Echocardiography in collaboration with the Society of Cardiovascular Anesthesiologists and the Society of Thoracic Surgeons. J Am Soc Echocardiogr. 2020;33(6):692-734. doi: 10.1016/j.echo.2020.03.002 [DOI] [PubMed] [Google Scholar]
- 2.MacKay EJ, Neuman MD, Fleisher LA, et al. Transesophageal echocardiography, mortality, and length of hospitalization after cardiac valve surgery. J Am Soc Echocardiogr. 2020;33(6):756-762.e1. doi: 10.1016/j.echo.2020.01.014 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.MacKay EJ, Zhang B, Augoustides JG, Groeneveld PW, Desai ND. Association of intraoperative transesophageal echocardiography and clinical outcomes after open cardiac valve or proximal aortic surgery. JAMA Netw Open. 2022;5(2):e2147820. doi: 10.1001/jamanetworkopen.2021.47820 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.MacKay EJ, Zhang B, Heng S, et al. Association between transesophageal echocardiography and clinical outcomes after coronary artery bypass graft surgery. J Am Soc Echocardiogr. 2021;34(6):571-581. doi: 10.1016/j.echo.2021.01.014 [DOI] [PubMed] [Google Scholar]
- 5.Metkus TS, Thibault D, Grant MC, et al. Transesophageal echocardiography in patients undergoing coronary artery bypass graft surgery. J Am Coll Cardiol. 2021;78(2):112-122. doi: 10.1016/j.jacc.2021.04.064 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.MacKay EJ, Talham CJ, Zhang B, et al. Testing clinical selection criteria for intraoperative transoesophageal echocardiography in isolated coronary artery bypass graft surgery. BJA Open. 2024;10:100278. doi: 10.1016/j.bjao.2024.100278 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.MacKay EJ, Werner RM, Groeneveld PW, et al. Transesophageal echocardiography, acute kidney injury, and length of hospitalization among adults undergoing coronary artery bypass graft surgery. J Cardiothorac Vasc Anesth. 2020;34(3):687-695. doi: 10.1053/j.jvca.2019.08.042 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.MacKay EJ, Zhang B, Neuman MD, Augoustides JG, Desai ND, Groeneveld PW. Impact of hospital practice and staffing differences on transesophageal echocardiography use in cardiac valve or coronary artery bypass graft surgery. J Cardiothorac Vasc Anesth. 2022;36(11):4012-4021. doi: 10.1053/j.jvca.2022.07.005 [DOI] [PubMed] [Google Scholar]
- 9.MacKay EJ, Zhang B, Shah RM, Augoustides JG, Groeneveld PW, Desai ND. Predictors of intraoperative echocardiography: analysis of the Society of Thoracic Surgeons database. Ann Thorac Surg. 2023;115(5):1289-1295. doi: 10.1016/j.athoracsur.2023.01.005 [DOI] [PubMed] [Google Scholar]
- 10.Gupta S, Wang W, Hayek SS, et al. ; STOP-COVID Investigators . Association between early treatment with tocilizumab and mortality among critically ill patients with COVID-19. JAMA Intern Med. 2021;181(1):41-51. doi: 10.1001/jamainternmed.2020.6252 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Hernán MA, Robins JM. Using big data to emulate a target trial when a randomized trial is not available. Am J Epidemiol. 2016;183(8):758-764. doi: 10.1093/aje/kwv254 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Hernán MA, Wang W, Leaf DE. Target trial emulation: a framework for causal inference from observational data. JAMA. 2022;328(24):2446-2447. doi: 10.1001/jama.2022.21383 [DOI] [PubMed] [Google Scholar]
- 13.Kent DM, Saver JL, Kasner SE, et al. Heterogeneity of treatment effects in an analysis of pooled individual patient data from randomized trials of device closure of patent foramen ovale after stroke. JAMA. 2021;326(22):2277-2286. doi: 10.1001/jama.2021.20956 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Society of Thoracic Surgeons . STS fact sheet. Accessed September 26, 2024. https://www.sts.org/sts-fact-sheet
- 15.von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP; STROBE Initiative . The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. Lancet. 2007;370(9596):1453-1457. doi: 10.1016/S0140-6736(07)61602-X [DOI] [PubMed] [Google Scholar]
- 16.Kim KM, Arghami A, Habib R, et al. The Society of Thoracic Surgeons adult cardiac surgery database: 2022 update on outcomes and research. Ann Thorac Surg. 2023;115(3):566-574. doi: 10.1016/j.athoracsur.2022.12.033 [DOI] [PubMed] [Google Scholar]
- 17.Alkhouli M, Alqahtani F, Cook CC. Association between surgical volume and clinical outcomes following coronary artery bypass grafting in contemporary practice. J Card Surg. 2019;34(10):1049-1054. doi: 10.1111/jocs.14205 [DOI] [PubMed] [Google Scholar]
- 18.Kim LK, Looser P, Swaminathan RV, et al. Outcomes in patients undergoing coronary artery bypass graft surgery in the United States based on hospital volume, 2007 to 2011. J Thorac Cardiovasc Surg. 2016;151(6):1686-1692. doi: 10.1016/j.jtcvs.2016.01.050 [DOI] [PubMed] [Google Scholar]
- 19.Squiers JJ, Schaffer JM, Banwait JK, Ryan WH, Mack MJ, DiMaio JM. Long-term survival after on-pump and off-pump coronary artery bypass grafting. Ann Thorac Surg. 2022;113(6):1943-1952. doi: 10.1016/j.athoracsur.2021.07.037 [DOI] [PubMed] [Google Scholar]
- 20.Khoury H, Sanaiha Y, Rudasill SE, Mardock AL, Sareh S, Benharash P. Readmissions following isolated coronary artery bypass graft surgery in the United States (from the Nationwide Readmissions Database 2010 to 2014). Am J Cardiol. 2019;124(2):205-210. doi: 10.1016/j.amjcard.2019.04.018 [DOI] [PubMed] [Google Scholar]
- 21.Barnett PG, Hong JS, Carey E, Grunwald GK, Joynt Maddox K, Maddox TM. Comparison of accessibility, cost, and quality of elective coronary revascularization between Veterans Affairs and community care hospitals. JAMA Cardiol. 2018;3(2):133-141. doi: 10.1001/jamacardio.2017.4843 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Buell KG, Spicer AB, Casey JD, et al. Individualized treatment effects of oxygen targets in mechanically ventilated critically ill adults. JAMA. 2024;331(14):1195-1204. doi: 10.1001/jama.2024.2933 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Goligher EC, Lawler PR, Jensen TP, et al. ; REMAP-CAP, ATTACC, and ACTIV-4a Investigators . Heterogeneous treatment effects of therapeutic-dose heparin in patients hospitalized for COVID-19. JAMA. 2023;329(13):1066-1077. doi: 10.1001/jama.2023.3651 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Zhang B, Small DS, Lasater KB, McHugh M, Silber JH, Rosenbaum PR. Matching one sample according to two criteria in observational studies. J Am Stat Assoc. 2023;118(542):1140-1151. doi: 10.1080/01621459.2021.1981337 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Jain S, Rosenbaum PR, Reiter JG, et al. Assessing the ambulatory surgery center volume-outcome association. JAMA Surg. 2024;159(4):397-403. doi: 10.1001/jamasurg.2023.7161 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Goldstone AB, Chiu P, Baiocchi M, et al. Interfacility transfer of Medicare beneficiaries with acute type A aortic dissection and regionalization of care in the United States. Circulation. 2019;140(15):1239-1250. doi: 10.1161/CIRCULATIONAHA.118.038867 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Rosenbaum PR. Modern algorithms for matching in observational studies. Annu Rev Stat Appl. 2020;7:143-176. doi: 10.1146/annurev-statistics-031219-041058 [DOI] [Google Scholar]
- 28.Rosenbaum PR, Ross RN, Silber JH. Minimum distance matched sampling with fine balance in an observational study of treatment for ovarian cancer. J Am Stat Assoc. 2007;102(477):75-83. doi: 10.1198/016214506000001059 [DOI] [Google Scholar]
- 29.Rosenbaum PR, Rubin DB. Constructing a control group using multivariate matched sampling methods that incorporate the propensity score. Am Stat. 1985;39(1):33-38. doi: 10.1080/00031305.1985.10479383 [DOI] [Google Scholar]
- 30.Silber JH, Rosenbaum PR, Trudeau ME, et al. Multivariate matching and bias reduction in the surgical outcomes study. Med Care. 2001;39(10):1048-1064. doi: 10.1097/00005650-200110000-00003 [DOI] [PubMed] [Google Scholar]
- 31.Rosenbaum PR. Overt bias. In: Rosenbaum PR. Observational Studies. Springer; 2002. doi: 10.1007/978-1-4757-3692-2 [DOI] [Google Scholar]
- 32.VanderWeele TJ, Ding P. Sensitivity analysis in observational research: introducing the E-value. Ann Intern Med. 2017;167(4):268-274. doi: 10.7326/M16-2607 [DOI] [PubMed] [Google Scholar]
- 33.Wickham H, François R, Henry L, Müller K, Vaughan D. dplyr: a grammar of data manipulation. 2020. Accessed March 2024. https://cran.r-project.org/web/packages/dplyr/index.html
- 34.Lawton JS, Tamis-Holland JE, Bangalore S, et al. 2021 ACC/AHA/SCAI guideline for coronary artery revascularization: executive summary: a report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines. Circulation. 2022;145(3):e4-e17. doi: 10.1161/CIR.0000000000001039 [DOI] [PubMed] [Google Scholar]
- 35.Heusch G, Libby P, Gersh B, et al. Cardiovascular remodelling in coronary artery disease and heart failure. Lancet. 2014;383(9932):1933-1943. doi: 10.1016/S0140-6736(14)60107-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Aimo A, Gaggin HK, Barison A, Emdin M, Januzzi JL Jr. Imaging, biomarker, and clinical predictors of cardiac remodeling in heart failure with reduced ejection fraction. JACC Heart Fail. 2019;7(9):782-794. doi: 10.1016/j.jchf.2019.06.004 [DOI] [PubMed] [Google Scholar]
- 37.Heusch G. Myocardial ischemia: lack of coronary blood flow, myocardial oxygen supply-demand imbalance, or what? Am J Physiol Heart Circ Physiol. 2019;316(6):H1439-H1446. doi: 10.1152/ajpheart.00139.2019 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Hilberath JN, Oakes DA, Shernan SK, Bulwer BE, D’Ambra MN, Eltzschig HK. Safety of transesophageal echocardiography. J Am Soc Echocardiogr. 2010;23(11):1115-1127. doi: 10.1016/j.echo.2010.08.013 [DOI] [PubMed] [Google Scholar]
- 39.Kallmeyer IJ, Collard CD, Fox JA, Body SC, Shernan SK. The safety of intraoperative transesophageal echocardiography: a case series of 7200 cardiac surgical patients. Anesth Analg. 2001;92(5):1126-1130. doi: 10.1097/00000539-200105000-00009 [DOI] [PubMed] [Google Scholar]
- 40.Freitas-Ferraz AB, Bernier M, Vaillancourt R, et al. Safety of transesophageal echocardiography to guide structural cardiac interventions. J Am Coll Cardiol. 2020;75(25):3164-3173. doi: 10.1016/j.jacc.2020.04.069 [DOI] [PubMed] [Google Scholar]
- 41.Kumar S, Brown G, Sutherland F, et al. The transesophageal echo probe may contribute to esophageal injury after catheter ablation for paroxysmal atrial fibrillation under general anesthesia: a preliminary observation. J Cardiovasc Electrophysiol. 2015;26(2):119-126. doi: 10.1111/jce.12575 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
eAppendix 1. List of Matching Covariates
eAppendix 2. Descriptive Analysis
eFigure 1. Density Plot of the Proportion of CABG Surgeries Using TEE by Hospital Surgical Volume
eTable 1. Covariate Distribution Among Training and Testing Data, Stratified by Receipt of TEE
eAppendix 3. Details on Statistical Matching Methodology
eAppendix 4. Methods for Identification of Subgroups That Benefit From TEE (Stage One)
eAppendix 5. Methods for Validating TEE Score (Stage Two)
eAppendix 6. Preliminary Results
eTable 2. 30-Day Mortality in the Training Data Set Stratified by Hospital Volume and TEE Receipt
eAppendix 7. Low Surgical Volume Hospitals
eTable 3. Low Volume Covariate Distribution Among Post-Match Testing Data
eTable 4. Low Volume Hospitals Training Data Selected Covariate Distribution by TEE Score
eAppendix 8. Medium Surgical Volume Hospitals
eTable 5. Medium Volume Covariate Distribution Among Post-Match Testing Data
eTable 6. Medium Volume Hospitals Training Data Selected Covariate Distribution by TEE Score
eAppendix 9. High Surgical Volume Hospitals
eTable 7. High Surgical Volume Hospitals Covariate Distribution Among Post-Match Testing Data
eTable 8. High Surgical Volume Hospitals Training Data Selected Covariate Distribution by TEE Score
eAppendix 10. Testing Data Results
eTable 9. Characteristics Associated With Predicted Individualized TEE Treatment Effects at Low Surgical Volume Hospitals
eTable 10. Characteristics Associated With Predicted Individualized TEE Treatment Effects at Medium Surgical Volume Hospitals
eAppendix 11. Additional Outcome Analyses
eAppendix 12. Risk Score Calculator
eAppendix 13. Code Documentation
eReferences
eAppendix 14. Two-stage Target Trial Matched Analysis
Data Sharing Statement