Intraoperative transit-time flow measurement and high-frequency ultrasound in coronary artery bypass grafting: impact in off versus on-pump, arterial versus venous grafting and cardiac territory grafted

Abstract

 

Open in new tabDownload slide

OBJECTIVES

Despite society guideline recommendations, intraoperative high-frequency ultrasound (HFUS) and transit-time flow measurement (TTFM) use in coronary artery bypass grafting (CABG) has not been widely adopted worldwide. This retrospective review of the REQUEST (REgistry for QUality assESsmenT with Ultrasound Imaging and TTFM in Cardiac Bypass Surgery) study assesses the impact of protocolled high-frequency ultrasound/TTFM use in specific technical circumstances of CABG.

METHODS

Three REQUEST study sub-analyses were examined: (i) For off-pump (OPCAB) versus on-pump (ONCAB) procedures: strategy changes from preoperative plans for the aorta, conduits, coronary targets and graft revisions; and for all REQUEST patients, revision rates in: (ii) arterial versus venous grafts; and (iii) grafts to different cardiac territories.

RESULTS

Four hundred and two (39.6%) of 1016 patients undergoing elective isolated CABG for multivessel disease underwent OPCAB procedures. Compared to ONCAB, OPCAB patients experienced more strategy changes regarding the aorta [14.7% vs 3.4%; odds ratios (OR) = 4.03; confidence interval (CI) = 2.32–7.20], less regarding conduits (0.2% vs 2.8%; OR = 0.09; CI = 0.01–0.56), with no differences in coronary target changes or graft revisions (4.1% vs 3.5%; OR = 1.19; CI = 0.78–1.81). In all REQUEST patients, revisions were more common for arterial versus venous grafts (4.7% vs 2.4%; OR = 2.05; CI = 1.29–3.37), and inferior versus anterior (5.1% vs 2.9%; OR = 1.77; CI = 1.08–2.89) and lateral (5.1% vs 2.8%; OR = 1.83; CI = 1.04–3.27) territory grafts.

CONCLUSIONS

High-frequency ultrasound/TTFM use differentially impacts strategy changes and graft revision rates in different technical circumstances of CABG. Notably, patients undergoing OPCAB experienced 4 times more changes related to the ascending aorta than ONCAB patients. These findings may indicate where intraoperative assessment is most usefully applied.

Clinical trial registration number

ClinicalTrials.gov: NCT02385344

INTRODUCTION

Surgical revascularization remains standard of care for complex multivessel coronary artery disease [1, 2]. Innovations in surgical techniques and perioperative care have resulted in excellent contemporary short- and long-term outcomes for coronary artery bypass grafting (CABG) [3]. However, targets remain for ongoing improvement in CABG [4]. For instance, perioperative graft failure occurs in up to 7% of grafts and 9% of patients and is associated with significant morbidity and mortality [5, 6]. Technical failure can easily go unrecognized without concerted interrogation of completed grafts, as patients often show no intraoperative haemodynamic, echocardiographic or electrocardiographic changes [6]. Additionally, while multiple factors contribute to perioperative stroke, one preventable technical aetiology is inadvertent cerebrovascular embolization during aortic manipulation [7].

Several techniques have been proposed to improve CABG outcomes [8]. Transit-time flow measurement (TTFM) provides quantitative intraoperative assessment of graft flow with values demonstrated to correlate with short- and mid-term graft patency and allows identification and correction of faulty grafts before closing the chest [6, 9]. High-frequency ultrasound (HFUS) provides morphologic detail and colour-flow assessment of completed anastomoses and can improve diagnostic accuracy of truly patent or failing grafts, reaching a positive predictive value of nearly 100% when added to TTFM [10]. Moreover, epiaortic HFUS has been associated with decreased stroke rates versus procedures without epiaortic scanning, as demonstrated in a recent meta-analysis (0.6% vs 1.9%) [11]. Accordingly, the European Society of Cardiology/European Association for Cardio-Thoracic Surgery 2018 Guidelines on Myocardial Revascularization provided class IIa, level B recommendations for intraoperative assessment of graft flow and class IIa, level C recommendations for epiaortic ultrasound scanning prior to aortic manipulation during CABG [1].

Despite these recommendations, worldwide adoption rates of these techniques remain low. Understanding technical circumstances in CABG where HFUS/TTFM scanning produces the greatest impact on operative course might guide efficient adoption of these techniques. Most preceding studies of intraoperative assessment with HFUS/TTFM came from single centres, providing limited generalizability. The REQUEST (REgistry for QUality assESsmenT with Ultrasound Imaging and TTFM in Cardiac Bypass Surgery) study was the first prospective multi-site study that examined the impact of protocolled intraoperative HFUS/TTFM assessment on strategy changes to preoperative surgical plans [12]. This study reviews REQUEST data to assess effects of protocolled HFUS/TTFM use in specific CABG techniques: on-pump (ONCAB) versus off-pump (OPCAB) CABG, arterial versus venous conduits and grafting to different cardiac territories.

MATERIALS AND METHODS

Ethical statement

This study is a post-hoc sub-analysis of data from the previously published REQUEST study [12]. Institutional review boards at participating centres approved the trial (Washington DC VA Medical Center IRB#01731; initial approval date 14 April 2015). Enrolled participants provided informed consent. The study was conducted in accordance with the Declaration of Helsinki. The REQUEST study was funded by Medistim ASA (Oslo, Norway). Principal investigators/authors had complete scientific freedom.

REQUEST study

The REQUEST study was an international, multicentre, prospective observational registry designed to capture information on changes to preoperatively proposed surgical plans during CABG based on protocolled intraoperative assessment with HFUS and TTFM using [MiraQ™] or [VeriQ C™] devices (Medistim ASA). Patients undergoing isolated CABG for ≥2-vessel coronary artery disease were enrolled. Exclusion criteria included emergent cases, concomitant surgical procedures, medical history of muscle disorders or psychological, developmental or emotional disorders.

All participating surgeons (n = 36) were experienced with HFUS/TTFM having used HFUS/TTFM in >20 CABG procedures. Surgeons and study coordinators were trained to use and interpret HFUS/TTFM results according to a structured study protocol to minimize bias in inter-operator experience and interpretation. Preoperatively, surgeons formulated surgical plans using coronary angiography and other imaging at the surgeon’s discretion (e.g. computed tomography), which included aortic cannulation, cross-clamp and proximal anastomosis sites, bypass conduits, number of anastomoses and coronary targets. Plans were later compared with operative conduct to determine occurrence of surgical changes. Protocol steps included HFUS assessment of: (i) the ascending aorta; (ii) in situ conduit arteries; (iii) location of optimal sites for anastomoses; and (iv) anatomy/flow in completed anastomoses/grafts. TTFM assessment was recommended at a mean arterial pressure of 80 mm mercury. Parameters prompting re-evaluation of completed grafts for possible revision included: (i) low mean graft flow: arterial grafts <15 ml/min and venous grafts <20 ml/min; (ii) increased pulsatility index (difference of maximum and minimum flow, divided by mean flow across 5 cardiac cycles) >5; (iii) decreased diastolic filling (<70% for left-sided and <50% for right-sided coronary vessels). Changes in plan and/or revisions were performed at the surgeon’s discretion. Adherence to HFUS/TTFM assessment protocols and post-revision reassessment was encouraged, but not mandatory. Surgeons provided details regarding surgical changes performed, including location and reason.

Study population

Between April 2015 and December 2017, 1046 patients undergoing isolated CABG for multivessel coronary artery disease at 7 centres in Europe and North America were enrolled in the REQUEST study. Thirty patients were excluded due to screening failure (n = 8), lack of training of all surgical team members according to REQUEST study protocol (n = 11) or unavailability of HFUS/TTFM images for analysis (n = 11). All 1016 patients meeting inclusion criteria for the REQUEST study were included in the current study.

Strategy changes-definitions

Strategy changes were defined as any alteration from the preoperative plan. These changes could be based primarily on HFUS/TTFM use or visual/tactile inspection. Changes related to the aorta included changes to the cannulation site, cross-clamp site or site for proximal anastomosis. Changes regarding in situ conduits occurred if an alternative conduit was used, either due to inadvertent injury to the intended conduit or insufficient conduit calibre or length. Changes to coronary targets included different locations of anastomoses due to calcification or insufficient calibre, detection of intramural vessels or need for endarterectomy. Changes to completed grafts were defined as primary anastomotic revision (i.e. revision of proximal and/or distal anastomosis due to technical problems), secondary anastomotic revision (revision of proximal or distal anastomoses due to graft kinking or inadequate length, but not due to problems with the anastomosis itself), primary conduit revision (without revision of either anastomosis) or need for additional grafts (Supplementary Material, Table S1).

Variables and outcome measures

Patients undergoing OPCAB versus ONCAB procedures were compared on an as-treated basis by the following measures: (i) Preoperative characteristics; (ii) Operative variables; (iii) Frequency of HFUS and/or TTFM use at each site of interrogation; (iv) Qualitative and quantitative data regarding the number and reason for changes in surgical strategy related to all areas assessed with HFUS and TTFM (the primary outcome of interest), including graft revision rates on both a per-patient and per-conduit basis; and (v) Rates of in-hospital major adverse cardiovascular and cerebrovascular events [composite of all-cause death, stroke or transient ischaemic attack, myocardial infarction or repeat revascularization] and other major perioperative morbidities.

For the whole REQUEST cohort, changes made to arterial and venous conduits were compared on a per-conduit basis. Revision rates of all completed grafts were compared by location of the distal anastomosis in either anterior, lateral or inferior cardiac territories.

Statistical analysis

Categorical data were presented as proportions. Continuous data were reported as either mean ± standard deviation for parametric variables or median (interquartile range) for non-parametric variables. Normality was assessed using Kolmogorov–Smirnov test. Preoperative demographic and clinical variables, procedure variables, incidence of surgical changes and in-hospital clinical outcomes were compared between OPCAB and ONCAB cohorts. Incidence of surgical changes was presented as odds ratios (OR) with 95% confidence intervals (CI) from corresponding univariable logistic regression models. All statistical analyses were performed using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA).

RESULTS

Off-pump versus on-pump

Patient demographics and comorbidities

There were 402 (39.6%) OPCAB and 614 (60.4%) ONCAB patients. Two patients were changed from ONCAB to OPCAB due to detection of plaque at the proposed aortic cannulation or cross-clamp site, while 7 patients were changed from OPCAB to ONCAB due to extent of coronary disease discovered intraoperatively. Demographic and preoperative clinical characteristics of OPCAB and ONCAB patients can be found in Table 1.

Operative variables

All procedures were performed via median sternotomy. Operative times were similar between ONCAB and OPCAB operations; however, Table 2 shows objective differences in operative characteristics between the procedures. Compared to ONCAB, OPCAB procedures had higher left internal thoracic artery use (98.3% vs 95.8%), higher radial artery use (37.8% vs 12.7%) but lower use of bilateral internal thoracic artery grafts (25.1% vs 34.0%) and sequential grafting (3.3% vs 11.5%). OPCAB procedures had higher rates of multi-arterial grafting (50.0% vs 38.8%), complete arterial grafting (31.8% vs 22.3%) and Y/T graft use (18.9% vs 12.2%). There were more arterial (62.7% vs 55.4%) and arterio-venous (0.8% vs 0.1%) conduits and arterial distal anastomoses (1.7 ± 0.9 per-patient vs 1.6 ± 1.0 per-patient) in OPCAB procedures.

Frequency of high-frequency ultrasound/transit-time flow measurement use

HFUS scanning was used more frequently for aorta evaluation in OPCAB versus ONCAB procedures (88.3% vs 73.5%). Surgeons used HFUS to interrogate coronary targets (36.6% vs 54.7%) and completed grafts (46.0% vs 67.9%) less in OPCAB versus ONCAB procedures. There was no difference in HFUS use for evaluation of in situ conduits or TTFM use to review completed grafts between groups (Supplementary Material, Table S2).

Strategy changes

Surgical strategy changes in OPCAB and ONCAB procedures and information regarding the sites and reasons for changes are reported in Table 3. There were more surgical changes to the ascending aorta in OPCAB versus ONCAB procedures (14.7% vs 3.4%; OR = 4.03; 95% CI = 2.32–7.20). The primary determinant of this difference was HFUS-related changes, accounting for 98.3% of surgical changes in OPCAB compared to 76.2% in ONCAB. Changes to proposed aortic manipulation in OPCAB procedures were most frequently due to identifying an alternative proximal anastomosis site (81.4%), which accounted for only 33.3% of aortic plan changes in ONCAB procedures.

Strategy changes for in situ conduits occurred less in OPCAB versus ONCAB (0.2% vs 2.8%; OR = 0.09; CI = 0.01–0.56). ONCAB group changes were primarily driven by HFUS findings (58.8%), with 23.5% of changes due to conduit injury/dissection and 23.5% of changes after identifying a conduit vessel of insufficient calibre.

Changes to coronary targets did not differ between OPCAB/ONCAB groups (10.4% vs 10.9%; OR = 0.95; CI = 0.62–1.46). However, when comparing only patients that underwent HFUS scanning, there were more target location changes in OPCAB versus ONCAB (28.6% vs 19.9%; OR = 1.63; CI = 1.02–2.62). HFUS accounted for more changes in ONCAB (71.6%) versus OPCAB (59.5%). The predominant anatomic factor inciting change in both groups was identification of calcified vessels (45.2% OPCAB, 47.8% ONCAB).

Completed graft revisions did not differ between OPCAB/ONCAB groups, on either a per-patient (8.5% vs 7.3%; OR = 1.17; CI = 0.71–1.90) or per-graft (4.1% vs 3.5%; OR = 1.19; CI = 0.78–1.81) basis. Most changes were due to HFUS and/or TTFM use in both OPCAB (73.5% of patients, 75.0% of grafts) and ONCAB (57.8% of patients, 51.8% of grafts). The distribution of change types was comparable between groups, with primary anastomotic revisions accounting for the most changes in both (OPCAB: 64.7% of patients, 59.1% of grafts; ONCAB: 44.4% of patients, 39.3% of grafts). Repeat measurement after completed graft revisions was performed in 93.9% and 100% of OPCAB and ONCAB procedures, respectively, with improved HFUS/TTFM parameters achieved in 87.1% of OPCAB grafts and 75.9% of ONCAB grafts (Supplementary Material, Table S3).

In-hospital adverse events

Complication rates in both groups were low, with no group differences in overall major adverse cardiovascular and cerebrovascular events (1.2% OPCAB vs 2.4% ONCAB), mortality (0.7% OPCAB vs 0.5% ONCAB), stroke/transient ischaemic attack (0.5% OPCAB vs 1.3% ONCAB), myocardial infarction (0% OPCAB vs 0.5% ONCAB), repeat coronary revascularization (0% OPCAB vs 0.2% ONCAB), requirement for new haemodialysis (0% OPCAB vs 0.5% ONCAB) or unplanned cardiothoracic reoperations (0.7% OPCAB vs 1.0% ONCAB) between OPCAB and ONCAB groups (Supplementary Material, Fig. S1). No deaths occurred in patients that underwent graft revision.

Arterial versus venous grafting

There were 1570 arterial conduits (58.7%) and 1105 venous conduits (41.3%) among the 1016 REQUEST patients (Fig. 1). Revisions were more common for arterial versus venous grafts (4.7% vs 2.4%; OR = 2.05; CI = 1.29–3.37). HFUS/TTFM evaluation accounted for the most changes in both conduit types (Arterial: 60.8%; Venous: 65.4%). The most common revision performed for both conduit types was primary anastomotic revision (arterial: 50.0%; venous: 42.3%). Secondary anastomotic revision was the next most frequent revision for arterial conduits (33.8%), while primary conduit revisions were the second most common revision in venous conduits (26.9%). Of the arterial grafts, graft revisions occurred in 16/252 (6.3%) radial artery grafts and 58/1318 (4.4%) internal thoracic artery grafts.

Cardiac territories

Of 2959 total distal anastomoses, there were 1467 (49.6%) to anterior, 843 (28.5%) to lateral and 649 (21.9%) to inferior cardiac territories. Inferior territory grafts were revised more than anterior (5.1% vs 2.9%; OR = 1.77; CI = 1.08–2.89) and lateral (5.1% vs 2.8%; OR = 1.83; CI = 1.04–3.27) grafts. There was no difference between anterior and lateral territory graft revisions. HFUS/TTFM evaluation accounted for the most changes in all territories (Anterior: 67.4%; Lateral: 50.0%; and Inferior: 63.6%). The distribution of types of changes was comparable across all territories (Fig. 2).

DISCUSSION

This REQUEST study sub-analysis found that CABG with protocolled intraoperative HFUS and TTFM use saw different patterns of surgical strategy changes and graft revisions associated with key technical aspects of CABG procedures. Patients undergoing OPCAB experienced 4 times more changes related to the ascending aorta than ONCAB patients. With respect to conduits, arterial grafts were revised twice as frequently as venous grafts. Finally, grafts to inferior cardiac territory were revised 1.8 times more often than grafts to anterior and lateral territories.

Surgical changes pertaining to the aorta occurred in 14.7% of OPCAB patients (81.4% of changes involved changing the site of proximal anastomosis) versus 3.4% of ONCAB patients. HFUS scanning accounted directly for 98.3% of these changes in OPCAB patients after identification of plaques not initially detected by visual or tactile inspection. The increased rate of lesion detection in OPCAB versus ONCAB procedures is likely influenced by increased rates of HFUS use in this cohort (88.3% vs 73.5%). An additional contributing factor is patient selection for OPCAB procedures (e.g. diabetics comprised 44% of OPCAB patients versus 37% ONCAB). It is possible that patients perceived as high risk for stroke, with preoperative imaging assessment showing a heavily atherosclerotic ascending aorta, may have been chosen for OPCAB given the known benefits of this approach for stroke risk reduction (i.e. fewer cannulation sites, potential for clampless techniques) [13]. HFUS might offer additional benefit in this population, as evidenced in a recent meta-analysis associating CABG strategy including epiaortic HFUS with decreased stroke rates versus procedures without epiaortic scanning [11]. Indeed, the 0.5% rate of perioperative stroke observed in our OPCAB cohort compares favourably to the 1.1% rate calculated for all CABG patients in a recent pooled analysis of 11 randomized trials [14]. Coupled with existing literature showing benefits of these 2 techniques, the low stroke rate in the OPCAB group (especially considering its high proportion of diabetics) suggests OPCAB with epiaortic HFUS scanning may help prevent strokes in appropriately selected patients and warrants further investigation.

This study also found that arterial conduits had more graft revisions than venous conduits. An abundance of evidence has shown benefits of multi-arterial grafting towards clinical outcomes and graft patency with revascularization for patients with multivessel disease [15, 16]. However, anecdotally, arterial conduits present greater technical challenges than saphenous vein grafts [17]. In addition to more difficult anastomotic construction, arterial conduits’ tendency towards vasospasm and susceptibility to competitive flow could make assessment of graft anastomoses challenging to interpret with TTFM measurement alone [18, 19]. Combining HFUS imaging and TTFM may then have a valuable role in the assessment of questionable arterial grafts. In addition to avoiding unnecessary graft revisions, combined HFUS/TTFM assessment might also prevent abandoning salvageable arterial grafts. Indeed, in this study, only 3 arterial graft revisions (4.1%) entailed replacement by, or addition of saphenous venous grafts.

Graft revisions to the inferior cardiac territory (5.1%) were 1.8 times more frequent than those to anterior and lateral territories. This observation implies increased difficulty grafting this region, potentially due to anatomic factors, such as a different atheromatous disease profile and calcification distribution in the right coronary artery circulation [20, 21]. Furthermore, right coronary artery territory grafts have been shown to have the highest rate of disease progression after revascularization [22, 23]. It is unclear whether this observation stems from the predominant use of venous grafts for this territory, or the greater tendency towards incomplete revascularization of the inferior territory [22]. While there is a steep learning curve using TTFM and HFUS for this region, and especially in OPCAB, even among experienced REQUEST study surgeons 5.1% of inferior territory grafts required revision, primarily due to technical problems with proximal or distal anastomoses (48.5% of revisions). 81.0% of inferior territory grafts revised with abnormal HFUS/TTFM findings showed improved objective parameters on TTFM reassessment after revision, suggesting these revisions likely reduced the risk of inadequate revascularization. Moreover, no operative mortalities occurred in patients undergoing graft revisions, suggesting that when a defect is intraoperatively confirmed, the fear of ‘doing more harm than good’ should not dissuade necessary revision.

This study has some limitations. The REQUEST study is not a randomized controlled trial, so comparisons of CABG procedures with and without use of HFUS/TTFM were not made. Also, as follow-up was limited to in-hospital events, it is unknown if differential rates of surgical changes and graft revisions impact long-term results of CABG with respect to clinical outcomes and graft patency [12, 24, 25]. Additionally, the REQUEST study was performed at 7 well-established centres with experienced CABG surgeons, with rates of OPCAB, bilateral internal thoracic artery use and complete arterial revascularization that exceed those reported in current literature [26, 27]. Therefore, generalizability to all other centres is uncertain.

CONCLUSION

In the prospective REQUEST study, protocolled HFUS/TTFM use in CABG surgery resulted in the most strategy changes towards handling of the aorta in OPCAB procedures. Revisions were more common for arterial versus venous grafts, and for inferior versus anterior and lateral territory grafts. Graft revision resulted in improved TTFM parameters in 82% of cases, with no added major adverse cardiovascular and cerebrovascular events in those patients whose graft was revised. Understanding these specific technical circumstances in CABG in which intraoperative assessment with HFUS and/or TTFM provides the greatest impact on operative course can help optimize utilization of these techniques. Additional, larger studies over longer follow-up periods are warranted to further characterize benefits of HFUS/TTFM use, particularly with respect to graft patency and long-term outcomes.

SUPPLEMENTARY MATERIAL

Supplementary material is available at EJCTS online.

Presented at the 5th International Coronary Congress, New York, NY, USA, 6–8 December 2019.

ACKNOWLEDGEMENTS

Thanks to Dr A. Pieter Kappetein and Dr Stuart J. Head for leading the Erasmus University Medical Centre study site and to clinical coordinators at participating study sites.

Funding

Work funded by Medistim ASA, Oslo, Norway.

Conflict of interest: Ethan S. Rosenfeld, Michael A. Napolitano and Andrew D. Sparks declare no conflicts. Gregory D. Trachiotis, Daniel Wendt, Teresa M. Kieser, John D. Puskas, and Gabriele DiGiammarco received travelling support and/or speaking fees from Medistim. David P. Taggart received research funding, speaking and travelling honoraria from and acts as an adviser to Medistim.

Author contributions

Ethan S. Rosenfeld: Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Writing—original draft; Writing—review & editing. Gregory D. Trachiotis: Conceptualization; Investigation; Supervision; Validation; Writing—review & editing. Michael Andrew Napolitano: Data curation; Validation; Visualization; Writing—review & editing. Andrew D. Sparks: Data curation; Formal analysis; Methodology; Software; Writing—original draft. Daniel Wendt: Conceptualization; Investigation; Validation; Writing—review & editing. Teresa M. Kieser: Conceptualization; Investigation; Validation; Writing—review & editing. John D. Puskas: Conceptualization; Investigation; Validation; Writing—review & editing. Gabriele DiGiammarco: Conceptualization; Investigation; Validation; Writing—review & editing. David P. Taggart: Conceptualization; Investigation; Methodology; Supervision; Validation; Writing—review & editing.

Reviewer information

European Journal of Cardio-Thoracic Surgery thanks Ikuo Fukuda, Jose Lopez-Menendez, Ardawan J. Rastan and the other anonymous reviewer(s) for their contribution to the peer review process of this article.

REFERENCES

ABBREVIATIONS

 
• CABG

  Coronary artery bypass grafting

 
• CI

  Confidence interval

 
• HFUS

  High-frequency ultrasound

 
• ONCAB

  On-pump coronary artery bypass grafting

 
• OPCAB

  Off-pump coronary artery bypass grafting

 
• OR

  Odds ratio

 
• REQUEST

  REgistry for QUality assessment with Ultrasound Imaging and TTFM in Cardiac Bypass Surgery

 
• TTFM

  Transit-time flow measurement