Structured Vs Conventional Endoscopic Vein Harvesting Training: Vein Integrity and Learning Curve Performance

Techniques and intervention

Four Surgical Care Practitioners had comparable prior experience (>250 OVH cases) but no EVH exposure. Variability in baseline knowledge and performance was observed and equivalence between trainees cannot be assumed. All CABG procedures were performed by consultant cardiac surgeons with varying levels of experience. Emergency and off-pump CABG surgery patients were excluded. Trainees with <12 months post-qualification experience, prior EVH experience, lack of smartphone access or unwillingness to participate in mobile games were excluded.

Structured training: (ST - MELT group)

Two trainees (T1 & T2) were recruited eight weeks before EVH training and completed four-step MELT training.⁷

Conventional training: (CT group)

Two trainees (T3 & T4) underwent conventional EVH training according to the Getinge™ company training protocol. To minimise bias, the company trainer was not provided with the study protocol. According to principal investigator records, both trainees completed the company e-learning platform approximately three weeks before EVH training and undertook limited leg-model simulation to familiarise themselves with EVH equipment. Trainees then progressed to supervised clinical cases with a company proctor present for an initial 2-week continuous training period, after which independent practice was expected.

Both groups

All procedures used the Maquet VasoView Hemopro 2 system with intraoperative heparinisation.^3,11 In the ST trainee group, progression to independent harvesting required ≥80% competency as assessed by the trainer using MELT. In the CT trainee group, progression was determined by the company trainer according to company protocol without a formalised competency threshold.

Vein quality was assessed using structured scoring protocols.³ CD34 immunohistochemistry quantified endothelial integrity (0–100%).^4,12 Smooth muscle damage (Masson’s Trichrome) and endothelial stretching/detachment (H&E) were graded (0–4 scale). Sections were independently scored by three blinded assessors and validated by a consultant histopathologist. Formal inter- and intra-observer variability analysis was not performed.

Trainee outcomes were evaluated using the Kirkpatrick framework¹³ (reaction, learning, behaviour, results). Technical performance was assessed using the Global Operative Assessment of Laparoscopic Skills (GOALS; 25-point validated scale),¹⁴ and anxiety was measured using the Beck Anxiety Inventory¹⁵ (BAI). Assessments were conducted pre-, during and post-training. “During training” refers to the supervised phase of EVH skill acquisition. “Post training” refers to performance following progression to independent practice or near-independent practice. Outcome assessors and data analyst were blinded.

Study endpoints

The primary outcome was histological conduit damage. Additional endpoints included the incidence of MACE and quality of life measured using EQ-5D-5L. MACE comprised repeat angina, dyspnoea, myocardial ischemia/infarction, reintervention, stroke and death and was determined through telephone interviews, outpatient records, general practitioner letters and coroner reports at 3-month intervals up to 12 months postoperatively.

Statistical analysis

Due to the absence of patient-related samples from trainees at STH hospital, the analysis was mainly descriptive and no comparative analysis between ST and CT trainees was performed. Data were summarised as frequencies/percentages for categorical variables and as median and interquartile range [IQR] for continuous variables. Endothelial integrity (CD34 expression) is presented as median percentage while other histological outcomes are presented as median scores.

Demographics

Four trainees and 57 patients provided written informed consent. Eleven patients (19%) were excluded due to surgery cancellation and performed by different operators ( Figure 1). EVH was performed by four trainees in 46 patients: 40 at MFT (ST group, T1 & T2) and six patients at STH hospital (CT group, T3 & T4). ( Table 1). Trainees within each group underwent the same training pathway within their respective centres. The observed differences between individual trainees likely reflect variation in learning and case exposure rather than systematic differences. ST trainees continued EVH practice (>400 cases each) whereas CT trainees discontinued EVH practice in June 2024. Histological evaluation included 138 veins (406 analysable sections after 8 exclusions) with proximal (H1), random/distended (H2) and distal (H3) sections ( Figure 1).

Figure 1. Consort patient and trainee flow diagram.

Endothelial integrity: CD34

Across all sections (H1-H3), the ST group demonstrated consistently high median endothelial integrity, generally above 90% in proximal (H1) and distal (H3) and slightly lower values in random sections (H2). Median endothelial integrity was lower in the CT group. ( Figure 2a; Figure 3). Table 2 provides detailed median scores for CD34, H&E and MT staining.

Figure 2. a: illustration of the median percentage of endothelial integrity and stretching on proximal, random and distal vein samples.

Figure 2 b: Illustration of the median percentage of muscular hypertrophy of circular and longitudinal muscle on proximal, random and distal vein samples.

Figure 3. Illustrates the CD34 staining of H1, H2, H3 sections.

a) represents full intact endothelium to minimal disruption indicated with black arrows; b) represents more than 50% loss of endothelium; c) represents 90 to 100% loss of endothelial layer with muscle migration towards lumen.

Endothelial stretching: H&E

Endothelial stretching was low across all sections in the ST group (median ≤ 1.1). In contrast, the CT group demonstrated numerically higher stretching values, particularly in distal sections (H3, median ≥ 1.2). ( Figure 4).

Figure 4. Illustrates the H&E staining of proximal (H1), random (H2) and Distal (H3).

All images are magnified from 500 to 800 μm. a) represents normal H1, H2, H3; b) represents mild to moderate muscle migration and hypertrophy towards the lumen; c) represents moderate to severe detachment, migration and hypertrophy towards the lumen. In H3 image (b) black arrow represents vein valves.

Muscle hypertrophy: masson trichrome

Muscle hypertrophy was consistently low in the ST group across all vein sections ( Figure 5) indicating preserved morphology. Longitudinal hypertrophy in the ST group was mild (median 0.8–2.0) with lower scores in T2. Circular hypertrophy was minimal in the ST group across all sections (median values predominantly 0.0). In contrast, CT group demonstrated greater variability and higher circular hypertrophy, particularly in distal samples (median 0.3–0.8) ( Figure 2b).

Figure 5. Illustrates the Masson Trichrome staining of H1, H2, H3 sections.

All images are magnified from 500 and 400 μm. a) represents normal to mild adventitial layer detachment indicated with black arrows; b) represents moderate to severe endothelial detachment and muscle hypertrophy; c) represents severe migration towards the lumen with detachment.

Trainee outcomes: pre EVH training tests

Conventional Training:

Baseline Single Best Answer (SBA) of LSV anatomy & physiology (8 questions) and EVH procedural knowledge (2 questions) demonstrated low pre-training scores in CT group (T3: 30%; T4: 20%), with most incorrect responses attributed to knowledge recall deficits. Post-training SBA scores showed modest improvement (T3: 40%; T4: 45%). CT trainees subsequently discontinued EVH practice in June 2024.

Structured training:

In ST group, SBA pre-training assessments of LSV anatomy & physiology (8 questions) and EVH procedural knowledge (2 questions) yielded scores of 36% (T1) and 20% (T2). Trainees then completed an online e-learning covering CABG, vessel preservation, EVH procedure and molecular mechanisms followed by post-training examination achieving scores of 80% (T1) and 63% (T2).

During step 2, trainees played Temple Run 2™ (Imangi studios, North Carolina, USA) for a minimum of five sessions daily over 14 days. Mean (±SD) performance metrics were: T1 (minutes played 25.42 ± 6.2, sessions/day 8.3 ± 3.4, highest daily score 336,992 ± 214,517) and T2 (minutes played 25.79 ± 7.5, sessions/day 9.6 ± 3.9, highest daily score 236,568 ± 108,578). Trainees were encouraged to continue gaming thereafter to reinforce hand-eye coordination and cognitive responsiveness.⁷

Additional outcomes: trainee

Median GOALS scores improved for all trainees from training to post-training, indicating increased technical competence. During training, median scores ranged from 15.0–21.0, rising post-training to 23.0–25.0 across all trainees. Median vein quality scores were generally good across trainees (median 0–2), improving slightly post-training (1–2). Median BAI scores were low across all trainees during and after EVH training, indicating minimal self-reported anxiety ( Table 3).

Kirkpatrick evaluation

Confidence scores across Kirkpatrick domains were high for anatomical knowledge and procedural step-up median [IQR] 5.0 [5.0, 5.0] during and post training. For advanced tasks (dissection, ligation, coagulation and management of complications) confidence increased from median 3.0 to 5.0 during training to 4.0–5.0 post-training across all trainees. Confidence in recognising when to see assistance was uniformly high ( Table 3).

Patient outcomes

Of 46 patients, 40 (87%) were male. Median age was 62.0–68.5 years across all groups.

Respiratory parameters

Overall, across four trainees (46 patients) median PCO₂ increased from baseline to 60 minutes, accompanied by increases in ETCO₂ at 60 minutes in all three trainees. Median pH remained stable in T1 and decreased from 7.4 at baseline to 7.3 at 60 minutes in the rest of the trainees. Minute volume increased over time in T1 and T4, remained stable in T2 and decreased at 60 minutes in T4 ( Table 4).

Wound outcomes

Wound complications were assessed at Days 1, 3, and 5 and sensory symptoms at six weeks. Bruising was more frequent in T2 than T1 at all early time points (Day 1: 45.0% vs 15.0%; Day 3: 45.0% vs 30.0%; Day 5: 45.0% vs 35.0%). All patients in T3 (n = 1) and T4 (n = 5) exhibited bruising. Hematoma was uncommon overall, observed only in isolated cases in T1 and T2 (5.0% at any time point). At six weeks, leg numbness was reported in 40.0% of T1 and 25.0% of T2 patients and in all patients in T3 and T4. Leg tenderness was more common in T1 (65.0%) compared with T2 (25.0%) and was reported in all T3 patients and 60.0% of T4 patients ( Table 3).

Major adverse cardiac events

MACE was assessed up to 12 months and survival to 24 months. Repeat angina was infrequent occurring in 5.0% of T1 patients at 3 and 6 months and in 5.0% of T2 patients at 12 months. No cases were reported in T3 or T4. Re-intervention was reported in T1 at 1 and 9 months (each 5.0%) and in T2 at 1 month (5.0%) with none in T3 or T4. Myocardial infarction occurred only in T1 (5.0% at 9 months). Survival in T1 was 100.0% to 9 months and 95.0% at 12 and 24 months. T2 and T4 demonstrated 100.0% survival throughout follow-up while T3 survival was 100.0% at 1 month.

CT angiography

CT angiography was scheduled at 12–24 months in which ST group (21/39 patients) and CT group (3/6) underwent imaging. One patient died at 11 months due to subdural haematoma and 18 did not attend due to work commitments. In the CT group, one did not attend following stroke and two were unable to undergo imaging due to lack of radiology support. All scans were independently reviewed by a cardiologist and a surgeon. One symptomatic patient underwent coronary angiogram at 6 months, all grafts patent.

Among asymptomatic patients, 13 vein grafts (T1), 23 (T2) and 2 (T4) were assessed. Flow-limiting lesions identified in 1/13 grafts (T1) and 4/23 (T2). Occlusion occurred in 2/13 (T1) and 8/23 (T2) and the remaining grafts were patent. Affected territories included Diagonal (D1) (2/13 T1; 5/23 T2), Obtuse Marginal (OM2) (1/13 T1; 3/23 T2), and Right Coronary Artery (RCA/PDA) (0/13 T1; 4/23 T2). All lesions in T1 and T2 were classified as ACC/AHA Type C and small-calibre diabetic vessels were reported in 7 (T1) and 14 (T2) ( Table 5).

Patient reported outcomes

Wound satisfaction was high (median 98–100). EQ-5D-5L domains improved post-operatively, with most patients reporting “no problem” across mobility, self-care, usual activities, pain, and anxiety ( Table 6).

General health visual analogue scores improved in all groups (T1: 90.0 vs 62.5 pre-operatively; T2: 85.0 vs 57.5; T4: 75.0 vs 48.0). SF-36 domains showed consistent favourable with most patients reporting their health as “much better” or “somewhat better” than one year earlier and improvements observed across functional capacity, pain, emotional wellbeing, social functioning and energy levels ( Table 7).

Limitations

This study has several important limitations. Early termination due to the COVID-19 pandemic substantially reduced the planned sample size and small number of trainees and absence of randomisation mean that differences may reflect individual trainee ability or centre-level factors rather than training method alone. Observer variability was not formally quantified. Most of the outcomes reflect those prespecified in the original CREST trial protocol but are reported descriptively. Trainee exposure differed markedly between groups (case volume and continuity), introducing potential performance and experience bias. A predefined competency threshold (≥80%) guided progression in the structured training group. However, formal learning curve modelling and case-number thresholds were not assessed. Incomplete CT angiography follow-up and the observational cluster design without randomisation restrict causal inference. All CT angiography studies were ECG-gated, however functional assessment such as FFR-CT was not performed. Future studies incorporating both anatomical and functional imaging may provide a more comprehensive evaluation of graft performance. CTA interpretation may have been affected by artefacts from surgical clips and adjacent grafts, potentially limiting assessment accuracy in some cases. Additionally, self-reported anxiety (BAI) may underestimate training-related stress, as objective measures were not employed.

Patient and public involvement and engagement (PPIE)

PPIE was integral to all stages of this study. Patient representatives from the Ticker Club, Manchester Heart Charity, Wythenshawe Hospital were involved from the beginning and helped shape the study design, recruitment approach and participant-facing materials. They reviewed and provided feedback on the study protocol, NIHR postdoc application, participant information sheets, consent forms, data collection materials and ethics application documentation to improve readability, accessibility and relevance to patients undergoing Coronary Artery Bypass Graft (CABG) surgery. Throughout the study, patient voluntary contributors provided advice on study conduct, recruitment of CABG patients, interpretation of findings and dissemination activities. Study findings were communicated through patient and public networks at Manchester Foundation Trust, Ticker club charity newsletters, annual charity educational events, Internationally Society of Minimally Invasive Surgery (ISMICS), Association of Cardiothoracic Surgical Care Practitioner (ACTSCP) conferences to maximise accessibility and impact.