Objective: To evaluate evolving selection criteria and angiographic outcome ('learning curve') for directional coronary atherectomy.
Setting: Tertiary referral, university-based hospital.
Patients: Initial 50 subjects undergoing directional coronary atherectomy of de novo left anterior descending stenoses at The Toronto Hospital from July 1990 to April 1991.
Interventions: Directional coronary atherectomy according to standard interventional techniques, with pre- and post procedure qualitative evaluation and quantitative coronary arteriography (Cardiac Measurement System; Leiden, The Netherlands) to define angiographic outcome.
Results: Comparing 'early' (group 1) versus 'late' (group 2) subjects, baseline demographics and clinical parameters were similar. Later subjects demonstrated increased coronary tortuosity (group 1, 1.40 versus group 2, 1.64, P < 0.01) and major side branch involvement within the stenosis (group 1, seven of 25 [28%] versus group 2, 18 of 25 [72%], P < 0.01). Regardless of experience, post procedure residual minimum stenotic diameters were equal (group 1, 2.75 +/- 0.55 versus group 2, 2.49 +/- 0.42 mm) in progressively longer lesions (group 1, 11.4 +/- 4.9 versus group 2, 13.3 +/- 5.5 mm, P < 0.1), with increased symmetry (group 1, 0.60 +/- 0.28 versus group 2, 0.73 +/- 0.19, P < 0.05). Analysis of consecutive pentiles (10 subjects per group) indicated gradual increases in post procedure residual lumen during early experience (the first 30 subjects), with an abrupt deterioration in outcome (fourth pentile), secondary to qualitative changes in coronary anatomy, before a return to satisfactory residual minimum stenotic diameters (fifth pentile).
Conclusions: This study defines a distinct 'learning curve' during the initial 30 patients undergoing directional coronary atherectomy, with subtle changes in case selection, predominantly reflected by qualitative indices (eg, tortuosity, dystrophic calcification), resulting in a transient deterioration in final outcomes (patient 31 to 40). Subsequently, optimal results were re-established after defining appropriate case selection criteria, in conjunction with progressive expertise.