Comparison of Multiple Drug-Eluting Stent Percutaneous
Coronary Intervention and Surgical Revascularization in Patients
with M

Since the introduction of drug-eluting stents (DES) in clinical practice, it has become clear that percutaneous revascularization of multivessel coronary artery disease (CAD) could be more effective in terms of event reduction (major adverse cardiac events [MACE] and target vessel revascularization [TVR]) and become comparable to surgery. Results of randomized, controlled trials of drug-eluting stents (DES) versus bare-metal stents (BMS) in patients with progressively more complex lesions^1–4 have confirmed the very good clinical and angiographic results of the first studies^5,6 and reinforced this thinking. At the same time, the possibility of conducting randomized clinical trials between percutaneous coronary intervention (PCI) and coronary artery bypass graft surgery (CABG) has become very difficult. All published studies report a low enrollment rate versus the total eligible population.⁷ Real-world practice involves treating increasingly complex lesions and patients with PCI and referring to surgery only those patients who are not suitable for the percutaneous option. Even the ARTS II trial⁸ is not a randomized trial because it compares DES-treated patients with historical cohorts of CABG- and BMS-treated patients.
The purpose of this observational study was to compare mid-term clinical results and total costs in multivessel CAD patients who received multiple DES with a population treated with CABG in a single, high-volume center during the same period.

Materials and Methods

Population. One-hundred and eleven consecutive patients with multivessel CAD treated exclusively with multiple DES and 95 similar patients treated with CABG were selected between November 2003 and January 2005 from a total of 1,560 coronary angiography and 946 PCI procedures at a single interventional center.
Multivessel PCI has been defined as the treatment of at least 2 lesions in 2 different epicardial main vessels amenable to DES implantation in order to obtain at least a functionally complete revascularization. Multivessel PCI strategy was determined based on the agreement of at least 2 expert interventionalists, while cases with discordant indications were collegially discussed with cardiac surgeons. The indication for surgical treatment was generated by weekly meetings between cardiologists and surgeons where the two different options were discussed together.
Procedure. Coronary angiography was performed during the index hospital stay or electively on an outpatient basis. All patients gave written informed consent for coronary angiography and PCI before the procedure and were pretreated withspirin and clopidogrel (loading dose of 300 mg and then 75 mg twice daily) or ticlopidine 250 mg twice daily at least 48 hours before catheterization. Thienopyridines were discontinued in the event that CABG was required. Surgery was performed within a week if urgent, or within 30 days if elective.
Multivessel PCI was performed in two stages only if the patient was considered at high risk (first step during acute myocardial infarction, large area of jeopardized myocardium, particular anatomical complexity). A dose of 100 U/kg of unfractionated heparin was administered immediately before PCI in order to achieve an activated clotting time between 250 and 300 seconds, or 70 U/kg in cases where a glycoprotein IIb/IIIa inhibitor was used, the latter being given at the discretion of the first operator. Ninety-nine patients received multiple Taxus and 12 Cypher stents (Taxus^®, Boston Scientific, Natick, Massachusetts, Cypher^™, Cordis, Miami Lakes, Florida). Dual antiplatelet therapy was then continued for 3 to 6 months after the index procedure.
Definitions. Complete angiographic success was defined as residual stenosis < 30% of all the treated lesions, while partial angiographic success was defined as residual stenosis < 30% in the main vessel(s).
Complete revascularization was defined as successful treatment of all stenotic lesions. Functionally complete revascularization referred to successful revascularization of all main branches with residual stenosis in small (< 2.0 mm) secondary vessels or residual collateralized chronic obstruction. Incomplete revascularization was defined as residual stenosis or obstruction of 1 main vessel or significant secondary branches.
Data collection. All demographic, clinical, anatomical, procedural and resource consumption data relative to the treated patients were collected in an electronic database at the time of the procedure. Procedural and inhospital complications, duration of coronary care unit (CCU) and ward stay were recorded during the index hospital stay. All events (target lesion revascularization [TLR], TVR, CABG, acute myocardial infarction [AMI], death and all other clinically relevant events), together with resource consumption data, were collected during the 12-month follow up.
Follow up. One-month clinical examination, and 6- and 12-month telephone interview follow up were performed. All clinical events and all data relative to resource consumption (visits, noninvasive and invasive examinations and all hospital admissions related to cardiological clinical problems) were recorded.
Events. Major cardiovascular and cerebrovascular events were defined as follows: (1) death by any cause; (2) postprocedural Q-wave and non-Q-wave MI (rise in CK-MB > 3 times the upper limit of normal); (3) clinical events (at least 2 elements among typical chest pain, typical ECG changes and a rise in myocardial-specific enzymes); (4) a new percutaneous myocardial revascularization (divided into TLR, TVR no-TLR, no-TVR in cases of repeat PCI in the same lesion, in the same vessel, but in another segment, and in a different vessel, respectively) or CABG; and (5) ischemic cerebral events (transient ischemic attack or stroke).
Cumulative MACE was defined as composite death and/or MI and/or TVR, while event-free survival was defined as survival in the absence of death, MI, total revascularization (TVR and no-TVR) or cerebrovascular accident.
Costs analysis. Calculation of PCI costs, comprising the initial cost of the procedure (materials, personnel and equipment mortgages) and of the hospital stay, have been described in detail previously.⁹ Briefly, for each procedure, the material used was recorded and the cost calculated on the purchasing cost provided by the hospital pharmacy including the value-added tax. The cost of DES was 1,800 € in 2003 and 1,650 € in 2004; for patients with multiple Taxus stents, the manufacturer, Boston Scientific Corporation (Milan, Italy), agreed on a discounting pricing scheme for consecutive stents placed in the same patient. The personnel and equipment costs were calculated on a per-examination basis relative to a 4-month activity period in 2003, following the mean paychecks referenced from tables compiled by the Health Administration and applying a depreciation rate of 12.5% to data provided by administration for equipment mortgages calculation. The following assumptions were made: that a session of both coronary angiography and PCI was equal to 1.5 times the average examination cost and that the staff present always consisted of 2 cardiologists, 2.5 cath lab nurses, 1 radiology technician and 0.5 ward nurses. Finally, the cost per patient stay in the cardiology unit, derived from Ravenna ASSR data tables,¹⁰ amounted to 428 €/day, while the cost in the coronary care unit was calculated applying a 1.9 incremental factor compared to the cardiology unit, yielding a total of 813 €/ day.
The costs for CABG included the sum of the initial hospital stay with cardiac catheterization (according to DRG reimbursement for clinical diagnosis), and the hospital stay with cardiac surgery (according to the DRG) with or without cardiac rehabilitation.
Follow-up costs consisted of noninvasive diagnostics (cardiological visits, echocardiogram, exercise test, myocardial scintigraphy, minor surgical procedures) calculated according to the regional reimbursement tables,¹¹ and DRGs for new hospital admissions¹² (angina, heart failure) or MACE (AMI, PCI, CABG).
DRG reimbursement. The DRG reimbursement levels reported in the “Tariffario Emilia Romagna”^11,12 for Type A hospitals are listed in Table 1.
Statistical analysis. Continuous variables were compared using the two-sample t-test and confidence interval (CI), while discrete variables were compared with the test and CI for 2 proportions. Event-free survival is expressed with Kaplan-Meier curves at 12 months.

Results

Clinical characteristics. Clinical characteristics of the patients are listed in Table 2. The mean age was higher in patients treated with CABG, but in both groups, the prevalence of patients > 75 years of age was similar. The prevalenceof patients with previous coronary revascularization was higher in the PCI group, as was that of acute coronary syndrome (non-Q-wave MI, acute or recent Q-wave or unstable angina). Although rather frequent in patients treated with PCI (71%), triple-vessel CAD prevalence was higher in patients treated with CABG, as was left main disease. The two groups did not differ with regard to risk factors such as diabetes, previous MI and left ventricular ejection fraction.
Procedural characteristics and inhospital outcomes. PCI group. Multivessel PCI was performed at the end of diagnostic angiography in 98% of patients and during the index hospital stay in 80%, while 20% of patients were scheduled on a day-hospital basis; the procedure was two-step only in 8 cases (7%). Table 3 shows principal procedural characteristics, resource consumption and duration of hospital stay. In each patient, 2.36 vessels and 2.82 lesions were treated, and 2.84 DES were implanted for a mean total length of 52 ± 16 mm. Out of 313 lesions treated, 51% were type B2/C and 15% were bifurcations with a side branch diameter > 2 mm; 7 chronic total obstructions (CTO) were treated (out of 27 patients with at least 1 CTO). Four lesions (1.2%) were treated with balloon angioplasty only, and 9 (2.8%) with BMS due to difficulty in deploying a DES. Revascularization was complete or functionally complete in 70% of cases.
Complete procedural success was achieved in 93% of cases, and partial success in 7%; inhospital events included 6 postprocedural non-Q-wave MIs (5.4%) and 2 retroperitoneal hemorrhages (1.8%) treated with blood transfusions. The mean postprocedural hospital stay was 2.5 days, and the total hospital stay was 5.3 ± 3 days.
CABG group. In the surgical group, CABG was performed urgently in 51 patients (54%, mean waiting time 5.9 ± 2.9 days) or electively (mean out-of-hospital waiting time 18 ± 23 days). Revascularization was complete in all patients, with a mean implantation of 3.9 ± 0.9 grafts and an arterial graft on the LAD in all but 1 patient. The index hospital stay duration comprehensive of diagnostic angiography was 6.7 ± 5.7 days, while mean hospital stay for CABG was 16.9 ± 7.4 days, comprehensive for 48 patients (50.5%) of a mean postprocedural rehabilitation of 8 days. Total mean index hospital stay for CABG was 23.5 ± 10.4 days. There were early postoperative complications in 16 patients (17%): 2 patients died early and 1 patient died 2 months later (postoperative septicemia), 1 patient had an embolic stroke and 1 repeated AITs, 2 patients had hemorrhagic complications requiring reintervention in one of them, 1 patient had to undergo reoperation for sternal suture problems, 6 patients had pulmonary complications with need of thoracentesis, or pleural drainage of pneumothorax in 3 cases, and 2 patients had neurologic dysfunction.
Follow-up events. The median follow up was 14 ± 2 months and 18 ± 6 months in PCI and CABG groups respectively; 12-month follow up was completed for all patients. Table 4 reports 12-month single events incidence and cumulative MACE in both groups.
There were 4 deaths in the PCI (1 of noncardiovascular cause) group and 5 deaths in the CABG group (2 postoperative, 1 at 2 months and 2 of cerebrovascular cause). One patient in the PCI group experienced Q-wave MI at 3 months due to subacute stent thrombosis treated with urgent re-PCI, and 1 patient suffered non-Q-wave MI at 6 months with severe stenosis of a marginal obtuse branch not treated in the index procedure. In the CABG group there were 4 non-Q-wave MIs; in 1 of these patients, PCI was performed at 3 and 5 months in 2 native vessels due to graft failure.
In the PCI group, there were 17 total revascularization procedures at 12 months, 14 TVRs (12.6%) and 3 no-TVRs. Out of the 14 TVR procedures, 2 patients underwent elective CABG, and 3 underwent PCI for disease progression in previously untreated segments (no TLR); 9 TLR PCIs (8%) were performed, 7 on DES and 2 on balloon angioplasty alone or BMS. The total revascularization rate at 12 months in patients who had a previously incomplete revascularization was 30%, while in patients with a previously complete or functionally complete revascularization, the rate was 9% (p = 0.010).
Cumulative MACE were 15.3%, while event-free survival was 83% (88.5% in patients with at least a functionally complete revascularization and 67% in incompletely revascularized patients).
In the CABG group, 3 patients underwent clinically-driven coronary angiography in the first 12 months, but only 1 patient underwent 2 PCI procedures; cumulative MACE were 9.5% and event-free survival was 87%.
There were no significant differences in MACE between the two groups. Figure 1 shows the Kaplan-Meier curves.
Table 5 shows resource consumption at 12-month follow up. Echocardiography and exercise tests were used more frequently in the CABG group, while PCI patients were preferentially monitored with myocardial scintigraphy. Nineteen PCI patients and 12 CABG patients were admitted to the cardiology unit due to angina, coronary angiography with or without revascularization and/or heart failure.
In the CABG population, postsurgical sequelae (pleural effusion, sternal suture or safenectomy infections) requiring minor surgical procedures or prolonged medical therapy were observed in 7 patients.
Costs. Table 6 shows initial, late and total costs in the two groups. The true calculated costs in the PCI group were 5,092 ± 1,058 € (Euros) for materials used, 5,518 ± 1,098 € for total procedural costs which included personnel and equipment expenses, and 3,473 ± 2,347 € for hospital stay, leading to a total initial cost of 8,992 ± 2,825 €, significantly higher than that provided by DRG reimbursement (7,216 €). In the CABG group, the cost of index hospital admission with coronary angiography, according to the DRG reimbursement, was 3,283 ± 1,083 €; the cost of the CABG procedure, which included a physical rehabilitation period for 50.5% of the patients, was 15,918 ± 1,383 €, leading to an initial total cost of 19,124 ± 1,912 €, significantly higher than that for multivessel PCI with DES, as was the 12-month total cost (20,050 ± 2,702 € CABG vs. 10,214 ± 4,184 € PCI with DES; p < 0.001).

Discussion

The comparison between PCI with stent implantation and CABG in multivessel CAD in randomized clinical trials has shown similar medium-term clinical hard events with a greater need for new revascularization procedures in the percutaneous arm.^13–15 With the diffusion of percutaneous procedures, the availability of more technologically advanced devices, more aggressive antithrombotic treatment and the willingness to treat more complex anatomical disease, it has become common clinical practice to treat patients with multivessel CAD, and to refer to surgery only those patients who are determined technically unsuitable for percutaneous treatment. With the advent of DES, the gap in terms of need for repeat revascularizations between PCI and CABG patients becomes narrower, thus increasing the population of patients who can benefit from percutaneous revascularization; specific studies estimated a 21% shift towards PCI¹⁶ and the possibility of complete percutaneous revascularization with a mean use of 3.6 DES/patient for up to 66% of patients for whom surgery is indicated.¹⁷ The ARTS II study⁸ showed an excellent performance of the sirolimus-eluting stent in multivessel PCI, with an incidence of MACE and cerebrovascular events at 12 months of 10.4%; these results are similar to those of the surgical arm of the ARTS I study (MACCE 11.6%), and significantly better than the BMS arm (MACCE 26.5%). It is of note that the comparison was made with a historical surgical cohort of patients and not with a simultaneous and randomized group of CABG patients due to the evolution of clinical practice and to the intrinsic difficulties in designing randomized clinical trials in this field. Observational studies¹⁸ and the results of the BARI registry¹⁹ show that clinical judgment must take into account all the clinical, anatomical and functional characteristics, as well as the periprocedural morbidity and mortality risks in order to choose the best revascularization modality for the individual patient. This usually translatesinto a surgical indication for higher-risk patients. Moreover, there is significant need for prolonged follow up and for study of very large populations. This would help show differences in outcomes between percutaneous and surgical revascularization in multivessel CAD patients, especially those not at high risk.²⁰
Our population of patients treated with multivessel PCI shows rather high-risk clinical and anatomical characteristics, which were partly exclusion criteria from previous studies. Despite this, acute results were excellent (complete procedural success in 93% of cases, partial in 7%), and the 12-month results were good (MACE 15.3%), which is comparable to those of single-vessel PCI with BMS, such as in the BMS arm RCT’s of single-vessel disease,^5,6 or unselected pre-SES population in the RESEARCH Registry.²¹ One should notice the high proportion of glycoprotein IIb/IIIa inhibitor utilization in a rather aggressive and early treatment scenario (ad hoc PCI in 98% of cases, single-stage in 93%), with very few inhospital events. Anatomical characteristics were also rather complex (LAD or LM treated in 99% of patients, with at least a B2/C lesion in 99/111 patients). According to the analysis of two of the authors (EV, MB) in the absence of DES, the alternative revascularization strategy would have been CABG in 46% and PCI with BMS in 54% of cases. Out of the 14 TVR at 12 months, 3 were due to progression of disease in different segments, 9/11 TLR were on a DEStreated segment and only 8 of them were clinically driven. A complete or functionally complete revascularization was obtained in 70% of our population; their 12-months eventfree survival was comparable to that of the CABG group (88.5% vs. 87%). It is of note that 60% of all revascularizations at follow up (TVR and no-TVR) were necessary in the group of patients who underwent an initial incomplete revascularization. The possibility of achieving complete revascularization seems to be fundamental in the choice of the type of revascularization for patients with multivessel CAD. This proved to be a significant prognostic factor in patients with multivessel disease and ACS.²²
The evaluation of a parallel population of patients treated with CABG in the same period was made in order to show a real-world setting of multivessel CAD revascularization in a high-volume center and to compare costs of the two different strategies. Our surgical group is characterized by an elevated incidence of left main disease associated with double- and triple-vessel disease. Early complications are rather high (17% overall), and hard events are concentrated in the early followup period, with a nonnegligible incidence of surgical sequelae (8%) and need for prolonged medical treatment.
Generally, the two groups of patients selected in current clinical practice for percutaneous or surgical revascularization do not differ as far as global clinical risk is concerned, but for some anatomical characteristics permitting at least a functionally complete percutaneous revascularization.
In the cost analysis, we evaluated the real costs of multivessel PCI with DES to make a comparison with actual reimbursement and CABG costs. A rough estimation of CABG expenses is based on the DRG reimbursement rate, recently evaluated as very near to the real costs,²³ and not linked to case-mix complexity, which is responsible for incremental cost increases due to prolonged hospital stay and the necessity for more intensive care. The initial and final costs of multivessel PCI with DES implantation are both significantly lower than those of CABG, even if follow-up costs are higher (but without statistical significance). The consequence is that percutaneous treatment appears to be economically very advantageous if technically feasible. The obstacle to the widespread implementation of this strategy appears to be the gap between real costs of multivessel PCI with DES use (mean of 8,992 €) and the actual reimbursement (7,216 €) (+25%), which does not take into consideration the use of DES instead of BMS. Ideally, national reimbursement tables must be updated in order to balance the differences in regional utilization of the two revascularization strategies, where overt preference is given for the treatment method that will receive the highest reimbursement.²⁴ The real cost calculation can be considered as a “minimum” cost because of the short duration of total hospital stay, the ability to perform ad hoc PCI, single-stage multivessel PCI and the medium-to-high volume of cases performed at our center.

Study Limitations

The obvious limitation of this study is that the comparison between multivessel PCI and CABG is nonrandomized and that there is a clear selection bias. Moreover, the population size is small, underpowered to show clinical outcome differences between the two groups and insufficient to apply multivariate analysis. The difficulty in avoiding a selection bias in the comparison of PCI and CABG in multivessel CAD patients, and the observational nature of this real-world scenario study, make the comparison between groups less important than the outcome in each group.

Conclusions

Multivessel PCI with multiple DES in a medium- to high risk population is safe, with good immediate and mediumterm results and a complication rate, comparable to those obtained with single-vessel PCI and to historical surgical patient cohorts.
Patients undergoing surgical revascularization have a rather high incidence of early complications and a nonnegligible occurrence of surgical sequelae, with significantly longer hospital stays than patients treated with PCI.
Initial and total 12-month costs of multivessel PCI with DES are significantly lower than those of CABG, but 25% higher in comparison to the DRG reimbursement actually provided. As a consequence, a recalculation of the DRG for DES PCI is advisable in order to allow better resource distribution in the field of myocardial revascularization.
From our experience, only multivessel patients who are not suitable PCI candidates should undergo surgical revascularization if DES are available and if complete percutaneous revascularization is feasible.

References

1. Schofer J, Schluter M, Gershlick AH, et al. Sirolimus-eluting stents for treatment of patients with long atherosclerotic lesions in small coronary arteries: Doubleblind, randomised controlled trial (E-SIRIUS). Lancet 2003;362:1093–1099.

2. Schampaert E, Cohen EA, Schluter M, et al. The Canadian study of the sirolimuseluting stent in the treatment of patients with long de novo lesions in small native coronary arteries (C-SIRIUS). J Am Coll Cardiol 2004;43:1110–1115.

3. Stone GW, Ellis SG, Cox DA, et al for the TAXUS-IV Investigators. One-year clinical results with the slow-release, polymer-based, paclitaxel-eluting TAXUS stent: The TAXUS-IV trial. Circulation 2004;109:1942–1947.

4. Dawkins KD, Grube E, Guagliumi G, et al on behalf of the TAXUS VI Investigators. Clinical efficacy of polymer-based paclitaxel-eluting stents in the treatment of complex, long coronary artery lesions from a multicenter, randomized trial. Support for the use of drug-eluting stents in contemporary clinical practice. Circulation 2005;112:3306–3313.

5. Moses JW, Leon MB, Popma JJ, et al for the SIRIUS Investigators. Sirolimuseluting stents versus standard stents in patients with stenosis in a native coronary artery. N Engl J Med 2003;349:1315–1323.

6. Colombo A, Drzewiecki J, Banning A, et al for the TAXUS II Study Group. Randomized study to assess the effectiveness of slow- and moderate-release polymerbased paclitaxel-eluting stents for coronary artery lesions. Circulation 2003;108:788–794.

7. Hoffman SN, TenBrook JA, Wolf MP, et al. A meta-analysis of randomized controlled trials comparing artery bypass graft with percutaneous transluminal coronary angioplasty: One- to eight-year out comes. J Am Coll Cardiol 2003;41:1293–1304.

8. Serruys PW, Ong ATL, Morice M-C, et al for the ARTS II Investigators. Arterial Revascularization Therapy Study Part II – Sirolimus-eluting stents for the treatment of patients with multivessel de novo coronary lesions. Euro Interv 2005;2:147–156.

9. Varani E, Balducelli M, Gatti C, et al. Cost of single-vessel and multivessel coronary drug-eluting stenting: Comparison to the DRG funding level. Ital Heart J 2005;6:52–58.

10. ASSR 2001 Data. Il Sole 24 Ore Sanità 2004;9:14–21

11. “Nomenclatore tariffario delle prestazioni ambulatoriali” del D.M. 22 luglio 1996. Data provided by web research on www.e-drg.it

12. “Determinazione delle tariffe per prestazioni di assistenza ospedaliera in strutture pubbliche e private accreditate della RER applicabili a decorrere dall’1-1-2005” Del. N^° 2126 del 19-12-2005. Data provided by web research on www.e-drg.it

13. Rodriguez A, Bernardi V, Navia J, et al, for the ERACI II Investigators. Argentine randomized study: Coronary angioplasty with stenting versus coronary bypass surgery in patients with multiple-vessel disease (ERACI II): 30-day and one-year follow-up results. J Am Coll Cardiol 2001;37:51–58.

14. The SOS Investigators. Coronary artery bypass surgery versus percutaneous coronary intervention with stent implantation in patients with multivessel coronary artery disease (the Stent or Surgery trial): A randomised controlled trial. Lancet 2002; 360:965–970.

15. Serruys PW, Unger FG. Sousa JE, et al. Comparison of coronary-artery bypass surgery and stenting for the treatment of multivessel disease. N Engl J Med 2001;344:1117–1124.

16. Ferreira AC, Peter AA, Salerno TA, et al. Clinical impact of drug-eluting stents in changing referral practices for coronary surgical revascularization in a tertiary care center. Ann Thorac Surg 2003;75:485–489.

17. Griffiths H, Bakhai A, West D, et al. Feasibility and cost of treatment with drug eluting stents of surgical candidates with multi-vessel coronary disease. Eur J Cardiothorac Surg 2004;26:528–534.

18. Hannan EL, Racz MJ, Walford G, et al. Long-term outcomes of coronary-artery bypass grafting versus stent implantation. N Engl J Med 2005;352:2174–2183.

19. Detre KM, Guo P, Holubkov R, et al. Coronary revascularization in diabetic patients: A comparison of the randomized and observational components of the Bypass Angioplasty Revascularization Investigation (BARI). Circulation 1999;99:633–640.

20. Gersh BJ, Frye RL. Methods of coronary revascularization – Things may not be as they seem. N Engl J Med 2005;352:2235–2237.

21. Lemos PA, Serruys PW, van Domburg RT, et al. Unrestricted utilization of sirolimus-eluting stents compared with conventional bare stent implantation in the “real world”: The rapamycin-eluting stent evaluated at Rotterdam cardiology hospital (RESEARCH) Registry. Circulation 2004;109:190–195.

22. Shishehbor MH, Lauer MS, Singh IM, et al. In unstable angina or non-ST-segment acute coronary syndrome, should patients with multivessel coronary artery disease undergo multivessel or culprit-only stenting? J Am Coll Cardiol 2007;49:849–854.

23. Campolo L. La scelta del miglior trattamento della coronaropatia con PCI ed i rimborsi dei DRG. Il paradosso della cardiochirurgia. Cardiologia Invasiva 2005;3(Suppl):210–203.

24. Arduini R, Colombo A, Genuini GF, et al. L’analisi dei DRG italiani e la pratica clinica in Italia. In: L’impatto economico dell’innovazione tecnologica nelle malattie dell’apparato cardiovascolare. Centro Studi Economia e Sanità- Business International Quaderno 16/2004, pp. 25–28.