Drug-eluting stents: a study of appropriateness and variations in practice

Austin, David (2010) Drug-eluting stents: a study of appropriateness and variations in practice. MD thesis, University of Glasgow.

Abstract

Summary
1) Background
Drug-eluting stents (DES) reduce in-stent restenosis, hitherto a frequent complication of percutaneous coronary intervention (PCI). Despite the reduction in recurrent symptoms and the need for repeat procedures, only highly selected patient groups with simple coronary lesions were studied in the initial trials, and long-term outcome or safety data were initially limited. The lack of long term data was compounded by reports of late stent thrombosis, potentially threatening the safety of the device. DES are more expensive than bare metal stents (BMS), so there is a significant increase the procedural cost of PCI. The uptake of DES in clinical practice created pressure on healthcare budgets leading to explicit rationing in some countries.

This thesis comprises a programme of original work addressing three related aspects of DES use: DES use in clinical practice (Scotland and internationally), “off-label” use of DES and appropriate use of DES (a modified Delphi consensus study).

2) Drug-eluting stent use in clinical practice
Relatively little was known about the actual application of DES in clinical practice. Anecdotal accounts suggested there were geographical practice variations within Scotland and internationally. Scottish practice was analysed with the aim of determining whether DES use varied between hospitals and operators within the Scottish NHS beyond the influence of clinical factors. In examining international practice the aim was to show whether the adoption and use of DES varied between countries and to determine whether practice changed after the stent thrombosis controversy.

a. Clinical practice variation within Scotland - methods and results
Using data from the Scottish Coronary Revascularisation Registry, multi-level logistic regression analysis was performed, analysing for variations in DES use at hospital, operator and patient level. Overall, DES were used in 47.6% of lesions, but varied between hospitals (range 30.6%-61.8%, χ2=341.6, p<0.0001). There was significant between-operator variation in the null model. This was attenuated by the addition of hospital as a fixed effect. Nonetheless, the final model demonstrated significant between-operator variability and between-hospital variation, after case-mix adjustment.
b. Drug-eluting stent use internationally - methods and results
This study involved collaboration between centres from four countries with established PCI registries: APPROACH Registry (Alberta, Canada), BWGIC (Belgium), Mayo Clinic PCI Registry (USA) and the Scottish Coronary Revascularisation Registry. Customised graphics software was employed to perform trend analysis examining variations in DES use over time, and by clinical sub-group. 178,504 lesions treated between January 2003 and September 2007 were included. In the Mayo Clinic Registry rapid adoption to a peak of 91% DES use for all lesions by late 2004 was observed. Alberta and Scotland showed delayed adoption with lower peak DES use, respectively 56 and 58% of lesions by early 2006. Adoption of DES in Belgium was more gradual and peak use of 35% lower than other registries. Reductions in DES use were seen in all datasets during 2006, though this varied in absolute and relative terms and by clinical sub-group.
c. Conclusions
Practice variations were found at operator and hospital level within Scotland and between countries internationally. Influences on stent choice in the “real world” are likely to be multi-factorial; on an international level, macro-economic forces exerting their influence through healthcare system regulation, payment systems, level of funding and central control are particularly important. It was also clear from the multilevel study of Scottish practice, however, that a clinical consensus does not exist.

3) “Off-label” use of drug-eluting stents
DES are often used for “off-label” indications, untested in RCTs, where observational studies demonstrate complications are higher when compared to “on-label” use. The aim was to determine whether clinical outcomes differ following DES and BMS implantation in a patient cohort defined by DES “off-label” indications.

a. Methods and results
Patients who underwent coronary stenting for an “off-label” indication between January 2003 and September 2005 in Scotland were included. Clinical outcomes were determined using linkage to national admission and death databases. Propensity scores were calculated using important baseline variables and DES were matched to BMS on a one-to-one basis to provide a fair comparison. The final study population comprised 1,642 well matched patients. Event-free survival was calculated over 24 months using the Kaplan-Meier method. All-cause death was more common following BMS during follow-up. No difference in the rates of MI was noted. Target vessel revascularisation was reduced in patients treated with DES.

b. Conclusions
The largely reassuring findings of this study should be seen in the context of a subsequent growing body of literature also suggesting similar risks for DES and BMS when compared for both on-label and off-label use. Although the benefits of DES were evident, the absolute reduction in TVR was lower than previously demonstrated in RCTs.

4) Appropriate use of drug-eluting stents - a modified Delphi consensus study
Best practice with respect to stent selection during PCI was poorly defined, as evidenced by the lack of detailed clinical guidance on the use of DES and wide practice variation demonstrated. It was not clear whether in any given setting there had been either underuse - potentially forfeiting the benefits of DES, or overuse - where benefits may be outweighed by risks. The aim was to use an expert panel to develop criteria for the appropriate use of DES using the modified Delphi method, to determine the extent to which current practice in Scotland met the appropriateness criteria and to validate the criteria by analysing clinical outcomes
a. Methods, results, conclusions
Consensus criteria for appropriate DES use were defined using a modified Delphi questionnaire. Expert panelists were used to define levels of appropriate use and were compared to clinical practice. The results suggested that current overall rates of DES use are acceptable. Better targeting of DES to the most appropriate lesions may be possible with the aims of reducing the known geographical inequities and maximising clinical benefit. Finally, using similar methods to chapter 4, it was shown that underuse of DES in appropriate patients was associated with higher levels of target vessel revascularisation without any difference in MI or death.

• Item Type: Thesis (MD)
• Qualification Level: Doctoral
• Keywords: drug-eluting stents, angioplasty, coronary artery disease, practice variations, multilevel modelling, propensity scores, Delphi consensus method
• Subjects: R Medicine > RC Internal medicine
• Colleges/Schools: College of Medical Veterinary and Life Sciences > School of Health & Wellbeing
• Supervisor's Name: Pell, Prof. Jill P. and Oldroyd, Dr. Keith G.
• Date of Award: 2010
• Depositing User: Dr David Austin
• Unique ID: glathesis:2010-2186
• Copyright: Copyright of this thesis is held by the author.
• Date Deposited: 23 Dec 2010
• Last Modified: 10 Dec 2012 13:52
• URI: https://theses.gla.ac.uk/id/eprint/2186