Send to:

Choose Destination
See comment in PubMed Commons below
Pharmacoeconomics. 2012 Nov 1;30(11):1067-84. doi: 10.2165/11595080-000000000-00000.

The cost effectiveness of genetic testing for CYP2C19 variants to guide thienopyridine treatment in patients with acute coronary syndromes: a New Zealand evaluation.

Author information

  • 1School of Population Health, University of Auckland, Auckland, New Zealand.



A recent clinical trial has demonstrated that patients with acute coronary syndromes (ACS) and the reduced function allele CYP2C19*2 (*2 allele), who are treated with thienopyridines, have an increased risk of adverse cardiac events with clopidogrel, but not with prasugrel. The frequency of the *2 allele varies by ethnicity and the Maoris, Asians and Pacific Islanders of New Zealand have a relatively high incidence.


Our objective was to evaluate, from a New Zealand health system perspective, the cost effectiveness of treating all ACS patients with generic clopidogrel compared with prasugrel, and also compared with the genetically guided strategy that *2 allele carriers receive prasugrel and non-carriers receive clopidogrel.


A decision-tree model consisting of five health states (myocardial infarction, stroke, bleeding, stent thrombosis and cardiovascular death) was developed. Clinical outcome data (two TRITON-TIMI 38 genetic sub-studies) comparing clopidogrel and prasugrel for both *2 allele carriers and non-carriers were combined with the prevalence of the heterozygosity for the *2 allele in New Zealand Europeans (15%), Maoris (24%), Asians (29%) and Pacific Islanders (45%) to determine the predicted adverse event rate for the New Zealand population. National hospital diagnosis-related group (DRG) discharge codes were used to determine alternative adverse event rates, along with the costs of hospitalizations during the 15 months after patients presented with an ACS. The primary outcome measure was the incremental cost per QALY (calculated using literature-reported weights). Monte Carlo simulations and alternative scenario analysis based on both clinical trial and national hospital incidence were used. Additional analysis considered the overall TRITON-TIMI 38 rates. Costs (in New Zealand dollars [$NZ], year 2009 values) and benefits were discounted at 3% per annum.


Actual hospital-based adverse event rates were higher than those reported in the TRITON-TIMI 38 randomized controlled trial and the genetic sub-studies, especially for myocardial infarction and cardiovascular death, and for Maoris and Pacific Islanders. For both sources of adverse event rates, treating the population with prasugrel was associated with worse outcomes (QALYs) than clopidogrel. However, prasugrel became cost effective ($NZ31 751/QALY) when the overall TRITON-TIMI 38 rates were used. A genetic test to guide the selected use of prasugrel was cost effective ($NZ8702/QALY versus $NZ24 617/QALY) for hospital and clinical trial incidence, respectively. Based on the hospital rates, the genetically guided strategy was especially cost effective for Maoris ($NZ7312/QALY) and Pacific Islanders ($NZ7041/QALY). These results were robust to the sensitivity analysis, except the genetically guided strategy under the 15-month clinical trial event rate scenario ($NZ168 748/QALY) did not remain cost effective under a $NZ50 000 threshold.


Use of a genetic test to guide thienopyridine treatment in patients with ACS is a potentially cost-effective treatment strategy, especially for Maoris and Pacific Islanders. This treatment strategy also has the potential to reduce ethnic health disparities that exist in New Zealand. However, the results comparing clopidogrel and prasugrel are sensitive to whether the genetic sub-studies or the overall TRITON-TIMI 38 rates are used. While the national hospital event rates may be more appropriate for the New Zealand population, many assumptions are required when they are used to adjust the genetic sub-studies rates.

[PubMed - indexed for MEDLINE]
PubMed Commons home

PubMed Commons

How to join PubMed Commons

    Supplemental Content

    Full text links

    Icon for Springer
    Loading ...
    Write to the Help Desk