Clinical and technical evidence

A literature search was carried out for this briefing in accordance with the interim process and methods statement. This briefing includes the most relevant or best available published evidence relating to the clinical effectiveness of the technology. Further information about how the evidence for this briefing was selected is available on request by contacting mibs@nice.org.uk.

Published evidence

Six studies are summarised in this briefing involving 3,930 patients. Studies include 2 randomised controlled trials, 3 prospective cohort studies, and a study reporting on outcomes from a subset of people enrolled in 1 of the randomised controlled trials.

The clinical evidence and its strengths and limitations is summarised in the next section on overall assessment of the evidence.

Overall assessment of the evidence

The evidence for Spartan RX suggests that the test is easy to use and clinically feasible. In most of the studies, the test could be done by a healthcare professional after short training and results were usually available within 1 hour to 2 hours after sample collection. Some studies reported that another sample had to be taken because of inconclusive results but it is not clear if this delayed treatment decisions or patient discharge. This highlights a potential need for extra training, particularly in taking time-sensitive buccal swabs.

The evidence suggests the test may mean clopidogrel is prescribed less often to people who poorly metabolise it because of their CYP2C19 genotype. But it does not show if treatment decisions made using the Spartan RX test alone significantly improve clinical outcomes for patients.

None of the included studies were done in the UK, and it is likely that clinical practice at the study sites varied. Therefore, the generalisability of the evidence to the NHS may be limited. No studies directly compared Spartan RX with other genotyping methods (such as TaqMan StepOnePlus assay) in a clinical setting so it is unclear if the test has advantages over laboratory-based testing, especially in centres where batchwise genotyping is routinely done.

Claassens et al. (2019)

Intervention and comparator(s)

Intervention: P2Y12 inhibitor therapy based on early CYP2C19 genetic testing (genotype-guided group).

Comparator: standard treatment with either ticagrelor or prasugrel (standard treatment group).

Key outcomes

At 12 months, adverse clinical events (death from any cause, myocardial infarction, definite stent thrombosis, stroke, or major bleeding as defined by Platelet Inhibition and Patient Outcomes criteria) occurred in 5.1% (n=63) of patients in the genotype-guided group and in 5.9% (n=73) of patients in the standard treatment group. CYP2C19 genotype-guided P2Y12 inhibitor therapy (in which patients without the CYP2C19 gene mutation had clopidogrel) was shown to be non-inferior to standard treatment with ticagrelor or prasugrel at 12 months with respect to thrombotic events (p<0.001; non-inferiority margin of 2% points for the absolute difference). Superiority analysis did not support the superiority of genotype-guided therapy (hazard ratio, 0.87; p=0.40). Genotype-guided therapy was associated with a lower incidence of bleeding (122 events [9.8%] compared with 156 events [12.5%]; p=0.04).

Strengths and limitations

It was a randomised multicentre study but had an open-label design so may be subject to bias. None of the study centres was based in the UK so generalisability of the results to the NHS is uncertain. The incidence of the primary combined outcome was lower than anticipated. Not all patients were genotyped using the Spartan RX test – TaqMan StepOnePlus assay at a central laboratory was also used. The company provided the Spartan RX system and reagents free of charge.

Bergmeijer et al. (2018)

Study size, design and location

Feasibility study reporting on data from 2 projects in which CYP2C19 genotyping was used to tailor antiplatelet treatment in patients admitted with ST-segment elevation myocardial infarction (STEMI) or having a planned PCI.

The study included all 1,283 patients who enrolled in the POPular Genetics study between July 2014 and 8 December 2017. It also reported on data from the Popular Risk Score project (n=2,556), but outcomes have not been included here because genotyping was done using the TaqMan StepOnePlus assay only.

Intervention and comparator(s)

Intervention: CYP2C19 genotyping strategies (Spartan RX, in-hospital TaqMan genotyping or central laboratory testing).

No comparator.

Key outcomes

Results from the study suggest that it is feasible to have genotyping results available within 24 hours to 48 hours for most patients. In the POPular Genetics study, the median time between randomisation and genotyping result was 2 hours 24 minutes (2 hours 4 minutes for in-hospital TaqMan genotyping, 2 hours 16 minutes for Spartan RX genotyping and 52 hours 32 minutes for central laboratory testing). One Spartan RX test result did not match with TaqMan genotyping result, and 8% of people had an inconclusive result with Spartan RX.

Strengths and limitations

The study did not involve any UK centres so the generalisability to the NHS is uncertain. Not all genotyping in the POPular Genetics study was done using Spartan RX; on-site TaqMan genotyping was used by 1 centre (487 patients), on-site Spartan RX genotyping by 7 centres (411 patients) and shipment to the central laboratory was used by 6 centres (140 patients). The company provided the Spartan RX system and test kit to 7 centres involved in the POPular Genetics study.

Cavallari et al. (2018)

Intervention and comparator(s)

Intervention: Spartan RX CYP2C19 genotype testing.

No comparator.

Key outcomes

The median turnaround time for genotype test results was 96 minutes. Genotyping was unsuccessful for the initial sample in 14% of people (56 inconclusive results and 73 device errors). Of these people, 123 agreed to have samples recollected; 6 of these had multiple inconclusive results. Forty-two percent of people who were genotyped had PCI (n=392). Genotype results were available for 99% of PCI patients before discharge. At discharge, 17% of poor metabolisers, 41% of intermediate metabolisers and 50% of patients without the CYP2C19 gene mutation were prescribed clopidogrel (p=0.110). At 6 months, clopidogrel was prescribed in 0% of poor metabolisers, 51% of intermediate metabolisers, and 63% of patients without a CYP2C19 gene mutation (p=0.008 across groups; p=0.020 for poor metabolisers compared with patients without a CYP2C19 gene mutation).

Strengths and limitations

The study did not account for confounding factors that may have influenced the decision to switch therapy, for example the affordability of treatment for patients and the presence of conditions that may increase bleeding risk. The study did not include a control group of patients not having genotyping. The study was done in 1 US centre so may not be generalisable to the NHS.

Zhou et al. (2017)

Intervention and comparator(s)

Intervention: Spartan RX CYP2C19 test.

Comparator: Verigene CYP2C19 test (laboratory validation study only).

Key outcomes

In the laboratory validation study, results were consistent between both assays (100% accuracy); both tests showed a sensitivity and specificity of 100% when results were obtainable. Samples tested on Verigene CYP2C19 produced a 20% no call rate and 1 processing error, whereas Spartan RX accurately identified the genotype on all 36 specimens. Of the 342 patients tested with Spartan RX in a clinical setting, 14 samples had inconclusive results, 10 failed controls, and there was 1 instrument failure. Results were available after retesting for 12 of these patients; 1 patient declined a retest, and 1 had a second inconclusive result. Overall, 27% (n=99) of the patients had a CYP2C19 gene mutation.

Strengths and limitations

The laboratory validation used a small number of unique samples. The study did not compare Spartan RX with Verigene in a clinical setting. The study was done in a single US centre so may not be generalisable to the NHS.

Koltowski et al. (2017)

Study size, design and location

Single-centre pilot study reporting interim results from a randomised controlled clinical trial (ONSIDE TEST project).

The study involved 50 people (aged between 18 and 75) with stable coronary disease scheduled for an elective PCI with stent implantation in 1 centre in Poland.

Intervention and comparator(s)

Intervention: Spartan CYP2C19 test (genotype-guided therapy) or phenotype testing with the VerifyNow P2Y12 assay (phenotype-guided therapy).

Comparator(s): No genotype testing (standard treatment with aspirin and clopidogrel).

Key outcomes

Five (32%) patients in the genotyping arm and 2 (13%) in the phenotyping arm were identified as poor metabolisers of clopidogrel and were given a loading dose of prasugrel at least 2 hours before the scheduled PCI. Genotyping took 1 hour. The test had to be repeated for 2 patients because of inconclusive results. Periprocedural platelet reactivity (pharmacokinetic response) was significantly lower in the genotyping and phenotyping groups compared with the control group (80 platelet reactivity units [PRU], range 49.0, p=0.01; 36.5 PRU, range 47, p=0.03; 176 PRU, range 67.8; respectively). There were no differences in the prevalence of periprocedural myocardial infarction, myocardial biomarker leak, and risk of major bleeding and major adverse cardiovascular events at the 30‑day follow up.

Strengths and limitations

The study involved a small number of patients and may have been underpowered for outcomes. It also had an open-label design because of the number of interventions over the course of the study. People in the control group had a longer duration of total vessel occlusion during intervention. The study was in Poland and so results may not be generalisable to the NHS.

Choi et al. (2016)

Intervention and comparator(s)

Intervention: Spartan RX CYP2C19 test.

Comparator: single nucleotide polymorphism (SNP) genotyping assay (TaqMan).

Key outcomes

Based on results from Spartan RX CYP2C19 genotyping, 3.3% of patients were classed as ultra-rapid metabolisers, 32.8% as extensive metabolisers, 45.4% as intermediate metabolisers, and 18.5% as poor metabolisers. There were 2 discrepancies (1.7%) with Spartan RX compared with SNP genotyping. The discrepancy appeared in the *17 allele analysis for both patients, and the result with Spartan RX was a false positive. The authors said that a false positive result for *17 allele could not affect the intermediate or poor metaboliser group (who were more likely to have thrombotic events) but it could affect ultra-rapid metaboliser group (who were more likely to have bleeding events).

Strengths and limitations

The study was done in Korea and so may not be generalisable to an NHS population.

Sustainability

The company has not submitted any sustainability claims.

Recent and ongoing studies