NICE commissioned an external assessment centre (EAC) to review the evidence submitted by the company. This section summarises that review. Full details of all the evidence are in the documents in the history section for this guidance.
3.1 The EAC assessed 19 studies. Eight were full text publications: 1 randomised controlled trial, 3 prospective studies (2 of which were comparative), and 4 retrospective studies (2 of which were comparative). One abstract reported results from a randomised controlled trial, and 8 abstracts and posters reported retrospective studies. Two posters reported results from the same acute kidney injury (AKI) reduction programme. Two unpublished papers were included: 1 retrospective comparative study and 1 prospective comparative study. For full details of the clinical evidence, see section 3 of the assessment report.
3.2 The evidence from comparative studies showed that using DyeVert Systems reduced the amount of contrast media injected by between 17% (Sattar et al. 2018) and 41% (Desch et al. 2018) compared with standard coronary angiography. A company meta-analysis estimated that contrast volume received by the patient was lowered by 39.43% when using DyeVert Systems compared with when they were not used (calculated from 5 single-arm studies and 3 comparative studies).
3.3 Seven published studies reported no loss to image quality with DyeVert Systems (Desch et al. 2018, Gurm et al. 2019a, Briguori et al. 2020, Bruno et al. 2019, Sapontis et al. 2017, Corcione et al. 2017, Zimin et al. 2020) and 2 abstracts (Amoroso et al. 2020, Rao et al. 2019). A company meta-analysis estimated that image quality was 98.2% of that when the system was not used (calculated from 6 published clinical studies and 1 abstract).
3.4 Evidence on the risk reduction in AKI was from 1 published paper (Briguori et al. 2020) and 5 abstracts and posters (Sattar et al. 2018, Kutschman et al. 2019, Bunney et al. 2019, Cameron et al. 2020, Turner and Tucker 2020). These studies reported outcomes from people who were identified as having chronic kidney disease (CKD) stages 2 and 3 before their angiography procedure. Briguori et al. (2020) was a single-centre, observational, non-randomised design, which used a control group of patients treated in the same centre. Propensity score matching was used to match the control group to the DyeVert group, resulting in 90 patients in each group. This study reported AKI in 8% of the DyeVert Systems group and 19% of the control group (p=0.047). A company meta-analysis estimated the relative risk of contrast-induced AKI in a DyeVert Systems group compared with a control group to be 0.59, calculated from Briguori et al. (2020) and 3 comparative studies reported as an abstract or poster.
3.5 The evidence presented was limited because measurements were only taken during the procedure in the studies that looked primarily at contrast volume used or reduced, and image quality. In the studies that followed up after the procedure, the methodology around collecting and reporting of the outcomes was not clear. Briguori et al. (2020) did report serum creatinine for 72 hours after the procedure, and AKI incidence and major adverse events within 1 month of the procedure. But the studies were limited by their retrospective design, which meant not all AKI incidents could be identified.
3.6 The only evidence on the DyeVert Power XT device was on the first version of the system and from 1 full text single-arm retrospective study (Bruno et al. 2019) and 1 retrospective study presented as an abstract (Amoroso et al. 2020). This was from a total of 35 people having angiography procedures. No studies included the current version of the Power XT system.
3.8 The company presented 1 published UK-based cost–utility analysis (Javanbakht et al. 2020) and 3 economic studies based in the US. Javanbakht et al. (2020) found DyeVert Systems were cost saving by £435 and estimated a quality-adjusted life year (QALY) gain of 0.028 over an individual's lifetime compared with current practice. The company's cost model was based on an updated version of this published model. For full details of the cost evidence, see section 4 of the assessment report.
3.9 The company's model included people with CKD stages 3 and 4. It used a decision tree for the first 3 months after the procedure, then a Markov model for the remainder of the individual's lifetime. The Markov model transitions between 6 health states in 3‑month cycles. The company's model had some differences to Javanbakht et al. (2020). The relative risk reduction of AKI because of DyeVert Systems use was increased from 21.4% to 41% (based on the company meta-analysis results). Unit costs were also updated, peripheral angiography procedures included in the population, and the age of the cohort entering the model was reduced from 72 years to 65 years.
3.10 The EAC agreed with the company's cost model overall. The EAC corrected small errors in the model and parameter inflation costs, lowered the relative risk of contrast-induced AKI, and differentiated the costs of fatal and non-fatal myocardial infarctions. The EAC found that DyeVert Systems remained cost saving by £22 with a QALY gain of 0.013.
If the baseline risk of contrast-induced AKI is below 8.2%, DyeVert Systems are no longer cost saving
3.11 The EAC updated the company's model and applied a baseline risk of contrast-induced AKI of 8.72% for people with CKD stage 3 and 4 aged 65. The EAC considered the company's stated risk of 30% to be too high if people are appropriately hydrated. The EAC estimated that, if the risk of contrast-induced AKI was below 8.2%, DyeVert Systems would no longer be cost saving.
3.12 The company's model used a relative risk reduction of contrast-induced AKI after DyeVert Systems use of 41% based on its meta-analysis results. The EAC accepted the statistical validity of the meta-analysis but noted that the strength of the included studies was low to moderate. The EAC did sensitivity analysis around the risk reduction of contrast-induced AKI from DyeVert Systems and found that the break-even relative risk reduction was 38.5%, assuming a baseline AKI risk of 8.72%.