Clinical and technical evidence | RIDASCREEN tests for monitoring infliximab in inflammatory bowel disease | Advice

Published evidence

The literature search identified 10 studies that reported on the use of the RIDASCREEN tests for monitoring infliximab in inflammatory bowel disease (IBD). These include studies reporting on the earlier, pre-commercialised versions of the tests (Leuven in-house enzyme-linked immunosorbent assay [ELISA] method), because they were considered comparable to the current versions of the test. Studies reporting on these tests under their apDia branding were also included.

Five of these studies (table 3; n=587 serum samples in total) reported relevant clinical outcomes. These comprise 1 prospective observational study (Van Stappen et al. 2016) and 4 studies that had relevant test comparisons which reported concordance data (Marini et al. 2017, Lee et al. 2016, Malickova et al. 2016, Schmitz et al. 2015).

This briefing excludes 1 randomised controlled trial (TAXIT; Vande Casteele et al. 2015) and 1 test comparison study (Vande Casteele et al. 2012) from full review, because they are summarised in the NICE diagnostics guidance on therapeutic monitoring of TNF-alpha inhibitors in Crohn's disease. Three bench-test studies (Van Stappen et al. 2015a, Van Stappen et al. 2015b, Gils et al. 2016) were also excluded because they did not include clinical outcomes.

Table 3 summarises the clinical evidence as well as its strengths and limitations.

Overall assessment of the evidence

Only 1 relevant additional study (Van Stappen et al. 2016) was identified since the publication of the NICE diagnostics guidance on therapeutic monitoring of TNF-alpha inhibitors in Crohn's disease in February 2016. This study included clinical outcome data on the use of a pre-commercialised version of the RIDASCREEN Anti‑IFX Antibodies test. However, clinical outcome data were limited because it was not the primary outcome of the study. No information was reported on how therapeutic drug monitoring had helped with the management of symptoms.

In the absence of a gold-standard reference method for therapeutic drug monitoring of infliximab, there is limited evidence on the accuracy of the RIDASCREEN tests. Therefore, the evidence base is largely focused on correlation studies with other ELISAs for therapeutic drug monitoring of infliximab.

Only the Lee et al. (2016) and Schmitz et al. (2015) studies clearly stated which platforms were used for analysis, or whether the analysis was automated or manual.

The evidence base would be improved by prospective studies evaluating the clinical outcomes associated with using the tests.

Table 3 Summary of the current evidence

Van Stappen et al. 2016
Study size, design and location	N=190 serum samples from 29 anti-TNF-naive patients with ulcerative colitis starting IFX induction. Serum samples were prospectively collected during IFX induction (weeks 0, 2, and 6) and maintenance therapy (week 14 onward for up to 2 years). N=221 serum samples from the 29 patients were taken during the 1-year follow-up for testing 2 ATI tests. Prospective observational study, Belgium.
Intervention and comparator(s)	Intervention: rapid lateral flow-based assay (LFA, RIDA QUICK IFX). Comparator: RIDASCREEN IFX Monitoring. A new drug-tolerant anti-infliximab assay was also compared with a drug-sensitive anti-infliximab assay (Leuven in-house ELISA test, pre-commercialised version of Anti‑IFX Antibodies), using 221 serum samples.
Key outcomes	Receiver operating characteristic analysis identified an IFX concentration threshold ≥2.7 micrograms/ml using the RIDASCREEN IFX Monitoring test (AUROC of 0.802, p=0.012, sensitivity 100%, specificity 50%) to be associated with mucosal healing. During the 1-year follow-up, the drug-sensitive anti-infliximab assay detected ATI in 14% (4/29) patients at at least 1 time point. In total, 5/221 (2%) samples were ATI-positive using the drug-sensitive assay. All 4 patients with positive ATI in the drug-sensitive bridging ELISA had at that time no detectable TC. Two out of 8 patients discontinued IFX maintenance treatment because of adverse events associated with ATI development. The drug-sensitive assay detected the ATI at 40 and 17 weeks before the adverse events.
Strengths and limitations	Strengths: the study provides useful clinical outcome data on the use of RIDASCREEN. The patients in this trial are likely to be representative of those who would have IFX in UK practice. Limitations: the analysis was post-hoc which may introduce potential bias. Samples were collected from a small cohort of patients with ulcerative colitis. Results may not be generalisable to Crohn's disease.
Lee et al. 2016
Study size, design and location	N=12 serum samples, comprising 3 control samples from 2 healthy men and 1 woman, and 9 physiological interference samples. Physiological interference samples were from patients with positive rheumatoid factor (n=3) and with para-proteins (n=6). IFX was added to each of the serum control and physiological interference samples to achieve final infliximab concentrations of 0, 0.2, 2.0, 3.0 and 7.0 micrograms/ml. Pharmacological interference samples were prepared using adalimumab and etanercept and were diluted to achieve therapeutic concentrations of 20 mg/ml and 25 mg/ml respectively. Assay correlation study, Australia.
Intervention and comparator(s)	Comparison of 3 ELISAs for the detection of serum IFX trough levels: RIDASCREEN IFX Monitoring (R-Biopharm), Promonitor-IFX (Grifols) and LISA-Tracker Infliximab (Theradiag).
Key outcomes	RIDASCREEN IFX Monitoring had poorer agreement compared with other methods and consistently produced readings lower than the known concentration of the prepared samples. RIDASCREEN IFX Monitoring produced accurate negative readings at an IFX concentration of 0 micrograms/ml. To calculate the CV, 7 samples were analysed at concentrations of 3 and 7 micrograms/ml. CVs at both were 5%. Low concentrations of ATI caused little effect (in vitro binding) on the IFX results, while higher concentrations resulted in reduced IFX drug levels in a dose-dependent fashion. Falsely low readings were observed in the presence of etanercept, while no significant interference was observed with adalimumab.
Strengths and limitations	Strengths: a blinded set of samples were aliquoted from the original samples before sending to the respective laboratories, therefore reducing potential for operator bias. Limitations: samples were sent to different laboratories for analysis and details of their analysis techniques (platforms used and automated or manual processes) were not described. The para-protein interference samples were considered not entirely relevant to the IBD population because they are not common findings in this patient group.
Marini et al. (2017)
Study size, design and location	N=56 serum test samples for IFX concentration. N=32 serum test samples for ATI. Samples were from IBD patients (with and without IFX exposure) and healthy volunteers. Samples based on normal human serum were spiked with an IFX concentration of 5 micrograms/ml. Assay correlation study, US.
Intervention and comparator(s)	Intervention: Janssen (in-house) IFX and ATI (original and drug-tolerant versions) assays used in clinical trials of IFX. Comparators: RIDASCREEN IFX Monitoring and an in-house version of the RIDASCREEN Anti‑IFX Antibodies test. Commercialised IFX and ATI assays from Sanquin (Level Infliximab and ADA Infliximab), LabCorp (MSD ECLIA), and Dynacare (Promonitor-IFX ELISAs).
Key outcomes	No cross-reactivity was observed for the RIDASCREEN IFX Monitoring test with other anti-TNFs drugs prepared at 5 micrograms/ml of (adalimumab, certolizumab pegol, golimumab or siltuximab). The presence of ATI affected the selectivity of the RIDASCREEN IFX Monitoring test in a titre-dependent manner. However, TNF‑alpha concentrations (0.5 to 50 ng/ml) did not affect IFX detection in the assay. The in-house version of the RIDASCREEN Anti‑IFX Antibodies test produced inconclusive results in the presence of ≥2 micrograms/ml IFX concentrations.
Strengths and limitations	Strengths: sample analysis of all ELISAs was done in a blinded manner, which would reduce reviewer bias. Limitations: the study only compared the Janssen methods to each of the commercialised tests; little comparison was made between each of the commercialised tests. The study was funded by Janssen and therefore reporting bias may be introduced.
Malickova et al. 2016
Study size, design and location	N=42 serum samples from IBD patients (n=22) having Remsima (biosimilar IFX). 22 serum samples were measured at week 2 of the induction treatment, 15 serum samples at week 6 and 5 samples at week 14. Assay correlation study, US.
Intervention and comparator(s)	Comparison of 3 ELISAs for the detection of serum IFX trough levels: RIDASCREEN IFX Monitoring (R-Biopharm), SHIKARI Q-Inflixi (Matriks Biotek) and LISA-Tracker Duo Infliximab (Theradiag).
Key outcomes	At the cut-off value of 3 micrograms/ml IFX, the RIDASCREEN IFX test showed good correlation with the LISA-Tracker Duo (Kappa=0.76, r=0.92). At the cut-off value of 3 micrograms/ml IFX, the SHIKARI Q-Inflixi showed good correlation with the RIDASCREEN IFX test (Kappa=0.76, r=0.86). IFX levels measured by RIDASCREEN IFX Monitoring did not predict treatment responsiveness at week 2 after the induction phase treatment (p=0.71). Similar results were obtained for week 6 measurements.
Strengths and limitations	Strengths: this is the only study to investigate the correlation of 3 IFX ELISAs using the biosimilar Remsima. Limitations: very limited data were described about the clinical performance evaluation of the assays. Insufficient details of sample blinding may introduce reviewer bias. There were discrepancies between the Spearman's rank correlation coefficient values reported in the text and figures for RIDASCREEN IFX compared with LISA-Tracker Duo. The correct sample dilution was also not used.
Schmitz et al. 2015
Study size, design and location	N=34 blood samples (30 samples from patients with Crohn's disease or ulcerative colitis, collected immediately before a new IFX infusion, and 4 spiked plasma samples of IFX concentrations 10, 5.0, 2.5 and 0.5 micrograms/ml). Comparative performance evaluation study, The Netherlands.
Intervention and comparator(s)	Comparison of 3 ELISAs for the detection of infliximab: Infliximab ELISA (apDia), Theradiag (LISA-Tracker Infliximab), Progenika (Promonitor-IFX). All ELISA tests were implemented on the automated DSX 4-plate ELISA Processing System (Dynex Technologies).
Key outcomes	Imprecision results were determined using patient samples at 3 levels by triplicate measurements on 5 different days. The within-run imprecision of the apDia Infliximab ELISA at low (0.68 micrograms/ml), middle (2.8 micrograms/ml) and high (5.9 micrograms /ml) levels of infliximab were 6.1%, 1.9% and 2.4%, respectively. The between-run imprecision of the apDia Infliximab ELISA at low (0.68 micrograms/ml), middle (2.8 micrograms/ml) and high (5.9 micrograms/ml) levels of infliximab were 6.3%, 7.1% and 6.0%, respectively. The agreement of the apDia ELISA to the target values of the spiked samples ranged between 96% and 108%.
Strengths and limitations	Strengths: robust methods were used to determine imprecision values as triplicate measurements were made on 5 different days. Results were anonymised following IFX measurement, reducing the potential for reporter bias. Limitations: there were a relatively small number of samples analysed.
Abbreviations: ATI, antibodies to infliximab; AUROC, area under receiver operating characteristic; CI, confidence interval; CV, coefficient of variation; ELISA, enzyme-linked immunosorbent assay; IBD, inflammatory bowel disease; IFX, infliximab; LFA, lateral flow assay; TC, trough concentration; TNF, tumour necrosis factor.

Recent and ongoing studies

Two of the specialist commentators highlighted an analytical evaluation study of the RIDASCREEN IFX Monitoring test. This was presented as a poster at the British Society of Gastroenterology Conference in 2016, and is due to be published as a paper in 2017.

Specialist commentator comments

Comments on these technologies were invited from clinical experts working in the field and relevant patient organisations. The comments received are individual opinions and do not represent NICE's view.

All 4 specialist commentators were familiar with the RIDASCREEN tests and 2 had used them for research purposes. Three specialist commentators routinely use therapeutic drug monitoring, but none uses the RIDASCREEN tests.

Level of innovation

Three specialist commentators considered the RIDASCREEN tests to have little innovation because similar technologies already exist and are readily available. One added that RIDASCREEN is only a minor variation on existing tests. The fourth commentator highlighted that the University of Leuven uses the RIDASCREEN tests, and this is one of the leading and reputable centres in this field.

One commentator considered that ELISA technologies will likely be superseded in the near future by fully automated immunoassays, which offer higher throughput and efficiencies of scale.

Potential patient impact

Two specialist commentators considered that therapeutic drug monitoring is widely recognised as best practice in managing IBD, but that this isn't currently reflected in national guidelines. One added that 3 centres in the UK already offering routine testing using tests other than RIDASCREEN.

Three commentators indicated that the RIDASCREEN tests allow a personalised approach to drug optimisation, which is better for patients than empirical dose escalation. One added that this is particularly useful in patients whose disease doesn't respond (primary non-response) or stops responding (secondary loss of response). Additionally, 2 commentators considered that the tests reduce exposure to unnecessary, potentially harmful drugs in certain patients.

One specialist commentator described an analysis in their practice which demonstrated that in most people with secondary loss of response, the amount of infliximab in the blood had reached the therapeutic level. This means that the lack of response was not caused by a shortage of infliximab, so an increased dose wouldn't help.

One commentator added that patients who test positive for antibodies against infliximab are at an increased risk of infusion reactions. They felt that therapeutic drug monitoring may help to achieve a better response to infliximab treatment. This could help eliminate disease-related symptoms and improve quality of life with as few side effects as possible.

Potential system impact

Three specialist commentators considered that no changes in facilities or infrastructure would be needed because similar technologies are already used in the NHS. One commentator disagreed, and felt that any centre choosing to adopt the RIDASCREEN tests would need some infrastructural changes before implementation. Additional training may also be needed for any extra equipment (such as the Dynex DS2).

Another commentator highlighted that interpreting the results of the tests needs considerable expertise and understanding of their limitations, and appropriate network facilities and software would be needed to transfer patient results.

All 4 specialist commentators considered that the RIDASCREEN tests could lead to cost savings for the NHS. Three added that cost analyses should take the cheaper costs of biosimilars into account.

One commentator described that from their experience, the test cost was negligible compared with the annual cost of infliximab per patient. Costs savings can be achieved by dose de-escalation, drug optimisation and better outcomes for patients. They considered that using the tests may also reduce future hospital admissions.

General comments

One specialist commentator explained that, for the RIDASCREEN Anti‑IFX Antibodies test kit, they run singlicate patient samples but at 2 different dilutions (1:25 and 1:200), so the maximum number of patient samples per plate in this instance is 40. Another commentator advised that sample dilutions for infliximab measurement may later affect the number of samples that will need repeat testing for absolute values.

One commentator highlighted that routine use of the RIDASCREEN tests is subject to the same limitations noted in the NICE diagnostics guidance on therapeutic monitoring of TNF-alpha inhibitors in Crohn's disease, and further research is needed to address this issue.

Two commentators highlighted that a full analytical verification of the RIDASCREEN tests on existing microplate readers would be needed before offering a clinical service. One added that laboratories using an automated or semi-automated ELISA platform would need to ensure acceptability of the in-built software, in use and implement sufficient quality control procedures and validation criteria for ongoing quality assurance. One commentator further advised that manual methods would not be recommended for routine services.