Blood-test monitoring strategies for patients established on immune-suppressing drugs: a mixed-methods study

The optimal monitoring strategy for adults with inflammatory conditions established on immune-suppressing treatment is unknown.

To ascertain the optimum strategy of monitoring blood tests for patients established on immune-suppressing drugs. The objectives were to: investigate the association between different strategies of monitoring and detection of pre-specified haematological, hepatic or renal side effects identify prognostic factors associated with pre-specified haematological, hepatic or renal side effects develop prediction models for pre-specified haematological, hepatic or renal side effects evaluate cost-effectiveness of testing strategies, and explore patients' and clinicians' experiences of current testing strategies, and acceptability of recommended strategies.

The study undertook systematic review, prognostic modelling, health economic modelling of alternate monitoring intervals, and qualitative study.

Primary and secondary care.

Data from the UK's Clinical Practice Research Datalink and British Association of Dermatologists Biologics Register were used for prediction modelling. Patients with inflammatory conditions treated with immune-suppressing drugs and health professionals were interviewed.

There were differences in testing strategies used, outcome definitions and follow-up precluding evidence synthesis. However, clinically significant pre-specified side effects were found to be uncommon or very uncommon during long-term treatment. There was high-quality evidence for elevated liver enzymes and folate non-supplementation being prognostic for methotrexate hepatotoxicity, mercaptopurine being prognostic for thiopurine hepatotoxicity and myelotoxicity; poor thiopurine metabolizer being prognostic for cytopenia; and elevated liver enzymes being prognostic for hepatotoxicity due to anti-tumour necrosis factor-alpha. Five prediction models were developed and validated. They had excellent performance characteristics. The calibration slope (95% confidence interval) in validation cohort was 0.94 (0.85 to 1.02) for methotrexate, 1.10 (0.84 to 1.36) for thiopurines, 0.91 (0.74 to 1.07) for leflunomide, 1.19 (0.96 to 1.43) for sulfasalazine, and 0.90 (0.61 to 1.19) for 5-aminosalicylates. Prediction modelling could not be undertaken for anti-tumour necrosis factor-alpha and mycophenolate due to few outcomes. Health economic modelling found it was cost-effective to increase the interval between monitoring blood tests except in very high-risk scenarios for some drugs (e.g. biennial testing in leflunomide). Eighteen patients from a range of risk profiles, and 13 health professionals were interviewed remotely. They found it acceptable to increase the interval between monitoring blood tests from 3-monthly to 6-monthly or annually depending on their risk profiles.

It was not possible to synthesise the association between different strategies of testing and detection of pre-specified side effects. A meta-analysis of prognostic factors was not possible.

Clinically significant myelotoxicity, hepatotoxicity and nephrotoxicity are uncommon during long-term immune-suppressing drug treatment and were predicted using readily available information. Extending the intervals between monitoring blood tests was more cost-effective than the current practice. It was acceptable to patients and health professionals to increase the interval between monitoring blood tests.

Test clinical effectiveness, undertake knowledge mobilisation, validate in non-UK populations. Research on other biologics.

The study is registered as PROSPERO CRD42020208042, CRD42020208042 and the interview studies were registered on ClinicalTrials.gov NCT05458531.

This award was funded by the National Institute for Health and Care Research (NIHR) Health Technology Assessment programme (NIHR award ref: NIHR130580) and is published in full in Health Technology Assessment; Vol. 30, No. 64. See the NIHR Funding and Awards website for further award information.