Test accuracy of drug and antibody assays for predicting response to antitumour necrosis factor treatment in Crohn’s disease: a systematic review and meta-analysis

Objective To present meta-analytic test accuracy estimates of levels of antitumour necrosis factor (anti-TNF) and antibodies to anti-TNF to predict loss of response or lack of regaining response in patients with anti-TNF managed Crohn’s disease. Methods MEDLINE, Embase, the Cochrane Library and Science Citation Index were searched from inception to October/November 2014 to identify studies which reported 2×2 table data of the association between levels of anti-TNF or its antibodies and clinical status. Hierarchical/bivariate meta-analysis was undertaken with the user-written ‘metandi’ package of Harbord and Whiting using Stata V.11 software, for infliximab, adalimumab, anti-infliximab and anti-adalimumab levels as predictors of loss of response. Prevalence of Crohn’s disease in included studies was meta-analysed using a random effects model in MetaAnalyst software to calculate positive and negative predictive values. The search was updated in January 2017. Results 31 studies were included in the review. Studies were heterogeneous with respect to the type of test used, criteria for establishing response and loss of response, population examined and results. Meta-analytic summary point estimates for sensitivity and specificity were 65.7% and 80.6% for infliximab trough levels and 56% and 79% for antibodies to infliximab, respectively. Pooled results for adalimumab trough levels and antibodies to adalimumab were similar. Pooled positive and negative predictive values ranged between 70% and 80% implying that between 20% and 30% of both positive and negative test results may be incorrect in predicting loss of response. Conclusion The available evidence suggests that these tests have modest predictive accuracy for clinical status; direct test accuracy comparisons in the same population are needed. More clinical trial evidence from test–treat studies is required before the clinical utility of the tests can be reliably evaluated.


Plain English Summary
Crohn"s disease is an uncommon long term disease involving painful and damaging inflammation of the gut lining. Damage can cause bloody stools, development of very narrow sections along the gut (strictures), and the formation of abnormal channels (fistulas) between different regions of the gut or between gut and body surface or between gut and nearby organs. Particularly distressing fistulas may occur between intestine and vagina in female patients. During a patient"s life the severity of Crohn"s disease fluctuates between remission (no symptoms) and relapse (active disease) and treatments aim to induce and maintain remission. Tumour necrosis factor (TNF) has been identified as a molecule important in the development of inflammation in Crohn"s disease. Medicines called anti-TNF agents have been developed that counteract the action of TNF and have been found to benefit Crohn"s disease patients; they are by far the most expensive medicines used for Crohn"s disease and, like all Crohn"s disease medicines, for some patients they are associated with unwanted side effects.
Unfortunately many patients eventually develop resistance to anti-TNF agents and remission fails.
One reason for failure is that some patients develop antibodies to anti-TNFs so that the amount of drug in the patient"s blood decreases below levels that are effective. Test kits have been developed and marketed that allow estimation of the levels of anti-TNF and of antibodies to anti-TNF in a patient"s blood sample. This information can aid clinicians and patients to decide on the best course of future treatment, and may help avoid continued use of expensive but ineffective medicine. The present project aims to examine evidence about the clinical and cost effectiveness of test kits. The current report will allow NICE to make recommendations about how well the kits work and whether the benefits are worth the cost of the tests for use in the NHS in England and Wales. The assessment will consider both potential for improvement in patients" symptoms associated with use of the tests and the cost of the tests.

Decision problem
The current report being undertaken for the NICE Diagnostics Assessment Programme examines the clinical and cost effectiveness of ELISA tests (LISA-TRACKER EISA kits, TNFα-Blocker ELISA kits, and Promonitor ELISA kits) for measuring patient blood levels of anti-TNF agents (Infliximab and Adalimumab; also known as TNF inhibitors) and of antibodies to these agents (i.e., anti-drug antibody levels, ADAbs) in people with Crohn"s disease whose disease responds to treatment with TNF inhibitor or who experience secondary loss of response during a maintenance course of TNF inhibitor therapy.

Anti-tumour necrosis factor alpha (anti-TNFα) agents
TNFα is a small cell-signalling protein (cytokine) involved in inflammatory responses primarily by influencing regulation of various effector cells of the immune system. TNFα has been shown to have a role in several inflammatory diseases including Crohn"s disease, ulcerative colitis, rheumatoid arthritis and ankylosing spondylitis. Therapies have been developed that are directed at blocking the actions of TNFα and thereby reducing inflammation. Such anti-tumour necrosis factor alpha (anti-TNFα) agents bind to cell surface TNFα and free TNFα and block its activity. Blocking of TNFα with anti-TNF drugs has been shown to successfully reduce the inflammation for some patients with inflammatory diseases including Crohn"s disease. As these drugs are expensive and can cause potentially serious adverse effects, in England, they are generally used as second or third line treatment in the management of Crohn"s disease and are employed when other drugs have not worked or have caused major side effects, and when surgery is not considered the appropriate treatment option. The anti-TNF agents recommended by NICE for the treatment of Crohn"s disease are infliximab (Remicade®, Schering-Plough) and adalimumab (Humira®, Abbott Laboratories). These are monoclonal antibodies introduced into the human body to bind and block TNFα. They are classed as monoclonal antibodies because they are derived from genetically engineered immune cells, which are all daughters of a single parent cell, so that in culture they generate and secrete antibodies that are all of identical structure and affinity for TNFα.

Infliximab
Infliximab is a chimeric (mouse-human) monoclonal antibody. It is said to be chimeric because the genetic code determining its amino acid sequences is partly derived from the mouse genome and partly from the human genome. Infliximab belongs to the IgG1 (immunoglobulin gamma type 1) group of antibody molecules ( Figure 1). It should be born in mind that IgG1 molecules are globular (not linear as in the diagram) and that they are glycoproteins that have carbohydrate chains attached (not shown in Figure 1). As infliximab is generated from cultured mouse cells, the carbohydrate part of the molecules corresponds to that of mouse rather than human glycoproteins.  Infliximab was the first anti-TNF agent that was approved and licenced for treating severe active

The molecule comprises two heavy chains (HC) and two light chains (LC); the HCs are joined together across disulphide bonds (S-S) and each LC is joined to a HC by S-S bonding. The LC and HC have a variable region (different from all other antibodies) at the amino (NH
Crohn"s disease and active fistulising Crohn"s disease in adults and children over the age of six. It is administered intravenously over 1-2 hours. Details of the licenced indication are given in Appendix 1.
Side effects of infliximab include:

Adalimumab
Adalimumab is a human IgG1 monoclonal antibody with Kappa light chains. It consists of purely human antibody polypeptide domains ( Figure 1). However, as adalimumab is generated from cultured Chinese hamster ovary cells, the carbohydrate part of the molecules corresponds to that of hamster rather than human glycoproteins. Adalimumab is a more recent anti-TNFα therapy that was approved for treating Crohn"s disease in adults only. It is administered as a subcutaneous injection by a doctor or nurse or can be self-injected by the patient or a family member. Details of the licenced indication are given in Appendix 1.
Side effects of adalimumab include: Many of the side effects are reversible if the drug is stopped.
They estimate the following molecules in patient blood sera:  Anti-adalimumab antibodies

Rationale
In some patients an initial or maintained response to anti-TNF therapy may disappear. This has been observed for all conditions in which these therapies have been used. The reasons for response failure may be various and are not fully understood, however loss of response has often been found to be associated with the generation of immune responses to the anti-TNF agent itself. In particular the patient may generate antibodies directed against the anti-TNF agent, these will bind to the administered anti-TNF agent, nullify its effectiveness and hasten its clearance from the circulation.
These effects may explain or partially explain the phenomena of loss of response experienced by some patients. The generation of antibodies against infliximab may not be surprising since about 30% of the molecule has mouse identity. Adalimumab, although termed a fully humanised antibody, has potential to be antigenic since its carbohydrate moieties are mouse derived and because its binding site for anti-TNF is unique and could, according to the network hypothesis of Jerne, 1 lead to generation of antibodies directed against this "idiotypic" region of the drug.
Other patients may respond well to an induction phase of treatment with a TNF inhibitor. However, these patients may lose response in the future, may benefit from optimising dosing or may require review after 12 months of treatment with a TNF inhibitor. Management of responders could benefit from knowing levels of anti-TNF drug and anti-drug antibodies in the patients" blood. Reagents are added with wash steps between additions. The final step involves quantifying the amount of a peroxidase label in the titre well, this amount being proportional to the amount of anti-TNF or ADAb in the patient"s sample or in the calibrator standard.
The amount of peroxidase present in the well is quantified using a timed incubation with excess substrates (hydrogen peroxide + 3,3",5,5"-tetramethylbenzidine). Peroxidase catalyses the following reaction: Tetramethylbenzidine + hydrogen peroxide → chromogen + water The incubation is stopped after an appropriate time by the addition of acid and the accumulated chromogen quantified by measuring optical density with a spectrophotometer.
The reagents used for coating the microtitre plate wells and the reagents used in subsequent steps of the assay procedure differ from each other according to manufacturer. The LISA-TRACKER assays for Infliximab and for Adalimumab are illustrated in Figure 2.   Serum samples from patients may contain soluble TNFα receptors; these could compete with anti-TNF for the immobilised TNFα on the well plate and may potentially interfere with the assay. The   assay quantifies free anti-TNF. Samples may contain anti-TNF bound to antibodies to anti-TNF,   especially in patients who have lost a response to treatment. These anti-TNF-antibody complexes will be washed away at the first wash step leaving only free anti-TNF bound to immobilised TNFα. The amount of anti-TNF lost at the wash step is likely to vary between patients and is unknown; the practical implications of this are uncertain.
TNFα-Blocker and Promonitor differ from LISA-TRACKER in employing a single step and one reagent for detecting well-bound anti-TNF, rather than two steps (C and D in Figure 2) and two reagents. Table 1 summarises the information currently available describing the principle of these assays. These are available as commercial kits and several "in house" methods are mentioned in the literature.
The majority of ELISAs only quantitatively measure "free" anti-TNF and "free" ADAbs and it is acknowledged that the level of the unmeasured "bound" anti-TNF and of "bound" ADAb may vary considerably between patients. The Immundiagnostik assays give semi-quantitative measurement of "total" ADAbs. Thus for some patient samples there is an unknown and unmeasured amount of anti-TNF and of ADAb present, in addition to the measured "free" levels.
Below the LISA-TRACKER methods are reported and differences to TNFα-Blocker and Promonitor are described. The LISA-TRACKER assays for antibodies to infliximab and to adalimumab are illustrated in Figure 3.  TNFα-Blocker and Promonitor differ from LISA-TRACKER in employing a single step and reagent for detecting well-bound anti-TNF rather than two steps (C and D in Figure 2) and two reagents.  There are no "gold standard" assays for measuring anti-TNF agents or for antibodies to anti-TNF agents which might provide a robust basis for comparisons between the performance of different assays. According to the US Medical Insurance assessments "candidate" gold standards have been insufficiently investigated to establish any as a gold standard, and according to Steenholdt et al.
(2013) 2 it is unknown if and how these different assays compare. [3][4][5][6][7] There appear to be four types of assay for measuring the levels of anti-TNF drugs and the levels of antibodies against TNF inhibitors in patient blood sera. which differ fundamentally from each other.
In addition to ELISAs (solid phase assays) these are: (a) Radioimmunoassays (RIA)liquid phase. They appear to measure total anti-TNF and total ADAb (probably as long as the ADAb light chain is lambda class). These RIAs use 125 iodine-labelled human TNFα and 125 iodine-labelled anti-TNFs. In these assays the patient"s sample is mixed with a solution containing a fixed amount of 125 iodine-labelled TNFα or 125 iodine-labelled anti-TNF further antibody (e.g., rabbit anti-human immunoglobulin λ-chain) which promotes the formation of immune complexes which are pelleted by centrifugation. Radio-iodine in the pellet is quantified in a gamma-counter. Characteristics of these assays include: i) radio-labelled reagents do not store indefinitely (125 iodine decays with a half-life of 59 days), ii) the laboratory needs to be equipped for handling hazardous (radioactive) material, iii) some staff training may be necessary, and iv) the laboratory requires a gamma counter (preferably automated for high throughput). In measuring ADAb the patient sample is subjected to an acid step which "unbinds" bound anti-TNF and ADAb so that all anti-TNF and ADAb are "free"; after neutralisation the sample is incubated with fluorescent-dye-labelled anti-TNF (D*) as described above. Some D* will form immune complexes with the sample ADAbs (D*-ADAb complexes) and these have a different mobility on SE-HPLC than D* thus the mobility of some of the D* is shifted, the proportion of D* shifted is dependent on the level of ADAb in the sample.

Timing and use of ELISAs
Scoping searches indicate that the anti-TNF and ADAb assays are most frequently administered just before the next administration of the anti-TNF agent. This is said to allow measurement of a "trough" level of anti-TNF and may have been adopted when ELISAs are used so as to minimise effects from the presence of anti-TNF-ADAb immune-complexes in samples. For patients whose response to therapy has waned, the results of the tests are frequently dichotomised using a cut off assay result.
Thus, on the basis of anti-TNF assays patients are classified as having therapeutic levels of anti-TNF or sub-therapeutic levels, and on the basis of ADAb assay results they are classified as having clinically significant levels of ADAbs or insignificant levels. Such classifications yield four categories of patient for whom different explanations of failed response are possible. Algorithms have been developed prescribing treatment pathways and / or further diagnostic tests (e.g., colonoscopy) based on such classification.

Target condition / indication
Anti-TNFα is commonly given to people with inflammatory bowel disease (IBD) including Crohn"s disease. The general background and treatment pathway for Crohn"s disease is summarised below.

Crohn"s disease
Crohn"s disease is a chronic fluctuating episodic inflammatory condition of the digestive tract; it is uncommon and is currently estimated to affect about 115,000 people in the UK. 8 Together with ulcerative colitis it comprises conditions classed as inflammatory bowel disease (IBD).

Aetiology and pathology
Crohn"s disease can affect adults, adolescents or children. Crohn"s disease manifests itself mainly during late adolescence or early adulthood. The first onset most commonly occurs between the ages of 16 and 30 with a second peak between the ages of 60 and 80. Women are slightly more frequently affected than men but in children it is seen more often in boys than in girls. The condition has highest prevalence among Jewish people with European descent.
Crohn"s disease follows a pattern of acute disease interspersed with periods of remission. Crohn"s disease causes inflammation of the lining of the digestive tract which, depending on the individual, occurs at any location from the mouth to the rectum, but most commonly affects the terminal ileum (35%) or the ileocaecal region (40%). Within individuals the disease location is fairly stable.
The main symptoms of Crohn"s disease are dependent on disease location and include chronic or nocturnal diarrhoea, abdominal pain, anal lesions, rectal bleeding and weight loss. Clinical signs include pallor, cachexia, abdominal mass or tenderness, or perianal fissures, fistulas or abscesses.
Systemic symptoms include malaise, anorexia or fever. [9][10][11] Extra-intestinal symptoms related to intestinal inflammation include spondyloarthritis (inflammatory rheumatic diseases which cause arthritis, most commonly ankylosing spondylitis), cutaneous manifestations or ocular inflammation. 11 In children, growth failure may be the primary manifestation of Crohn"s disease. 12 Classification of Crohn"s disease disease states and measurement of disease activity Several classification systems of Crohn"s disease have been proposed. The Montreal 13 and Vienna 14 systems are summarised in Tables 3 and 4.  values above 450 are associated with very severe disease. 16 Some investigators have arbitrarily labelled CDAI scores of 150-219 as mildly active disease and scores of 220 to 450 as moderately active disease. 15 The CDAI has been criticised for having limitations since it fails to encompass aspects of quality of life such as psychological, social, sexual wellbeing and occupational functioning. A patient with a low CDAI score may still be severely limited by these factors. 17 Substantial variability exists when different observers review the same case histories and calculate the CDAI score, although this can be reduced after discussion and education about the terminology. The calculation is based in part on a daily diary kept by the patient for seven days before the evaluation. In practice some investigators and study coordinators assist the patient to complete the diary retrospectively at the time of an evaluation visit; there is no information on the prevalence of this practice. The CDAI score may be low in patients whose primary symptom is drainage of enterocutaneous fistulas, presumably because the presence of an actively draining fistula contributes only 20 points to the score. The CDAI is therefore not an appropriate instrument for assessing the activity of draining abdominal or perianal enterocutaneous fistulas. The CDAI has been criticised for giving too much weight to "general wellbeing" and "intensity of abdominal pain" because these are relatively subjective items. However these aspects of disease are important to patients. 18 A paediatric CDAI has been developed. 18,19 The HBI or Simple Index is a modified/simplified version of the adult CDAI. It uses a single day"s reading for diary entries and excludes three variables (body weight, haematocrit and use of drugs for diarrhoea). Code values are added together rather than summing the products of code values and coefficients. Scores range from 0 to 20. The CDAI can be predicted reasonably well from the HBI. 20 Other instruments derived from the CDAI are: the Cape Town Index (CTI), which includes parameters on subjective symptoms, physician clinical findings and laboratory data; the three-variable version of the CDAI used for survey research; and the Van Hees Index (VHI), which includes laboratory parameters, sex (male or female) and seven clinical features and excludes subjective patient related items such as well-being and pain. Anti-TNF monitoring in Crohn"s disease Crohn"s disease is associated with elevated levels of the immune-regulatory protein TNFα. The reasons for this elevation in Crohn"s disease is still largely unknown. Anti-TNF therapies have been shown to block the action of TNFα and to improve outcomes for some patients. Patients receive anti-TNF therapy after failed attempts to improve the condition with first line glucocorticosteroids, 5aminosalicylates, antibiotics and second line treatment (e.g., methothrexate). These patients have severe symptoms and they are at the end of the patient pathway with the only alternative option being surgery.
Like other treatment regimens anti-TNF treatment aims to induce remission (induction therapy) and prevent relapse (maintenance therapy). However failure to induce a response and relapse or loss of response are common. Approximately 10% of patients per year loose response to anti-TNF drugs. 24 The annual risk of response loss per patient has been estimated at about 13%. 25 During "episodic" infliximab therapy about 37-61% lose response. 26 Mechanisms of loss of response to anti-TNF agents and of failure to respond are still mainly unclear, however the fact that some patients generate immune responses to therapy offers one plausible contributory explanation. However other pharmacodynamics mechanisms may reduce the drug below therapeutic levels, furthermore there may be alternative secondary pathways of inflammation independent of TNFα that operate in some patients rendering anti-TNF of little use.
During scheduled infliximab therapy the incidence of antibodies is 6-16%. 27,28 Anti-TNF antibody formation in patients treated with Infliximab has been shown to be as high as 37-61%. 29 Concomitant immunosuppressive therapy may decrease the formation of ADAbs. 26,27,29 Candidate risk factors for ADAb production include hereditary predisposition, a dysfunctional immune system, experience of infection(s) that trigger an abnormal response, smoking, environmental factors such as sanitation.
The ELISA assays could be used in good responders (i.e., those responding to initial induction course of anti-TNF treatment) as well as in patients with secondary loss of response (i.e., those initially responding to anti-TNF treatment but loosing this response over time). The use of these technologies provides a clinician with potentially useful information that may guide individual patient"s future treatment. Such information may aid in anticipating the loss of response in responders, while for nonresponders such analyses may help in estimating the likelihood of various candidate reasons for primary non-response or secondary loss of response. For example in non-responders with low levels of drug and high levels of ADAbs the loss or lack of response may be surmised to be due to rapid  Once remission has been achieved, maintenance therapy can be considered following assessment of the course and extent of Crohn"s disease, effectiveness and tolerance of previous treatments, presence of biological or endoscopic signs of inflammation, and potential for complications. Surgery should be considered as an alternative to medical treatment early in the course of the disease for people (adults, children, and young people) whose disease is limited to the distal ileum or have growth impairment despite optimal medical treatment and/or refractory disease (children and young people).

Maintenance of remission
People with Crohn"s disease in remission can be managed with or without maintenance treatment. The options for maintenance therapy (including treatment or no treatment) need to be discussed with patients, their parents, and/or carers. The discussion should include risk of inflammatory exacerbations (with and without drug treatment) and the potential side effects of drug treatment.
People who decline to receive maintenance treatment should agree with follow-up plans (e.g., frequency and duration of visits) and receive information on symptoms related to relapse (e.g., unintended weight loss, abdominal pain, diarrhoea, general ill-health) to ensure timely consultations with their healthcare professional.
People with Crohn"s disease in remission who choose to receive maintenance therapy may be offered azathioprine or mercaptopurine monotherapy if their remission was induced using a conventional glucocorticosteroid or budesonide. Methotrexate can be offered to people whose remission was induced by methotrexate or people who did not tolerate azathioprine or mercaptopurine for maintenance therapy or those who have contraindications to azathioprine or mercaptopurine.
Treatment with 5-ASA can be recommended to maintain remission after surgery.
If remission has been achieved with anti-TNF medication, then maintenance with anti-TNF with or without combination with another immunomodulator can be recommended. Continuation of treatment with infliximab or adalimumab during remission is advised only if there is evidence of ongoing active disease given clinical symptoms, biological markers, including endoscopy if necessary. The balance between harms and benefits of ongoing treatment should be taken into account. People who relapse after treatment is stopped have the option to start this treatment again.

Decision questions
The decision questions for this project are shown in the box below:

Objectives
Given these decision questions the four main objectives for this report are: A) To provide a technical description, and (where evidence allows) an evaluation, of the listed intervention tests used for Crohn"s disease in therapeutic monitoring of TNF inhibitors (infliximab and adalimumab) and their respective antibodies. This will include what the assays measure and the mechanisms of the assays.
In addition, published studies which include a comparison (including relative test performance) of two or more intervention tests, or which compare an intervention test with a test method which can be used to perform a linked evidence assessment will be reviewed and critiqued. Data submitted by the manufacturers will be used to supplement published studies if deemed of sufficient detail and quality. Where evidence exists on the comparison of standard care with other test assays used in conjunction with an algorithm, this will be assessed and critiqued and test performance will be compared with that of the study interventions (LISA-TRACKER ELISA kits, TNFα-Blocker ELISA kits, and Promonitor ELISA kits) (see Objective A). D) To assess the cost-effectiveness of employing anti-TNF monitoring with LISA-TRACKER ELISA kits, TNFα-Blocker ELISA kits, and Promonitor ELISA kits in patients with Crohn"s disease compared with standard care (no anti-TNF monitoring). Where direct evidence is unavailable for this comparison, or where such a comparison is not well supported with evidence, a linked approach to evidence will be considered (see Objective C above) in which evidence of clinical effectiveness is taken from studies using alternative test methodology and an assessment is made of the relative performance this methodology relative to the intervention assays.

Methods for assessing clinical effectiveness
Systematic review methods will follow the principles outlined in the Centre for Reviews and Dissemination (CRD) guidance for undertaking reviews in health care 38  All retrieved papers will be screened for potential inclusion.
The search strategy will comprise the following main elements: The reference lists of included studies and relevant review articles will be checked. Citation searches of selected included studies will be undertaken using Scopus. Identified references will be downloaded in Endnote X7 software. Included papers will be checked for errata using PubMed.

Inclusion and exclusion of relevant studies
Inclusion of relevant studies to address Objective A Detailed information will be sought from manufacturers regarding mechanisms and reactants (in particular specificities and properties of antibodies and other reagents) employed in ELISA tests and radioimmunoassay, mobility shift assays and cell reporter tests (if used for a linked evidence approach).
In addition published studies which describe the intervention tests and tests used for a linked evidence approach will be identified. Those providing useful information about test mechanisms that is different or additional to that supplied by manufacturers of tests will be included. Assessment of inclusion will be based on the judgement of two reviewers.
Studies which compare test performance of two or more tests will be included either if they compare two or more intervention tests, or compare an intervention test with a test method which can be used to perform a linked evidence assessment.
All study designs will be considered for inclusion.
Inclusion criteria for studies to address Objective B Studies that report an algorithm with the use of one of the intervention tests for the management of patients with Crohn"s disease following measurement of serum levels of anti-TNF drug and anti-drug antibodies (infliximab or adalimumab). All study designs will be considered for inclusion.
Inclusion criteria for studies to address Objective C Studies that satisfy the following criteria will be included:

Population
Crohn"s disease patients (adults and children) receiving infliximab or adalimumab. If the evidence on Crohn"s disease patients is limited, mixed patient groups containing Crohn"s disease and ulcerative colitis patients will be included even if results are not reported separately. The limitations following from this will be discussed.

Intervention
Use of LISA-TRACKER ELISA kits, TNFα-Blocker ELISA kits, and Promonitor ELISA kits to estimate plasma or sera levels of anti-TNF agents and / or of ADAbs in which test results are employed in conjunction with a treatment algorithm (Table 6). Other assay methods will be considered should a linked evidence approach be adopted (Table 6).

Comparator
Standard care (Treatment decisions made on clinical judgement without measuring levels of TNF inhibitor and antibodies to TNF inhibitors).

Outcome
Any patient outcome (e.g., CDAI score based response rate, any measure of

Study design
All study designs will be considered for inclusion.
Healthcare setting Secondary and tertiary care.
Meeting abstracts will be included if they provide sufficient data on type of ELISA assay, patient group, algorithm, measurements from assays and clinical outcomes. For Objective C test methods that are not included as an intervention but have evidence comparing it to an intervention test and evidence reporting clinical outcomes, should be included for the purpose of performing linked evidence modelling only (including: radioimmunoassays, cell reporter assays, liquid-phase mobility shift assays and in-house ELISAs).

Review strategy
The general principles recommended in the PRISMA statement will be considered. 41 Records rejected at full text stage and reasons for exclusion will be documented. Two reviewers will independently screen the titles and abstracts of all records identified by the searches and discrepancies will be resolved through discussion. Disagreement will be resolved by retrieval of the full publication and consensus agreement. Full copies of all studies deemed potentially relevant, will be obtained and two reviewers will independently assess these for inclusion; any disagreements will be resolved by consensus or discussion with a third reviewer.

Data extraction strategy
Data will be extracted by one reviewer, using a piloted, data extraction form. A second reviewer will check the extracted data and any disagreements will be resolved by consensus or discussion with a third reviewer. Examples of data extraction sheets for patient-based and diagnostic accuracy studies are provided in Appendix 4.

Quality assessment strategy
Where appropriate, the quality of diagnostic accuracy studies will be assessed using QUADAS-2 (see Appendix 5). 42 As a broad range of study designs have been identified in the scoping searches, the use of a single checklist, in contrast to individual checklists for each study design, is considered appropriate. The Downs and Black checklist 43 will therefore be used to assess the quality of nonrandomised studies meeting the inclusion criteria (see Appendix 5). This 27-item checklist provides both an overall score for study quality and a profile of scores not only for the quality of reporting, internal validity (bias and confounding) and power, but also for external validity. RCTs will be quality appraised using the Cochrane risk of bias tool (see Appendix 5). 44 The results of the quality assessment will provide an overall description of the quality of the included studies and will provide a transparent method of recommendation for design of any future studies. Quality assessment will be undertaken by one reviewer and checked by a second reviewer, any disagreements will be resolved by a third reviewer through discussion.

Methods of analysis/synthesis
Objective A Narrative descriptions of tests in tables and texts will be undertaken.

Objective B
Algorithms will be narratively described and compared to the algorithm used in the TAXIT study (for good responders), 36 and the algorithm adapted from Scott and Lichtenstein (2014) (for secondary loss of response). 37 Non-compliant patients may be considered additionally in the algorithms. Time of testing, sequence of testing (drug and antibodies), sequence of analysis as well as thresholds used in the algorithms will be considered to address the research questions.

Objective C
Depending on the available evidence, analyses will be stratified according to the type of ELISA assay, type of drug (infliximab or adalimumab) and patient group (patients with secondary loss of response and patients with good response to anti-TNF treatment).
Study, treatment, population, and outcome characteristics will be summarised and compared qualitatively and, where possible, quantitatively in text, graphically and in evidence tables. Pooling studies results by meta-analysis will be considered. Where meta-analysis is considered unsuitable for some or all of the data identified (e.g., due to the heterogeneity and/or small numbers of studies), we will employ a narrative synthesis. Typically, this will involve the use of text, graphs and tables (as appropriate) to summarise data. These will allow the reader to consider any outcomes in the light of differences in study designs and potential sources of bias for each of the studies being reviewed.
Studies will be organised by objective addressed. A detailed commentary on the major methodological problems or biases that affected the studies will also be included, together with a description of how this may have affected the individual study results.
For Objective C we aim to identify studies that compare treatment decisions made on clinical judgement without measuring levels of TNF inhibitor and antibodies to TNF inhibitors with treatment decisions based on measurement of TNF inhibitor and antibodies to TNF inhibitors. We will consider using a linked-evidence approach 45 in which studies report patient management informed by measurement of anti-TNF and antibodies by other methods (e.g., radioimmunoassay, liquid-phase mobility shift assay, in-house ELISAs); this will require an assessment of evidence relating to the comparable performance of ELISA assays with radioimmunoassay, liquid-phase mobility shift assays and in-house ELISAs.
In studies where an ELISA has been used but there is no comparator arm, or the comparator arm is a convenience sample (retrospective/historical population), outcomes will be listed and appraised.
Time of testing, sequence of testing (drug and antibodies) and sequence of analysis will be considered to address the research questions.

Identifying and reviewing published cost-effectiveness studies
Published cost-effectiveness studies will be reviewed. All papers which present findings on the costs and outcomes of LISA-TRACKER ELISA kits, TNFα-Blocker ELISA kits, and Promonitor ELISA kits for measuring levels of TNF inhibitors and of anti-drug antibodies will be reviewed in detail.
Information on assay procedures additional to ELISA methods will be sought for the purposes of providing data for a linked approach to evidence synthesis should this be required. Additional searches will be performed where necessary to identify other relevant information to support the development of an economic model for this project, these may be directed towards -costs, utilities and transition probabilities as required.

Search strategy and data extraction
Data will be extracted by one reviewer and checked by a second, using a standardised data extraction form for the economic studies; this will be developed to summarise the main characteristics of the studies and to capture useful data that can inform the economic model. Any discrepancies will be resolved by discussion. If this is not feasible, a third reviewer will be consulted.
The quality of any full economic evaluation studies will be assessed using the CHEERS checklist (see Appendix 5). 46 Any studies containing an economic model will be further assessed using the framework for the quality assessment of decision analytic modelling (see Appendix 5). 47

Model structure, time horizon and transition probabilities
In developing the economic model we will consult the previous Health Technology Assessment report (HTA) conducted by Dretzke and colleagues (2011). 48 The main aim of this HTA report was to assess the cost-effectiveness of anti-TNFs in the management of moderate-to-severe Crohn"s disease in the Promonitor ELISA kits) will be compared with standard care in the following populations:  In patients with secondary loss of response to anti-TNF treatment  In patients who respond well to anti-TNF treatment The following comparisons will be made where possible:  Concurrent versus reflex testing  Testing conducted every 3 to 4 months versus testing conducted at 3 to 4 months then yearly (in patients who respond well to anti-TNF treatment) If data permits, we will compare the different LISA-TRACKER ELISA kits, TNFα-Blocker ELISA kits, and Promonitor ELISA kits with each other. In the absence of sufficient clinical data for specific ELISAs we will assume equal assay performance and compare ELISAs on the basis of cost only.
If data permits, a linked evidence approach will be adopted to compare LISA-TRACKER ELISA kits, TNFα-Blocker ELISA kits, and Promonitor ELISA kits with standard care in which clinical outcomes for the intervention arm are taken from studies in which the assay procedure was not one of the intervention assays; this will involve an assessment of the comparability of LISA-TRACKER ELISA kits, TNFα-Blocker ELISA kits, or Promonitor ELISA kits performance with that of the alternative procedure.
The model will have a one-year time horizon in line with the previous HTA report 48 and other studies we have found during our initial scoping search (e.g., Velayos et al., 2013). 49 It is anticipated that information from the clinical effectiveness analyses will help inform the probabilities for each of the clinical pathways. Sensitivity analyses will be conducted in areas of uncertainty.

Resource use and costs
Resource use and costs will be estimated in line with the DAP programme manual. Information on resource use and costs associated with the different patient pathways (e.g., comparing clinical pathways followed when LISA-TRACKER ELISA kits, TNFα-Blocker ELISA kits, or Promonitor ELISA kits are employed, versus standard care pathway etc.) will be collected from systematic reviews of the literature, discussions with individual manufacturers and hospitals and if need be, by eliciting expert clinical advice. Any remaining gaps for resource use parameters will be filled by assumptions made by the research team.
Unit costs data will be based on national data were possible. For the different LISA-TRACKER ELISA kits, TNFα-Blocker ELISA kits, and Promonitor ELISA kits, costs will be from published list prices from the NHS supply chain, from the NHS reference costs, 50 or discussions with individual manufacturers or hospitals. Costs of consultations with secondary care staff will be drawn from Unit Costs of Health and Social Care 51 and drug costs will be obtained from the British National Formulary. 34

Health outcomes
Health outcomes and utility data will be derived from the literature review including the previous HTA report and other sources. If direct measurements of utility or choice-based multi-attribute utility scales (such as the EQ-5D or SF-6D) suitable for calculation of QALYs for the economic model are not reported, we may need to use one of the algorithms for mapping from a clinical measure (e.g. CDAI) to a measure of utility. If insufficient information is available for utilities it may have to be elicited from an expert clinical panel or by assumptions made by the research team.

Cost-effectiveness analysis
The results of the cost-effectiveness analysis will be presented as an incremental cost per QALY gained for LISA-TRACKER ELISA kits, TNFα-Blocker ELISA kits, and Promonitor ELISA kits compared with standard care. If the data allows us to compare LISA-TRACKER ELISA kits, TNFα-Blocker ELISA kits, and Promonitor ELISA kits with each other, then we will undertake a rank comparison and exclude any options which are dominated or extended dominated. It may be necessary, in the absence of suitable clinical outcome data, to rank ELISAs on the basis of cost only.
We will use both simple and probabilistic sensitivity analysis to explore the robustness of the results and to estimate the impact of uncertainty over model parameters. The simple sensitivity analysis will be used to assess the robustness of the results to changes in deterministic parameters such as costs, and utilities. The results from the probabilistic sensitivity analysis will be presented as costeffectiveness acceptability curves. Decisions regarding mutually exclusive alternatives will be reflected using cost-effectiveness planes and cost-effectiveness acceptability curves or frontiers.
If a longer time horizon is chosen (more than one year), both costs and outcomes will be discounted using the recommended 3.5% discount rate by HM Treasury.

Handling of information from manufacturers
All data submitted by the manufacturers/sponsors will only be considered if received by the External Assessment Group before 27 January 2015. Data arriving after this date will not be considered. Any data that meets the inclusion criteria stated will be extracted and quality assessed as stated in the methods section of this protocol.
Any "commercial in confidence" data provided by manufacturers, and specified as such, will be highlighted in blue and underlined in the assessment report (followed by company name in parentheses). Any "academic in confidence" data provided by manufacturers, and specified as such, will be highlighted in yellow and underlined in the assessment report. All confidential data used in the cost-effectiveness models will also be highlighted.

Competing interests of authors and advisors
None of the authors have any competing interests.

Appendix 1. Licenced indications for Infliximab and Adalimumab in Crohn's disease
The licence indication for Crohn"s disease detailed in the European Medicines Agency Summary of Product Characteristics (Remicade) 52 is as follows: "Adult Crohn"s disease: Remicade is indicated for:  treatment of moderately to severely active Crohn"s disease, in adult patients who have not responded despite a full and adequate course of therapy with a corticosteroid and/or an immunosuppressant; or who are intolerant to or have medical contraindications for such therapies;  treatment of fistulising, active Crohn"s disease, in adult patients who have not responded despite a full and adequate course of therapy with conventional treatment (including antibiotics, drainage and immunosuppressive therapy).

Paediatric Crohn's disease
Remicade is indicated for treatment of severe, active Crohn"s disease, in children and adolescents aged 6 to 17 years, who have not responded to conventional therapy including a corticosteroid, an immunomodulator and primary nutrition therapy; or who are intolerant to or have contraindications for such therapies. Remicade has been studied only in combination with conventional immunosuppressive therapy.

Moderately to severely active Crohn's disease
5 mg/kg given as an intravenous infusion followed by an additional 5 mg/kg infusion 2 weeks after the first infusion. If a patient does not respond after 2 doses, no additional treatment with infliximab should be given. Available data do not support further infliximab treatment, in patients not responding within 6 weeks of the initial infusion.
In responding patients, the alternative strategies for continued treatment are:  Maintenance: Additional infusions of 5 mg/kg at 6 weeks after the initial dose, followed by infusions every 8 weeks or Although comparative data are lacking, limited data in patients who initially responded to 5 mg/kg but who lost response indicate that some patients may regain response with dose escalation. Continued therapy should be carefully reconsidered in patients who show no evidence of therapeutic benefit after dose adjustment.
In Crohn"s disease, experience with re-administration if signs and symptoms of disease recur is limited and comparative data on the benefit/risk of the alternative strategies for continued treatment are lacking.

Crohn's disease (6 to 17 years)
5 mg/kg given as an intravenous infusion followed by additional 5 mg/kg infusion doses at 2 and 6 weeks after the first infusion, then every 8 weeks thereafter. Available data do not support further infliximab treatment in children and adolescents not responding within the first 10 weeks of treatment.
Some patients may require a shorter dosing interval to maintain clinical benefit, while for others a longer dosing interval may be sufficient. Patients who have had their dose interval shortened to less than 8 weeks may be at greater risk for adverse reactions. Continued therapy with a shortened interval should be carefully considered in those patients who show no evidence of additional therapeutic benefit after a change in dosing interval." The Adalimumbab licence indication for Crohn"s disease detailed in the European Medicines Agency Summary of Product Characteristics (Humira) 53 is as follows:

Paediatric Crohn's Disease
Humira is indicated for the treatment of severe active Crohn's disease in paediatric patients (from 6 years of age) who have had an inadequate response to conventional therapy including primary nutrition therapy, a corticosteroid, and an immunomodulator, or who are intolerant to or have contraindications for such therapies.
Paediatric Crohn's disease patients < 40 kg: The recommended Humira induction dose regimen for paediatric subjects with severe Crohn's disease is 40 mg at Week 0 followed by 20 mg at Week 2. In case there is a need for a more rapid response to therapy, the regimen 80 mg at Week 0 (dose can be administered as two injections in one day), 40 mg at Week 2 can be used, with the awareness that the risk for adverse events may be higher with use of the higher induction dose.
After induction treatment, the recommended dose is 20 mg every other week via subcutaneous injection. Some subjects who experience insufficient response may benefit from an increase in dosing frequency to 20 mg Humira every week.
Paediatric Crohn's disease patients ≥ 40 kg: The recommended Humira induction dose regimen for paediatric subjects with severe Crohn's disease is 80 mg at Week 0 followed by 40 mg at Week 2. In case there is a need for a more rapid response to therapy, the regimen 160 mg at Week 0 (dose can be administered as four injections in one day or as two injections per day for two consecutive days), 80 mg at Week 2 can be used, with the awareness that the risk for adverse events may be higher with use of the higher induction dose.
After induction treatment, the recommended dose is 40 mg every other week via subcutaneous injection. Some subjects who experience insufficient response may benefit from an increase in dosing frequency to 40 mg Humira every week.  Index test(s): Reference standard: Phase 2: Draw a flow diagram for the primary study

Phase 3: Risk of bias and applicability judgements
QUADAS-2 is structured so that four key domains are each rated in terms of the risk of bias and the concern regarding applicability to the review question (as stated in Phase 1). Each key domain has a set of signalling questions to help reach the judgements regarding bias and applicability.

A. Risk of bias
Describe methods of patient selection: Was a consecutive or random sample of patients enrolled?
Did the study avoid inappropriate exclusions?

Could the selection of patients have introduced bias?
Risk:

B. Concerns regarding applicability
Describe included patients (prior testing, presentation, intended use of intervention test and setting): Range of drug / antibody concentrations: Is there concern that the included patients or range of drug / antibody concentrations do not match the review question? Concern:

A. Risk of bias
Describe the intervention test and how it was conducted and interpreted: Were the number of failed results and measurement repeats reported?

Could the conduct or interpretation of the intervention test have introduced bias?
Risk:

B. Concerns regarding applicability
Describe the preparation and storage of the sample before the intervention test was applied: Is there concern that the intervention test, its conduct, or interpretation differ from the review question? Concern: