Monoclonal antibody BTT1023 targeting vascular adhesion protein 1 for treating primary sclerosing cholangitis: BUTEO single-arm Phase II trial

Phase I: dose confirmation

The run-in component of the trial incorporated a conventional 3 + 3 cohort design to confirm the therapeutic dose, with decisions regarding continuation based on toxicity and pharmacokinetic (PK) data (see Report Supplementary Material 1, Figure S1). The trial began with the recruitment of six patients, all of whom received the starting dose of 8 mg/kg. Recruitment was paused while awaiting the results of trough blood serum levels of circulating BTT1023 at visit 7 (i.e. day 50) from all six patients and until the dose-limiting toxicity (DLT) reporting period was completed for each patient [visit 10 (i.e. day 99)].

If results from the first cohorts showed an acceptable DLT rate (see later) and trough levels of BTT1023 met the stipulated success criteria, the trial was to continue into the expansion component (see Phase II: Simon’s two-stage minimax design). Acceptable trough levels were set at 3 µg/ml free-circulating BTT1023 at 8 weeks from the first infusion, which is approximately 100-fold the dissociation constant (K_d) of BTT1023 from VAP-1 and will result in target occupancy of approximately 90%. In the event that the DLT rate was acceptable but the PK values did not meet the success criteria, then the trial was to move into a conventional 3 + 3 cohort design, using escalating doses of BTT1023. In this event, the original cohort of six patients would no longer be evaluated, but a new cohort of three patients would be recruited to receive the newly identified test dose of 12 mg/kg. If there were no DLTs at visit 10 (i.e. day 99), an additional cohort of three patients would be recruited at the new test dose. If the DLT rate remained acceptable but the PK values still did not meet the success criteria, a further 3 + 3 patients would be recruited at the highest dose of 16 mg/kg. If this was found not to result in sufficient blood levels of BTT1023, then the trial would be stopped. If the PK values were found to be too high (e.g. if trough levels consistently exceeded 100 µg/ml), then there was potential to de-escalate the dose in agreement with the Data Monitoring Committee (DMC) and after regulatory approvals. The trial was to be stopped at any stage if patient safety was compromised. Individual patients would receive only one dose level. Once a confirmed dose had been established, the trial was expanded until a total of 37 patients had received treatment with that dose, including those patients who had previously received this dose during the confirmatory period. Those patients not receiving the confirmed dose were not included in the final evaluation.

Figure S1 in Report Supplementary Material 1 shows the trial design.

Phase II: Simon’s two-stage minimax design

The Phase II Simon’s two-stage minimax design centres around the hypothesis test for the trial’s primary outcome. The null hypothesis was that the response rate for PSC patients receiving BTT1023 is < 15% (P0). The alternative hypothesis was that the response rate for PSC patients receiving BTT1023 is > 30% (P1).

A response was considered to be a reduction in inflammation, indicated by reduction in serum ALP activity of ≥ 25% (a comparison of measured ALP activity from baseline to day 99).

The baseline measurements were recorded at pre-infusion visit 3, on day 1 of treatment and on day 99 (i.e. 21 days following last infusion).

The two-stage design incorporated an interim analysis of the accumulating data after 18 patients had received the trial treatment. At this point, the design required that at least 3 of 18 patients had a successful response (i.e. a reduction in ALP activity of ≥ 25%) to allow the trial to continue. If the stage 1 criterion was not met, and fewer than three evaluable patients had a successful response, then the trial was to stop early. However, if the criterion was met, then further patient recruitment was to continue until 37 evaluable patients were recruited. The final design required at least 9 of 37 patients to have a successful response to conclude that the trial treatment successfully met the trial’s primary outcome.

Patients with particularly elevated ALP activity are predicted to be more at risk of progressive disease. Therefore, these patients were selected for trial participation. Six UK academic hospital centres were involved, with four actively recruiting. (Those centres that were actively recruiting were based in Birmingham, Nottingham, Oxford and London, and the centres that failed to recruit were based in Newcastle and Cambridge.) Informed consent was obtained by appropriately trained members of the research team at each site.

A single-arm, rather than placebo-controlled, design was chosen to allow efficient enrolment of patients into the trial because of its intensive nature, as a significant chance of being allocated to the placebo group may have acted as a substantial barrier to enrolment. There was no concurrent control group as the proposed primary end point (ALP activity) was a biochemical measurement and, therefore, not open to subjective bias as a clinical assessment would be.

Given the unpredictable nature of PSC and natural variation of ALP activity, the process of screening for eligibility took place in two stages, with one screening visit 5–7 weeks before beginning infusion and another within 1 week of beginning infusion. Potential participants were eligible for the trial if their ALP activity did not vary by > 25% over this period. During screening, patients underwent routine blood screening and other non-invasive markers of liver fibrosis measured, including Mayo PSC risk score, model for end-stage liver disease (MELD) score, enhanced liver fibrosis (ELF) testing, a FibroScan (Echosens, Paris, France) and magnetic resonance cholangiopancreatography with additional LiverMultiScan^® (Perspectum^® Ltd, Oxford, UK) imaging. These assessments were repeated during treatment and in the follow-up period to determine if any change had occurred. During all seven treatment visits, patients received premedication with 10 mg of cetirizine plus 400 mg of ibuprofen orally (in the absence of any contraindications) plus 100 mg of i.v. hydrocortisone, 1–2 hours pre infusion (the last for the first three doses only). The first infusion was given over 2 hours, with a 4-hour monitoring period post infusion. Provided that no adverse reactions were seen, the infusion time dropped to 1 hour with an initial 3-hour observation period (for the second dose) and then down to 2 hours’ monitoring post infusion (for all subsequent doses). Safety investigations were completed pre and post infusion, and included haematological/biochemical testing of blood samples, electrocardiography, clinical assessment and physical examination. The full trial schema can be seen as Figure S2 in Report Supplementary Material 1. An aliquot (0.5–1.0 ml) of the BTT1023 infusion solution was taken at the end of every infusion and refrigerated. These samples could be used for analysis of BTT1023 concentration if anomalies in PK data that could be due to errors in BTT1023 preparation were observed.

Dose-confirmatory phase

The overall aim of the dose-confirmatory phase was to determine a dose of BTT1023 that provides an acceptable level of PK activity and was deemed to be safe, meeting the acceptable DLT level.

This was to be achieved by evaluating:

• the number of DLTs per dose level of BTT1023

• PK data per dose level of BTT1023.

Phase II

Dose-confirmatory phase

Phase II

Exploratory end points

The following exploratory biomarkers of liver fibrosis were collected and analysed:

• C3M

• C4M2

• C5M

• COL-18N

• ELM7

• EL-NE

• P4NP7S

• PRO-C3

• PRO-C5.

The outcome measure was to calculate any change (improvement or worsening) from first treatment (trial visit 3 overall) to day 99 (visit 10).

The following exploratory biomarker ratios were also explored:

• C3M/PRO-C3

• C4M2/P4NP7S

• C5M/PRO-C5.

Sample size

The overall total sample size needed to accommodate both the dose-confirmatory stage and Phase II of the trial design.

Statistical analysis

The BUTEO trial was an early-phase trial of BTT1023 in immune cell-mediated liver disease, with the rationale being to identify biochemical efficacy (i.e. reduction in ALP activity) and safety in an orphan disease that presently lacks any other effective medical therapy. The trial design, therefore, focused on identifying early biochemical efficacy signals to justify larger-scale randomised controlled trials of a longer duration.

Our primary outcome measure was patient response to treatment at day 99, as measured by a reduction in serum ALP activity of ≥ 25% from baseline to day 99. Our data on stability of ALP activity in PSC suggest that such responses occur very seldom during the natural course of the disease and we can, therefore, reliably assess changes from baseline and response rates for this proof of concept trial to evaluate the therapeutic potential of BTT1023. In addition, we excluded patients in whom ALP values changed significantly naturally by > 25% between the screening visit at 5–7 weeks before infusion and the screening visit within 1 week before infusion.

Secondary outcome measures included safety and tolerability as determined by treatment compliance, patient withdrawal, frequency of SAEs/AEs, change in quality-of-life questionnaires (i.e. EQ-5D, FSS, pruritus VAS score and IBD diaries) and change in quality of BTT1023 efficacy (as determined by tests of liver fibrosis including ELF, FibroScan and liver biochemistry). In addition, liver MRI was an emerging method for monitoring liver disease and its treatment. We evaluated changes in MRI scans pre and post therapy using the LiverMultiScan protocol (or equivalent methodology, at sites where this is possible). Finally, we evaluated changes in levels of sVAP-1/SSAO as a biomarker of liver disease activity across the trial period.

Statistical analyses were conducted on a modified intention-to-treat (mITT) basis, whereby only patients who had received at least one infusion at the confirmed dose of BTT1023 were to be analysed. Descriptive statistics are presented as mean, median, interquartile range (IQR) and minimum and maximum range for numerical variables, with the frequency and percentage given for categorical variables. Statistical methods used are described at the beginning of each relevant analysis section. All analyses were carried out using either Stata^® (at least version 14.0 for design and interim analyses) (StataCorp LP, College Station, TX, USA) or R (version 3.6.0 for final analyses) through RStudio (version 1.2.1335) (The R Foundation for Statistical Computing, Vienna, Austria).

Interim analysis

An interim analysis was carried out once 18 patients had been evaluated for the primary outcome (i.e. ALP response). If three or more responses were observed, then the trial would continue. If not, then the trial would cease. If adequate response was seen, then a further 19 patients were to be recruited to obtain the required sample size of 37 patients. Allowing for 10% patient dropout during trial duration, this number could reach a total of 41 patients recruited. The final success criterion chosen maintains power and, in doing so, can result in increased type I error rates (note that the stage 1 criterion is fixed, as the final sample size is unknown at stage 1). These were calculated by the trial biostatistician and independently verified by a CRCTU biostatistician. Patients who could not be evaluated for the primary outcome (e.g. because of withdrawal or lost to follow-up) were treated as non-responders.

Final analysis

Final analyses of the primary outcome were performed when all patients had been followed to day 120 and once the database had been locked. If, overall, there were nine or more responses from 37 evaluable patients, then it could be concluded that the treatment warrants further investigation. Only patients treated at the confirmed dose contributed to the total patient requirement.

Dose-limiting toxicity

Dose-limiting toxicity is defined as an AE that meets the criteria of grade 3 cytokine release syndrome or the criteria for any other adverse event at grade 4 or 5, as defined in the CTCAE v4.0. Although previous studies have shown no DLTs with BTT1023, toxicity monitoring was ongoing throughout the trial, and any concerns were reported to the trial office within 24 hours of the investigator becoming aware of the event. During the confirmatory stage, the DLT reporting period was defined as the treatment period from visit 3 (first infusion) to visit 10 (i.e. day 99 follow-up visit). An acceptable DLT rate was established for the trial as a maximum of one incident in six patients (≈ 17%). If the DLT rate increased to two or more at any stage during the DLT reporting period [from visit 3 (i.e. day 1) to visit 10 (i.e. day 99)], the trial would have been halted after consultation with the DMC.

Recording and reporting of serious adverse events

The collection and reporting of AEs was in accordance with The Medicines for Human Use (Clinical Trials) Regulations 200417 and its subsequent amendments. The CTCAE v4.0 criteria were used to grade each AE. Any pre-existing conditions were reported in the medical history and were not reported as an AE unless the condition worsened by at least one CTCAE grade during the trial. The reporting period for AEs commenced on the date of consent (visit 1) and continued until the final follow-up visit (visit 11 on day 120) or, alternatively, to 45 days post last infusion if the patient withdrew from the trial prior to completion of all seven trial drug infusions. All trial patients continued to receive standard concomitant clinical care throughout the trial.

The sponsor, appropriate regulatory authority (e.g. Medicines and Healthcare products Regulatory Agency) and the Research Ethics Committee (REC) were informed of all SAEs, as required by current regulations. SAEs judged to have a reasonable causal relationship to the drug were recorded as serious adverse reactions or as suspected unexpected serious adverse reactions, as appropriate, and were reported to the Medicines and Healthcare products Regulatory Agency and the REC within 7 days. The independent DMC also reviewed all SAEs.

In the event that a patient or their partner became pregnant during the SAE reporting period, this was recorded, reported and followed up, subject to required patient/partner approvals.

Baseline patient characteristics

In the pooled population there was an uneven sex divide, with 82.61% (19/23) of all registered participants being male. Table S8 shows the breakdown of recruitment by sex according to registration centre. In Table S8, it can be seen that, despite the sex imbalance, the John Radcliffe Hospital (Oxford, UK) recruited equal numbers (n = 2) of male and female patients.

The majority of patients (87.0%) identified as Caucasian (n = 20), with the two remaining patients identifying as South Asian (8.7%). Details regarding ethnic identification were missing for a patient who was found to be ineligible post registration. Full details regarding the breakdown of ethnicity by sex can be found in the Additional Material 2 – Patient Information document, Table S9.

Table 1 shows descriptive characteristics regarding patient age, both generally and stratified by sex. From Table 1, it is apparent that the mean age was similar for both males and females (approximately 46 years). However, there is more variability in the age distribution of males, with both a wider range (22–69 years) and a wider IQR (36.0–52.5 years). The observed wider range of age values for males may be attributable to the larger number of male patients registered to the trial. Table S10, Additional Material 2 – Patient Information, shows the number of subjects enrolled per age group, categorised according to EudraCT age categories.

Table 2 gives a line listing of patient baseline characteristics.

Table 2 also shows that patient 6 had an absolute ALP percentage change value that would make him ineligible for the trial. The following deviation was listed: ‘Patient had minor fluctuation in ALP. It was just > 25% and was an oversight and “not” clinically significant’. Following a review by the Trial Management Group, it was decided to allow this individual to participate. For full details on the deviation, see Table S333 in Report Supplementary Material 8.

Patient medical histories

See Report Supplementary Material 9.

Baseline/screening biochemistry and haematology

Dose-limiting toxicity evaluation

Version 5.0 of the BUTEO protocol defines a DLT as follows:

. . . an AE that meets the criteria of grade 4 or grade 5 as defined in the Common Terminology Criteria for Adverse Events (CTCAE v4.0) (Protocol Appendix 2). However, if a definite diagnosis of cytokine release syndrome was made, it was classed as a DLT at grade 3 and above, and if considered at least possibly related to BTT1023.

Report Supplementary Material 10, Table S39, lists all grade 3 or higher AEs experienced by all patients during the dose confirmation stage. The table includes grade 3 AEs to ensure that any occurrences of grade 3 cytokine release syndrome were identified.

As evident from Report Supplementary Material 10, Table S39, during the dose-confirmatory phase of the trial, the seven-patient cohort recorded 17 grade 3 AEs and one grade 4 AE. Investigation into the grade 4 AE experienced by patient 2 revealed that this AE was recorded at screening visit 1, prior to any treatment with BTT1023. Consequently, the event was not considered to be a DLT. Moreover, during the dose confirmation phase, there were no instances of confirmed cytokine release syndrome at grade 3 or higher. Therefore, it was concluded that no DLTs were experienced during the dose-confirmatory period.

Pharmacokinetic evaluation of BTT1023

In addition to the incidence of DLTs, the success criteria for transition from the dose confirmation period to the Simon’s two-stage design have been approximated as reaching a trough concentration of 3 µg/ml of free circulating BTT1023 at 8 weeks from first infusion, which is about 100-fold the dissociation constant of BTT1023 from VAP-1 and should result in a target occupancy of approximately 90% with either the starting dose of 8 mg/kg or following dose escalation up to a maximum of 16 mg/kg.

If PK evaluation showed the levels of BTT1023 to be too low, then dose escalation was considered to be appropriate. If the results of PK evaluation showed the levels of BTT1023 to be too high, then dose reduction was required. Blood samples were collected from patients and sent to Biotie Therapies Ltd/Envigo (Bicester, UK) for analysis.

The dose-confirmatory population was taken to be the first six patients who were eligible for PK evaluation 8 weeks from first infusion. As patient 2 discontinued treatment and withdrew from the study after one infusion, this patient was excluded from this analysis. All patients in this cohort were allocated 8 mg/kg. As PK assessments were performed at different time points of the trial visit, pre-infusion values are used.

Report Supplementary Material 10, Table S40, presents descriptive statistics of the circulating BTT1023 levels (µg/ml) recorded at each treatment visit. The laboratory test has a lower limit of quantification (104 ng/ml), and results below this level were set to zero. Moreover, the laboratory test also has an upper limit of quantification. Results above the limit of quantification were treated as missing.

The criterion for continuation to the Simon’s two-stage expansion was for BTT1023 circulating trough levels to reach 3 µg/ml 8 weeks from the first infusion. For the trial to continue to the Simon’s two-stage expansion, circulating levels of BTT1023 8 weeks from infusion trial needed to reach a trough level of 3 µg/ml. As 8 weeks from the first infusion falls between visit 7 (i.e. day 50) and visit 8 (i.e. day 64), the forthcoming analysis presents results at visit 7 and assumes this to be 8 weeks from the first infusion. Summary statistics of circulating levels of BTT1023 at visit 7 pre infusion can be found in Report Supplementary Material 10, Table S40. Table 5 shows patient-specific estimates.

Report Supplementary Material 10, Figure S5, shows a repeated measures plot of circulating BTT1023 levels recorded at each treatment visit in the dose-confirmatory population.

From Report Supplementary Material 10, Figure S5, despite the significant amount of heterogeneity between patients, at visit 7 the circulating levels of BTT1023 exceeded the minimum requirement of 3 µg/ml for all patients in the dose-confirmatory phase. Therefore, the minimum threshold criterion for circulating level was met.

Safety analysis

Adverse event data were recorded for each patient throughout their time on the trial. Any AEs that fulfilled the DLT criteria have been presented in Dose-limiting toxicity evaluation. As the patients from the dose-confirmatory cohort received the final selected dose of BTT1023 for further testing in the expansion part of the trial, the safety data will be presented for the overall total patient cohort (see Safety and toxicity reporting).

Primary outcome measure: response at day 99

The primary outcome of the BUTEO trial was to determine the activity of the anti-VAP-1 antibody BTT1023 in patients with PSC, as measured by a decrease in ALP activity (i.e. the primary end point), with the outcome measure being response to BTT1023 at visit 10 (i.e. day 99). This required patients to exhibit a reduction in serum ALP activity of at least 25% from baseline (i.e. pre-infusion visit 3) to day 99. In an attempt to reduce any between-laboratory analytical variability, blood samples were taken from all patients and processed at a centralised laboratory in Birmingham [see schedule of events in v5.0 of protocol URL: www.journalslibrary.nihr.ac.uk/programmes/eme/1216531/#/documentation (accessed 9 August 2021)]. This ensured that all samples were analysed using the same equipment and procedures. Note that, for the centralised processing, individuals who were treated in Birmingham had their samples recorded on a biochemistry form and may have had their samples recorded on the ALP sample form. For individuals who were not treated in Birmingham, values on the biochemistry form relate to the analysis of the sample performed at the individual treatment centre and values on the ALP sample form relate to records from the centralised processing. Therefore, for individuals treated in Birmingham, data relating to the primary outcome measure were extracted from the biochemistry form. For individuals not treated in Birmingham, data were extracted from the ALP sample form.

Percentage change in ALP was calculated using Equation 1:

Table 6 shows the ALP measurement both at pre-infusion visit 3 and at follow-up visit 10, along with the difference and percentage change in values for those patients constituting the interim data set, ordered according to percentage. In Table 6, patient 19 was missing their follow-up visit 10 measurement. Patient 19 was not treated at the Queen Elizabeth Hospital (Birmingham, UK) and so their sample was to be transported for centralised processing. However, the ALP sample form shows that the ‘sample was never received by the UHB [University Hospitals Birmingham] laboratory’ and that, although tracking was available, it could not be ‘identified where the sample was delivered’. Therefore, as treatment was administered, patient 19 was included in the interim analysis data set, but their follow-up visit 10 value was missing.

From Table 6 it can be seen that nine patients (50%) experienced a reduction in ALP activity (i.e. a negative percentage change value), with eight patients (44.44%) experiencing an increase.

In accordance with the Simon’s two-stage design of the trial, for the trial to continue to stage 2, at least three successful responses to BTT1023, out of 18 evaluable responses, were required at the interim assessment. From Table 6, it is immediately apparent that only 2 of the 18 evaluable patients (11.11%) achieved a decrease of ≥ 25%. Following the trial’s interim assessment analyses, and the subsequent DMC and TSC meetings, it was recommended that, despite there being no safety concerns regarding treatment, the trial should be closed because of a lack of efficacy.

Safety data

All AEs and safety reporting can be found in Safety and toxicity reporting.

Primary outcome measure: response at day 99

Using definitions given in Primary outcome measure: response at day 99, the percentage change in ALP between visit 3 (pre infusion) and visit 10 (follow-up) can be calculated for the entire study population. Table 7 shows the ALP measurement at both pre-infusion visit 3 and follow-up visit 10, along with the difference and percentage change in values for all evaluable patients, ordered according to percentage.

Table 7 shows that there appeared to be a high level of variability in ALP activity, with 10 patients (45.5%) experiencing a decrease in ALP post treatment.

Report Supplementary Material 11, Table S41, shows that only 9.09% of patients achieved the target reduction in serum ALP of 25%. Figure 2 shows a repeated measures plot of all centrally processed serum ALP.

FIGURE 2.

Repeated measures plot of centrally processed ALP activity for the mITT population. A mean locally weighted smoothing (LOESS) trend line is shown in dark blue (thicker line), with uncertainty depicted by the shaded region.

Safety and toxicity reporting

The following section contains AE information pertinent to the safety population of the study. All AEs were graded in accordance with CTCAE v4.0. As per version 2.0 of the statistical analysis plan, the safety population contains all patients who received treatment with BTT1023 at any dose level. Therefore, the safety population included 22 patients.

In total, there were 1133 AEs experienced by 22 patients. Further details regarding the number of patients who experienced AEs can be found in Report Supplementary Material 11, Tables S44 and S45.

All AEs are categorised according to how related to the investigational medicinal product they are on the following scale:

1. unrelated

2. unlikely to be related

3. possibly related

4. probably related

5. definitely related.

Moreover, the following definitions are assumed throughout this section:

• affected – the number of patients who experienced this event

• occurrences – the total number of each event reported by those affected

• related – the number of occurrences that were classified as related to the trial treatment

• fatal – the number of occurrences for which the outcome was reported as death

• related and fatal – the number of occurrences that were related to the trial treatment and had death as the outcome.

An event was considered to be related to BTT1023 if it was categorised as at least possibly related.

Report Supplementary Material 11, Table S43, shows the number of AEs experienced by grade and treatment visit and Report Supplementary Material 11, Table S44, shows the number of patients experiencing AEs by grade and treatment visit. Recorded AEs were overwhelmingly either grade 1 (73.02%) or grade 2 (19.49%), with no grade 5 events being reported throughout the trial period. The distribution of AEs by visit was fairly even across all treatment visits. AEs were most commonly reported at screening visit 1 (15.96%). As can be seen in Report Supplementary Material 11, Table S43, eight events were not categorised as happening during a visit. The reason ‘TNO15 missed visit 7, however experienced 8 adverse events on the dates between the dates of visit 6 and visit 8’ was provided. As there was no treatment date that corresponds with the date of these AEs, they were categorised as ‘other’.

Report Supplementary Material 11, Table S45, gives a listing of all AEs that occur in at least one patient, ordered by category and toxicity. The most common toxicities, experienced by 22 patients (100%), were an increase in activity of ALP and GGT, both falling under the CTCAE category of investigations. Moreover, 136 AEs each occurred in only one patient.

Report Supplementary Material 11, Tables S46 and S47, show the overall and patient-level incidence of AEs by toxicity and grade.

Report Supplementary Material 11, Table S48, shows the number of AEs experienced by treatment cycle and relatedness. Generally speaking, AEs were deemed to not be related to BTT1023, with 76.19% being categorised as unrelated and 11.11% being unlikely to be related.

Report Supplementary Material 11, Table S49, gives information on all grade 3 and higher AEs that were deemed to be possibly, probably or definitely related to BTT1023. Report Supplementary Material 11, Table S49, shows that there were 15 instances of a related AE of grade 3 or higher, with only one event being grade 4 and only two events being definitely related. Of the 15 related events, eight were experienced by patient 22.

Report Supplementary Material 11, Table S50, shows the number of AEs of any grade categorised according to CTCAE category by treatment visit. The most common category was investigations, with 802 (70.72%) AEs.

Report Supplementary Material 11, Table S51, shows the duration of AE by CTCAE grade. When the duration of the AE is missing, in all 271 cases this was attributable to the AE being ongoing. Further details regarding which visits these were attributable to are given in Report Supplementary Material 11, Table S52. Moreover, from Report Supplementary Material 11, Table S51, it was apparent that both the mean and median duration of an AE were noticeably longer for grade 4 AEs (mean 90.0 days; median 74 days) than for grade 1–3 AEs, while the range of AE duration was greatest for grade 1 AEs (range 0–2411 days).

Report Supplementary Material 11, Table S52, shows the visits on which the ongoing AEs occurred. Only 23.61% of ongoing AEs occurred during the treatment period (42.44% occurred either pre-screening or at screenings 1 or 2, whereas 33.95% occurred during one of the follow-up visits).

Finally, Report Supplementary Material 11, Table S53, shows the AE outcome by visit. Overwhelmingly, AEs were resolved without sequelae (76.72%), with the second most common outcome being unresolved (23.19%).

Serious adverse events

In total, four SAEs were experienced by four patients, with no one individual experiencing more than one SAE. Table 8 gives a full line listing of all SAEs reported during the BUTEO trial and Report Supplementary Material 11, Table S54, lists the toxicity that the site identified as the AE that prompted the SAE report. The population for the incidence of SAEs was taken to be the safety population.

Secondary outcome measures

In this section, when percentage change calculations were made, the response was calculated as the difference between the first follow-up visit (visit 10) and either screening visit 2 or visit 3 (i.e. the first treatment visit), depending on when each outcome was measured. The stipulation is given at the start of each section.

Calculation of percentage change was always performed with reference to the baseline period, as:

Therefore, a positive percentage change indicates an increase in the measurement between baseline and follow-up, and a negative percentage change indicates a decrease.

Serum vascular adhesion protein 1/semicarbazide-sensitive amine oxidase liver disease activity biomarker: semicarbazide-sensitive amine oxidase enzyme activity

Semicarbazide-sensitive amine oxidase assays were performed on serum sent to the laboratories at the University of Birmingham, and the laboratories forwarded the results of the completed case report form to the CRCTU. Enzyme activity was assessed at all trial visits [i.e. both screening visits (trial visits 1 and 2), all infusion visits (trial visits 3–9) and both follow-up visits (trial visits 10 and 11)].

Semicarbazide-sensitive amine oxidase monitors the enzyme activity of VAP-1 and is expressed as the pool of H₂O₂ generated per minute. The difference in SSAO activity between follow-up visit 10 and the first infusion visit (i.e. visit 3) is calculated as:

⁽¹⁹⁾ Difference = SSAO activity at follow-up visit 10 − SSAO activity at visit 3 ,

and the percentage change is calculated as:

⁽²⁰⁾ Percentage change = SSAO activity at follow-up visit 10 − SSAO activity at follow-up visit 3 SSAO activity at visit 3 .

In total, SSAO enzyme activity was assessed in only 14 patients.

Figure 3 shows a repeated-measures plot of all SSAO activity at all trial visits in the mITT population. A LOESS trend line has been added to aid identification of any potential trend. Table 12 shows a summary of SSAO enzyme activity over all trial visits. The internal controls (positive control, human serum control and human serum plus VAP control) are also shown.

FIGURE 3.

Repeated-measures plot of SSAO enzyme activity at all trial visits in the mITT population. A LOESS trend line is shown in dark blue (thicker line) with uncertainty depicted by the shaded region.

TABLE 12 - Semicarbazide-sensitive amine oxidase enzyme activity (H2O2) over all trial visits (N = 14 at all visits)

HS CTL, human serum control; HS + VAP CTL, human serum plus VAP control; POS CTL, positive control.

Measure	SSAO enzyme activity (pmol H2O2 generated per minute)											POS CTL	HS CTL	HS + VAP CTL
	Visit 1	Visit 2	Visit 3	Visit 4	Visit 5	Visit 6	Visit 7	Visit 8	Visit 9	Visit 10	Visit 11
Mean	0.22	0.26	0.26	0.29	0.25	0.21	0.19	0.18	0.17	0.17	0.18	0.51	0.10	0.18
Median	0.21	0.22	0.23	0.24	0.22	0.19	0.17	0.14	0.13	0.10	0.08	0.53	0.10	0.18
Range	0.11–0.42	0.13–0.59	0.12–0.6	0.17–0.76	0.14–0.46	0.11–0.37	0.13–0.38	0.09–0.41	0.09–0.42	0.08–0.72	0.03–1.33	0.24–0.92	0.08–0.12	0.13–0.25
IQR	0.17–0.23	0.17–0.31	0.16–0.28	0.21–0.31	0.17–0.31	0.16–0.24	0.14–0.21	0.12–0.20	0.11–0.20	0.09–0.17	0.07–0.10	0.41–0.59	0.09–0.11	0.16–0.19

Report Supplementary Material 11, Table S212, presents descriptive analyses of SSAO enzyme activity at visit 3 (i.e. the first treatment visit) and at visit 10 (i.e. the follow-up visit), and the difference between them. Report Supplementary Material 11, Table S213, then shows the patient-level absolute difference and percentage change in SSAO activity. Finally, Report Supplementary Material 11, Table S214, shows summary measures pooled across the whole mITT population.

Figure 3 shows a slight downwards trend over all trial visits and, therefore, the repeated-measures analysis is performed. i is the index patient number and j is the index trial visit number, for j = 1, 2, . . ., 11. Then consider the following mixed-effects model:

⁽²¹⁾ SSAO ij = α + ai + βtj + εij .

The above model was fitted using the lme4 package in R using restricted maximum likelihood to attain the following estimates:

⁽²²⁾ SSAO ij = 0.2742 + ai − 0.009654 tj + ε ij ,

where ai ~ N (0, 0.1163²) and εij ~ N (0, 0.009487²).

Report Supplementary Material 11, Figure S49, shows a repeated-measures plot of all SSAO enzyme activity over all trial visits, with a line depicting the fixed effects from the above mixed-effects model.

Report Supplementary Material 11, Figure S50, shows a repeated-measures plot of SSAO enzyme activity as a percentage of the screening 1 value across all trial visits in the mITT population.

Exploratory analysis

The exploratory biomarkers of fibrosis were analysed by Nordic Bioscience (Herlev, Denmark), with external data uploaded to the database.

Analysis of the exploratory biomarkers and biomarker ratios are given in Report Supplementary Material 11.

Additional analyses

Additional analyses have been performed on the following experimental variables derived from laboratory analyses:

• total bilirubin (biochemistry)

• sodium (biochemistry)

• potassium (biochemistry)

• urea (biochemistry)

• creatinine (biochemistry)

• calcium (biochemistry)

• total protein (biochemistry)

• eGFR (biochemistry)

• haemoglobin (haematology)

• platelets (haematology)

• red blood cells (haematology)

• white blood cells (haematology)

• haematocrit (haematology)

• mean cell volume (haematology)

• mean cell haemoglobin (haematology)

• neutrophils (haematology)

• lymphocytes (haematology)

• monocytes (haematology)

• eosinophils (haematology)

• basophils (haematology)

• APTT ratio (haematology).

Results are presented in Report Supplementary Material 11.

Sensitivity analyses

Sensitivity analyses were performed in the per-protocol population and regarded the primary analysis only. Therefore, using centrally processed samples, the sensitivity analyses evaluate the percentage of patients who completed the treatment as originally allocated, with no dose modifications or missing doses (i.e. patients who received all seven infusions as scheduled in the protocol at the minimum effective dose) and with a clinically meaningful reduction in ALP (a reduction of ≥ 25%). In total, 19 patients were compliant with all treatment visits and, therefore, constituted the per-protocol population. Table 14 shows the number of patients in the per-protocol population achieving a target reduction in ALP.

One patient was compliant with all treatment visits but was unevaluable for the sensitivity analysis (see Table 14). This is because the patient was not treated at the Queen Elizabeth Hospital and, therefore, their ALP was sent for central processing. However, the sample was lost and, therefore, the participant was treated as a non-responder, contributing to the denominator only.

Figure 4 shows a repeated measures plot of centrally processed ALP for sensitivity analyses in the per-protocol population.

FIGURE 4.

Repeated-measures plot of centrally processed ALP activity for sensitivity analyses in the per-protocol population.

Ethics and dissemination

The trial was performed in accordance with the Recommendations of Guiding Physicians in Biomedical Research Involving Human Subjects,21 adopted by the 48th World Medical Association General Assembly, as well as the Research Governance Framework for Health and Social Care, the applicable UK statutory instruments (which include the Medicines for Human Use Clinical Trials 200417 and subsequent amendments and the Data Protection Act 199822) and guidelines for Good Clinical Practice. The protocol was approved by the REC (reference number 14/EM/1272) (see Report Supplementary Material 12 and 13). The first REC approval date was 6 January 2015, with subsequent amendments on 18 March 2015 (non-substantial), 27 November 2015 (substantial), 16 March 2016 (substantial), 27 March 2018 (substantial) and 31 July 2018 (substantial) (see Report Supplementary Material 14). All active sites obtained local research and development department approval and are up to date with the latest protocol amendment.

Report Supplementary Material 14, Table S334, shows the trial milestones and Report Supplementary Material 14, Table S335, shows a summary of approved protocol versions.

Standard regulations for data-handling and patient confidentiality were followed in accordance with the Data Protection Act 201823 for health and care research and General Data Protection Regulation. 24 During the trial, patients were identified using only their unique trial number, initials and date of birth on the case report form, and this information was also used in correspondence between the trial office and the participating site. Data quality was maintained in accordance with trial-specific guidance and was consistent with the source data. All essential trial documentation and source records will be securely retained for at least 25 years after the end of the trial to allow the results of the trial to be verified. Results will be disseminated via peer-reviewed publication and presentation at international conferences, and additional summaries will be provided to patients and patient support groups.