1 INTRODUCTION

Haemophilia A is a congenital bleeding disorder characterised by a deficiency in coagulation factor VIII (FVIII) that results in an increased tendency for spontaneous or traumatic bleeding events.1 Prophylaxis with clotting factor concentrates or other haemostatic agents is the standard of care for patients with severe haemophilia A to prevent bleeding.1

The major complication in the treatment of haemophilia A with clotting factor concentrates is the development of alloantibodies (inhibitors) against exogenous FVIII.2 Inhibitors neutralise the clotting activity of FVIII therapy, which may result in an increased risk of severe bleeding and, subsequently, higher risk of morbidity, mortality and disability, and decreased quality of life.3, 4 Switching FVIII concentrates in previously treated patients (PTPs; > 75–150 exposure days [EDs]) is not associated with a higher risk of developing inhibitors5; however, data for PTPs with a history of inhibitors are sparse because they are usually excluded from Phase 1–3 clinical studies.6, 7

Turoctocog alfa (NovoEight®) is a third-generation, recombinant, B-domain-truncated human coagulation FVIII molecule developed by Novo Nordisk A/S and approved for prophylaxis and treatment of bleeding episodes in haemophilia A in all age groups.8-10 The safety and efficacy of turoctocog alfa have been demonstrated across several Phase 3 clinical studies in PTPs (guardian 1, 2, 3 and 7) and previously untreated patients (guardian 4) with haemophilia A.6, 7, 11-13

Real-world studies help to understand treatments in broader, more representative patient populations than those used in interventional clinical studies,14 which may exclude patient groups such as those with a history of inhibitors. Additionally, treatment practices may vary in the real-world setting due to differences in regional procedures and prescribing patterns. Issues with adherence can arise outside a controlled clinical study setting.14

The aim of the guardian 5 study was to evaluate the real-world immunogenicity, safety and clinical efficacy of turoctocog alfa for prophylaxis, treatment of bleeding episodes and for surgery in PTPs with severe and moderately severe haemophilia A.

2 METHODS 2.1 Study design

Guardian 5 was a prospective, multinational, non-interventional, post-authorisation safety study investigating the long-term safety and efficacy of turoctocog alfa in routine clinical practice (NCT02035384). The study design followed European Medicines Agency guideline recommendations for FVIII product investigation.15 Patients received intravenous prophylaxis or on-demand treatment as per their and the treating physician's choice, and clinical assessment was done according to local practices. Patients were enrolled at 31 sites in 13 countries (12 European countries and the USA; see Supplementary Table 1).

For each patient, data were planned to be collected until they reached a minimum of 100 EDs (counted from the baseline visit) with turoctocog alfa, or until at least 50 patients had a minimum of 100 EDs after inclusion in the study, whichever came first. The study was completed in 67.5 months between June 2014 and January 2020. The study was approved by all relevant independent ethics committees and institutional review boards and conducted in accordance with the Declaration of Helsinki,16 Guidelines for Good Pharmacoepidemiology Practices17 and Good Pharmacovigilance Practice.18 All patients or representatives provided written consent.

2.2 Participants

Patients were considered for inclusion if it had already been decided that they would switch from their previous FVIII product to turoctocog alfa, typically because of tendering. Patients were eligible if they were male, diagnosed with congenital severe or moderately severe haemophilia A (FVIII level ≤ 2%), previously FVIII treated with at least 150 EDs at the time of first dosing with turoctocog alfa, and had a negative FVIII inhibitor test not more than 4 weeks prior to first dosing.

Patients with a history of FVIII inhibitors were eligible for inclusion. Patients who had previously participated in any clinical study with turoctocog alfa were not eligible.

2.3 Treatment

Product usage was in accordance with clinical daily practice and generally in agreement with the approved product label and national guidelines.10 Commercially available product was used as part of this study.

2.4 Study endpoints and clinical assessments

The primary endpoint was the proportion of patients developing FVIII inhibitors (≥.6 BU, or above the specific local laboratory reference range). The physician was encouraged to perform clinical evaluation and blood testing for FVIII inhibitors during visits to the clinic or when there was a lack of therapeutic effect. Inhibitor testing was in accordance with the local practice and not mandatory, due to the non-interventional nature of the study. Secondary safety endpoint was the number of adverse reactions reported during the study period. Secondary efficacy endpoints included haemostatic effect of turoctocog alfa in the treatment of bleeds and during surgical procedures, annualised bleeding rate (ABR) for preventive and on-demand treatment, and consumption of turoctocog alfa.

Detailed information related to treatment and bleeding episodes was captured in a paper patient diary by the patient or parent/caregiver. If a patient was unable to enter a treatment or bleeding episode in the diary, or was hospitalised, information was reported by the physician. It was possible to use the patient's own diary to capture details of bleeding episodes (e.g. location, type, date of occurrence), where some of the data points (haemostatic response and severity of bleed) were not captured.

2.5 Statistical methods

Data evaluation was mainly based upon descriptive statistics. The proportion (percentage) of patients developing FVIII inhibitors was calculated; a one-sided 97.5% upper confidence limit based on an exact calculation for a binomial distribution was provided. In the analysis, all patients with at least one inhibitor test performed after baseline visit were included.

Bleeds were classified as mild/moderate or severe; haemostatic responses for the treatment of bleeds and during surgery were evaluated on a four-point scale (see Supplementary Table 2). Haemostatic response assessment was repeated with the inclusion of missing responses (counted as treatment failure) in the sensitivity analysis. ABR was analysed using a negative binomial model and estimated ABR with confidence intervals (CIs). A Poisson model allowing for over-dispersion was applied in the sensitivity analysis of ABR.

3 RESULTS 3.1 Patient disposition and baseline characteristics

In total, 70 patients were screened for this study (see Figure 1); 68 patients were exposed to FVIII during study participation and therefore included in the full analysis set and safety analysis set. One patient with FVIII level > 2% at inclusion could not be classified into ‘severe haemophilia (FVIII level < 1%)’ or ‘moderately severe haemophilia (FVIII level 1–2%)’ categories; however, this patient was included in the total number of patients in the analyses by severity of haemophilia A to represent the full analysis set.

Participant flow. FAS, full analysis set; SAS, safety analysis set. aPatient forgot diary and did not wish to continue in study. bEvaluations until the moment of withdrawal were included in the analyses. c8 patients were considered as study completers without recording a minimum of 100 EDs but confirmed by the physician as completers.

During the study, 68 patients were exposed to turoctocog alfa for a total of 87.8 patient-years and 8967 EDs. The mean EDs (standard deviation [SD]) per patient was 131.9 (99.0) days/patient; 50 patients reported 100 EDs. Of the 68 exposed patients, 63 patients were on a prophylaxis regimen (14 patients < 12 years; 49 patients ≥ 12 years) and five patients received on-demand treatment (all patients ≥ 12 years). No patients switched from on-demand to prophylaxis (or vice versa) after entering the study. A total of 58 (84.1%) patients (9 patients < 12 years; 49 patients ≥ 12 years) completed the study, of which eight patients (all ≥ 12 years) were considered as study completers without recording a minimum of 100 EDs but confirmed by the physician as completers.

The study population, disposition and baseline demographics are shown in Table 1. The study population included 14 paediatric patients (< 12 years) and 55 adolescent/adult patients (≥ 12 years), of which one patient was not exposed to turoctocog alfa. Before study entry, five patients (all ≥ 12 years) received an on-demand regimen. Overall, 63 patients (14 patients < 12 years; 49 patients ≥ 12 years) were on a prophylaxis regimen. Historical frequency of dosing for prophylaxis is reported in Table 1; prophylaxis dose before the trial was not collected.

TABLE 1. Patient demographics and baseline characteristics Full analysis set Age at inclusion Haemophilia severity at inclusiona < 12 years ≥ 12 years Severe Moderately Severe Total Number of patients 14 54 58 9 68a Age (years) Mean (SD) 7.4 (2.3) 33.6 (15.5) 26.2 (15.5) 38.6 (24.8) 28.2 (17.4) Severity of haemophilia, N (%) Severe (< 1%) 13 (92.9) 45 (83.3) 58 (100.0) – 58 (85.3) Moderately severe (1–2%) 1 (7.1) 8 (14.8) – 9 (100.0) 9 (13.2) FVIII level > 2% – 1 (1.9) – – 1 (1.5) Current treatment prior to study, N (%) On-demand – 5 (9.3) 3 (5.2) 1 (11.1) 5 (7.4) Prophylaxis 14 (100.0) 49 (90.7) 55 (94.8) 8 (88.9) 63 (92.6) FVIII product given as prophylaxis or on-demand, N (%)b N 13 53 56 9 66 Plasma FVIII product 1 (7.7) 6 (11.3) 5 (8.9) 2 (22.2) 7 (10.6) Recombinant FVIII 12 (92.3) 47 (88.7) 51 (91.1) 7 (77.8) 59 (89.4) Average number of bleeds per year (prophylaxis regimen) N 14 44 51 7 58 Mean (SD) 3.2 (6.2) 5.3 (7.8) 4.6 (7.3) 5.9 (9.0) 4.8 (7.5) Average number of bleeds per year (on-demand) N 0 5 3 1 5 Mean (SD) – (–) 22.6 (11.2) 17.6 (11.0) 36.0 (–) 22.6 (11.2) Dose level for patients on prophylaxis at inclusion N 14 49 55 8 63 Mean (SD) 30.2 (12.1) 27.0 (12.2) 28.2 (12.2) 24.1 (11.7) 27.7 (12.1) Frequency of dosing for patients on prophylaxis at inclusion N 14 (100.0) 48 (100.0) 55 (100.0) 7 (100.0) 62 (100.0) Once weekly 1 (7.1) 5 (10.4) 5 (9.1) 1 (14.3) 6 (9.7) Every second day 1 (7.1) 10 (20.8) 9 (16.4) 2 (28.6) 11 (17.7) Three times weekly 11 (78.6) 23 (47.9) 30 (54.5) 4 (57.1) 34 (54.8) Other 1 (7.1) 10 (20.8) 11 (20.0) – 11 (17.7) Severe: FVIII level < 1%, Moderately Severe: FVIII level 1–2%. FVIII, factor VIII; SD, standard deviation. aOne patient with FVIII activity > 2% at inclusion is included in the ‘Total’ column to represent the full analysis set. bPatients can consume more than one product.

Six patients (one patient < 12 years; five patients ≥ 12 years [oldest patient: 41 years old]) reported a history of clinical suspicion of inhibitors. Information on the historical inhibitor titre, duration and treatment was not consistently captured. Five of the six patients were tested for FVIII inhibitors during the study; one patient (< 12 years) was not tested. In total, 41 (60.3%) patients had a family history of haemophilia A. Fifty-one patients had a history of switching FVIII products.

3.2 Safety 3.2.1 FVIII inhibitor development

None of the 68 exposed patients was reported to be positive for FVIII inhibitor during treatment with turoctocog alfa (proportion of patients: 0%; 97.5% upper confidence limit: 6.5). Of the 68 exposed patients, 55 had at least one inhibitor test post-baseline; all of them reported negative FVIII inhibitor test results.

Six patients completed the study without any inhibitor test taken, and for seven patients an inhibitor test result was reported only at baseline visit; therefore, these 13 patients were not included in the primary analysis. For these patients, reported data did not indicate clinical signs of possible inhibitor development, such as an increase in bleeding frequency, increased consumption for prophylaxis and bleed treatment, or lack of response to hemostatic treatment. Four patients did not have an inhibitor test done within the 4 weeks prior to starting treatment with turoctocog alfa. Of these, three patients had an inhibitor test taken during the study (these patients were included in the primary analysis).

3.2.2 Adverse events

One serious adverse reaction of angina pectoris was reported. A 38-year-old patient (body mass index 40.1 kg/m2) with severe haemophilia was enrolled into the study and treated with a mean dose regimen of 16.2 IU/kg every alternate day. The patient had a medical history of heart disease and non-ST segment elevation myocardial infarction. Concomitant medication included platelet aggregation inhibitors, anti-hypertensive and cholesterol-lowering therapy. After reaching 110 EDs in the study, 2 hours after a prophylactic injection of turoctocog alfa, the patient developed symptoms of angina pectoris and elevated troponin. A coronary angiography with percutaneous transluminal coronary angioplasty and right coronary artery stent (bare metal) was performed. The serious adverse reaction was reported to be possibly related to the study drug by the physician because of the temporal relationship. It was resolved, and the patient continued with turoctocog alfa. No other adverse reactions were reported during the study.

3.3 Efficacy 3.3.1 Prevention of bleeding episodes

Of the 63 patients who received prophylaxis, 62 had > 1 ED reported and were included in the bleeding rate calculation. Additionally, one patient reported one bleeding episode during treatment with turoctocog alfa with unknown number of EDs and was excluded from the ABR analyses. Eight patients had missing diary periods; those periods were not accounted in the ABR analyses. Overall, the negative binomial estimates of ABR among patients on prophylaxis regimens and on-demand treatment were 3.65 (95% CI: 2.53–5.25) bleeds/patient/year and 20.28 (12.09–34.01) bleeds/patient/year, respectively. Overall median (interquartile range) ABR was 1.97 (6.19) (Table 2). Estimated ABR on prophylaxis regimen for spontaneous bleeds was 2.39 (95% CI: 1.50–3.81) bleeds/patient/year and for traumatic bleeds was 1.15 (95% CI: .76–1.73) bleeds/patient/year.

TABLE 2. Annualised bleeding rates by age and haemophilia severity for patients on prophylaxis Age at inclusion, N (%) Haemophilia severity at inclusion, N (%) Baseline age < 12 years Baseline age ≥ 12 years Severe Moderately Severe Total Number of patients 14 49 55 8 63 Na 14 48 54 8 62 Number of patients with bleeds 8 (57.1) 32 (66.7) 35 (64.8) 5 (62.5) 40 (64.5) Mean observation period per patient (years) .89 1.24 1.20 .96 1.17 Observation period per patient, min.; max. (years) .21; 1.71 .21; 5.06 .21; 5.06 .21; 1.97 .21; 5.06 Total observation period (years) 12.52 59.73 64.61 7.64 72.25 Annualised bleeding rate Negative binomial analysis 2.75 3.90 3.67 3.37 3.65 95% CI 1.35, 5.61 2.57, 5.92 2.47, 5.46 1.41, 8.06 2.53, 5.25 Poisson estimate 2.56 3.48 3.36 3.01 3.32 95% CI 1.36, 4.82 2.29, 5.29 2.25, 5.00 1.43, 6.36 2.31, 4.78 Median (IQR) 1.34 (4.02) 2.14 (6.56) 1.93 (6.19) 2.76 (7.04) 1.97 (6.19) Severe: FVIII level < 1%, Moderately Severe: FVIII level 1–2%. CI, confidence interval; FVIII, factor VIII; IQR, interquartile range; max., maximum; min., minimum; SD, standard deviation. aNumber of patients with more than one exposure day reported: only these patients are included in the analysis. 3.3.2 Treatment of bleeding episodes

Overall, 469 bleeds were reported in 46/68 patients (Table 3). Most were mild or moderate (n = 387, 82.5%) in severity. The most frequent location was in the joints (69.7% of bleeds). Of the 469 bleeds, 308 (65.7%) were spontaneous, 103 (22.0%) were traumatic and 58 (12.4%) bleeds had unknown cause.

TABLE 3. Details of bleeding episodes and haemostatic response to turoctocog alfa treatment Age at inclusion, N (%) Haemophilia severity at inclusion, N (%) Baseline age < 12 years Baseline age ≥ 12 years Severe Moderately Severe Total Number of patients with bleeding episodes 8 38 39 6 46 Number of bleeding episodes 32 437 412 50 469 Cause of bleed, N (%) 32 (100.0) 437 (100.0) 412 (100.0) 50 (100.0) 469 (100.0) Spontaneous 8 (25.0) 300 (68.6) 283 (68.7) 19 (38.0) 308 (65.7) Traumatic 24 (75.0) 79 (18.1) 97 (23.5) 5 (10.0) 103 (22.0) Missing – 58 (13.3) 32 (7.8) 26 (52.0) 58 (12.4) Site of bleed, N (%) 32 (100.0) 437 (100.0) 412 (100.0) 50 (100.0) 469 (100.0) Joint 18 (56.3) 309 (70.7) 294 (71.4) 29 (58.0) 327 (69.7) Mucocutaneous 4 (12.5) 33 (7.6) 35 (8.5) 2 (4.0) 37 (7.9) Muscular 3 (9.4) 48 (11.0) 38 (9.2) 11 (22.0) 51 (10.9) Other/unknown 7 (21.9) 47 (10.7) 45 (10.9) 8 (16.0) 54 (11.6) Classification of bleed, N (%) 32 (100.0) 437 (100.0) 412 (100.0) 50 (100.0) 469 (100.0) Mild/moderate 27 (84.4) 360 (82.4) 335 (81.3) 45 (90.0) 387 (82.5) Severe 1 (3.1) 12 (2.7) 9 (2.2) 4 (8.0) 13 (2.8) Unknown 4 (12.5) 65 (14.9) 68 (16.5) 1 (2.0) 69 (14.7) Infusions to treat the bleed, from start to stop of the bleed, N (%) 32 (100.0) 437 (100.0) 412 (100.0) 50 (100.0) 469 (100.0) 1 infusion 12 (37.5) 222 (50.8) 206 (50.0) 28 (56.0) 234 (49.9) 2 infusions 12 (37.5) 87 (19.9) 93 (22.6) 3 (6.0) 99 (21.1) ≥ 3 infusions 8 (25.0) 128 (29.3) 113 (27.4) 19 (38.0) 136 (29.0) Number of patients with bleeding episode with diary data a 7 30 31 5 37b Haemostatic response, c