Personalized Prophylactic Treatment With Advate® in Severe or Moderate Haemophilia A Patients

Research Question:

• Is standard prophylaxis effective for all hemophilia A patient?
• Which are the causes for poor clinical response in some patients on prophylactic treatment?
• Can differences in PK impact the clinical outcome and effectiveness of prophylaxis?
• For a given group of poor controlled patients in standard prophylaxis, may personalized prophylaxis improve clinical outcome when acting on the specific cause(s)?

Study Background \& Rationale:

Haemophilia A is an inherited bleeding disorder caused by a deficiency of factor VIII (FVIII). Patients with severe hemophilia A have a FVIII plasma concentration less than1 IU/dL and experience spontaneous and trauma-induced bleeds. Joint bleeds lead to hemophilic arthropathy resulting in progressive disability. Patients with moderate hemophilia (FVIII level between 1-5 IU/dL) are characterized by fewer hemarthroses, usually trauma-induced, and a decreased likelihood of developing arthropathy. This clinical observation led to the use of prophylactic FVIII infusions to convert patient´s bleeding phenotype from severe to moderate with the result of decreasing or preventing arthropathy.

Prophylactic regimens may be effective when based on standard fixed-dose protocols (that assumes one approach fits all patients) or phenotypic dosing determined by bleeding patterns, but do not protect all patients with severe haemophilia from joint damage caused by spontaneous or activity-triggered bleeding.

Individualized treatment in haemophilia A takes into consideration all available information about the patient, not only his phenotypic bleeding pattern. Some of the factors that contribute to the observed interpatient variability include baseline or residual FVIII activity, the pharmacokinetic (PK) profile of the replacement factor, the individual's level of physical activity and perceived risk of traumatic bleeding, the presence or absence of joint disease, genotype, presence of comorbidities and adherence to the dosing regimen.

The PK response to FVIII varies between patients and this has important clinical implications for treatment. Although PK is affected by patient characteristics, this relationship is too weak to infer a result for an individual and, if required, PK must be measured. An important determinant of the efficacy of prophylaxis is the length of time an individual spends with a low level of coagulation factor.

According to the existing International Society on Thrombosis and Haemostasis (ISTH) guidelines measurement of PK in clinical practice requires in adult patients a total of 8 samples (5 in children) to be taken over a period of 48 h. It requires significant commitment in time from the patient, and family and overnight hospital admission may be required.

The Bayesian estimation method uses a population PK model based on FVIII levels from a large population of patients as a mathematical/ statistical framework to estimate the PK in an individual patient from minimal data. Several studies employed this technique for FVIII in a limited number of patients. Using this strategy, a patient´s coagulation factor half-life may be calculated from two or three time points, reducing the inconvenience to the patient, the discomfort of venipuncture and the cost of sample handing and assays. This methodology can facilitate measurement of PK in routine clinical practice.

Knowledge of the individual's PK response to the replacement factor helps to determine both the dosing level as well as the frequency of administration needed to achieve optimal levels of the deficient hemostatic factor. In particular, achieving ideal peak levels helps reduce the risk of bleeding related to repetitive physical activity, whereas minimizing the time spent at trough levels below 1 IU/dL helps to reduce breakthrough bleeding events . The effect of patient´s FVIII half-life will potentially have a significant impact of prophylactic regimens, whereas changing the frequency of dosing and increasing the dose/kg of FVIII has a smaller effect on the through level.

Different trough levels may be targeted, depending on circumstances: higher levels may be desired to manage target joints, highly active patients, or those more prone to bleeding; alternatively, lower levels may be allowed in a patient who has not bled for a long time.

Last studies demonstrated that the utility of prophylaxis dose tailoring with individual PK with similar results than phenotypic dosing but with fewer infusions, and maybe, this option could increase treatment adherence. Besides, the PK monitoring could be more cost-effective dosing compared to standard dosage.

Adherence to prophylaxis regimens is another patient-specific factor that influences plasma levels and bleeding risk. There are two definitions of adherence in haemophilia, adherence as the percentage of infusions within a specified dose range (adherence to dose) and the percentage of weeks without missed doses (adherence to frequency). Different studies showed that bleedings rate was higher in nonadherent patients, being older patients more likely to miss doses. Strategies to improve adherence would be expected to decrease the number of bleeds, whereas poor adherence make PK dose tailoring irrelevant.

Use of sparse blood sampling and Bayesian analysis for measuring pharmacokinetics, and for the use of this information to tailor doses in prophylaxis has been previously reported. Clinical implementation of these methods with all the available information of the patient (bleeding pattern, joint status, genetic mutation, adherence, etc.), is lacking. Exploratory studies on the use of PK tailored prophylaxis are required to establish the safety and efficacy of this approach. This procedure could allow individualization of treatment by the routine determination of individual pharmacokinetics in the clinic, potentially making prophylaxis more cost-effective.

Primary Endpoint:

Identify and analyze cause(s) of poor bleeding control in patients on prophylaxis treatment and study the clinical impact of a "personalized pilot program" with a 1 year follow up to act on the specific causes.

1. Describe PK parameters in patients on prophylaxis treatment with Advate®
2. Analyze differences in PK parameters in non-controlled vs well controlled patients
3. Identify causes of poor clinical outcome in non-controlled patients. Patients' individual variables that influence bleeding risk will be studied (individual PK, bleeding pattern, joint status, physical activity, life style and patient's adherence).
4. Study the improvement in clinical outcomes (ABR and Joint status) of a 1 year Personalized Prophylaxis Program that acts specifically on the previously identified causes of bleeding in non-controlled patients (named: short half-life, high bleeding pattern, joint damage, high risk physical activity, active life style and poor patient's adherence).

Duration of subject participation: One year

Total number of subjects to be enrolled: 25-30 patients

Specific details of Treatment/Intervention:

This prospective, open-label, observational study will be performed at the Hospital Universitari i Politecnic La Fe in Valencia, Spain. This study will include patients with severe and moderate haemophilia A in prophylactic treatment with rFVIII (Advate®).

1. The study will start with the determination of individual PK by sparse blood sampling (2-3 blood samples) and Bayesian analysis (using myPKFiT Medical Device). The PK parameters of the cohort of patients in prophylaxis with Advate® will be described.
2. Patients will be divided in two groups according to the clinical outcome (Bleeding Episodes) in the previous year:
3. Well controlled patients: Patients with an adequate disease control with the routine clinical practice (AJBR≤2, non-severe ABR≤5 and severe ABR≤2) (Ministerio de Sanidad, Servicios Sociales e Igualdad. Gobierno de España; 2012)
4. Poor controlled patients: Patients with poor disease control, based on the international guidelines: AJBR>2, ABR>2 for severe BE or ABR >5 for non-severe BE
5. Differences in PK parameters in these 2 groups will be analyzed and compared adjusting by age and weight:
6. Half-life (h) b. Clearance (dl/h) c. Volume on steady state (L) d. Time to FVIII level below 1% (h)
7. Non-controlled patients will be studied and interviewed to identify causes of poor disease control. Patients' individual variables that influence bleeding risk will be studied:
8. Individual PK parameters (See bullet point 3)
9. Bleeding pattern: Haemophilia Severity Score
10. Joint status (Gilbert score and HJHS)
11. Physical activity: Category Risk 1, 2 or 3
12. Patient's adherence: Adherence index (AI)
13. After identifying in each patient the individual cause(s) of poor bleeding control, a "Pilot Program on Personalized Prophylaxis" will be designed to act on the specific factors that may cause bleeding despite prophylaxis.
14. Shorter half-life compared to Advate PK Population (graph displayed in myPKFiT)

• Prophylaxis will be adjusted to individual PK using the recommendation of myPKFiT Medical Device*.

1. Patients with bleeding phenotype: HSS >0.79 (for severe HA) or HSS>0.47 (for moderate HA)

• Prophylaxis will be adjusted to increase through level according to bleeding phenotype*

1. Patients with joint damage and at least one target joint (at least score of 3 in the Gilbert scale and /or at least score of 3 in the Pettersson scale and/or Chronic synovitis evidenced by ultrasound.) • Prophylaxis will be adjusted to increase through level according to Joint status*
2. Physical activity: Category >1

• Prophylaxis will be adjusted to:

<!-- -->

1. Match the day of physical activity with the day of infusion
2. Increase through level according to physical activity*
3. Patient's with poor adherence:

• Monthly sessions together with the psychologist and pharmacist of the HTC will be scheduled to address adherence issues. myPKFiT displayed graphs will be used to explain availability and elimination of FVIII from the blood. *All these dose adjustments will be made not only based on clinical data, but also based on the knowledge and experience of the responsible physician of the patient.

1. Using a prospective design, patients will be studied during 12 months following the start of the "Personalized Prophylaxis Program". Patient´s clinical state will be assessed at least 2 times per year (normally 2-4 visits per year depending on disease severity).
2. AJBR and ABR (severe, non-severe, spontaneous or traumatic bleeding episodes) will be collected at month 6 and 12.
3. Joint status, Gilbert and HJS will be measures at month 12.
4. Adherence index (AI) will be calculated at month 12
5. Quality of Life will be measures at month 12.

• This study will be conducted without significant changes in medical treatment of patients.
• The study is currently in approval process by the local Clinical Research Ethics Committee and will be conducted in accordance with the declaration of Helsinki and its amendments. An informed consent will be obtained from the patients included in the study.

Statistical Methods:

Statistical analyses will be performed with the R software (version 3.2.2). Data will be described with the mean, standard deviation, median and interquartile range for quantitative variables and relative and absolute frequencies for categorical variables. P-values < 0.05 will be considered statistically significant. For all the analyses, the Institution's Biostatistical Unit will support this study.

A non-parametric statistical method (Wilcoxon test) will be used for simple comparisons between PK parameters of both groups of patients (clinically controlled and uncontrolled with prophylaxis treatment).

The effects of time below 1 IU/dL, Vss, Cl and half-life on annual bleeding rates (ABR and AJBR) will be analyzed by multivariate analyses in both groups. Multivariate analysis on annual bleeding rates (as a dependent variable) will be performed separately for each of the potential factors associated with bleed rate such as PK parameters, physical activity, adherence to treatment schedule and quality of life. In all instances a regression model utilizing the negative binomial distribution will be used.

Annualized FVIII consumption and rates of treatment related AEs for each prophylaxis group will be compared using a Mann-Whitney U-test.

The sample size of 25-30 patients has been calculated to provide 90% power to detect a mean treatment difference of four ABR or AJBR episodes with two-sided a = 0.05, assuming a standard deviation (SD) of 6 ABR or AJBR episodes in six months. The sample size assumed an ABR variance of at least that observed for compliant subjects in a previous study. Sample size has been calculated using NQUERY, version 5.0, module MOT 1-1 (Statistical Solutions, Saugus, MA, USA).