Abstract

Background:

Prophylaxis with recombinant factor VIII (rFVIII) is considered the optimal treatment for severe or moderate haemophilia A (HA). Bleeding into joints determines a proportional chronic joint damage to its frequency and severity. Knowledge of the individual's pharmacokinetics (PK) using Bayesian analysis helps to individualize prophylaxis therapy with recombinant factor VIII (rFVIII) in severe or moderate haemophilia A (HA) and minimize the risk of bleeding, extracting only 2-3 samples.

Methods

Retrospective study in HA patients with rFVIII (Advate®) prophylaxis from January 2014 to May of 2016. Bayesian model (myPKFit®) was employed to perform an individual PK profile using the retrospective data of rFVIII levels. PK parameters analyzed were: clearance (Cl); steady state volume (Vss); plasma half-life (t1/2); and time to reach rFVIII levels <1% (T1%). Intraindividual and interindividual coefficients of variation (CV) of the t1/2 were calculated. Besides, Kruskal-Wallis test (R® version 3.1.2) was employed in comparisons between t1/2 and clinical variables: annualized bleeding rate (ABR), annualized joint bleeding rate (AJBR), Gilbert score (GS), Pettersson scale (PS) & lower extremity affected joints detected by NMR.

Results

Nineteen patients were analyzed, with a mean age of 32 years (SD: 11.3; range 11-46) and 86 PK monitoring (4.5 per patient). Two patients with <15 years were excluded because t1/2 was lower than adult patients. The mean PK values in adult patients were: Cl 2.9 (0.40) mL/h/kg; Vss 50.0 (<0.001) ml/kg; t1/2 14.1 (2.1) h; and T1% 74.4 (14.4) h. The mean intraindividual CV in t1/2 was 3.6% (SD 0.02; range 0.3-6.6), whereas interindividual CV was 14.8%. We categorized t1/2 in short (<p25: 12.3 h), normal (p25-p75) and long (>p75 14.4 h), with average age of 19.5; 39.4 & 35.8 years, respectively.

We detected significant differences between t1/2 and median values of joint state, but not in ABR and AJBR. The mean values of joint scores were: PS (5.5, 22.1; 17; p=0.028), GS (1, 25.9, 17.6; p= 0,008) and NMR (1.2, 2.6, 2.0; p=0.042) for short, normal and long t1/2, respectively.

The limited number of patients only allows observe significant differences in patients with short t1/2, patients that also have significantly lower age (p=0.007). The younger age of these patients also justifies the lower joint damage observed. After excluding two patients <15 years significant differences in the scores of the joint state disappeared, showing that age could be a confounding variable.

Conclusion

PK monitoring showed a low intraindividual variability in t1/2, but significant interindividual CV. Age could modify PK parameters, so it should be assessed in an integrated manner with other clinical variables. Bayesian estimate with MyPKFit® allows know the PK profile of each patient and could be a useful tool to individualize prophylaxis adjusting by the physical activity and the bleeding pattern. We are performing a personalized one-year program to identify and treat the specific causes of poor bleeding control in prophylaxis therapy, and these are our preliminary results.

Acknowledgments: This study will be supported by Baxalta grant "H15-29403".

Disclosures

Cid:Baxalta Innovations GmbH, now part of Shire: Other: Investigator Clinical Studies.

Author notes

*

Asterisk with author names denotes non-ASH members.