Traditional PK estimates individual parameters on the basis of the concentrations of the drug obtained at different sequential sample collections following administration of one dose of the replacement factor. No population-based models are used. According to the ISTH, measuring the individual pharmacokinetic profile of FVIII and FIX requires 9-11 adult patient samples (4 collected during the distribution phase and 5-7 during the excretion phase) and at least five samples in children6. A washout period of five t1/2 is also needed. The procedure also requires a firm commitment from the patient and their family given the large number of samples and the amount of time necessary, which makes its application difficult in clinical practice. These kinds of analyses are normally restricted to clinical trials that include small and homogeneous groups of hemophilic patients, where the reference technique selected must be subject to a low intrinsic error rate.

Bayesian analysis is a statistical procedure used to adjust patient data to a previously proposed general model. It uses the experimental information obtained from the individual (individual information) plus the information known ex ante about a drug's performance in a given population (population-based information) whose physiopathological characteristics are similar to those of the patient under analysis. If experimental information is limited, the influence of population-based values will be high. However, the influence of population-based values decreases as more experimental data become available. PopPK is capable of estimating individual parameters without the need of the exhaustive sample collection required by traditional PK4.

The first few PK studies on hemophilic patients undergoing surgery were published in the 1980's7–9. At that time, a retrospective study analyzed the PK values obtained in 350 surgical procedures in patients with hemophilia A7. A pre-infusion sample was collected at baseline, followed by a post-infusion one al 10 minutes to calculate the t1/2 and IVR of the replacement factor. In 1986, Ruffo et al. published the first software ever based on a PopPK model applied to surgery8, modifying previous models used for prophylaxis10. The new PopPK, based on FVIII, used a non-linear single-compartment strategy that assumed that the t1/2 of FVIII peaked immediately after surgery and gradually decreased over the next few days. The model suggested obtaining two samples, one at 3 hours and the other at 9 hours post-infusion. Sometime later, a nomogram was put together based on the PK data from 20 patients with hemophilia A undergoing surgery, which allowed determination of the required maintenance dose on the basis of the FVIII concentration observed 10 hours after application of the loading dose for three target FVIII steady-state concentrations (30, 60 or 90 IU/dL)9.

Durisová et al. estimated the PK of FVIII intraoperatively by using the “frequency-response” method, based not only on post-surgical FVIII levels as previous models, but also considering pre-surgical FVIII concentrations11. The sheer complexity of the model combined with the lack of a software that facilitated its application and the failure to titer the dose in some patients meant that the model was soon neglected.

Bolon-Larger et al. developed a two-compartment PopPK model using a non-linear mixed effects modeling (NONMEM) strategy based on data from 33 patients with hemophilia A12. Of the different samples analyzed, the ones exhibiting the greatest accuracy and the fewest biases were those obtained at 0.5 and at 6-8 hours post infusion. Body weight, body surface area, and the baseline FVIII concentration were the covariates with the greatest influence on the distribution volume (Vd).

Intraoperative use of continuous infusion (CI) of coagulation products was for some time promoted over intermittent injections with a view to achieving more stable levels of FVIII/FIX and reducing overall factor consumption. However, continuous infusion has lately fallen into disuse. The literature search performed as part of this study detected 8 studies on the use of CI during surgery, 6 of which with FVIII13–18 and 2 with FIX19,20. These studies usually included a preoperative individual PK analysis with nine sample collections and using non-compartmental models to titer the dosage of the clotting factors, following the recommendations of the ISTH6. These studies confirmed that clearance decreases over the first five days post-op, making it possible to adjust the dosing schedule and reduce factor consumption. Use of CI with extended t1/2 factor (EHL) products during surgery was only reported in the case of one patient, who was being treated with efmoroctocog alfa18. In that case, WAPPS-Hemo® was used to analyze previous PK values21.

A study by Suzuki et al. compared five different methods for calculating the clearance of nonacog alfa in patients on CI22. The method, based on IVR and t1/2 distribution, obtained similar clearance rates as direct CI calculations, while clearance calculations using terminal t1/2 and AUC values underestimated the clearance rate. The simulated single-compartment model also obtained good correlations.

The advent of the new EHL factor products has resulted in the performance of new research into the intraoperative behavior of PK, both for EHL FVIII23–25 and FIX26–28. Most of these studies carry out a preoperative individual non-compartmental PK analysis to estimate the PK parameters. The exception is a study on eftrenonacog alfa (rFIXFc) that used the PopPK model in patients on prophylaxis29 and compared the estimated levels with the real-life levels, obtaining an excellent correlation between them28.

The OPTICLOT trial was designed to create a FVIII PopPK model for surgical use and compare it to the results of standard dosing30. A bicompartmental surgical model was developed using the NONMEM technique with data from 140 procedures on 75 adult patients and 58 procedures on 44 children with hemophilia A31. Covariates in this PopPK analysis included age, blood type and type of surgery. The model was validated through a cohort of 87 pediatric patients, and a new model comprising a total of 206 patients was generated32. This new model significantly improved the available predictions, with the estimation accuracy improving from a median underestimation of 17 IU/dL to a median overestimation of 2 IU/dL. Similarly, a three-compartment surgical PopPK analysis was developed using the NONMEM technique, with data from 255 procedures on 118 patients with hemophilia B33. Body weight, age and type of FIX were the main covariates.

Few studies have compared the PK values of the new EHL factors with those of standard t1/2 (SHL) factors or other EHL factors during surgery. Noteworthy among them is a study on nonacog beta pegol (N9-GP). The authors developed a specific bicompartmental PopPK and analyzed its performance in different scenarios against recombinant FIX (rFIX) and plasma-derived FIX (pdFIX)34. Intraoperative simulations were carried out to compare the dosing regimens of N9-GP, rFIX and pdFIX needed to achieve the target FIX levels established by the WFH (100-120 IU/dL pre-op, 40 IU/dL at days 1-3; 30 IU/dL at days 4-6; and 20 IU/dL at 7-14 post-op)35. Use of N9-GP, as measured in IUs/kg, was 80% lower than that of rFIX and pdFIX; the number of infusions required was also lower (2 vs.16).

With the help of the results of a cross-sectional clinical trial comparing the PK of EHL FIX products with that of N9-GP and rFIXFc, specific PopPK (single- and three-compartment, respectively) models were designed for surgical and on-demand applications36. The model was used to make estimations on the basis of the recommendations of the WFH for factor administration during surgery37, with N9-GP requiring the lowest number of infusions (67% and 55% in major and minor surgery, respectively) and the lowest product consumption (67% and 58% in major and minor surgery, respectively).

The FVIII PopPK model developed under the OPTICLOT trial showed that clearance decreased with age and was 26% higher in patients with blood type 031. Moreover, a 7% decrease in clearance was observed in major surgeries as compared with minor ones. Two case reports on PK-based FVIII dose titration during surgery concluded that the recommended weight is the ideal body weight both for obese patients38 and in cases of extreme weight loss39.

FIX PopPK showed that clearance and the distribution volume in the central compartment (V1) grew gradually with age and increasing body weight until the age of 2033. Patients treated with pdFIX showed lower clearance and V1 levels than those treated with rFIX (11% and 17%, respectively). Similarly, V1 in patients with moderate hemophilia B was 10% lower than in those with severe hemophilia B.

Other studies analyzed the influence of the covariates on PK during surgery. A study on pdFIX compared IVR before and after surgery and demonstrated an age, weight and pdFIX type-dependent effect on dosing40. A retrospective study analyzed the variables that influenced IVR and clearance in patients on CI of FVIII during surgery and detected differences in the clearance rate depending on the subjects’ body mass index and the type of FVIII used17.

Another study analyzed the factors capable of predicting FVIII under- and overdosing in 198 surgeries on 119 patients41. Blood type 0 and major surgery turned out to be predictors of overdosing, while underdosing was associated with increasing age, plasma-derived FVIII and intermittent infusion. Blood type 0 was also associated with an increased bleeding rate.

To ensure bleeding control during surgery, the clinical guidelines recommend maintaining FVIII/FIX levels above specific thresholds for specific time periods. Exact target levels will depend on the type of surgery performed. These recommendations have undergone certain changes over time3,35,37. Table 2 contains the recommendations of the latest WFH guidelines3.

Another approach is the one followed by the Delphi expert consensus, which redefined the target plasma levels of FVIII, replacing the traditional target level of 1 IU/dL by 8 different target levels42. Four of these target levels are related to surgery, with thresholds being established depending on the different surgical stages and the complexity of the surgery (Figure 2).

Figure 2.

Target intraoperative FVIII plasma levels. Adapted from lorio et al. 201743.

(0.11MB).

Against this background, hemophilic patients undergoing surgery should have their peak levels measured 15-30 minutes after the replacement factor has been infused; trough levels should also be regularly measured. It has also been suggested that a full preoperative PK or a PopPK analysis be conducted. The former would require a larger number of samples (9-11) than the latter to adjust the preoperative dose and, if needed, the continuous infusion rate, based on the calculated clearance rate3. These preoperative PK estimations must be adjusted during surgery by regularly measuring peak and trough factor concentrations. Several studies have shown that the FVIII/FIX concentrations obtained with this method tend to fall outside the established target range, leading to under- or overdosing41,43. This is the reason why the OPTICLOT group is promoting the use of surgery-specific PopPK models to make these kinds of adjustments31,33.