Replacement therapy during surgery in von Willebrand disease needs personalization

In this issue of Haemophilia, Hazendonk and colleagues1 report results from a multicentre retrospective study in which current perioperative management with a dual VWF/FVIII concentrate (Haemate-P) in von Willebrand disease (VWD) was evaluated. The study was conducted in 5 academic centres in the Netherlands, and subjects were selected as follows: a clinical and laboratory diagnosis of VWD using cut-off levels of VWF:Antigen (VWF:Ag) ≤0.30 IU mL−1 and/or VWF:Activity (VWF:Act) ≤0.30 IU mL−1 and/or FVIII ≤0.40 IU mL−1. Patients undergoing minor or major surgery with use of replacement with Haemate-P during the years 2000 through 2015 were eligible. The objective of the study was to evaluate current perioperative management by specification of concentrate administration and analysis of subsequently achieved peak and trough levels of VWF:Act and FVIII in comparison with the target VWF and FVIII levels prescribed by national guidelines. In addition, predictors of high/low levels of VWF and FVIII (eg blood group, mutation type) as well as surgical variables associated with factor consumption were collected. The study population consisted of 103 subjects undergoing 148 surgical procedures. A broad range of procedures and VWD subtypes was included. Most patients (90%) received bolus injections of concentrate. No differences were observed in achieved VWF:Act and FVIII levels among patients with type 1, type 2 and type 3 VWD after replacement therapy, and most perioperative levels of VWF:Act and FVIII were well above predefined targets. Postoperatively, accumulation of FVIII was observed after repeated dosing, resulting in increased FVIII in comparison with VWF:Act. Thirteen (8%) of FVIII trough levels were above 2.70 IU mL−1. In 54 patients with type 1 VWD undergoing major surgery, 84% of trough and steady-state VW:Act levels in the first 36 hours after surgery were above predefined target level with a median deviation of 0.80 IU mL−1 (IQR 0.38-1.11), whereas 7 were below target. Similar results were recorded for FVIII with a median trough of 1.46 IU mL−1 (IQR 0.43-1.07). Only 5 patients were below the predefined level. Overall, no differences were observed for minor versus major surgery, blood group non-O versus O or adults versus children. In the 43 patients with type 2 and 6 with type 3 VWD, the same pattern was seen as for those with type 1. Clinically relevant bleeding only occurred in 5 (3.4%) surgical procedures. Despite excessive FVIII levels, no thrombotic complications were reported. Of the 18 patients reaching very high (>2.70 IU mL−1) levels, 61% received thromboprophylaxis with low molecular weight heparin. Predictors of high/low factor levels could only be evaluated in type 1 and type 2 VWD. Only blood group O was predictive of high VWF:Act levels but not FVIII. Most patients were treated with VWF/FVIII concentrate only. In a few patients, desmopressin, with or without concentrate, was used. Interestingly, concentrate treatment did generally not differ between type of surgery or type of VWD with the exception of the group with type 1, where the maintenance dose during day 1 after surgery in fact was slightly, but significantly, higher for minor surgery (33 IU kg−1 for minor and 26 IU kg−1 for major surgery). The authors discuss obvious limitations of their study such as retrospective design, different assays used at the centres and lack of application of a clear definition of regular monitoring. However, the study cohort is relatively large and reflects real-life performance of surgery in VWD. They conclude that perioperative replacement therapy in VWD is successful with few bleeding complications. They also conclude that the patients in the study were overtreated with accumulation of FVIII as a consequence. Fortunately, no thrombotic complications were seen, but about one-half of the patients were on thromboprophylaxis. The authors hypothesize that population PK models may be a promising tool for personalization of replacement therapy in all patients with VWD. The study by Hazendonk et al is important as it provides real-life data on performance of surgery in VWD. Reports of surgery with replacement therapy are more accessible for haemophilia than for VWD, and there are several reasons for that. These include the following: (i) the number of patients with haemophilia in need of replacement therapy substantially exceeds that with VWD, likely by a factor of 10 (approximation by the author); (ii) haemophilia is a more homogenous disease (eg if restricted to FVIII deficiency only mild, moderate and severe forms) compared with VWD types which are more complex as both plasma coagulation (FVIII) and primary haemostasis (VWF) are compromised and interaction between VWF/FVIII further complicates prediction of levels, especially hours after a bolus injection; and (iii) concentrates used for replacement therapy in VWD are far more diverse than those used in treating haemophilia. All these reasons hamper the ability to report findings from large and well-defined cohorts of surgical procedures in VWD, although a few such studies have been published for specific concentrates.2-6 Altogether, the 5 studies report 257 procedures with excellent or good efficacy ranging from 91% to 100%. Two studies reported 21 paediatric cases, all with good or excellent efficacy. The concentrates used were different in VWF/FVIII ratio. One product was almost devoid of FVIII with a ratio of 1:≤0.2 (Wilfact/Wilfactin); another (in 2 studies) had a 1:1 ratio (Wilate), and in 2 studies, the same brand was used as in the study reported by Hazendonk et al with a ratio of 1:0.4 (Humate-P) (all ratios given are approximate). Accumulation of FVIII after repeated dosing was not seen using the 1:1 ratio concentrate or the purified concentrate,2, 6 whereas the studies using the product with 1:0.4 ratio reported some degree of accumulation after multiple dosing,3, 5 although not to the same degree as reported in the study by Hazendonk et al. Lethagen and colleagues5 reported one case of a pulmonary embolism without sequelae in a patient with multiple risk factors. It is clear that different compositions of VWF/FVIII concentrates produce different pharmacokinetic (PK) profiles in patients with VWD. In a head-to-head PK study between Wilate and Humate-P,7 a clear difference in FVIII decay was seen. The level of FVIII was sustained using the latter product due to recruitment of endogenous FVIII. Thus, theoretical considerations, substantiated by comparative data, support the notion that different concentrates may confer different thrombotic risk levels of FVIII in the peri- and postoperative management of surgery in VWD. Let us consider this in the light of the study by Hazendonk et al. A few statements might be formed if their results are considered as a foundation: A conclusion that can be drawn from the statements is that the dosing in the study by Hazendonk is not optimal. The doses many patients received were too high which raises safety and economic issues. Although no serious adverse events occurred, this may be explained by the relatively high use of thromboprophylaxis. We do not know how thrombogenic procedures were implemented or the nature of the risk factors among those subjects not receiving thromboprophylaxis. We recently published a paper in which the “Dos and Don'ts” in surgery in the presence of VWD were discussed, and important issues for consideration were listed.10 The relevance of these issues is underscored by the paper by Hazendonk and colleagues. A take-home message from these papers, as well as other reports of surgery in patients with VWD, may well be: plan surgery at experienced centres having close collaboration between the surgery and haemophilia departments and be sure that you know the characteristics of your patient (VWD type, factor levels, thrombotic risk), the concentrate (VWF/FVIII ratio, pharmacokinetic properties, as these may differ between brands7) and the surgical procedure (minor, major, risks of thrombosis or bleeding). Hazendonk et al raise the issue of PK models as tools for personalization of replacement therapy in VWD. In view of the issues discussed above, it seems, indeed, very important for a better personalization of replacement in VWD. This is not as “easy” as it is in haemophilia, where population PK models have been published and are implemented in clinical practice.11-13 The complex variation between VWD concentrate brands and VWD types makes population PK and personalized care more difficult to develop in VWD and, so far, prediction of levels during surgery using parameters such as blood group or pharmacokinetics does not seem reliable.1, 5, 10 This is, consequently, a field of clinical research with a high priority for patients, health care and society. The author has received research grants, or honoraria for lectures and advisory boards, from CSL Behring, LFB, Octapharma and Shire.