Ten-year experience of recombinant activated factor VII use in surgical patients with congenital haemophilia with inhibitors or acquired haemophilia in Japan

Subjects and data collection

All patients with congenital haemophilia with inhibitors to FVIII or FIX or acquired haemophilia who received rFVIIa for the management of bleeding during surgical procedures at the 18 sites participating in the postmarketing surveillance conducted by Novo Nordisk Pharma Ltd. from May 2000 to March 2010 were included in this study.

During the surveillance period, the following data were collected for patients who received rFVIIa treatment: haemophilia type, age, sex, body weight, medical history and complications, date and type of surgery, transfusions, use of rFVIIa, haemostatic efficacy of rFVIIa treatment, laboratory test results before and after surgery, adverse events and concomitant administration of other drugs. Observations were continued for 6 months after surgery. Adverse events were defined as all medically unfavourable events that occurred during or after administration of rFVIIa (adverse drug reactions, abnormal laboratory test results and unexpected symptoms). Events for which a causal relationship with rFVIIa could not be ruled out by the participating investigators were classified as adverse drug reactions.

The study protocol was approved by the institutional review boards of all participating institutions, and was conducted in accordance with the principles of the Good Post-Marketing Study Practice 7.

Haemostatic efficacy of rFVIIa treatment

The haemostatic efficacy of rFVIIa treatment during and after surgery was evaluated by the participating investigators. Intraoperative bleeding was judged subjectively based on the experience of the investigators as less than, the same as, or more than expected for patients with haemophilia undergoing surgery (including those who receive other haemophilia inhibitor products or blood coagulation factor products), regardless of the inhibitor status of individual patients. Postoperative bleeding control (up to 3 days after surgery) was judged as effective (bleeding stopped or considerably reduced), slightly effective (bleeding slightly reduced) or not effective (bleeding not reduced). Maintenance of haemostasis of the surgical wound (until suture removal) was classified as yes or no. ‘Complete haemostasis’ was defined as complete control of bleeding that satisfied all of the following criteria: blood loss judged as less than expected or the same as expected, postoperative bleeding control judged as effective, and maintenance of haemostasis classified as yes. Haemostatic efficacy results are presented in a descriptive manner.

Subjects

Thirty-seven patients received rFVIIa treatment during the study period, including 24 with congenital haemophilia A, seven with congenital haemophilia B and six with acquired haemophilia. All 37 patients (66 procedures) were included in the safety analysis. The 34 patients for whom haemostatic efficacy data were available were included in the efficacy analysis, including 38 procedures in 22 patients with congenital haemophilia A (median age: 18.5 years, age range: 1–65 years), 13 procedures in seven patients with congenital haemophilia B (median age: 18 years, age range: 10–37 years) and five procedures in five patients with acquired haemophilia (median age: 62 years, age range: 39–79 years) (Table 1). Twelve of the patients with congenital haemophilia with inhibitors were also included in previous studies 8-10.

The 38 surgical procedures in patients with congenital haemophilia A included 16 orthopaedic procedures (42%), six central venous catheter or port placement procedures (16%), six dental/oral procedures (16%), four gastrointestinal procedures (11%), four suture/incision procedures (11%) and two thoracic procedures (5%). Continuous rFVIIa infusion was administered after bolus injection for nine procedures (24%). Tranexamic acid was used concomitantly for 24 procedures (63%) (Tables S1 and S2).

The 13 surgical procedures in patients with congenital haemophilia B included eight orthopaedic procedures (62%), two central venous catheter or port placement procedures (15%), 1 dental/oral procedure (8%), one gastrointestinal procedure (8%) and one suture/incision procedure (8%). Continuous rFVIIa infusion was administered following bolus injection for six procedures (46%). Tranexamic acid was used concomitantly for nine procedures (69%) (Tables S3 and S4).

The five surgical procedures in patients with acquired haemophilia were two gastrointestinal procedures (40%), one dental/oral procedure (20%), one obstetric procedure (20%) and one neurosurgical procedure (20%). Tranexamic acid was used concomitantly for three procedures (60%) (Tables S5 and S6).

rFVIIa doses

Table 2 shows the doses of rFVIIa administered for common procedures in patients with congenital haemophilia A or B with inhibitors. Five of the six port placement procedures were performed with bolus injections and one was performed with continuous infusion of rFVIIa. In the patients who received bolus injections, the median dose per injection was 109 μg kg⁻¹ (range: 96–117 μg kg⁻¹) and the median number of injections was 15 (range: 3–64). The initial infusion rate in the patient who received a continuous infusion was 28 μg kg⁻¹ h⁻¹. In these six patients, the median duration of treatment was 4.5 days (range: 3–28 days) and the median total dose was 1765 μg kg⁻¹ (range: 330–7460 μg kg⁻¹).

Five of the 11 arthroscopic synovectomy procedures were performed with bolus injections and six were performed with continuous infusion of rFVIIa. In the patients who received bolus injections, the median dose per injection was 125 μg kg⁻¹ (range: 107–141 μg kg⁻¹) and the median number of injections was 39 (range: 29–55). In the patients who received continuous infusion, the median initial infusion rate was 32 μg kg⁻¹ h⁻¹ (range: 17–60 μg kg⁻¹ h⁻¹). In these 11 patients, the median duration of treatment was 13 days (range: 5–16 days) and the median total dose was 5441 μg kg⁻¹ (range: 1690–7340 μg kg⁻¹).

One of the three joint replacement procedures was performed with bolus injections and two were performed with continuous infusion of rFVIIa. In the patient who received bolus injections, the mean dose per injection was 120 μg kg⁻¹ and 72 injections were administered. The median initial infusion rate in the patients who received continuous infusions was 32.5 μg kg⁻¹ h⁻¹ (range: 29–36 μg kg⁻¹ h⁻¹). In these three patients, the median duration of treatment was 12 days (range: 7–15 days) and the median total dose was 5000 μg kg⁻¹ (range: 3890–8640 μg kg⁻¹). The doses and haemostatic efficacy for all procedures are available in the online data supplements (Tables S1–S6).

Efficacy

Table 3 shows the haemostatic efficacy of rFVIIa treatment. In patients with congenital haemophilia A, the blood loss was judged as less than expected for 20/38 procedures (53%), the same as expected for 15/38 procedures (39%) and more than expected for 3/38 procedures (8%). Postoperative bleeding control was judged as effective for 34/38 procedures (89%), slightly effective for 4/38 procedures (11%) and not effective for 0/38 procedures (0%). Two of the three procedures judged as having more blood loss than expected were also among those judged as having slightly effective postoperative bleeding control. Maintenance of haemostasis was classified as yes for 36/38 procedures (95%) and no for 2/38 procedures (5%). ‘Complete haemostasis’ was achieved for 32/38 procedures (84%).

In patients with congenital haemophilia B, the blood loss was judged as less than expected for 7/13 procedures (54%), the same as expected for 2/13 procedures (15%) and more than expected for 4/13 procedures (31%). Postoperative bleeding control was judged as effective for 10/13 procedures (77%), slightly effective for 3/13 procedures (23%) and not effective for 0/13 procedures (0%). Maintenance of haemostasis was classified as yes for 11/13 procedures (85%) and no for 2/13 procedures (15%). ‘Complete haemostasis’ was achieved for 8/13 procedures (62%).

In patients with acquired haemophilia, the blood loss was judged as less than expected for 1/5 procedures (20%), the same as expected for 2/5 procedures (40%) and more than expected for 2/5 procedures (40%). Postoperative bleeding control was judged as effective for 4/5 procedures (80%) and not effective for 1/5 procedures (20%). Maintenance of haemostasis was classified as yes for 3/5 procedures (60%) and no for 2/5 procedures (40%). ‘Complete haemostasis’ was achieved for 1/5 procedures (20%).

Rescue medication

Haemostasis was judged to be insufficient with rFVIIa treatment alone for 2/21 minor surgical procedures (teeth extraction and port placement), and these patients received prothrombin complex concentrate (PCC) or neutralization/immune tolerance induction with FVIII products as rescue medication (Table S1, surgical procedures 5 and 12). Haemostasis was also judged to be insufficient with rFVIIa treatment alone for 5/35 major surgical procedures (one cholecystectomy, two arthroscopic synovectomy procedures, one sigmoidectomy and one lumbo-peritoneal shunt placement). These patients received activated PCC (aPCC) or sequential therapy with rFVIIa and aPCC as rescue medication (Table S2, surgical procedure 38; Table S4, surgical procedures 45 and 46; Table S6, surgical procedures 55 and 56).

Safety

Three adverse drug reactions were reported among 66 procedures in 37 patients. A 11-year-old male with congenital haemophilia A who underwent elbow arthroplasty (Table S2, surgical procedure 19) developed mild thrombophlebitis at the site of the peripheral venous line, 5 days after the initiation of rFVIIa treatment by continuous infusion without dilution. This event was judged to be related to rFVIIa treatment. A 1-year-old male with congenital haemophilia A who underwent port implantation (Table S1, surgical procedure 14) developed a mildly decreased platelet count and a causal relationship with rFVIIa treatment could not be ruled out. A 79-year-old male with acquired haemophilia who underwent sigmoidectomy (Table S6, surgical procedure 55) developed hepatic dysfunction with a mild increase in the serum alanine aminotransferase level and a causal relationship with rFVIIa treatment could not be ruled out. These adverse events all resolved without treatment.

rFVIIa doses for minor and major surgery

For the six port placement procedures in the present study, the total number of rFVIIa injections ranged from three to 64, the treatment period ranged from 3 to 28 days, and the total dose ranged from 330 to 7460 μg kg⁻¹ (Table 2). The patient who received the smallest total dose (330 μg kg⁻¹) received supplementary rFVIIa in addition to the main treatment with aPCC (Table S1, surgical procedure 11). In the patient who received the smallest number of injections over the shortest period (three injections over 3 days), use of rFVIIa was stopped when immune tolerance induction with rFVIII was started (Table S1, surgical procedure 12). The patient with the longest treatment period (28 days) required longer-term use of rFVIIa because of surgical wound dehiscence (Table S3, surgical procedure 42). For the remaining three port placement procedures, only rFVIII was used, the number of injections ranged from 11 to 64, the treatment period ranged from 3 to 11 days, and the total dose ranged from 1080 to 7460 μg kg⁻¹. For these three procedures, the treatment was similar to the previously reported treatment for eight Broviac catheter or port placement procedures in patients who received 90 μg kg⁻¹ per injection, where the number of injections ranged from 26 to 98 and the treatment period ranged from 3 to 13 days 6.

Concomitant use of aPCC, PCC or rFVIII makes it difficult to assess the efficacy of rFVIIa treatment alone. In the six patients who underwent arthroscopic synovectomy and received only rFVIIa, the treatment period ranged from 7 to 15 days and the total dose ranged from 3090 to 7340 μg kg⁻¹. In the three patients who underwent joint replacement and received only rFVIIa, the treatment period ranged from 7 to 15 days and the total dose ranged from 3890 to 8640 μg kg⁻¹ (Table 2). The doses administered for arthroscopic synovectomy and joint replacement were therefore similar in the present study.

Continuous infusion

In the present study, 24% of patients with congenital haemophilia A and 46% with congenital haemophilia B received a continuous rFVIIa infusion. Continuous rFVIIa infusion is used when physicians prefer to avoid frequent injections. The initial infusion rate in these patients was 29–36 μg kg⁻¹ h⁻¹. In previous studies, continuous infusions were often started at 15–30 μg kg⁻¹ h⁻¹, but this dose was not always sufficient to achieve haemostasis in patients undergoing major surgical procedures 11, 12, and a higher dose such as 50 μg kg⁻¹ h⁻¹ may be required for optimal haemostasis 13, 14. Bolus injection may be useful during active bleeding because it induces a thrombin burst 15. Once the bleeding has stopped, continuous infusion may be effective for maintaining haemostasis because it maintains the plasma rFVIIa level and avoids low trough levels until the wound has healed 14, 15. As continuous infusion of undiluted rFVIIa may result in superficial thrombophlebitis, it is preferable to administer the infusion via a saline line 16. Mild superficial thrombophlebitis was reported in one patient in the present study who received a continuous infusion of undiluted rFVIIa.

Efficacy

Among patients with acquired haemophilia in the present study, postoperative bleeding control was effective in 80% of cases but ‘complete haemostasis’ was achieved in only 20% of cases (Table 3). This may be because most patients with acquired haemophilia did not undergo their surgical procedures at specific haemophilia treatment centres, and the attending staff therefore lacked experience in the management of haemostasis in such patients. Surgical procedures in patients with acquired haemophilia have rarely been reported 17, 18.

The efficacy rate of rFVIIa treatment observed in the present study is similar to that of previously reported studies (Table 3). A review published in 2011 identified 395 reported surgical and other medical procedures for which rFVIIa had been used (261 surgical procedures, 89 dental procedures, 45 other medical procedures) in 263 patients with haemophilia with inhibitors, including data from the randomized comparative trial described above, the Hemophilia Research Society/Hemophilia and Thrombosis Research Society database and other studies 6, 10, 19. The reported overall efficacy rate was 84% (333/395), but the definitions of efficacy varied among studies.

Safety

In the present study, mild superficial thrombophlebitis was reported in one patient who received a continuous infusion of undiluted rFVIIa. In the previously published review article, serious thrombotic events (not including superficial thrombophlebitis) were reported in 2/263 patients 19. One of these patients was a 4-year-old who underwent port placement and developed thrombosis of an internal jugular vein 6, and the other patient underwent left knee arthroplasty and developed thrombosis of the popliteal and fibular veins 13.