Long-term prophylaxis: what are our options and how to define success?

Hemophilia A and B are genetic disorders resulting in deficiencies of factors VIII (FVIII) and IX (FIX), respectively. Historically, severe hemophilia was associated with crippling arthritis in 90% of affected individuals and with intracranial hemorrhage in up to 12%.^1-3 

Dr Inge Marie Nilsson began prophylactic infusions of low-dose FVIII every 5 to 12 days in 1958 to prevent bleeding and subsequent joint damage.⁴  The safety and efficacy of prophylaxis were later established in a prospective randomized controlled trial, called the Joint Outcome Study (JOS), in which children were randomized prior to age 2.5 years to receive either 25 IU/kg of a standard half-life recombinant FVIII (rFVIII) every other day or treatment with rFVIII only at the time of clinical bleeding.⁵  By age 6 years, the children on prophylaxis had significantly less functional and structural joint disease, as determined by sensitive joint physical examination and magnetic resonance imaging (MRI), compared with children treated for bleeding only. Clinically recognized bleeding was also greatly decreased with prophylaxis; however, physical exam and MRI evidence of joint damage was present in some joints of children on early prophylaxis without clinical bleeding events, suggesting that subclinical bleeding on prophylaxis can promote joint damage.⁵  In the JOS continuation study that followed JOS participants to age 18 years, those on early prophylaxis showed evidence of progressive joint disease on MRI, although it was less than those who had started on prophylaxis after age 6. At age 6, 7% of those who had started prophylaxis prior to 2.5 years of age had osteochondral damage, and 27% had osteochondral or synovial abnormalities, while at age 18, 35% had osteochondral damage, and 86% had osteochondral or synovial abnormalities in at least 1 joint despite 92% treatment adherence.⁶ 

Regarding quality of life, older adolescents on lifelong prophylaxis complained about the inconvenience of every-other-day venipuncture and lack of access to certain occupations, such as the military, but chronic pain and disability were not noted. In addition, no child on early prophylaxis developed intracranial or other life-threatening hemorrhage.⁶ 

The JOS and its continuation clearly demonstrated that prophylaxis with a standard half-life rFVIII product was a significant improvement over episodic treatment for bleeding and arthropathy but was not adequate to completely prevent joint disease. Additionally, the JOS supported earlier German data that prophylaxis must be started very early in life,⁷  when the technical challenges of frequent venipuncture are greatest.

Efforts have been made to develop more effective products for prophylaxis of joint damage in hemophilia, as listed in Table 1.

During the first 5 years of the JOS, trough FVIII levels in half of the participants were less than 1%.^5,8  Other studies have also shown rapid FVIII clearance in young children.⁹  The promise of full joint bleeding prevention seemed to rest on the ability to maintain higher trough levels; attention turned to extended half-life (EHL) products. Hermans and colleagues published a comparison of 8 pivotal trials including 900 participants using 4 EHL rFVIII products (rFVIIIFc, efmoroctocog alfa [Eloctate]; rurioctocog alfa pegol [Adynovate]; turoctocog alfa pegol [Esperoct]; damoctocog alfa pegol [Jivi]).¹⁰  In this study, median total annualized bleeding rates (ABRs) were reported at 1.3 to 1.9; achieved trough levels of 1% to 3% were judged insufficient to prevent joint bleeding or bleeding overall.^10-14  Real-world experience using the same 4 EHL rFVIII products was reported by Dettoraki et al. It included 23 children, half 3 to 12 years old and half 13 to 18 years old, with pharmacokinetic studies performed on both standard and EHL FVIII products.¹⁵  On EHL FVIII, mean trough levels rose from 2.3% to 4.1%; mean ABR decreased from 1.0 to 0.3, and mean joint ABR decreased from 0.8 to 0.2. Dosing frequency decreased from every 2.3 to every 3.5 days. The results of Dettoraki et al are similar to the individual licensure trial reports.^11,14 

A novel FVIII product in clinical trials, efanesoctocog alfa (BIVV001, rFVIIIc-VWF-XTEN), has been reported to increase the geometric mean FVIII half-life 4-fold to 42.5 hours using a dose of 65 IU/kg.¹⁶  At this dose, the plasma FVIII activity was above 50% for 4 days and was 17% at 7 days, suggesting potential joint protection with once-weekly dosing.

The EHL FIX products rFIX-FP (albutrepenonacog [Idelvion]), rFIXFc (coagulation FIX, Fc fusion protein [Alprolix]), and N9-GP (nonacog beta pegol [Rebinyn]) have achieved greater circulating plasma durations and levels. In the rFIXFc clinical trials, the geometric mean half-life was 82.1 hours (vs 33.8 hours standard rFIX) for adults and 68.6 hours in children under 12 years.^17-18  In the pediatric study Kids-B-LONG, in which all children were on standard FIX prophylaxis 1 to 3 times per week prior to initiating rFIXFc prophylaxis weekly, the overall bleeding rate was similar (2.5 [0.0-5.0] vs 2.0 [0.0-3.1]; median [interquartile range; IQR] for standard FIX vs rFIXFc), with minimal joint bleeding on rFIXFc prophylaxis (0 [0-1.1]; median [IQR] joint ABR).¹⁸  rFIX-FP clinical trials showed a mean half-life of 92 hours in adults and 91.4 hours in children,^19,20  with a similar median ABR in the pediatric studies (3.0 [1.0, 6.0], median [IQR] vs 3.12 [0.91, 5.91], median [IQR]) in standard vs rFIX-FP prophylaxis with fewer infusions and few joint bleeds (0.99 [0, 2.33], median [IQR]).²⁰  Nonacog beta pegol (N9-GP, Rebinyn) is a recombinant FIX product in which the FIX gene is linked to a 40 kD PEG moiety at the FIX activation peptide. In phase 3 studies, 29 adults with moderate to severe hemophilia B given 40 IU/kg showed the product to have a single-dose half-life of 93 hours, while in 25 children it had a half-life of 70 hours in those younger than 7 years of age and 76 hours in those 7 to 11 years; trough levels were higher than 15%.^21,22  The mean ABR was 1.04 (IQR, 0-4) in adults, with 67% of adults having no bleeding; mean ABR was 1.44 (95% CI, 0.92-2.26) in children. Although the study design and analytic methods to calculate half-life differed among these studies, in clinical practice these 3 EHL FIX products perform similarly, and most children have no spontaneous bleeding on once-weekly dosing. These long-acting products could be competitive with emerging nonfactor and gene therapy alternatives.^23,24 

The FVIII mimetic emicizumab has presented another major advance in replacement therapy for hemophilia. Emicizumab is a bivalent monoclonal antibody that binds FIXa and activates FX in the absence of FVIII. Bleeding outcomes for up to 3.5 years on emicizumab have been reported.²⁵  Bleeding on weekly or every-other-week dosing was reduced to less than 1 per year. Bleeding rates continued to decrease over time, with a calculated median joint ABR of 0. To date, there are no long-term outcome data on joint structure and function in persons using emicizumab prophylaxis, although the HAVEN 7 study of infants starting on emicizumab does include joint structure and function evaluations.²⁶  The FVIII mimetic currently known as Mim8 is in clinical trials.²⁷ 

Hemostatic rebalancing is an approach to maximize thrombin generation in persons with hemophilia by decreasing thrombin regulation. Although there are no currently licensed products for hemostasis rebalancing, clinical trials are in progress for a small-molecule mRNA silencer of antithrombin (fitusiran) as well as antibodies to tissue factor pathway inhibitor (TFPI; concizumab, marstacimab), with both showing promise.²⁸  Additional potential targets also include protein C and protein S.²⁹  The hemostatic rebalancing agents have the potential downside to be thrombogenic in high-risk situations, although thrombosis has been rare in the clinical trials to date. Fitusiran has the advantage that it can be overridden with antithrombin concentrates should the need arise. In the report of a phase 1 study, 17 adults with hemophilia A or B with inhibitors given 50 or 80  mg of fitusiran had an 82% (SD 2.2) and 87.4% (SD 0.7) reduction in antithrombin, respectively. Eleven of 17 participants had no bleeding during a mean observation period of 69.4 days. Thrombin generation was increased, but there were no clinical thromboses.³⁰  Concuzimab phase 2 clinical trials have been reported with good safety profiles. The joint ABR decrease was greater for hemophilia A and B with inhibitor (mean ABR 2.7 and 3.8, respectively) than for hemophilia A without inhibitor (mean ABR 4.9). All participant groups showed an elevation of thrombin generation into the normal range and elevated prothrombin fragment 1 plus 2 (particularly in noninhibitor patients) with free TFPI suppression to approximately 25% of baseline.³¹  A phase 1b/2 clinical study of marstacimab has also been recently reported.³²  Participants on marstacimab experienced an 86% to 90% reduction in ABR. Prothrombin fragment 1 plus 2 and dimerized plasmin fragment D increased without clinical thrombosis. Mildly elevated troponins in 3 participants were judged not to be related to marstacimab and resolved without a change in treatment. To date, the 3 hemostasis rebalancers appear to have similar safety and efficacy profiles, but the trials are not yet completed.

Gene therapy holds the ultimate promise for prevention of joint disease by the continuous endogenous release of FIX or FVIII to normal levels. Although to date no FVIII or FIX gene therapy products are approved by the US Food and Drug Administration or European Medicines Agency, great progress has been made, and 3 trials are in the late stages. Table 2 displays gene therapies under investigation in the US and Europe. Biomarin's FVIII AAV-5 trial has reported 3-year outcomes.³³  FVIII levels tended to decrease over the first year, with an apparent slower decline thereafter. At 3 years, 3 of 7 participants who received 6 × 10¹³  vector genomes (vg)/kg achieved an FVIII level of 20% to 36%, and 1 participant achieved 100%. Participants reported an almost complete cessation in clinical bleeding and FVIII infusions, even with low assayed plasma FVIII activity. A 2022 report of 132 participants in a phase 3 trial showed a median FVIIII activity of 23.5% at 3-year follow-up. The FIX trial of Spark Therapeutics reported 10 treated patients on a phase 1/2a study who achieved a mean FIX activity of 33.7% (range, 14-81), with a follow-up of 28 to 78 weeks.³⁴  These patients experienced a reduction in ABR from 11.1 (range, 0-48) to 0.4 (range, 0-4). Nine of the 10 reported no bleeding events, and 8 of the 10 used no factor replacement. The UniQure trial, AMT-061, reported 3 patients with a mean 26-week level of 47% (range, 33.2%-57%).³⁵  Both of these gene therapies as well as the St. Jude, University College London, and Royal Free Hospital trial utilize the Padua variant of FIX with increased activity relative to wild-type FIX. All of the gene therapies have shown variability in interindividual response. There are no long-term reports of joint structure or function following gene therapy. It is very important to document joint outcomes over time in these patients, as many reported no bleeding and no use of prophylaxis with a steady-state FVIII or FIX level less than 10%, and the occurrence or progression of arthropathy in the absence of clinical bleeding, clearly documented in the JOS, could occur in the absence of symptoms.

Today there is no clear winner in the race for optimal prophylaxis of bleeding and arthropathy in hemophilia. Maintenance of a factor level within a reasonable hemostatic level with weekly injections is likely possible for 1 or more FVIII and FIX EHL products. Emicizumab has already achieved wide acceptance with the elimination of most spontaneous bleeding and a consistent hemostatic capacity of approximately 20% FVIII activity equivalence on chromogenic FVIII assay using human substrates.³⁶  The hemostasis rebalancers continue in clinical trial, and interim data are encouraging. Finally, 1 or more gene therapies is expected to be presented for regulatory approval within the year.

All of the novel therapies decrease ABR to 1 or fewer per year, with median joint bleeding rates at or near 0. It is very difficult to distinguish novel therapy efficacy based upon joint or other bleeding rates. More subtle outcomes of joint dysfunction and pain caused by subclinical bleeding now assume primary importance.

The success of prophylaxis can be considered from multiple angles: achieving a normal factor level; decreasing the number of bleeds; decreasing the risk of catastrophic bleeding; improving joint health; improving quality of life; preventing or decreasing pain; and improving the ability to participate in employment, physical activity, sports, and life.³⁷ Table 3 displays outcomes of importance for hemophilia prophylaxis. Outcome instruments for hemophilia prophylaxis have been recently reviewed.³⁷  A wider community perspective is also essential. The cost of a successful prophylaxis therapy must be such that it can be disseminated to all affected persons and all national economies, without selecting recipients based upon extraordinary resources.

In assessing the effectiveness of hemophilia prophylaxis, factor level continues to be a critical outcome, as a large body of experience exists relating arthropathy risk to baseline or trough factor level. Gene therapy is evaluated, in large measure, by the level of factor achieved. Many of the novel factor products do not perform equivalently to standard factors in standard laboratory assays. Thus, a menu of assays, including 1-stage and chromogenic assays, are needed. Nonfactor products cannot be evaluated in specific factor assays, and thus global assays such as thromboelastogram, thrombin generation, and dynamic (eg, microfluidic assays) are required.

Because even short durations of blood within a joint can cause long-term damage, many efforts have focused on optimizing prophylaxis by finding joint damage early.³⁸ 

In the initial prophylaxis protocols of Sweden, which started in the late 1950s with long-term outcomes published in the 1990s, those who started on low-dose factor had joint changes visible on plain x-ray and physical exam, with those starting on higher prophylaxis doses showing less damage.³⁹  However, the techniques used were less sensitive, and the imaging and physical exam scores used for those studies had a higher threshold for defining damage than current scores. For example, in the orthopedic joint scoring system used by Lofqvist et al, 1 point is attributed for a loss of range of motion of 11% in the ankle or knee (a loss of 15 degrees of flexion in the elbow), while in the Hemophilia Joint Health Score (HJHS) commonly used today, 1 point is attributed for a loss of range of motion of 1 to 4 degrees.⁴⁰ 

The early physical exam and imaging scores were critically important to begin to evaluate the success of hemophilia treatment, but they were limited because they were not able to detect early joint abnormalities, which are more likely to be improved with augmented prophylaxis than more severe joint damage.⁴¹  Part of the data safety monitoring plan for the JOS included the monitoring of hemophilia-related joint health every 6 months using physical examination, prompting the development of a more sensitive joint physical examination tool to detect very early joint changes.⁴²  A sensitive MRI scale to image bone, cartilage, and soft tissue changes was also developed and validated for the JOS trial.⁴³  Around that time, the International Prophylaxis Study Group was formed to develop, validate, optimize, and disseminate sensitive joint assessment tools.^41,44-46 

The International Prophylaxis Study Group continues to refine the HJHS for reliability, sensitivity, and ease of training and use.⁴⁷  However, more functional and dynamic assessments are required to detect very subtle early changes. Exciting advances in the application of gait analysis to hemophilia joint monitoring promise techniques that are more sensitive as well as more functional to detect early changes and plan interventions.^48,49  The gold standard for detecting joint damage with imaging is MRI, but MRI is less accessible and more time-consuming and expensive. Ultrasound evaluation of joints is logistically much simpler and more accessible. The Hemophilia Early Arthropathy Detection with Ultrasound (HEAD-US) and Joint Tissue Activity and Damage Examination (JADE) scores have been developed to systematically evaluate joint damage and bleeding by ultrasound.^50,51  To date there are no long-term outcome studies reported using ultrasound. Cross-sectional studies show concordance of HEAD-US with HJHS,^52-54  although the HEAD-US was better at detecting structural changes and the HJHS was better at detecting functional changes, suggesting a role for each.^53,54 

The standard outcome in the development of all new hemophilia medications is annualized bleeding rate. For many study participants, bleeding events can at times be difficult to define. For example, it is challenging to distinguish between arthritis pain and bleeding pain.^55,56  Although both are critically important to the patient experience, and both should ideally be prevented, the implications are different in a clinical trial. In addition, clinically, not all bleeds are equal. An ankle bleed that is recognized immediately and treated with a single additional dose of factor is a very different patient experience compared to a more serious ankle bleed that takes months to fully resolve, but they are counted equally in clinical trials. Days without bleeding or days without pain may be the preferred outcome as prophylaxis methods continue to improve. Pain can only be prevented if joint damage is detected early and mitigated with physical therapy and augmentation of prophylaxis.

Patient outcomes are of increasing interest in assessing prophylactic efficacy. Ultimately, it is the patient who decides what his or her goals are in health care and how a particular therapy meets those goals. Relevant patient outcomes are displayed in Table 4. Quality-of-life assessments include those that are disease agnostic, such as the SF-36 and the Patient-Reported Outcomes Measurement Information System (PROMIS) measures^57-59 ; those specific to children, such as the Peds-QL⁶⁰ ; and those specific to hemophilia, such as Patient Reported Outcomes Burdens and Experiences (PROBE) and Canadian Hemophilia Outcomes-Kids' Life Assessment Tool (CHO-KLAT).^61,62  We prefer the PROMIS measures because they were developed using item response theory in which computerized adaptive testing methods can be employed, shortening the number of questions needed to assess an outcome. Customized questionnaires can be created from existing item banks to focus measurement on particular aspects of an outcome that are important to a specific group of patients, and scores on a particular measure can be calculated even when some data points are missing. In addition, PROMIS measures are publicly available, and they provide standardized scores using a T-score metric based on normative data from the US general population, allowing comparisons of scores across domains and disease states.³⁷ 

Pain, chiefly joint pain, comprises the dominant adverse patient outcome of hemophilia.⁶³  Pain intensity constitutes the severity of pain experienced, while pain interference denotes the impact of pain on one's life activity and behavior. There are several validated scales for quality of life as well as pain. The McGill Pain Scale as well as the Brief Pain Inventory have been used extensively in hemophilia.^64,65  The PROMIS is a US National Institutes of Health initiative that led to the development of PRO instruments, collections of validated surveys to assess for a number of domains, including pain intensity, interference, and behavior.^58,59  The PROBE questionnaire was developed for persons with hemophilia, with extensive patient engagement, and has been rigorously validated.⁶¹  PROBE covers both pain occurrence and interference using patient-derived definitions for acute and chronic pain and captures use of pain medication. The CHO-KLAT pediatric hemophilia outcomes score has been developed with significant input from hemophilia patients and has been cross-culturally validated.⁶⁶  The PROMIS and PROBE validated instruments are now being used longitudinally and within hemophilia clinical trials.^67-69 

Currently, we are at an enviable place in hemophilia treatment. Although full prophylaxis with standard half-life recombinant or plasma-derived factor concentrates has been definitively shown to be inadequate for full protection against bleeding and arthropathy, a number of novel therapies with improved hemostatic enhancement are clinically available or in promising clinical trials. In order to compare outcomes of very efficacious therapies, it is necessary to have sensitive tools, employed in long-term follow-up for several years, in participants with no or minimal joint disease. The tool kit must be comprehensive, with outcomes of bleeding, factor level restoration or hemostatic capacity, joint structure, joint function, pain, quality of life, and patient satisfaction.

Marilyn Jean Manco-Johnson: no competing financial interests to declare.

Beth Boulden Warren: consultancy: Novo Nordisk, Hema Biologics, Genentech.

Marilyn Jean Manco-Johnson: nothing to disclose.

Beth Boulden Warren: nothing to disclose.