Prevalence of selected bleeding and thrombotic events in persons with hemophilia versus the general population: A scoping review

Life expectancy for persons with hemophilia has increased over recent decades due to advances in treatment practice and patient care. Those with hemophilia are now more likely to be affected by conditions associated with aging, such as myocardial infarction, hemorrhagic/ischemic stroke, deep vein thrombosis, pulmonary embolism, and intracranial hemorrhage. Here, we describe the results of a literature search designed to summarize current data on the prevalence of the above selected bleeding and thrombotic events in persons with hemophilia vs the general population. A total of 912 articles published between 2005 and 2022 were identified in a search of BIOSIS Previews, Embase, and MEDLINE databases conducted in July 2022. Case studies, conference abstracts, review articles, studies focusing on hemophilia treatments or surgical outcomes, and studies examining patients with inhibitors only were excluded. After screening, 83 relevant publications were identified. The prevalence of bleeding events was consistently higher in hemophilia populations vs reference populations (hemorrhagic stroke, 1.4%-5.31% vs 0.2%-0.97%; intracranial hemorrhage, 1.1%-10.8% vs 0.04%-0.4%). Serious bleeding events showed a high rate of mortality with standardized mortality ratios for intracranial hemorrhage ranging from 3.5 to 14.88. Although 9 studies reported lower prevalence of arterial thrombosis (myocardial infarction/stroke) in hemophilia vs general populations, 5 studies reported higher or comparable prevalence in hemophilia. Prospective studies are therefore needed to understand the prevalence of bleeding and thrombotic events in hemophilia populations, particularly with the observed increases in life expectancy and availability of novel treatments.


| I N T R O D U C T I O N
Hemophilia A and hemophilia B are rare bleeding disorders caused by congenital deficiencies in coagulation factors VIII or IX, respectively.
Advances in treatment and management over recent decades have led to vast improvements in the quality of life for persons with hemophilia [1]. Prophylactic treatment with recombinant or plasma-derived factor concentrates, the current standard of care for many patients, has been advanced by the availability of extended half-life factor products that require less frequent intravenous dosing. Nonfactor therapies such as emicizumab, concizumab, marstacimab, and fitusiran are either available or are under clinical trials as novel therapies for patients with and without inhibitors, with the major advantage of subcutaneous administration.
Life expectancy for persons with hemophilia has increased as a result of these advancements, and an increasing number of patients are now more likely to be affected by conditions typically associated with an aging population [2]. This includes a range of cardiovascular (CV) complications such as myocardial infarction, hemorrhagic/ ischemic stroke, deep vein thrombosis, pulmonary embolism, and intracranial hemorrhage (ICH). However, currently, there are gaps in our knowledge regarding the prevalence of these events in hemophilia populations. For example, it is unclear whether the prevalence of these events is comparable with that of the general population and whether persons with hemophilia are protected from thrombosis due to the nature of their condition despite the advances in replacement therapy regimens. Historically, persons with hemophilia have been thought to be less likely affected by thrombotic conditions due to hypocoagulability, which is associated with reduced thrombin generation [3]. However, the literature appears to present conflicting evidence regarding the prevalence of CV events in hemophilia populations vs the general population. Although many studies support the idea that prevalence of CV events is generally higher in the general population [4,5], conversely, several studies have found that, at least in some populations, prevalence of certain thrombotic CV conditions is in fact higher in those with hemophilia [6][7][8]. There is therefore uncertainty regarding the protective effect of hemophilia under thrombotic conditions.
In addition to the studies investigating the prevalence of CV events, studies have also evaluated the mortality associated with these conditions in persons with hemophilia compared with that in the general population [2,9,10]. Although ICH is one of the leading causes of death in persons with hemophilia, mortality due ischemic heart disease has been reported to be lower in persons with hemophilia [9]. The question remains as to whether all-cause or cause-specific mortality is decreasing over time for persons with hemophilia and whether this can be concluded based on the currently available data.
A scoping review was designed to extract and summarize currently available data on the prevalence of the selected bleeding and thrombotic events and mortality in populations of persons with hemophilia. Special attention was given specifically to publications that compared the prevalence in hemophilia with that in matched controls of the general population. Scoping reviews differ from systematic reviews in that they do not aim to answer a specific research question or understand the mechanisms that explain the identified data [11]. This approach is used when it is unclear whether more specific research questions can be answered with the currently available data (eg, whether persons with hemophilia are protected from thrombosis). The aim of this review was therefore to provide a comprehensive overview of the current state of the literature in these areas (bleeding, thrombosis, and mortality) to identify and discuss relevant research gaps that could be addressed in the future. Analyses extension for scoping reviews checklist [12] was used to guide drafting of the manuscript.

Essentials
• We compare the prevalence of bleeding/thrombotic events in hemophilia vs general population.
• Bleeding event prevalence was consistently higher in persons with hemophilia than those without hemophilia.
• Data suggest that the prevalence of thrombotic events is not always lower in populations with hemophilia.
• Prospective cohort studies are needed to examine bleeding/thrombosis prevalence in hemophilia.

| Screening
A flowchart of the process used to screen the abstracts identified in the search is provided in the Figure. To identify peer-reviewed studies reporting primary data on the prevalence of the selected events in the populations of persons with hemophilia, the following publication types were excluded: case studies, conference abstracts, posters, guidelines, consensus statements, editorials, comments, and reviews.
The following study types were also excluded as they were deemed highly unlikely to contain relevant data on the prevalence of CV conditions in persons with hemophilia: studies not directly related to hemophilia, studies focusing on cancer (eg, those comparing prevalence of cancer types in hemophilia), preclinical/animal studies, studies investigating persons with acquired hemophilia, studies investigating specific hemophilia treatments (eg, safety, efficacy, and pharmacokinetic and pharmacodynamic studies), cost-effectiveness studies, studies examining patients with inhibitors only (studies including both inhibitor and noninhibitor patients were considered), and studies with a focus on surgical outcomes. Studies investigating specific hemophilia treatments were excluded as they had narrow inclusion criteria and reported pharmacokinetic/ pharmacodynamic/efficacy/safety data instead of comparing the event prevalence in large populations of persons with hemophilia with that in the general population. Studies on surgical outcomes were excluded due to confounding issues, such as the wide range of procedures and patients included and the use of different replacement and thromboprophylaxis regimens. Abstracts not excluded based on these criteria were assessed based on the likelihood that they could potentially contain data on the prevalence of the selected events in persons with hemophilia. Those abstracts with a clear indication that no relevant data were to be found were assigned to a "not relevant" category. In cases where there was an indication that relevant data could potentially be included (either specifically mentioned or alluded to in the text), full-text articles were reviewed for data extraction.

| Data extraction
Full-text publications of selected abstracts were screened to identify data on the prevalence of the selected adverse events in persons with hemophilia. For each full-text publication, review of the results section and any accompanying figures or tables was performed. For those containing relevant data, this information was then extracted and documented in 3 separate spreadsheets (bleeding, thrombosis, or mortality).
Within each spreadsheet, data were sorted into 2 tabs: 1) Publications comparing prevalence/mortality between hemophilia and general populations and 2) publications reporting prevalence/mortality in hemophilia populations only. Some publications were included in more than one spreadsheet. Those studies not containing any relevant prevalence or mortality data were assigned again to the "not relevant" category.

| Search results
A total of 912 references were identified in the literature search, 774 of which were excluded from further screening based on the exclusion criteria ( Figure). Full-text publications for the remaining abstracts (n = 138) were then screened to extract relevant data on the prevalence of the selected events in persons with hemophilia. A further 55 full-text publications were excluded based on the screening criteria, leaving 83 relevant publications that were divided into 3 categories: bleeding, thrombosis, and mortality.

| Bleeding
There were 47 publications reporting data on the prevalence of the selected bleeding events (hemorrhagic stroke, ICH, and gastrointestinal bleeding) in persons with hemophilia. Of these, 10 publications included comparisons with the general population (Table 1) [

| Thrombosis
Prevalence of thrombotic events in hemophilia populations was re-  Similarly, Faghmous et al. [27] concluded that the prevalence of myocardial infarction and pulmonary embolism was comparable between both populations, and there was a slightly higher prevalence of deep vein thrombosis and ischemic stroke in persons with hemophilia. There was considerable variation in reported prevalence of the selected events between studies. In the 2 studies that split persons with hemophilia by severity [4,5], the prevalence of myocardial infarction and ischemic stroke was lower in patients with severe hemophilia than in persons with nonsevere hemophilia. As expected, studies assessing older patients showed higher prevalence of cardiovascular disease (CVD) than those that assessed patients of all ages [20,21,23,24,27]. When comparing the prevalence of arterial and venous thrombotic events in persons with hemophilia, most studies appear to only report arterial thrombotic events (eg, myocardial infarction and ischemic stroke). Four of 14 studies combined the prevalence of venous or arterial thrombosis into general "venous thrombosis" or "arterial thrombosis" categories [6,7,15,25]. When comparing these studies, there were differences in the prevalence of arterial vs venous events between populations.
For arterial thrombosis, prevalence ranges from 0% to 12.1% in hemophilia vs 0.5% to 5.9% in the general population. For venous thrombosis, prevalence ranges from 0.19% to 4.9% in hemophilia vs 0.09% to 5.9% in the general population

| Mortality
There were 29 studies containing mortality data, with 9 studies reporting standardized mortality ratios (SMRs) comparing mortality in the hemophilia population with that in the general population (

| Comparison of studies identified in the search
As described in the following section, the conclusions of this study align with several published reviews examining bleeding, thrombosis, and mortality in persons with hemophilia [35][36][37][38][39].

| Bleeding
ICH remains a severe bleeding event for persons with hemophilia that results in high mortality. ICH is considered to occur more frequently in 2 age groups: children ≤2 years and adults ≥60 years with associated risk factors such as hypertension [40]. In our searches, 2 publications reported ICH prevalence in newborns with hemophilia (3.1%-3.4%), which was higher compared to the general population (0.04%-0.11%) [14,16]. As newborns with hemophilia are at a high risk of ICH, current guidelines recommend planned cesarean section for delivery of the affected or potentially affected infants [41]. Two publications reported higher prevalence of nonfatal ICH in adults with hemophilia aged ≥30 years (1.6% vs 0.4%) [4] or ≥40 years (7.4% vs 0.4%) [5]. In both studies, approximately one-fourth of the adult patients were receiving prophylactic treatment (23% and 27%, respectively), highlighting the importance of life-long prophylaxis to prevent serious bleeding events as well as the need to optimize prophylaxis regimens to ensure best coverage.
A recent systematic review and meta-analysis provides an indepth analysis of the incidence and mortality of ICH in persons with hemophilia [39]. As described in the current study, incidence rates of ICH were found to be higher compared to the general population across all age groups, and the ICH risk was especially high in hemophilic neonates. Although neonates are especially vulnerable, it is important to monitor patients of all ages. Underlying risk factors such as hypertension, which has a higher prevalence in hemophilia vs the general population [42], should be addressed in accordance with current treatment guidelines [43] to mitigate the occurrence of ICH.

| Thrombosis
Interestingly, although several studies report that the prevalence of thrombotic events is higher in the general population, a small number of studies appear to show that the prevalence of some thrombotic events is comparable or in fact higher in those with hemophilia [6][7][8]27]. Three publications from Pocoski et al. [7] and Humphries et al. [6] were identified in the search, with 2 reporting a higher prevalence of thrombotic events in the general population and the third instead reporting a higher prevalence in those with hemophilia [25]. The first study, which was published in 2014, showed that the prevalence of cardiovascular comorbidities was found increased in persons with hemophilia A included in the MarketScan Commercial and Medicare Research Databases [7]. In 2016, these findings were confirmed in a second patient population from the US [6]. However, a third publication in 2018 that used medical records from the Henry Ford Health System was unable to replicate these findings [25]. As alluded to in the latter publication, differences in databases used between these 3 studies and the fact that at least 2 medical visits were required in the third study meant that controls could have had a higher medical burden and subsequently also have had higher rates of CVD [25]. The third study also had a comparably low sample size of persons with hemophilia (n = 74). Work from other groups has also reported higher or comparable levels of thrombotic events when comparing hemophilia and general populations [8,26,27].
Based on the data identified, the prevalence of arterial thrombotic events in persons with hemophilia (myocardial infarction and ischemic stroke) appears to be greater when compared with that of venous thrombotic events (deep vein thrombosis and pulmonary embolism).
However, this was not observed in the general population. Patients with hemophilia had a slightly reduced prevalence of venous thrombosis than the general population, which is in agreement with the data from a previous study that suggested that persons with hemophilia may potentially exhibit some protection from venous thrombosis, but not arterial thrombosis [44]. However, the paucity of data makes these comparisons difficult, and more research is needed to understand these differences. As discussed previously in the literature, cardiovascular risk factors such as atherosclerosis that contribute to arterial thrombosis are found equally in persons with hemophilia and general population [3]. It is therefore increasingly important that the underlying cardiovascular risk factors are investigated and addressed in persons with hemophilia as they are treated with replacement products and encounter new and improved hemostatic therapies.
Current treatment guidelines state that persons with hemophilia should receive the same screening and management of cardiovascular risk factors as the general population [43].
Previous reviews have also addressed this topic. Rizwan et al. [37] published a scoping review examining the prevalence of CVD in general and cardiovascular risk factors in persons with hemophilia 12based on 30 original articles and reviews published between 1983 and 2012. There was also conflicting evidence supporting the possibility that hemophilia exhibits a protective effect against CVD, which aligns with the overall conclusion of the current study. A recent narrative review also summarizes the prevalence of cardiovascular risk factors in persons with hemophilia and promotes increased awareness of CVD in an aging hemophilia population [38]. Two recent systematic reviews have examined mortality and causes of death in persons with hemophilia [35,36]. Hay et al. [36] concluded that persons with hemophilia A have a higher mortality rate than the general population, but incomplete reporting of data limits evidence on mortality. The systematic review and meta-analysis published by Alem et al. [35] provided an overall SMR of 1

| Implications for future clinical practice and research
Non-factor-replacement therapies are now either available (emicizumab) or are under clinical trials (concizumab, marstacimab, and fitusiran) as novel therapeutics for persons with hemophilia. One major advantage associated with these treatments relative to factor replacement therapy is the improvement in the quality of life due to subcutaneous administration, flexible dosing, and consistent levels with the avoidance of peaks and troughs. There has been a low occurrence of treatment-emergent reactions and a low incidence of neutralizing/nonneutralizing antibodies observed with the use of nonfactor-replacement therapies [46][47][48][49][50][51][52][53][54][55]. Nonreplacement therapies have been associated with some thrombotic effects in clinical trials, often with the concomitant use of other hemostatic agents or other specific thromboembolic risk factors [47,56,57]. Relevant amendments were made to trial protocols for emicizumab [47], concizumab [57], and fitusiran [58] to mitigate these risks by providing more detailed guidance on the management of breakthrough bleeds. Further research and data are needed to elucidate these clinical and nonclinical risks and benefits associated with these therapies.

| Strengths and limitations
Over 900 published abstracts were screened to identify relevant prevalence and mortality data of persons with hemophilia and the general population. This analysis therefore provides a comprehensive overview of the currently available prevalence data of several individual adverse events in persons with hemophilia vs nonhemophilic controls. However, direct comparison of data between the studies identified is limited due to the variation in the data sources used, the sizes and geographic distributions of the studied patient populations, and statistical methods used for analysis. For example, prevalence data were often derived from large patient databases from different sources, including medical records [4,5,8,20,21,[23][24][25][26] or commercial/ insurance databases [6,7,15,19]. Although some studies searched records using predefined criteria for hemophilia and comorbidities [4,5,8,[21][22][23][24], others used International Classification of Diseases codes to select individuals who had experienced adverse events [6,7,15,19,20,25,26]. These factors, therefore, greatly limited the ability to accurately compare studies to evaluate if persons with hemophilia are protected against thrombosis and whether the prevalence of cardiovascular events in persons with hemophilia is increasing over time.
SHAPIRO ET AL.

FUNDING
This study was supported by Novo Nordisk Healthcare AG.

AUTHOR CONTRIBUTIONS
All authors contributed to the design of the literature search, interpretation of the results, and manuscript drafting/critical revision of the manuscript. All authors read and approved the final version of the manuscript.