Factor VIII deficiency does not protect against atherosclerosis


Sara Biere-Rafi, Department of Vascular Medicine, F4-140, Academic Medical Center, Meibergdreef 9, 1105 AZ Amsterdam, the Netherlands.
Tel.: +31 20 5667050; fax: +31 20 5669343.
E-mail: s.rafi@amc.uva.nl


See also Makris M, van Veen JJ. Reduced cardiovascular mortality in hemophilia despite normal atherosclerotic load. This issue, pp 20–2; Zwiers M, Lefrandt JD, Mulder DJ, Smit AJ, Gans ROB, Vliegenthart R, Brands-Nijenhuis AVM, Kluin-Nelemans JC, Meijer K. Coronary artery calcification score and carotid intima–media thickness in patients with hemophilia. This issue, pp 23–9.

Summary. Background: Hemophilia A patients have a lower cardiovascular mortality rate than the general population. Whether this protection is caused by hypocoagulability or decreased atherogenesis is unclear. Objectives: To evaluate atherosclerosis and endothelial function in hemophilia A patients with and without obesity as well as in matched, unaffected controls. Methods: Fifty-one obese (body mass index [BMI] ≥ 30 kg m−2) and 47 non-obese (BMI ≤ 25 kg m−2) hemophilia A patients, and 42 obese and 50 matched non-obese male controls were included. Carotid and femoral intima–media thickness [IMT] and brachial flow-mediated dilatation (FMD) were measured as markers of atherogenesis and endothelial function. Results: The overall population age was 50 ± 13 years. Carotid IMT was increased in obese subjects (0.77 ± 0.22 mm) as compared with non-obese subjects (0.69 ± 0.16 mm) [mean difference 0.07 mm (95% confidence interval [CI] 0.02–0.13, = 0.008)]. No differences in mean carotid and femoral IMT between obese hemophilic patients and obese controls were found (mean difference of 0.02 mm [95% CI − 0.07–0.11, = 0.67], and mean difference of 0.06 mm [95% CI − 0.13–0.25, P = 0.55], respectively). Thirty-five per cent of the obese hemophilic patients and 29% of the obese controls had an atherosclerotic plaque (P = 0.49), irrespective of the severity of hemophilia. Brachial FMD was comparable between obese hemophilic patients and obese controls (4.84% ± 3.24% and 5.32% ± 2.37%, P = 0.45). Conclusion: Hemophilia A patients with obesity develop atherosclerosis to a similar extent as the general male population. Detection and treatment of cardiovascular risk factors in hemophilic patients is equally necessary.


Studies assessing the role of hemostasis in ischemic cardiovascular disease (CVD) indicate that hypercoagulability increases the risk of CVD, whereas a bleeding tendency seems to be associated with a lower risk. Previous studies have shown that high levels of factor VIII are associated with an increased risk of both venous and arterial thrombosis [1], and numerous other coagulation factors have also been related to an increased thrombotic risk [2–6]. The opposite also holds true, as patients with a hereditary deficiency of FVIII (hemophilia A) experience considerable protection against mortality caused by CVD [7–9]. A recent meta-analysis showed that, as compared with the general population, hemophilic patients have a 50% reduction in mortality caued by ischemic heart disease [10].

Two processes are required for an arterial thrombotic event to occur: atherogenesis, which gradually leads to the development of an atherosclerotic plaque, and atherothrombosis, the acute formation of an occluding thrombus. A role of hemostasis in the formation of an occluding thrombus is evident, but coagulation factors such as FVIII may also be involved in atherogenesis [11,12].

Whether hemophilia A protects against atherogenesis is unclear. Although FVIII deficiency seemed to protect against atherogenesis in some animal and human studies [13–16], other studies found no association [17–20]. However, the relatively low incidence of cardiovascular events in this group of patients requires large population-based studies. A major drawback of previous studies in hemophilic patients is the low prevalence of cardiovascular risk factors. Measurement of the carotid and femoral intima–media thickness (IMT) and endothelial function by means of brachial flow-mediated dilatation (FMD) allows early detection of atherosclerosis or functional vessel wall abnormalities, but comparing subjects with a low prevalence of risk factors may not be the best way to determine differences in subclinical atherosclerosis. Studies assessing the prevalence of atherosclerosis in hemophilic patients with proatherosclerotic risk factors are lacking. Obesity is such an established and major cardiovascular risk factor [21,22]. In addition, obesity is as prevalent in the hemophilic population as in the general population [23].

To test our hypothesis that a lifelong hypocoagulable state (i.e. FVIII deficiency) reduces the formation of atherosclerosis, we investigated the relationship between hemophilia A and the extent of atherosclerosis in a multicenter cross-sectional study. To assess subclinical atherosclerosis and (impaired) vascular function, we measured IMT and brachial FMD in hemophilia A patients with and without obesity and in controls who were matched for age, sex, and body mass index (BMI).


Participating centers

Hemophilia A patients were recruited from various hemophilia treatment centers across The Netherlands and Belgium. Enrollment in the study took place in three study centers: the Academic Medical Center in Amsterdam (AMC), The Netherlands; the University Medical Center Utrecht in Utrecht, The Netherlands; and the University Hospital in Leuven, Belgium. The study was approved by the local ethics committees, and inclusion took place after informed consent had been obtained.

Study population

Hemophilia A patients older than 18 years who had a BMI ≥ 30 kg m−2 were eligible for inclusion, irrespective of the severity of hemophilia. These patients were matched for age and severity of disease with non-obese hemophilia A patients (BMI ≤ 25 kg m−2). Hemophilia A patients with a history of symptomatic atherosclerotic disease (i.e. ischemic heart disease, stroke, or peripheral vascular disease) or HIV infection were excluded.

Controls were matched to hemophilic patients for BMI, age, and sex. These controls were recruited through placement of an advertisement in local newspapers or were healthy volunteers who had participated in other studies on CVD at the Department of Vascular Medicine of the AMC. To improve accrual, an obesity clinic in The Netherlands was also approached to identify suitable control subjects.

Study regimen and assessment of risk factors

All study subjects were invited to one of the centers after an overnight fast. Patients were instructed to refrain from consuming food and drinks, except water, in the 10 h prior to each measurement. Furthermore, to avoid any influence of a vena puncture on the brachial FMD measurements, patients were asked to refrain from prophylactic infusion of FVIII prior to the visit. Patients with a severe form of hemophilia, however, received FVIII prophylaxis after blood collection, to avoid any bleeding complications as a result of the prolonged blood pressure cuff inflation during the brachial FMD measurement. The study visit included measurement of carotid and femoral IMT and brachial FMD by means of ultrasonography; a vena puncture to assess fasting glucose and lipid levels; and a physical examination, including measurement of weight, length, and waist and hip circumference. BMI was estimated prior to enrollment by using data on weight and length from the hemophilia treatment centers, and was calculated after length and weight had been obtained during the physical examination. In all subjects, the ultrasound measurements preceded blood collection. Blood pressure was measured three times with the patient in a supine position during the IMT assessment, and the last measurement was registered. Additionally, a medical history was obtained. Levels of total cholesterol, LDL cholesterol and triglycerides were considered to be increased when they exceeded the 95th percentile of the reference values for the relevant age categories. HDL levels were considered to be low when they were below the fifth percentile of the reference values for the relevant age categories. Dyslipidemia was defined as the use of lipid-lowering drugs and/or any of the cholesterol levels exceeding the reference values for the relevant age categories. Diabetes mellitus was defined as the use of antidiabetes medication and/or fasting glucose levels higher than 7.1 mmol L−1. Hypertension was defined as the use of antihypertensive medication, a systolic blood pressure of > 140 mmHg, and/or a diastolic blood pressure of > 90 mmHg.


Assessment of carotid and femoral IMT  For assessment of carotid and femoral ultrasound IMT measurements, Acuson Sequoia instruments (Siemens Medical Solutions, Erlangen, Germany) equipped with linear-array ultrasound transducers (L7, 5–12 MHz) were used in all three study centers. Sonographers were trained and certified. Instrument application and scanning protocols were standardized as described previously [24]. In each subject, three arterial wall segments of the right and three segments of the left carotid artery were scanned; in each of the femoral arteries, one segment was scanned. In each center, a maximum of two sonographers performed the ultrasound procedures. High-resolution images of each of the segments were saved with the use of Digital Imaging and Communication in the diastole of the vessel. From all of the centers, scans were transmitted by secure file transfer protocol to the AMC Vascular Imaging core laboratory. Image analysis was performed by one certified ultrasound analyst (reader). For carotid and femoral IMT analyses, etrack (Vascular Imaging and Department of Physiology, AMC, Amsterdam, The Netherlands) was used. The reader was blinded to demographic and clinical information of subjects.

The primary ultrasound outcome was the per subject mean IMT of the six carotid arterial wall segments (mean carotid IMT). All other IMT outcomes, such as the per subject average maximum IMT of the six carotid segments (maximum carotid IMT), used to assess the presence of plaque, as well as the femoral IMT, were secondary endpoints. To assess intrasonographer reproducibility and for quality control (QC) purposes, repeat scans were assessed in 17 subjects. The observed mean difference in mean carotid IMT was 0.077 mm, which was well within the predefined intrasonographer QC limits of 0.2 mm.


As there is no clear consensus on the definition of plaques in the femoral artery, we only analyzed the presence of atherosclerotic plaque in the carotid artery. All segments of the carotid artery were assessed for the presence of atherosclerotic plaques. A plaque was predefined as a maximum IMT ≥ 1.3 mm of any given segment of the carotid artery [25].

Brachial FMD

Assessment of endothelial function by means of brachial FMD  Instrument application and scanning protocols were standardized as described previously [26,27]. Sonographers were trained and certified. Each study subject underwent measurement of endothelium-dependent vascular responses of the left brachial artery by B-mode ultrasound imaging. Acuson Aspen (Siemens, Mountain View, CA, USA) ultrasound systems equipped with L7, 5–10-MHz linear arrays were used. Prior to start of the brachial FMD scan, subjects rested for at least 10 min in a quiet and temperature-controlled (21–23 °C) examination room. Subsequently, the subjects’ left arm was placed in a custom-made transducer holder with arm support, and a blood pressure cuff was placed on the left forearm from the medial epicondyle downwards. A straight, non-branching segment of the brachial artery above the antecubital fossa was identified and scanned longitudinally. Following optimization of depth and gain settings, end-diastolic brachial artery diameters were recorded at a beat-to-beat interval for 1 min (baseline measurement). The cuff was then inflated to 250 mmHg on the forearm for 5 min, after which the cuff was deflated and the segment of the brachial artery was recorded continuously for another 3 min. Clips were stored on magnetic optical disk. Brachial artery diameter was analyzed qualitatively and quantitatively offline by certified image analyst at the AMC Vascular Imaging core laboratory. For image assessments, a validated automatic edge-detection system (Brachial Analyser; Medical Imaging Applications LLC, Coralville, IA, USA) was used. Brachial FMD was expressed as the percentage of the difference between the maximum post-cuff release brachial diameter and the average pre-cuff inflation (‘baseline’) diameter. For brachial FMD QC purposes, 19 repeat scans were available. To ensure reliable measurements, the intrasonographer QC limit was set at a mean difference in brachial FMD of less than 2%. The mean difference of the paired repeat brachial FMD scans was less than the predefined criteria of 2% (0.79%).

Statistical analysis

Data are presented as mean (± standard deviation) for continuous variables, median and ranges for variables with a skewed distribution, and frequencies or percentages for categorical variables. Differences in mean values were assessed with t-tests, after log transformation in cases of skewed data, and adjusted by Bonferroni correction for multiple testing. Differences in mean IMT and brachial FMD between the various subgroups were also compared by the use of t-tests. Categorical variables were compared by use of chi-square tests. Carotid and femoral IMT were stratified for age and risk factors for CVD to assess the influence of these variables on outcome. To assess the influence of severity of hemophilia on atherogenesis, IMT, brachial FMD and plaque data of patients with severe and moderate hemophilia were combined and compared with those of controls.


After the initial selection of potentially eligible hemophilic patients meeting the BMI criteria, two patients were excluded because of the presence of HIV, seven patients were excluded because of a history of CVD, and 27 patients did not want to participate, owing either to difficulty in traveling to the study center, to prior participation in recent studies, or to having no interest in participation in studies. A total of 205 study subjects were enrolled at the three study centers. Of these 205 subjects, 15 (including both hemophilic patients and controls) were excluded because they did not meet the BMI inclusion criteria (BMI ≥ 30 kg m−2 or BMI ≤ 25 kg m−2) during the study visit. The remaining study population (= 190) consisted of 51 (26.8%) obese hemophilic patients, 47 (24.7%) non-obese hemophilic patients, 42 (22.1%) obese controls, and 50 (26.3%) non-obese controls. As all hemophilic patients are male, the study population consisted entirely of males. Severity of hemophilia was equally distributed among the obese and normal-weight patients (Table 1). The use of prophylactic or on-demand treatment with FVIII concentrate and the prevalence of hepatitis C infection were also not different between the two groups (Table 1). None of the patients had an inhibitor against FVIII.

Table 1.   Characteristics of hemophilia patients
BMI ≥ 30 kg m−2
(= 51)
No. (%)
BMI ≤ 25 kg m−2
(= 47)
No. (%)
  1. BMI, body mass index.

 Severe (FVIII < 1%)17 (33)16 (34)
 Moderate (FVIII 1–5%)8 (16)8 (17)
 Mild (FVIII 6–40%)26 (51)23 (49)
 Prophylactic16 (31)12 (26)
 On demand35 (69)35 (74)
Hepatitis C infection
 Never28 (55)25 (53)
 Current10 (20)15 (32)
 In the past13 (25)7 (15)

Cardiovascular risk factors

Table 2 shows the presence of cardiovascular risk factors in hemophilic patients and controls. As expected, the subgroups were well matched for age and BMI. History of smoking in pack-years was similar in all subgroups (P = 1.00). Mean levels of systolic and diastolic blood pressure were significantly different between obese controls and non-obese controls (mean difference in systolic blood pressure of 10 mmHg, P = 0.003, and mean difference in diastolic blood pressure of 8 mmHg, P = 0.001). However, there was a higher prevalence of hypertension in hemophilic patients than in controls (43% and 25%, P = 0.01). Fasting glucose levels were higher in obese than in non-obese subjects in both hemophilic patients and in controls (P = 0.014 and P = 0.020, respectively).

Table 2.   Cardiovascular risk factors in hemophilia A patients and control subjects
BMI ≥ 30 kg m−2
(= 51)
BMI ≤ 25 kg m−2
(= 47)
BMI ≥ 30 kg m−2
(= 42)
BMI ≤ 25 kg m−2
(= 50)
  1. BMI, body mass index; CVD, cardiovascular disease; SD, standard deviation. P-values by t-test (after log transformation if necessary) adjusted by Bonferroni correction for multiple testing or chi-square tests. *P < 0.05 for the comparison of obese hemophiliacs with non-obese hemophiliacs. †P < 0.05 for the comparison of obese controls with non- obese controls. §P < 0.05 for the comparison of non-obese hemophiliacs with non-obese controls.

Age (years), mean ± SD50.0 ± 13.748.8 ± 13.950.7 ± 12.049.0 ± 13.6
Weight (kg)*†, mean ± SD107.8 ± 16.175.4 ± 7.5)109.1 ± 13.277.1 ± 7.7
BMI (kg m−2)*†, median (range)32.5 (30.1–50.2)23.5 (18.7–25.0)32.4 (30.0–50.2)23.2 (18.5–25.0)
Waist circumference (cm)*†, mean ± SD115.7 ± 12.289.9 ± 6.6113.7 ± 11.487.0 ± 6.9
Smoking (pack-years), median (range)10 (0–55)9 (0–45)6.5 (0–47)6.5 (0–50)
Systolic blood pressure (mmHg)†, mean ± SD135.6 ± 15.8130.9 ± 16.5138.2 ± 14.8126.4 ± 9.8
Diastolic blood pressure (mmHg)†, mean ± SD83.0 ± 9.979.1 ± 10.084.4 ± 9.076.2 ± 7.8
Hypertension*†§, no. (%)28 (55)14 (30)17 (41)6 (12)
Glucose (mmol L−1)*†, median (range)5.4 (3.9–11.4)5.2 (3.3–10.8)5.3 (4.4–15.5)5.1 (4.2–6.2)
Diabetes mellitus*†, no. (%)7 (14)1 (2)4 (10)0
Total cholesterol (mmol L−1)5.01 ± 1.015.09 ± 1.075.29 ± 1.055.20 ± 1.06
HDL cholesterol (mmol L−1)†1.22 ± 0.511.42 ± 0.421.21 ± 0.311.61 ± 0.64
LDL cholesterol (mmol L−1)3.17 ± 0.943.22 ± 0.963.40 ± 0.993.16 ± 0.96
Triglycerides (mmol L−1)†, median (range)1.20 (0.34–4.89)0.85 (0.27–2.50)1.35 (0.26–4.49)0.81 (0.28–2.19)
Dyslipidemia, no. (%)10 (20)5 (11)7 (17)4 (8)
Family history of premature CVD, no. (%)11 (22)8 (17)15 (36)9 (18)

Mean HDL levels were lower and triglyceride levels were higher in obese subjects than in non-obese subjects; this difference reached statistical significance in controls, but not in hemophilic patients. Levels of total cholesterol and LDL cholesterol did not differ significantly between the groups. Dyslipidemia was also equally prevalent among hemophilic patients and controls (15% and 12%, P = 0.50).

Carotid IMT

The mean carotid IMT in all hemophilic patients (IMT 0.74 ± 0.21 mm) was not different from that in all controls (IMT 0.72 ± 0.18 mm) [mean difference of 0.02 mm (95% confidence interval [CI] − 0.03–0.08 mm, P = 0.45)]. Interestingly, mean carotid IMT was increased in obese subjects (IMT 0.77 ± 0.22 mm) as compared with non-obese subjects (IMT 0.69 ± 0.16 mm) (mean difference 0.07 mm [95% CI 0.02–0.13, P = 0.008]). On comparison of obese hemophilic patients with obese controls, no difference in carotid IMT was apparent (IMT 0.78 ± 0.23 mm and 0.76 ± 0.22 mm, respectively [mean difference of 0.02 mm, 95% CI − 0.07–0.11, P = 0.67]) (Fig. 1A). The mean carotid IMT was not different in patients with severe and moderate hemophilia from that in controls (mean difference of − 0.03 mm [95% CI − 0.09–0.04, P = 0.44]).

Figure 1.

 (A) Mean intima–media thickness (IMT) of the carotid arteries. (B) Mean IMT of the femoral arteries. CI, confidence interval.

Femoral IMT

The mean femoral IMT in all hemophilic patients (IMT 0.87 ± 0.42 mm) was not different from that in all controls (IMT 0.85 ± 0.38) (mean difference of 0.02 mm [95% CI − 0.09–0.14 mm, P = 0.72]). The effect of obesity on femoral IMT is shown in Fig. 1B. The mean femoral IMT tended to be higher in obese subjects (IMT 0.90 ± 0.45 mm) than in non-obese subjects (IMT 0.82 ± 0.33 mm) (mean difference of 0.08 mm [95% CI − 0.03–0.20, P = 0.16]), although this difference was not statistically significant. The overall mean femoral IMT in obese hemophilic patients (IMT 0.92 ± 0.50 mm) was not different from that in obese controls (IMT 0.87 ± 0.40 mm) (mean difference of 0.06 mm [95% CI − 0.13–0.25, P = 0.55]). The mean femoral IMT in patients with severe and moderate hemophilia seemed to be lower than that in controls, but the difference did not reach statistical significance (mean difference in femoral IMT of 0.06 mm [95% CI − 0.06–0.18, P = 0.33]).

Adjusted IMT

Figure 2 shows the mean carotid and femoral IMT stratified for age. Overall, a similar trend was observed in hemophilic patients and in controls, with a gradual increase in IMT of both the carotid and femoral arteries with increasing age. Stratification for the presence of cardiovascular risk factors, such as hypertension, also showed no significant differences in mean carotid or femoral IMT.

Figure 2.

 (A) Mean carotid intima–media thickness (IMT) stratified for age. (B) Mean femoral IMT stratified for age.

Plaques in carotid artery

The prevalence of atherosclerotic plaques (carotid IMT ≥ 1.3 mm) was assessed in all six segments of the carotid artery. Of the hemophilic patients, 33% had a plaque in one or more of the six segments of the carotid artery, as compared with 25% of controls (P = 0.25). The prevalence of plaques was similar in obese subjects and in non-obese subjects (33% and 25%, respectively, P = 0.23) and in obese hemophilic patients and obese controls (35% and 29%, respectively, P = 0.49). The presence of plaques in patients with severe and moderate hemophilia was also similar to that in controls (27% and 25%, respectively, P = 0.79).

Brachial FMD

Table 3 shows the mean baseline brachial diameter and the mean peak postocclusion artery diameter, which were similar between hemophilic patients and controls (4.50 ± 0.67 mm and 4.47 ± 0.61 mm, P = 0.75, and 4.70 ± 0.65 mm and 4.68 ± 0.60 mm, P = 0.83, respectively). The mean brachial FMD in hemophilic patients was also similar to that in controls (4.75% ± 2.84 and 4.93% ± 2.39, P = 0.66). No effect of obesity on brachial FMD could be detected in the subgroups. The brachial FMD in obese subjects was similar to that in non-obese subjects (5.19% ± 2.79 and 4.51% ± 2.41, respectively, P = 0.09). The brachial FMD in obese hemophilic patients was also similar to that in obese controls (4.84% ± 3.24 and 5.32% ± 2.37, respectively, P = 0.45). On assessment of the influence of severity of hemophilia on brachial FMD, patients with severe and moderate patients had a similar brachial FMD as controls (5.15% ± 3.26 and 4.93% ± 2.39, respectively, P = 0.66).

Table 3.   Baseline and postocclusion hemodynamics
 Mean baseline brachial artery diameter (mm) ± SDMean peak brachial artery diameter (mm) ± SDMean brachial FMD (%) ± SD
  1. SD, standard deviation; FMD, flow-mediated dilatation.

All controls4.47 ± 0.604.68 ± 0.604.93 ± 2.40
All hemophiliacs4.50 ± 0.704.70 ± 0.604.75 ± 2.80
Obese hemophiliacs4.67 ± 0.704.88 ± 0.604.84 ± 3.20
Non-obese hemophiliacs4.35 ± 0.704.55 ± 0.704.68 ± 2.50
Obese controls4.45 ± 0.604.69 ± 0.605.32 ± 2.40
Non-obese controls4.48 ± 0.604.68 ± 0.604.55 ± 2.40
All obese4.53 ± 0.604.76 ± 0.605.19 ± 2.80
All non-obese controls4.43 ± 0.704.63 ± 0.604.51 ± 2.40


The present study indicates that obesity leads to increased formation of carotid atherosclerosis, but that this process is not affected by a lifelong hypocoagulable state, namely hemophilia A. The overall mean carotid and femoral IMT, the prevalence of atherosclerotic plaques, and endothelial dysfunction, as measured by brachial FMD, did not differ between obese hemophilic patients and obese controls. Moreover, our study shows that atherosclerotic plaques are also prevalent in obese hemophilic patients, predisposing them to future cardiovascular events.

Previously, a protective effect of hemophilia on mortality caused by ischemic heart disease was observed [7–9]. Although the standardized mortality ratio varied (0.20 and 0.62), an overall 50% reduction in the ischemic heart disease mortality rate was observed in hemophilia A patients as compared with the general population [10]. This beneficial effect of hemophilia on fatal artherothrombotic events could be the result of reduced thrombin formation. Thrombin is the key player in both fibrin formation and platelet activation [28]. Thrombin cleaves fibrinogen to form fibrin, but can also trigger platelet activation through protease-activated receptor (PAR)1 and PAR4 [29]. This may lead to the formation of thrombi, and ultimately to vascular occlusion. Importantly, thrombin may also influence the process of atherosclerosis. Tissue factor and PARs are expressed at high levels in human atheroma, and are induced in response to injury in animal models [30]. In vitro, PAR activation induces leukocyte chemotaxis, and smooth muscle cell proliferation and migration, which may lead to arterial remodeling and stenosis [31]. In addition, coagulation factors and PARs are also involved in inflammatory responses and repair after injury [32]. Although patients with hemophilia, who have decreased thrombin formation, may be relatively protected from these atherosclerotic processes, this was not the case in our study.

In previous studies, no clear association between hypocoagulability and IMT was shown [17–19]. In 59 hemophilia A and B patients, carotid IMT was not different from that in controls (mean carotid IMT 0.76 mm [95% CI 0.71–0.80 mm] and 0.77 mm [95% CI 0.75–0.80 mm], respectively) [17], and this was confirmed by Sartori et al. [18]. In patients with severe type III von Willebrand disease, similar results were obtained [19]. Also for femoral IMT, no differences between hemophilic patients and controls were shown [17]. A small protective effect of hemophilia was, however, observed in patients with moderate and severe types of hemophilia, whose mean femoral IMT was somewhat smaller than that in controls [17]. Brachial FMD, as a measure of endothelial dysfunction, seemed to be impaired in hemophilic patients as compared with controls (3.8% ± 5.2% and 20.3% ± 13.0%, respectively, P = 0.0001), but brachial FMD in healthy controls was remarkably high [18].

The previous studies had the same drawbacks. Patients were relatively young and had a low prevalence of cardiovascular risk factors. Therefore, a potential protective effect of hypocoagulability on atherosclerosis would have been difficult to detect. In our study, hemophilic patients all had a major risk factor for atherosclerosis, namely obesity. The strengths of this study include the careful selection of obese hemophilic patients and controls, as well as the use of validated surrogate markers for atherosclerosis [33,34]. In addition, the ultrasound measurements were of high quality, which was confirmed by the low variation found between the repeated measurements. Furthermore, the number of refusals to participate was very low in this population, as the majority of patients have a good relationship with the hemophilia nurses and physicians who generally recruited these patients. A potential limitation was that patients with severe hemophilia received regular prophylactic treatment, which changes their phenotype to moderate. Nevertheless, if a low FVIII level influences the formation of atherosclerosis, this should have become apparent in our study.

Our study has important clinical implications. We can conclude that hemophilia A patients with cardiovascular risk factors develop atherosclerosis to a similar extent as the general male population. This implies that detection and treatment of these risk factors in hemophilic patients is mandatory. Next, although the cardiovascular mortality rate is lower in hemophilic patients, the increasing life-expectancy will lead to more cases of CVD. The anticoagulant treatment of patients with hypocoagulability and consequently a higher bleeding risk is a major challenge [35].

In conclusion, we show that patients with hemophilia and obesity, which is a major risk factor for atherosclerosis, have the same degree of subclinical atherosclerosis as obese control subjects.


S. Biere-Rafi and P. W. Kamphuisen: designed the study, performed the statistical analysis, interpreted the data, and drafted the manuscript; E. de Groot, M. Peters, V. E. A. Gerdes, and H. R. Büller: participated in designing the study and critically revised the manuscript for important intellectual content; A. Tuinenburg, R. Huijgen, B. W. Haak, P. Verhamme, K. Peerlinck, F. L. J. Visseren, M. J. H. A. Kruip, B. A. P. Laros-van Gorkom, and R. E. G. Schutgens: made substantial contributions to the acquisition of the data and critical revision of the manuscript for important intellectual content.


We would like to express our gratitude to all study participants, the hemophilia nurses (M. C. Valk, M. Beijlevelt, K. van Leeuwen, J. VandeSande, G. Mulders van der Meer, F. J. Balkestein, and M. van der Linden), the sonographers (J. Gort, D. V. M. Klappe-Banse, E. Rijff, G. Pieters, C. A. M. Joosten, and I. Klaassen), and to those who provided technical and administrative support (I. van de Paverd, C. Goddart, W. Hanselaar, L. Nagel, R. Breet, S. Mitra, T. A. Postma, N. Sons, J. Wientjes, and W. Scholten).

Disclosure of Conflict of Interests

This study was supported by an unrestricted grant from Pfizer.