Key points are not available for this paper at this time.
Hemophilia A (HA) is an X-linked bleeding disorder caused by a mutation in the factor VIII (FVIII) gene. When a male is diagnosed with hemophilia his mother may or may not be a carrier. Mutation analysis is the gold standard to identify female carriers. Unfortunately, the causative mutation may be unknown or genetic testing may not be available. Therefore, an alternative tool to determine carrier status would be helpful. FVIII levels in carriers demonstrate considerable variability and a normal FVIII level (≥0.5 U/mL) does not rule out the possibility of a female being a carrier.1, 2 FVIII circulates bound to von Willebrand factor (VWF) and plasma FVIII levels are influenced by VWF levels. In carriers of HA, only the FVIII level is prone to be influenced, while the VWF level is unaffected. Therefore, we questioned whether the FVIII/VWF:antigen (Ag) ratio could be a useful test to identify carriers. In our center, we undertake mutation testing on all boys with HA and offer testing (genetic and phenotypic FVIII and VWF:Ag) to their available female relatives who may be carriers. The absence of bleeding symptoms does not exclude them from being tested. Here, we report the data on 328 potential/known hemophilia carriers. These include 242 mothers of 282 boys with HA, 65 female relatives of these boys (39 sisters, 10 grandmothers, 12 aunts, and 4 cousins), 17 obligate carriers (daughters of men with hemophilia) unrelated to any of our patients and 2 females with excessive bleeding but without a family history of hemophilia, along with their mothers; all 4 of these were confirmed to be carriers of HA by genetic testing. Of the 242 mothers; 133 (55%) were obligate carriers while 102 (45%) underwent genetic testing to determine carrier status. There were an additional 7 mothers in whom carrier status could not be proven/excluded: In 3 cases there was no family history of hemophilia, and, no mutation was identified despite genetic testing; 2 cases did not undergo genetic testing, and, in 2 cases the child been adopted and as such there was no biological mother to test. Excluding the 7 untested mothers, carrier status was determined in 235 mothers. Of these, 227 (96.6%) were determined to be carriers. As such, between 94% (227/242; assumes the 7 untested mothers are all non-carriers) and 97% (234/242; assumes the 7 untested mothers are all carriers) of mothers of children with HA in our large cohort are carriers of hemophilia. Even when there is no previous family history of hemophilia up until the first born child with hemophilia between 88% (93/106) and 92% (98/106) of their mothers are carriers, which indicates that it is rare for new spontaneous mutations to arise in males with HA. In addition, 29/65 female relatives (45%) and the 21 remaining tested females were found to be carriers. Thus, in total, we identified 277 carriers and 44 non-carriers (including 8 mothers of boys with hemophilia and 36 female relatives found not to be carriers). Consistent with previous reports,1-5 we found that the median FVIII level in carriers (0.68 U/mL; range 0.19-2.00 U/mL) was significantly lower than in non-carriers (1.28 U/mL; range 0.49-2.30 U/ml) (P < .001) (Table 1). Yet, 77% (158/206) of tested carriers had FVIII levels ≥0.5 U/mL. Both FVIII and VWF:Ag levels were available in 147 carriers and 25 non-carriers. The median FVIII/VWF:Ag ratio was significantly lower in carriers (0.71; range: 0.18-2.20) than in non-carriers (1.39; range: 0.58-2.48) (P < .001) (Table 1). In the majority of carriers (123/147; 84%), the FVIII/VWF:Ag ratio was <1. ROC analysis confirmed that determining the FVIII/VWF:Ag ratio is an acceptable, but not perfect tool, to predict carrier status (AUC = 0.9273) (Supporting Information Figure 1). A FVIII/VWF:Ag ratio ≤0.9 was rarely found in non-carriers (1/25; 4%) versus in 76.9% of carriers 113/147 (P < .0001). As such, if a cut-off of ≤0.9 was used to label a person as likely being a carrier, only 1 of 114 females (0.9%) would be falsely declared to be a carrier; the other 113 females are true carriers. However, 23% of the carriers would still remain undiagnosed. On the other hand, a FVIII/VWF:Ag ratio ≥1.2 was found in 9/147 carriers (6.1%) versus 19/25 (76%) non-carriers. Consequently, a FVIII/VWF:Ag ratio ≥1.2 would misdiagnose only 6.1% of carriers as being non-carriers. All other tested cut-off levels had worse performance in predicting the carrier versus non-carrier status. We evaluated the impact of age at time of testing, blood type, and type of mutation on FVIII/VWF:Ag ratios. The median age at testing for carriers was 36 years (range 3.6-61.1 years) which was not statistically different from the median age at testing for non-carriers (29 years; range 4.4-72.6 years) (P = .18). As both FVIII levels and VWF:Ag levels increase with age, we did not find any effect of age on FVIII/VWF:Ag ratios. It has been established that healthy blood type O individuals have lower FVIII and VWF:Ag levels compared to blood type non-O individuals.4-6 In our study, blood type O carriers had significantly lower VWF:Ag levels (median 0.79 U/mL; range 0.47-2.35 U/mL) and lower FVIII levels (median 0.61 U/mL; range 0.22-1.39 U/mL) than blood type non-O carriers median VWF:Ag level 1.13 U/mL (range 0.61-2.56 U/mL; P < .001) and median FVIII level 0.72 U/mL (range 0.19-1.65 U/mL; P = .02). However, the median FVIII/VWF:Ag ratio was not significantly different between blood type O (0.72; range 0.3-1.64) versus blood type non-O carriers (0.70; range 0.18-1.34) (P = .2) (Supporting Information Figures 2 and 3). Conflicting data are available on whether type of mutation influences FVIII levels in carriers of HA.7, 8 In our study, median FVIII levels in carriers of a null mutation (n = 52) were not statistically different from FVIII levels in carriers of missense mutations (n = 90): 0.64 U/mL (range 0.19-1.53 U/mL) versus 0.69 U/mL (range 0.21-1.31 U/mL) (P = .42). Unexpectedly, the median FVIII/VWF:Ag ratio was significantly lower in null mutation carriers versus missense mutation carriers (0.61; range 0.18-1.34 versus 0.77; range 0.25-2.20) (P = .03) due to higher VWF:Ag levels in the former (1.14 U/mL; range 0.53-2.35 U/mL versus 0.91; range 0.25-2.56 U/mL) (P = .04). At this time, we have no explanation for this finding (Supporting Information Figures 2 and 3). Our results represent data from an almost completely unselected group of carriers: most women (except 2) were identified not because they were referred because of increased bleeding but because they were mothers/relatives of patients with hemophilia. This is different from most studies on carriers. Our study showed that the FVIII/VWF:Ag ratio performed much better than FVIII level alone in identifying carriers but it is not 100% sensitive even in women with a high a-priori likelihood of being carriers. Genetic testing therefore remains the gold standard for carrier determination and should be pursued to correctly label women as carriers/non-carriers. It should be noted that the FVIII/VWF:Ag ratio would likely have very poor specificity in testing women at random for hemophilia carrier status and should not be used for that purpose. This work was conducted when Dr. Veerle Labarque was a clinical/research fellow in the pediatric thrombosis and hemostasis program in the Division of Hematology/Oncology at the Hospital for Sick Children, Toronto, Canada and a recipient of a 2010 Bayer Haemophilia Clinical Training Award. Nothing to report. V. Labarque was responsible for data collection, analysis and manuscript writing. V. Perinparajah contributed to data collection. V. Bouskill, AM. Stain, C. Wakefield and C. Manuel participated in data acquisition. V. Blanchette, P.D. James and D. Lillicrap critically revised the manuscript. M.D. Carcao supervised the study and critically revised the manuscript. The authors stated that they had no interests which might be perceived as posing a conflict or bias. Additional Supporting Information may be found in the online version of this article. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.
Labarque et al. (2017) studied this question.