Discussion
Previous investigations of adult hemophilia carriers, including an investigation of the ATHNdataset, have shown that roughly one-fourth meet criteria for having hemophilia.1,4 Our investigation of pediatric carriers enrolled in the ATHNdataset showed nearly half of pediatric carriers meet current criteria for having hemophilia. There are several potential explanations for the high proportion of pediatric carriers in this study having hemophilia. One is that there could be ascertainment bias in the ATHNdataset. Pediatric carriers with hemophilia (and more likely to have abnormal bleeding) may be more likely to be seen at a HTC, thereby skewing the data. However, of those reporting a BLS, the majority reported a normal BLS for age, and the median BLS was 0. This suggests that recruitment of subjects with a higher bleeding risk may not be skewing the data to a significant extent. Also, similar ascertainment bias may exist in studies of adult carriers. Pre-analytical or analytical errors are another possible explanation. However, these specimens were handled by HTCs well versed in handling blood specimens for hemophilia patients. Also, many of the subjects in this study are seen at the same HTCs that also see adults. Factor VIII is an acute phase response protein and if anything, the stress of phlebotomy in a child might lead to falsely elevated factor VIII activity levels. Pre-analytical and analytical issues are unlikely to explain the high proportion of pediatric carriers with hemophilia enrolled in the ATHNdataset compared to adult carriers with hemophilia enrolled in the same database. It is known that factors VIII and IX levels rise with age, and the mean factor activity level in child and adolescent carriers was lower than that reported for adults.2,8,9 This could explain the difference in proportion of those reporting a factor activity below 40 IU/dL in adults compared to pediatric carriers.
Of those reporting a BLS, a minority (9.7%) of pediatric carries with factor activity levels >40 IU/dL in the ATHNdataset report BLS that were abnormal for age. Abnormal bleeding with normal factor activity levels is a consistent finding in multiple studies of adult carriers as well and seems to be more prevalent than in children.1-3 Again, it is possible that bias could explain this finding. A very small, single institution study showed a higher BLS for carriers that knew their carrier status prior to obtaining their BLS compared to carriers that did not.10 In addition, cultural difference in the reporting of BLS in carriers have been reported.1,11,12 It is possible that carriers with normal factor activity levels do not truly have abnormal bleeding, and the apparent abnormal BLS are due to methodologic/cultural issues used to determine the BLS. However, this should not be assumed to be the explanation. To date, studies demonstrating a strong correlation between factor VIII and IX activity levels and bleeding risk have been done exclusively in males. It is not unreasonable to propose that factors VIII and/or IX have gender related hemostatic functions that are not measured by aPTT based assays. Factor VIII, at least, is known to have a non-hemostatic function that is not measured by an aPTT based assay.13 The factor VIII:Von Willebrand complex is important for bone homeostasis, and female carriers have recently been shown to have higher rates of osteoporosis and fractures compared to an unaffected population.13,14 In factor XI deficiency, there is a no correlation between factor activity levels and clinical bleeding.15 It has been proposed that this is due to hemostatic effects of factor XI that are not measured by aPTT based assays.16,17 As shown in table 3, moderate to weak correlation between factor activity and clinical bleeding is seen in other coagulation factor deficiencies.15,18 Additional investigation into possible hemostatic functions of factor VIII and IX outside those measured by aPTT based assays are needed.
Another finding of our study was that most pediatric females with hemophilia that reported a BLS had a normal BLS for age. This is also seen in adult carriers with hemophilia.1 However, an overwhelming majority with hemophilia had factor activity levels between 16 IU/dL and 40 IU/dL, levels that are expected to have a lower risk of abnormal bleeding.19 Alternatively, this could be another example of the inadequacy of aPTT based factor assays in predicting bleeding symptoms in this population.
Several guidelines recommend against testing potential carriers for genetic diseases during childhood.20.21 A principal reason for this stems from ethical concerns regarding the loss of the child’s future autonomy. However, these guidelines have not met with universal acceptance.22 A more recent policy statement from British Medical Health allows for carrier testing so long as no harm comes to the child.22 The American Academy of Pediatrics and American College of Medical Genetics most recent guideline also recommends against carrier testing unless a child is at risk for childhood onset conditions.21 Our study suggests that hemophilia carriers are at significant risk for having hemophilia and/or abnormal bleeding. Because factor activity levels may not be diagnostic of carrier status and may not be predictive of bleeding risk, genetic testing of potential/obligate carriers seems indicated.
This study has several strengths, even compared to other investigations of adult carriers, including data submitted from multiple institutions, genotyping of a significant proportion of subjects, standardized measure of bleeding symptoms, and a sample size that is larger than most studies involving adult hemophilia carriers.2,3,23 Even with this large sample size, only a small number of pediatric carriers reported an abnormal BLS for age. Thus, we could not make valid analysis of other factors that might contribute to an abnormal BLS. Further improvement of the study would have been inclusion of additional details regarding clinical bleeding such as joint or menstrual bleeding. However, this type of data is not consistently submitted by participating HTCs and would lead to an underestimation of bleeding symptoms.
Our investigation suffers from a weakness common to many previous studies of hemophilia carriers, namely, they only investigate carriers seen at a HTC. It is unknown to what degree carriers seen at a HTC are representative of all hemophilia carriers, most of whom are not seen at a HTC. The majority of subjects in this study did not submit a BLS, and of those that did, nearly all were participants in the MLOFgenotype initiative. Despite this limitation, the sample size of pediatric subjects that did submit a BLS was larger than most studies of adult carriers that report BLS, thereby making this data of value.
We relied on the ISTH-BAT BLS to describe bleeding symptoms. Bleeding scores remain the best objective measure of clinical bleeding and have been validated for carriers.23 However, there are limitations, which may be particularly relevant for pediatric hemophilia carriers. For instance, a 2-year-old carrier with a single untreated joint bleed and no other abnormal bleeding would get an ISTH-BAT BLS score of 2. This would not flag as abnormal for age in this study, but most clinicians would consider this abnormal bleeding. Unlike Von Willebrand disease, investigations of carriers have noted a poor correlation between factor activity levels and bleeding symptoms.1, 24-26 It is unknown if this is due to the inadequacies of the BLS, factor activity levels, or both.
The benefits of a large sample size can be offset by errors associated with large databases, including enrollment, data collection and submission, and query errors. We attempted to limit this by restricting this investigation to continuous variables (factor activity and BLS) and categorical variables that are audited by ATHN. This study was limited to pediatric carriers residing in the United States who participated in the ATHNdataset and may not be reflective of pediatric carriers residing in other areas of the world.
Despite the limitations of this study, as one of the few investigations of hemophilia carriers focusing exclusively on the pediatric population, we believe it makes important contributions. Significant knowledge gaps remain regarding clinical bleeding in this population and additional investigation is needed. Also needed are investigations into the potential hemostatic functions not measured by aPTT based assays. This would have applicability beyond hemophilia carriers and include all patients with coagulation factor deficiencies.
Acknowledgments : The authors would like to acknowledge ATHN and all the participating hemophilia treatment centers including physicians, nurse practitioners, nurses, and data managers for collecting and submitting data.
Conflicts of Interest: The authors report no conflicts of interest for this project.