The Discordance between Offsite to Onsite Testing for Von Willebrand Disease in Post-Menarchal Females: A Multi-Center Study

Introduction

Because of an increased awareness by the medical community of von Willebrand Disease (VWD) in females with heavy menstrual bleeding (HMB), diagnostic laboratory testing is often sent by primary care clinicians and obstetricians/gynecologists. Furthermore, insurers often necessitate testing be performed at an outside laboratory, instead of a hospital or academic based on-site coagulation laboratory. Falsely low von Willebrand factor (VWF) levels can occur due to pre-analytical variables, such as delayed processing and mishandling of VWF specimens. This study is the first multi-institutional analysis evaluating VWF testing to determine the percentage of post-menarchal females with discordant testing results when comparing off-site to on-site coagulation laboratories.

Methods

This retrospective, multi-center study compared VWF antigen, VWF activity/ristocetin cofactor (RCo) and factor VIII (FVIII) assays performed at off-site compared to on-site laboratories at 17 U.S. institutions. Data were collected in females aged 12 to 50 years who had VWF testing performed by their referring physician at off-site labs and then repeated under the guidance of a hematologist at their on-site lab. Data regarding clinical bleeding symptoms, final diagnosis and coagulation testing were also collected. The hematologist determined the final diagnosis. An off-site lab was defined as a clinic, community hospital or private facility where the phlebotomy and lab processing was not performed at the same center. Low VWF was defined as a VWF:RCo or VWF:Ag of 30-50% while type 1 VWD was defined as VWF:Ag or VWF:RCo <30%.

Results

A total of 264 subjects were included in the analysis and were evaluated for VWD between July 2013 and October 2017. Median age at first consultation visit was 15 years (range 12-50). The majority, 217 (83%), had heavy menstrual bleeding followed by easy bruising (n=96; 37%) and epistaxis (n=80; 30%). Testing was conducted while on hormonal contraception in 99 (38%) subjects at the initial consultation.

Seventy-three subjects had normal or elevated VWD testing prior to hematology consultation. There were 251 (95%) subjects who had VWF antigen testing, and 96 (38%) had low VWF antigen (<50%) at the off-site lab, but only 63 (25%) had low VWF antigen by on-site testing. Among 223 subjects (84%) with VWF activity/RCo results, 122 (55%) had low VWF activity (<50%) from off-site labs, while only 74 (33%) were confirmed by on-site testing. Similarly, 229 (87% subjects had FVIII assay results, 72 (31%) were low off-site and less than half, 33 (14%), were confirmed low by on-site testing. Higher proportions of patients demonstrated low levels of VWF antigen, VWF activity/RCo, and/or FVIII at off-site laboratories compared to on-site (McNemar's test p-value < 0.0005, for all assays).

Less than 40% (n=100) of subjects were given a final diagnosis of VWD (type 1, 2 or 3), and 7% were diagnosed with low VWF. Approximately 40% of the patients had a normal evaluation for disorders of hemostasis.

Conclusions

This study represents the largest multicenter evaluation comparing off-site to on-site laboratory testing of VWF assays in post-menarchal women. From the evaluable subjects, one-third to half of the VWF assays normalized when drawn and processed at a hospital or academic center coagulation lab. While a proportion of these patients with initially low levels may have had a "stress" induced response normalizing their levels at the time of on-site retesting, it appears the main discordance between assay results is most likely secondary to delayed processing of the specimen. This can lead to decay of factor activity if the plasma is not promptly separated and frozen prior to transport to the reference laboratory. Furthermore, transport could also affect sample integrity. These pre-analytical laboratory variables can lead to false positive results for VWD, leading to potential inappropriate interventions or inadequate therapy, in addition to added health care costs. These results highlight the need for VWD testing to be both drawn and processed on-site at a coagulation laboratory under the guidance of a hematologist. Since this may not always be practical in large geographic areas, further work in reducing the delay in processing is in order.

Acknowledgements: This is a study was conducted by members of the Foundation for Women and Girls with Blood Disorders, supported by a grant from CSL Behring.