Abstract

BACKGROUND: Although von Willebrand disease (VWD) is common, considerable uncertainty accompanies its diagnosis. Recently, Lipton described the phenomena of “misdiagnosis by milk box” where low levels von Willebrand factor (VWF) are observed with testing done at send-out laboratories when the specimen was shipped to the lab from a physician’s office or draw station, and the abnormal results could not be confirmed when repeated on site at a hemophilia treatment center (HTC) lab (

Lipton RA. Misdiagnosis by milk box.
Haemophilia
2003
;
9
:
235
). The consequences of misdiagnosis of VWD are significant including inappropriate therapy, missed alternative diagnoses, anxiety, and possible insurance discrimination. However, the frequency of this phenomenon remains undefined.

OBJECTIVE: To estimate the frequency of “milk box” misdiagnosis of VWD.

METHODS: We retrospectively reviewed the records of 105 consecutive patients referred to a single HTC for laboratory evaluation of suspected VWD. We compared the results of outside testing to repeat testing at the HTC. Specimens were drawn on site at the HTC and transported directly to the laboratory at room temperature. If not assayed immediately, blood was spun down twice and plasma frozen within 2 hours.

FINDINGS: Forty-four of these 105 patients had prior outside testing at send out labs. All 44 had repeat testing at the HTC, 23 of them once and 21 on more than one occasion. Twenty-seven patients did not have outside testing, but had testing at the HTC on more than one occasion. Thirty-four patients had testing on only one occasion at the HTC and no outside testing and were considered non-informative. When patients were referred with outside testing showing low levels of VWF antigen, repeat testing at the HTC was normal in 19/26 cases, low levels of ristocetin cofactor were normal in 15/19 cases and low levels of Factor VIII:C were normal in 12/12 cases. In contrast, when patients had testing on more than one occasion at the HTC, there was much higher consistency of results which was statistically significant using the McNemar test for matched pairs. When the first VWF antigen was low, the second was found to be normal on repeat testing in 4/28 patients, low levels of ristocetin cofactor were normal in 0/14 cases and low levels of factor VIII:C were normal in 1/2 cases. In patients having outside testing, an overall clinical diagnosis of VWD based on symptoms and initial laboratory findings could not be confirmed by repeat testing at the HTC in 20/44 patients (45%). In contrast, in patients tested twice at the HTC, repeat testing failed to confirm a clinical diagnosis of VWD in only 1/27 patients (4%). (p<0.01, Mantel-Haenszel chi-square).

CONCLUSION: Milk box misdiagnosis is common.

DISCUSSION: While our data do not address the mechanism of this phenomena, studies by Favoloro et al suggest the common practice of refrigeration of blood samples prior to centrifugation and freezing the plasma which results in significant degradation of factor VIII and von Willebrand factor in vitro, a phenomena known as “cold activation, ” as a plausible explanation (Favaloro EJ, Soltara S, McDonald J. Potential laboratory misdiagnosis of hemophilia and von Willebrand disorder owing to cold activation of blood samples for testing.

Am J Clin Pathol
2004
;
122
:
668
–692
). Given the significant incidence of “milk box” misdiagnosis of VWD, we suggest that physicians view an unconfirmed laboratory diagnosis of VWD critically. We recommend that, when possible, the laboratory testing for VWD be performed at a specialized coagulation laboratory with the specimen drawn on site and processed according to validated protocols. When this is not possible, the specimen should be handled by methods that avoid cold activation.

Disclosures: No relevant conflicts of interest to declare.

Author notes

Corresponding author