Normal Serum-Erythropoietin (S-epo) Level at Diagnosis of Polycythemia Vera (PV) Correlates with LowJAK2^V617F Mutant Allele Burden and Indicates Mild Phenotype.

Abstract 4978

A low serum-erythropoietin (S-epo) level is a minor criterion of the World Health Organization (WHO) recommendations for diagnosing polycythemia vera (PV) even though previous studies indicate that a normal level does not always rule out PV. We wished to determine whether a normal S-epo level correlates with a low JAK2^V617F mutant allele burden and a low phlebotomy (PHL) rate, since the relationship between S-epo level, JAK2^V617F allele burden, and PHL rate heretofore has not been reported.

At diagnosis, we grouped the S-epo levels of 26 PV patients (pts) as follows: low (<5 U/l), normal (5-10 U/l), and high (>10 U/l). Of the 26 pts, 4 (15.4%) had normal S-epo levels, and 22 had low S-epo levels. The diagnosis of PV in pts with normal S-epo levels was made by quantitative JAK2^V617F analysis and elevated Cr⁵¹ red cell mass (RCM). All pts had a bone marrow biopsy consistent with PV.

We determined the number of phlebotomies required to maintain normal hematocrit (hct) levels ('42% women, '45% men) as an assessment of disease severity during the period prior to myelosuppressive therapy. We then examined the number of phlebotomies per year (PHL/yr) in relation to S-epo level at diagnosis. Of the 26 pts, 14 were treated with PHL-only for a mean of 15 months (mos) prior to beginning other therapies including anagrelide, hydroxyurea, imatinib, and rIFNa-2b. The number of PHL/yr were grouped into 4 categories: 0 (none), 1 – 5 (low), 6 – 10 (moderate), and >10 (high). Of the 14 pts, 3 had a normal S-epo level at diagnosis, all of whom had a low PHL requirement, median 3/yr. The other 11 pts had low S-epo levels at diagnosis and a median PHL requirement of 6/yr (7 moderate and 4 low). Thus, while a low S-epo level at presentation was usually associated with moderate PHL requirement, a normal S-epo level at diagnosis was always associated with a subsequently low PHL requirement during the period of observation (mean 15 mos).

Of the 26 pts, 14 had quantitative JAK2 analyses done at diagnosis. The other 12 were diagnosed prior to 2002 and quantitative JAK2 determinations were not available. Of the 14 pts, 2 had a normal S-epo level and a low JAK2^V617F mutant allele burden (1-25%). These 2 pts also had a low PHL requirement. Thus, these pts who had normal S-epo levels, had JAK2^V617F allele burdens in the lowest quartile, and had low PHL rates. Of the remaining 12 pts, 5 had JAK2^V617F allele burdens in the 2^nd (25-50%) quartile, 5 in the 3^rd (50-75%) quartile, and 1 each in the 1^st and 4^th quartiles.

We conclude (1) as previously reported, about 15% of pts with PV present with normal S-epo levels at diagnosis, suggesting a limitation of this WHO criterion, and (2) those pts with a low JAK2^V617F allele burden and a normal S-epo level required fewer PHL/yr, all suggesting a more benign phenotype. This may provide a useful prognostic tool for patients with PV.