Is the recommended hematocrit target  in polycythemia vera evidence‐based?

Tiziano BARBUI, MD

Ospedali Riuniti, Bergamo, Italy

“We have generated a transgenic mouse line that reaches a hematocrit concentration of 0.85 due to constitutive overexpression of human erythropoietin in an oxygen‐independent manner. Unexpectedly, this excessive erythrocytosis did not lead to thrombembolic complicationsin all investigated organs at any age”.

Elevated hematocrit, risk of thrombosis, and polycythemia vera ( Prchal, Blood 2003,letter)

In 

conditions 

associated 

with 

elevated 

hematocrit  such as polycythemia of high altitude, erythropoietin receptor mutations,  Chuvash 

polycythemia, 

hemoglobin 

mutants 

with 

high 

oxygen 

affinity, 

and 

2,3‐BPG 

deficiency, 

the 

rates 

of 

thrombotic  complications 

are 

far 

below 

those 

seen 

in 

patients 

with PV. 

These 

observations 

suggest 

that 

the 

thrombotic 

complications  of 

PV 

may 

be 

induced, 

at 

least 

in 

part, 

by 

other 

unique 

features 

of 

PV, 

such 

as 

quantitative 

and 

qualitative  defects of platelets and neutrophils; these myeloid cells are, after all, the  progeny of the same mutated hematopoietic stem cell as the red cells.

Fatal and nonfatal cardiovascular events  n=226

• Myocardial infarction ……………………

0.9

• Stroke/TIA…………………………….......    3.4

• Peripheral arterial thrombosis….......  1.2

• Deep vein thrombosis and PE...........3.1

Thrombosis during follow‐up in 1638 PV  patients (Eclap)*

Marchioli et al, JCO 2005

. Risk of stroke in adults with cyanotic congenital 

heart disease. Perloff JK, Marelli AJ, Miner PD.

Circulation.

1993 Jun;87(6):1954‐9

• 112 cyanotic patients 19‐74 years old (mean, 36 +/‐

11.7  years) in the UCLA Adult Congenital Heart Disease Center  Registry were selected for study by virtue of continuous 

observation for 1‐12 years (total, 748 patient‐years).

• No patient with either compensated or decompensated  erythrocytosis, irrespective of hematocrit level, iron 

stores, or the presence, degree, or recurrence of cerebral  hyperviscosity symptoms, progressed to clinical 

evidence of a complete stroke (cerebral arterial  thrombosis with brain infarction).

Results from the Tromsø

study* in the general population  (survey in Tromsø, 

Norway) show  that  a hematocrit in the upper 20th percentile was found to be 

associated with a 1.5‐fold increased risk of venous thrombosis.

The link found between high hematocrit and risk of venous thrombosis is real, 

but there is uncertainty on whether the relation is causal

or explained by other 

diseases. No clear data from multivariate analysis.

Do high levels of hematocrit interact with other risk factors for venous 

thrombosis and, therefore, increase the risk of venous thrombosis evenmore? Could it indeed explain part of the link between arterial and venous 

thrombosis, as the authors suggest?* Brækkan et al, Haematologica 2012 

44

30

10

24

32 34

1

7

50

25

13

29

0

10

20

30

40

50

60

Death Thrombosis AML Malignancy

Perc

enta

ge

32-PPhlebotomyChlorambucil

PVSG – 01Treatment by phlebotomy alone resulted in a higher incidence of

thrombosis during the first 3 years

Evidence for recommendation to maintain Ht<45%

Hematocrit was not a risk factor for  thrombosis in the PVSG‐01 trial

In multivariate analysis, pre-treatment hematocrit was not associated with thrombosis.

In a case-control analysis, Ht measured at the closest observation prior to thrombotic event (up to 52%) - and compared with that in the matched control at the corresponding time on study - was not associated

with thrombosis

ECLAPECLAPCentral Coordination

Central registryRCT

518ASA Uncertain

benefit/risk ratioClear Contraindication

for ASAClear Indication

to ASA

Cohort for the Evaluation of Natural History (1,638 pts)

0,550

0,506 0,500 0,4980,484

0,4680,453 0,453 0,453 0,455 0,455

0,400 0,396 0,388 0,382 0,3900,407

0,500

0,30

0,35

0,40

0,45

0,50

0,55

0,60

0 12 24 36 48 60

Months

Pack

ed C

ell V

olum

e (L

/L)

90° 75° 50° 25° 10°

Hematocrit

1,431 1,240 1,023 678 379 82No. of patients

Di Nisio , Barbui, et al., Brit J.Haematol.2007

*Model adjusted for: age, gender, time from PV diagnosis to recruitment, thrombotic or hemorrhagic events prior to recruitment, smoking, history of diabetes, 

hypertension, claudicatio intermittens, erythromelalgia, splenomegaly, circulating immature cells, leukocyte count, total blood cholesterol, phlebotomy use, 

interferon use, hydroxyurea use, antiplatelets use, anticoagulants use, 32P use, busulfan use, chlorambucil use, and pipobroman use

Time‐dependent multivariate

analysis

on the  relative risk of major thrombosis among men and 

women with Polycythemia Vera (N = 1,638)* 

Time-dependent multivariate analysis on the relative risk of major thrombosis among men and women with Polycythemia Vera (N = 1,638)*

Di Nisio , Barbui et al, 2007

Hazard ratio (95% CI), P-valueHematocrit (%) 45 (N=556) 1 (Reference)

46-50 (N=530) 0.89 (0.6-1.3), 0.6> 50 (N=345) 1.04 (0.6-1.8), 0.9

Platelet number (x109/l) 300 (N=592) 1 (Reference)301-500 (N=622) 0.78 (0.5-1.2), 0.2

> 500 (N=407) 0.67 (0.4-1.1), 0.1

*Model adjusted for: age, gender, time from PV diagnosis to recruitment, thrombotic or hemorrhagic events prior to recruitment, smoking, history of diabetes, hypertension, claudicatio intermittens, erythromelalgia, splenomegaly, circulating immature cells, leukocyte count, total blood cholesterol, phlebotomy use, interferon use, hydroxyurea use, antiplatelets use, anticoagulants use, 32P use, busulfan use, chlorambucil use, and pipobroman use

N Engl J Med, 2004;350:114-124

EFFECT OF ASPIRIN ON THE RISK OF MAJOR ARTERIAL OR VENOUS EVENT,OR DEATH FROM CARDIOVASCULAR CAUSES IN VARIOUS SUBGROUPS

« unless the red cell mass is appropriately reduced, antiplatelet therapy or chemotherapy will be futile «

Polycythemia vera: myths, mechanisms, and  management  Jerry L. Spivak

(Blood 2002 ) 

A large‐scale trial testing the intensity of  CYTOreductive

therapy to prevent cardiovascular 

events In patients with Polycythemia

Vera (PV)  CYTO‐PV

(Clinical Trials Gov NCT 01645124

Study Chair:  Tiziano

BARBUI

Sponsor:  Consorzio

Mario Negri

Sud

Study funded by Agenzia

Italiana

del Farmaco

(AIFA), 

Project #FARM6YNXAN

Eligible patientsConfirmed diagnosis of PV

All inclusion criteriaNo exclusion criteria

Entry visit*Written informed consent

A patient can be randomized in the trial, provided she/he meets all recruitment criteria.Clinical visits at 3*, 6*, 12*, 18, 24*, 30, 36*, 42, 48*, 54, 6 0* months.* the following laboratory tests must be performed: hematocrit, hemoglobin, red and white cell count,total cholesterol, HDL cholesterol, LDL cholesterol, triglycerides, uricoemia, glucose levels, total CK,ALT, AST, creatinine, potassium, sodium, fibrinogen, spleen and liver ultrasonography (specific testsfor PV to added).

Figure 1 – Study design

Standard cytoreduction(HCT 40-45%)

Experimental cytoreduction(HCT 45-50%)

Hematocrit (%) during the study

46,9

44,744,2 44,4 44,4 44,3 44,3

45,0

47,447,0 47,2 47,5 47,7

47,2

48,3 48,3

42

43

44

45

46

47

48

49

50

51

52

0 6 12 18 24 30 36 42

75° Pct <45% Median <45% 25° Pct <45%

75° Pct 45‐50% Median 45‐50% 25° Pct 45‐50%

Months

364 303 201No of patients 356 260337 125 48