Long-term outcomes of polycythemia vera patients

The Hematology Journal (2003) 4, 198–207 All rights reserved 1466-4680/03 $25.00

& 2003 The European Hematology Association

www.nature.com/thj

Long-term outcomes of polycythemia vera patients treated with pipobroman as initial therapy Jean-Jacques Kiladjian*,1,6, Claude Gardin1, Michel Renoux2, Franck Bruno3 and Jean-François Bernard1,4,5 1 Service d’Hematologie Clinique, Hopital Beaujon, Clichy, France; 2Service d’Hematologie, Hopital de la cote Basque, Bayonne, France; 3STAT’aids, Paris, France; 4SEQASS, Assistance-Publique Hopitaux de Paris, Paris, France; 5Faculte de Medecine Xavier Bichat, Paris, France; 6Service d’Hematologie Clinique, Hopital Avicenne, Bobigny, France

From 1968 to 1993, 179 newly diagnosed patients with polycythemia vera (PV) were enrolled in a prospective study using pipobroman as first chemotherapy. Among them, 140 fulfilled the Polycythemia Vera Study Group criteria for PV, and 39 patients (22%) can be considered as idiopathic erythrocytosis (IE). Vascular events occurred in 10% of IE and 20% of PV patients and solid tumors in 7.7% of IE and 12.8% of PV patients. There were no differences between PV and IE patients with regard to progression to myelofibrosis (MF), leukemic events and overall survival. Overall, 98.3% of patients initially responded to pipobroman, with very mild toxicity. A total of 164 PV patients who received more than 1 year of pipobroman were analyzed for long-term evolution. The actuarial risk of thrombosis was 15.6 and 23.8% at 10 and 18 years, respectively. In all, 21 patients developed a solid tumor during follow-up, added and/or switched drugs being a risk factor. Actuarial risk of MF was as low as 4.9 and 9.4% at 10 and 15 years, respectively. Actuarial risk of leukemia was 14.4 and 18.7% at 10 and 15 years, respectively. Hyperleukocytosis at diagnosis was the only variable significantly associated with higher risk of leukemia. The median survival was 15.5 years, with two initial adverse prognostic factors: age above 60 years and hyperleukocytosis. Despite an increasing risk of leukemia with time, survival was not lower when compared to the French matched population. Only age and hyperleukocytosis at diagnosis were found to have a prognostic value in PV. The Hematology Journal (2003) 4, 198–207. doi:10.1038/sj.thj.6200250 Keywords:

polycythemia; myeloproliferative disorder; leukemia; myelofibrosis; pipobroman

Introduction Polycythemia vera (PV) is an acquired myeloproliferative disorder characterized by the expansion of the red cell mass unrelated to any pulmonary, cardiac, renal or neoplastic disease. Treatments including phlebotomy1–3 and/or cytoreduction with either radiophosphorus,2–5 alkylating1–7 or nonalkylating agents such as hydroxyurea (HU)2,8–11 have been largely utilized in the last decades. However so far, a standardized therapeutic strategy for PV patients has not been defined,12,13 in part because of the suspected long-term leukemogenic risk associated with exposure to currently available drugs. The efficacy of pipobroman, a piperazine derivative, was first established in 1962 and 1964 in the USA14 and confirmed in 1968 by European investigators.15,16 Only three groups recently reported long-term results of PV patients treated with pipobroman.11,17,18

*Correspondence: J-J Kiladjian, Service d’Hematologie Clinique, Hopital Beaujon, 100 Boulevard du General Leclerc, 92110 Clichy, France; Tel: þ 33 140875110; Fax: þ 33 140875487; E-mail: [email protected] Received 23 October 2002; accepted 25 February 2003

Since 32P and alkylating agents were considered to be the main cause of leukemic transformation in PV,19,20 and because long-term phlebotomies were rarely tolerated and were reported to favor thrombotic events and myelofibrosis (MF),21 we initiated in January 1968 a prospective nonrandomized study of pipobroman as the primary treatment in newly diagnosed PV patients. The aim of this study was to evaluate the benefit of pipobroman in terms of survival, with the expectation that there would be a lower incidence of major complications when compared to data available on 32P-treated patients. Periodic evaluations of this cohort allowed us to include new patients up until December 1993, as long as neither better treatment nor increased long-term risk of pipobroman was reported.22,23 We report the long-term outcomes of 179 PV patients enrolled in this study, expanding and updating our previous findings,23,24 with particular emphasis on risk factors for major vascular thrombosis, progression to MF, myelodysplastic syndromes (MDS) and acute leukemia (AL) or solid tumors.

Long-term outcomes of polycythemia vera J-J Kiladjian et al

199

Materials and methods

Adjuvant treatment

From January 1968 to December 1993, two hematological institutions included 179 consecutive previously untreated patients with PV in this study: 132 from the Clichy center and 47 from the Bayonne center. They all received pipobroman as first chemotherapy.

Phlebotomy with volumetric compensation was performed as an emergency therapy when needed. Busulfan (BU) at low dose could be temporarily added in case of initial or persistent elevated platelet count (above 600 109/l). The long term use of aspirin was left open to the decision of the physician.

Diagnostic criteria (1) Primary polycythemia: Among the 179 patients enrolled, 140 fulfilled the Polycythemia Vera Study Group (PVSG) criteria for PV.25,26 The remaining 39 patients (22%) did not fulfil the PVSG criteria for PV, but can be considered as idiopathic erythrocytosis (IE) as they presented an elevated red cell mass with normal platelet and leukocyte counts, no splenomegaly and no evidence of secondary erythrocytosis.27 (2) MF with myeloid metaplasia: Diagnostic criteria for MF were cytopenia and/or enlarged spleen and/or presence in the blood smear of tear drop poikilocytes, nucleated erythrocytes and immature granulocytes and, in all cases, a bone marrow biopsy showing at least an abnormal reticulin fibrosis. (3) Leukemic events: The diagnosis of the so-called ‘leukemic events’ was based on the French–American– British guidelines.28 It included MDS and AL. Among the MDS, we observed refractory anemias with ring sideroblasts (RARS), refractory anemias with excess of blasts (RAEB), RAEB in transformation (RAEBt) and chronic myelomonocytic leukemias (CMML). For AL, we observed both acute myeloid leukemias (AML) and acute lymphoblastic leukemias (ALL).

Study design The induction dosage of pipobroman was 75 mg/day for 2–4 weeks and then 50 mg/day until the objective response was achieved. Complete remission (CR) was defined by no palpable spleen, Ht p48% in males or p45% in females, white blood cells (WBC) p10 109/l, platelets p400 109/l; partial remission (PR) was defined by persistent palpable spleen with normal counts of the three myeloid lineages. After the initial response, maintenance therapy was necessary, the useful dosage of pipobroman being determined for each patient according to blood cell counts. All patients were seen at one of our hematological units at least every 6 months. They saw their family doctor every 3 months. Treatment changes were left open and were decided by the consultant hematologist in case of toxicity or insufficient efficacy. The primary objective of this study was to determine the response rate to pipobroman; secondary objectives were evaluation of toxicity, occurrence of vascular events, solid tumors, evolution to MF, MDS or leukemia, and survival.

Statistical analysis The last clinical and biological information was collected at the reference date of 30 November 2000. The end points for each case were death or loss of follow-up at the time. The analysis was performed in March 2001 on SAS system (6.12 version) for Windows. Data are presented as mean, standard deviation, median, minimum and maximum for continuous variables and as absolute and relative frequencies for categorical variables. Kaplan–Meier estimates were used to calculate and plot cumulative probabilities of survival and event-free survival. The following end points were considered for the analysis: time to (1) vascular event, (2) solid tumor, (3) MF, (4) leukemic events and (5) death, with age and sex stratification and Ht, WBC and platelet count stratification. The standardized mortality ratio (SMR) was calculated on the basis of the observed mortality compared with the probability of death in an age- and sex-matched French population.29 ANOVA or w2 tests were performed when needed and a Wilcoxon test was used for survival analysis. For long-term outcomes, we decided to analyze only the 164 patients (130 PV and 34 IE) who received pipobroman for more than 1 year, in agreement with previously published Pi studies.11,17,18

Results Patients The initial data and clinical course of the 179 PV patients enrolled in this study are shown in Tables 1 and 2. The male/female ratio was 1.7 and the median age was 60 years (range: 24–90). Splenomegaly was present in 50% of patients; WBC count was above 12 109/l in 37% and platelet count above 400 109/l in 50% of cases. PV and IE patients treated with pipobroman in this study had similar evolution (Tables 1 and 2) leading us to suggest, as others,30,31 that IE is an early stage of PV. This was discussed elsewhere32 but emphasizes the need, as in chronic myelogenous leukemia, of a specific and sensitive positive marker for PV diagnosis.33 Pipobroman treatment was rapidly discontinued in 15 patients (8.4%) because of three early deaths in elderly patients, four lost to follow-up, and eight unsatisfactory responses or poor tolerance. The Hematology Journal


200 Table 1 Patient characteristics at inclusion and response to pipobroman All patients (n=179) IE (n=39)

PV (n=140)

At least 1 year of treatment (n=164) All

IE (n=34)

PV (n=130)

All

Age at diagnosis (years) Median Range

57.0 35.8–90.2

61.8 24.2–86.9

Sex M/F Ratio

30/9 3.33

83/57 1.46

113/66 1.71

Enlarged spleen

0

89 (63.3%)

89 (49.7%)

0

81 (62.3%)

81 (49.4%)

58.2 6.2

57.6 5.9

57.7 6.0

58.0 6.2

57.7 6.0

57.8 6.0

28 (82.4%) 6 (17.6%) 0

45 (33.1%) 28 (20.6%) 63 (46.3%)

73 (42.9%) 34 (20.0%) 63 (37.1%)

24 (82.8%) 5 (17.2%) 0

43 (34.1%) 27 (21.4%) 56 (44.4%)

67 (43.2%) 32 (20.6%) 56 (36.1%)

51 (37.0%) 47 (34%) 40 (29.0%)

85 (49.4%) 47 (27.3%) 40 (23.3%)

48 (37.5%) 45 (35.2%) 35 (27.3%)

77 (49.0%) 45 (28.7%) 35 (22.3%)

128 (91.4%) 10 (7.1%) 2 (1.4%)

166 (92.7%) 10 (5.6%) 3 (1.7%)

123 (94.6%) 7 (5.4%) 0

156 (95.1%) 7 (4.3%) 1 (0.6%)

Hematocrit (%) Mean Standard deviation WBC (109/l) o10 [10, 12[ X12 Platelets (109/l) o400 [400, 600[ X600 Response Complete Partial Failure

34 (100%) 0 0 38 (97.4%) 0 1 (2.6%)

At the last time point, among the 164 patients who received pipobroman for more than 1 year, 19 were lost to follow-up. In all, 89 patients (54%) had died at a median age of 74.4 years (range: 45–93) (among them 74 PV patients at 74.6 years (range: 48–93) and 15 IE patients at 72.9 years (range: 56–91)). In all, 56 patients were still alive (42 PV patients and 14 IE patients). Median follow-up for these 164 patients was 11.4 years (range: 1–28).

Response to treatment CR was achieved in 166 patients (92.7%), PR in 10 (5.6%), and only three patients failed (1.7%). Until measurable response was reached, pipobroman was used at the mean dosage of 0.970.3 mg/kg/day for a median time of 60 days (range: 21–240). Patients in CR/PR received continuous maintenance pipobroman treatment at the mean dosage of 0.3770.22 mg/kg/day. Pipobroman was discontinued (without further treatment for PV) in nine patients. In three cases, it was discontinued because of progressive pancytopenia despite pipobroman dosage reduction after 6.6, 14 and 15 years. None of these three patients relapsed after 4.2, 9.5 and 5.5 years respectively. Six patients stopped pipobroman themselves and relapsed within a year. In each case, a second CR was obtained after pipobroman reintroduction. Among the 164 patients evaluated for long-term evolution, 32 received more than one cytoreductive The Hematology Journal

60.6 24.2–90.2

56.8 35.5–80.1

61.3 24.2–86.9

27/7 3.86

77/53 1.45

29 (100%) 0 0 33 (97.1%) 0 1 (2.9%)

60.0 24.2–86.9 104/60 1.73

treatment during evolution of their disease. Of these 32 patients, 18 received BU for initial or progressive elevated platelet count at a median cumulative dose of 222 mg (range: 20–1536) and during a median time of 12 months (range: 1–111). Treatment was changed for HU in 14 patients (12 PV, two IE) after a median time of 6.4 years (range: 1.1–16): nine patients with insufficient platelet count control after a median time of 7 years (range: 1.4–16); two with insufficient hematocrit control despite escalation of pipobroman dosage after 5.2 and 7.4 years; one with persistent diarrhea after 1.4 years; and in two cases, treatment change was made at the physician’s discretion after 2.5 and 10 years. These 14 patients received HU during a median time of 6 years (range: 1.8–12.7) and at median cumulative dose of 2184 g (range: 327.5–4805). Nine patients (64%) responded to HU and were in durable CR, but five failed to respond satisfactorily (two received interferon a without success, two other progressed to MDS/AL after 7 and 8 years of HU and one developed MF after 12 years of HU).

Clinical and hematological tolerance In total, 13 patients (7.3%) had transient digestive tract side effects (gastric pain, diarrhea) during the induction phase, leading to pipobroman discontinuation in three cases within the first year. Skin intolerance was noted in three patients leading to a treatment change in one. Mild macrocytosis (MCV: 100–108 fl) was noted in 26


201 Table 2 Complications and survival All patients (n=179)

IE (n=34)

PV (n=130)

32 (17.8%)

4 (11.8%)

24 (18.5%)

10 14 3 1

12 14 5 1

2 0 2 0

8 12 3 1

10 12 5 1

18 (12.8%)

21 (11.7%)

18 (13.8%)

21 (12.8%)

9 3 4 2

9 4 4 4

9 3 4 2

9 4 4 4

11 (7.8%)

14 (7.8%)

11 (8.4%)

14 (8.5%)

2 0 1 0

4 0 6 1

6 0 7 1

2 0 1 0

4 0 6 1

6 0 7 1

6 (15.4%)

26 (18.6%)

32 (17.8%)

6 (17.6%)

2 2 1 1

6 11 4 5

8 13 5 6

2 2 1 1

15.5 [13.7, 18.1]

16.0 [13.8, 18.2]

59 (33.0%) 27 (15.1%) 93 (52.0%)

14 (41.2%) 5 (14.7%) 15 (44.1%)

Vascular events (N (%)) Years from inclusion o7 [7; 12[ [12; 18[ X18

4 (10.2%)

Solid tumors (N (%)) Years from inclusion o7 [7; 12[ [12; 18[ X18

3 (7.7%)

Myelofibrosis (N (%)) Years from inclusion o7 [7; 12[ [12; 18[ X18

3 (7.7%)

Leukemic events (N (%)) Years from Inclusion o7 [7; 12[ [12; 18[ X18 Overall survival (years) Median 95% CI Patient’s status Alive Lost to follow-up Dead

At least 1 year of treatment (n=164) All

IE (n=39)

2 0 2 0

0 1 0 2

16.0 [13.8, 18.2] 14 (35.9%) 9 (23.1%) 16 (41.0%)

PV (n=140) 28 (20%)

19.2 [12.0, NC] 45 (32.1%) 18 (12.9%) 77 (55.0%)

patients (16%). Thrombocytopenia below 80 109/l was observed in 10 patients (6%). Among them, nine were older than 60 years (P ¼ 0.02). In each case, platelet count returned to normal value after pipobroman dosage reduction. Two patients had severe pancytopenia and bone marrow aplasia, without evidence for an overdosage: one after 13.8 years of pipobroman and the second after only 3 weeks of treatment. Both patients fully recovered without major complication in 2 and 3 months, respectively, after pipobroman discontinuation. Both patients had polycythemic relapses and reached again CR with lower doses of pipobroman.

Vascular thrombosis In total, 28 patients (17%) had a major vascular event, including 4/34 IE (11.8%) and 24/130 PV (18.5%). Four patients had two thrombotic events. Five patients had deep vein thrombosis of the lower limb, including two with pulmonary embolism. Of the 28 patients who presented vascular events, 23 had arterial thrombosis and 19 died: 11 from stroke at a median age of 82.8 years (range: 71.4–88.7); four from myocardial infarc-

3 (8.8%) 0 1 0 2 3 (8.8%)

26 (20%) 6 11 4 5 19.2 [12.0, NC] 42 (32.3%) 14 (10.8%) 74 (56.9%)

All 28 (17%)

32 (19.5%) 8 13 5 6 15.5 [12.9, 18.1] 56 (34.1%) 19 (11.6%) 89 (54.3%)

tion; three from proximal femoral artery thrombosis; and one from mesenteric infarction. Median age at occurrence of the vascular event was 74.5 years (range: 39–88.7) and the male/female ratio was 1.5. Median time from diagnosis of PV was 11.5 years (range: 1–20). The actuarial risk of thrombosis was 4.3, 15.6, 21.8 and 23.8% at 5, 10, 15 and 18 years, respectively (Figure 1). Actuarial risk of thrombosis was significantly higher for patients older than 60 years at diagnosis (20/83 versus 8/81, P ¼ 0.003). It was independent of any other initial parameter, including WBC count above 10 109/l (P ¼ 0.75). Interestingly, the occurrence of major thrombosis in these patients was independent of platelet count: elevated platelet count at diagnosis was not correlated to a higher risk of vascular event (P ¼ 0.42), and 27 patients out of 28 had a normal platelet count when thrombosis occurred.

Solid tumors In total, 21 patients (12.8%) developed solid tumors: 3/34 IE (8.8%) and 18/130 PV (13.8%). Solid tumor occurred after a median time of 8 years (range: 0.5–20) and at a median age of 72.5 years (range: 50–86). The The Hematology Journal


202

Figure 1 Vascular events and age. Kaplan–Meier curves for vascular events in the 164 patients who received pipobroman for more than 1 year. Patients older than 60 years are compared to patients younger than 60 years at diagnosis (P ¼ 0.003 by the Wilcoxon test).

Table 3 Numbers of solid tumors by site (by age at inclusion) o60 years

X60 years

All patients

Breast Colon Prostate Brain Bladder Lung Stomach Pancreas Kidney

2 2 1 2

3 2 2

5 4 3 2 2 2 1 1 1

Total

9

Tumor site

2 2 1 1 1 12

after a mean time of 0.97 year after occurrence of MF. Nine patients died early without evolution to leukemia (six from pancytopenia-related complications) after a median time of 1.2 years (range: 0.3–2.5). Only two patients were still alive after 4.3 and 18 years from MF diagnosis. Actuarial risk of MF (Figure 3) was 2.6, 4.9, 9.4 and 16.2% at 5, 10, 15 and 18 years, respectively. It was independent of any tested parameter including age 460 years (P ¼ 0.61), added and/or switched drugs (P ¼ 0.28) or initially enlarged spleen (P ¼ 0.09).

21

Leukemic events sites of solid tumors are shown in Table 3. Numbers are too small to allow comparison with the relative frequency in France or other countries,29,34 but no skin cancer was observed. In all, 13 patients died from their solid tumor at a median age of 75 years (range: 66–88). The actuarial solid tumor risk was 3.75, 8.4, 14.3 and 22.5% at 5, 10, 15 and 18 years, respectively. Age X60 years at diagnosis (12/83 versus 9/81) and added and/or switched drugs (9/32 versus 12/132) (Figure 2) were the only observed risk factors for solid tumor (P ¼ 0.009 and 0.016, respectively).

Myelofibrosis A total of 14 patients (8.5%) developed MF during follow-up: 3/34 IE (8.8%) and 11/130 PV patients (8.4%). MF occurred after a median time of 12.5 years (range: 2.6–19.4) and at a median age of 70 years (range: 53–83). Outcome of patients after evolution toward MF was poor. Three patients died from transformation to AML The Hematology Journal

In all, 32 patients (19.5%) developed AL/MDS (six in the IE group (17.6%) and 26 in the PV group (20%)): 26 AML (including six preceded by MDS and three by MF), two ALL and four MDS. The six MDS preceding AML included five RAEB-t and one CMML. Those 32 leukemic events occurred after a median time of 9.6 years (range: 2.3–27.8) and at a median age of 60 years (range: 45–92). Bone marrow cytogenetics at the time of transformation to MDS or AL were available in too few patients to be analyzed. All but two of the 28 AL patients died within a year. Four patients developed MDS without evolution to AL (two RAEBt, one RAEB, one RARS); three died after 0.3, 0.9 and 2.5 years of cardiac failure at 62, 72 and 93 years, and one (RARS) was alive after 4.6 years and was 79 years old. Two patients developed ALL without the Philadelphia chromosome. One died during induction treatment, but the second, a 66-year-old woman, was still alive in remission of the ALL which had developed 5.9 years after PV diagnosis. Polycythemia relapsed and was further controlled by HU for 2 years.


203

Figure 2 Solid tumors. Kaplan–Meier curves for solid tumors in the 164 patients who received pipobroman for more than 1 year, patients who received pipobroman alone and those with added busulfan (BU) and/or switch to hydroxyurea (HU) (P ¼ 0.016 by the Wilcoxon test).

Figure 3

Myelofibrosis. Actuarial risk of evolution toward myelofibrosis in the 164 patients who received pipobroman for more than 1 year.

Actuarial risk of leukemia (Figure 4) was 2% (CI 95: [0.3–6.4]), 14.4% (CI 95: [5.4–19.2]), 18.7% (CI 95: [12.4–24.7]) and 24.3% (CI 95: [18.7–31.1]) at 5, 10, 15 and 18 years, respectively. This risk was independent of added or switched drug (P ¼ 0.65), of cumulative dose or duration of pipobroman, or of age at diagnosis (P ¼ 0.96). It was also independent of initial splenomegaly (P ¼ 0.64), Ht 460% (P ¼ 0.54) or elevated platelet count (P ¼ 0.35). Initial WBC above 10 109/l was the only variable associated with a higher risk of leukemia (P ¼ 0.032). Mortality owing to leukemia was significantly higher in younger patients (24.7% in patients o60 years versus 11.7% in patients X60 years, P ¼ 0.023).

Causes of death and survival Overall, 89 patients (54.3%) died and 56 (34.1%) were still alive at the last time point. The different causes of

death are illustrated in Table 4. Among the unrelated deaths there were two acute pneumonia, one multiple myeloma, one acute hyperparathyroidism,35 one suicide and one car crash. AL/MDS in patients younger than 60 years (51.4%) and cardiovascular diseases in older patients (42.6%) were the major causes of death. Median actuarial survival of the 164 patients (Figure 5) was 15.5 years (95 CI: 12.9–18.1), with no statistical difference between PV and IE patients (respectively 19.2 years, 95 CI: [12–NC*] and 16 years, 95 CI: [13.8–18.2], P ¼ 0.61) (*NC: upper value could not be calculated because survival of all PV patients is greater than median level). Only two major prognostic factors for survival were identified: age X60 versus o60 years (median of 11.2 versus 19.2 years, P ¼ 0.0001) and hyperleukocytosis, that is, WBC above 10 109/l versus below 10 109/l (median 14 years, CI 95: [10.6–19] versus 18 years, CI 95: [15.9–19.7], P ¼ 0.02). Added and/or switched drugs (median 15.9 years, CI 95: The Hematology Journal


204

Figure 4 Leukemic events and initial WBC count. Kaplan–Meier curves for leukemic events in the 164 patients who received pipobroman for more than 1 year, and patients with initial WBC count above or below 10 109/l (nine observations with missing value) (P ¼ 0.032 by the Wilcoxon test).

Table 4 Causes of death (by age at inclusion) Causes of death n (%)

o60 years

X60 years

All patients

Acute leukemia/MDS Vascular events Solid tumors Cardiac failure Myelofibrosis Unrelated Unknown

18 (51.4%) 2 (5.7%) 5 (14.3%) 3 (8.5%) 1 (2.8%) 5 (14.3%) 1 (2.8%)

8 (14.8%) 17 (31.5%) 8 (14.8%) 6 (11.1%) 4 (7.4%) 2 (3.7%) 9 (16.7%)

26 (29.2%) 19 (21.3%) 13 (14.6%) 9 (10.1%) 5 (5.6%) 7 (7.9%) 10 (11.2%)

35

54

89

Total

[13.8–22]) did no affect survival when compared to patients who only received pipobroman (median 16 years, CI 95: [12.5–18], P ¼ 0.12). The standardized mortality ratios (SMR) were used to compare this cohort of patients to an age- and sexmatched French population.29 They were not statistically different for the entire cohort (SMR ¼ 1.095, P ¼ 0.75) and for the different subgroups: the 31 women younger than 60 years (SMR ¼ 1.1, P ¼ 0.77), the 73 men younger than 60 years (SMR ¼ 1.05, P ¼ 0.62), the 29 women older than 60 years (SMR ¼ 1.15, P ¼ 0.16) and the 31 men older than 60 years (SMR ¼ 1.08, P ¼ 0.69).

Discussion The PVSG developed the diagnostic criteria for PV that are widely used and cited.25,26 This combination of major and minor criteria proved very robust because it overlooks IE as an early-stage PV. In the present study, IE patients represented 22% of the PV cases at the time of diagnosis and had a long-term evolution similar to PV patients. The diagnostic difficulties of IE can easily The Hematology Journal

be solved by including bone marrow histopathology as a clue to early-stage PV and to differentiate it from secondary erythrocytosis.36 The natural history of 30 untreated patients with IE was featured by cerebrovascular death in 13% and progression to overt PV in 40% during a few years follow-up.27 Only four out of 34 IE patients (11%) treated in the present study developed a nonfatal major thrombotic event during long-term follow-up. This study of a large cohort of 140 PV and 39 IE patients treated with pipobroman has the longest follow-up reported to date, and therefore allows better evaluation of long-term outcomes. Vascular events occurred in 10% of IE and 20% of PV patients and solid tumors in 7.7% of IE and 12.8% of PV patients. There were no differences between PV and IE patients with regard to progression to MF, leukemic events and overall survival. The present study shows that incidence of leukemic evolution is higher than previously reported in pipobroman studies, and that survival of all (early and overt) PV patients is not shortened when compared to an age- and sexmatched population despite the increased risk of leukemia.


205

Figure 5 Survival and initial WBC count. Kaplan–Meier estimates of survival for the 164 patients who received more than 1 year of pipobroman treatment, and patients with initial WBC count above versus below 10 109/l (nine observations with missing value) (P ¼ 0.02 by the Wilcoxon test).

Overall, 98.3% of 179 PV patients responded to pipobroman. We confirm the efficacy of pipobroman in inducing a high response rate within a few weeks. Toxicity was mild, leading to treatment discontinuation in only four patients during the induction phase. As outlined in our first report,23 and by others,11,17,18,22 maintenance treatment was necessary to avoid fluctuations of Ht and platelet values. The incidence of thrombotic complications in our study was 15.6% at 10 years, clearly lower than observed in patients treated by phlebotomy alone.2,3 Furthermore, the incidence of thrombosis in this cohort of patients appears to be lower than observed in a French-matched control population.29 This can readily be explained by the good control of hematocrit and platelet counts (CR rate of 95%), but may also be due in part to the broad prescription of low-dose aspirin in patients over 60 years of age since the mid-1980s. The actuarial risk of solid tumor was 8.4% at 10 years, that is, about 0.8% per year, when the observed incidence in a French-matched population for age and sex is 0.9% per year. This result is not different from the 1.1 and 0.8% per year reported elsewhere.11,18 In all, 13 patients out of 21 died of solid tumor in accordance with other reports.4,11,18,37 Neither cumulative dose nor duration of pipobroman therapy did increase the risk of solid tumor, but this risk was significantly higher in patients who received more than one drug (P ¼ 0.016). Similar findings were reported in PV patients treated with 32P followed by HU.4 The actuarial risk of MF was as low as 4.9% at 10 years in this study. It was 9% in the study reported by Petti et al.17 However, in their study, 71 out of 199 patients (35%) with low-dose pipobroman maintenance had iterative phlebotomies, a known factor of increased risk for both thrombosis and MF.3 The relatively low incidence of MF observed in our study may be because of the good control of the platelet count during follow-

up (91%), suggesting an efficient control of megakaryocytic proliferation and therefore a lower release of fibrogenic cytokines, especially platelet-derived growth factor.38,39 This MF risk is lower than those reported in patients treated by either phlebotomy alone, 32P or HU.3,4,11,20 Evolution to MF in our study was not correlated with a higher risk of transformation to leukemia. However, median time of evolution toward MF was longer than median time of evolution to AL (12.5 versus 9.6 years), and most of our MF patients died of pancytopenia complications within a year. These observations could explain that MF per se does not appear to precede AL in this study. An important end point in PV patients is the evaluation of the risk of leukemia, especially with a very long follow-up. The actuarial risk of 14 and 17% at, respectively, 10 and 13 years is higher than previously reported in other pipobroman trials. Petti et al.17 reported a low incidence of secondary leukemia in PV patients treated with pipobroman (6.5% at 10 years). However, this retrospective study reported results of only 199 patients out of the 366 initially registered, and in addition only 64% of patients received pipobroman as the sole maintenance treatment. The pipobroman trial recently reported by Passamonti et al.18 showed a similar relatively low risk of evolution to leukemia (4% at 10 years), but the median follow-up was shorter (10 years) and the number of events smaller when compared to this study. The high incidence that we found is probably due in part to the very long follow-up of this study. In fact, the numbers of leukemic events regularly increase with time, despite the fact that leukemia risk was not influenced by duration of pipobroman therapy or by age at diagnosis. As shown in Figure 4, there is no evidence for a plateau, 11 cases of AL out of 32 (34%) being observed after 12 years of evolution. Updated results of the randomized study of Najean and Rain11 now show a similar The Hematology Journal


206

incidence of leukemia in pipobroman-treated patients: 10.2% at the 10th year and, 12.6% at the 12th year (personal communication). Median follow-up of other published series being less than 10 years, previously reported incidence of AL could be underestimated. Initial hyperleukocytosis (WBC above 10 109/l) was the only significant risk factor for leukemia (P ¼ 0.032). The prognostic value of hyperleukocytosis has been raised before,17,20 but only reached statistical significance in this prospective study. Long-term survival of patients with a ‘chronic’ clonal disease such as PV is the most important goal. The actuarial median survival in this study is 15.5 years, similar to those reported in other pipobroman trials,11,17,18,22 and longer than the survival reported after phlebotomy alone (12.5–13.5 years),2,3,37 or 32P (9.1–11.2 years).2,3,40 In our study, the standardized mortality ratio for the various age and sex subgroups was not elevated. Life expectancy of PV patients, when compared to matched controls, was significantly lower in most previously published studies11 (for example, the SMR was 1.7 (0.3– 4.6) in a recently published Italian study18). The absence of increased mortality in our study could be related to the low incidence of fatal thrombosis or MF, but also in

part to an increased surveillance of this aged population because of their inclusion in a prospective study with frequent medical evaluations. In this cohort of patients, the actuarial risk of secondary malignancy (cumulating AL, MDS and solid tumors) was 5.6, 21.2, 28.5 and 35.4% at 5, 10, 15 and 18 years, respectively, and they represented 44% of the causes of death. Despite this increased incidence, the lower risk of death from vascular event could also explain that overall survival is similar to the control population. Age and WBC count at diagnosis were the only two prognostic factors for survival identified in this study. Hyperleukocytosis above 20 109/l has previously been suspected to be an unfavorable factor for survival,17 but this is, to our knowledge, the first prospective study showing that initial WBC count above normal value is of major prognostic significance in PV.

Acknowledgements We thank Professor Yves Najean and Professor Jean-Didier Rain for communication of updated results of their randomized trial, Dr Jean-Pierre Chauvin for critical discussions about the results of this study, and Mrs Claude Pichard for preparation of the manuscript.

References 1 Messinezy M, Pearson TC, Prochazka A, Wetherley-Mein G. Treatment of primary proliferative polycythaemia by venesection and low dose busulfan: retrospective study from one center. Br J Haematol 1985; 61: 657–666. 2 Berk PD, Goldberg JD, Donovan PB, Fruchtman SM, Berlin NI, Wasserman LR. Therapeutic recommendations in polycythemia vera based on Polycythemia Vera Study Group protocols. Semin Hematol 1986; 23: 132–143. 3 Najean Y, Rain JD. The very long-term evolution of polycythemia vera: an analysis of 318 patients initially treated by phlebotomy or 32P between 1969 and 1981. Semin Hematol 1997; 34: 6–16. 4 Najean Y, Rain JD. Treatment of polycythemia vera: use of 32 P alone or in combination with maintenance therapy using hydroxyurea in 461 patients greater than 65 years of age. The French Polycythemia Study Group. Blood 1997; 89: 2319–2327. 5 Haanen C, Mathe G, Hayat M. Treatment of polycythaemia vera by radiophosphorus or busulfan: a randomized trial. Br J Cancer 1981; 44: 75–80. 6 Gilbert HS. Problems relating to control of polycythemia vera: the use of alkylating agents. Blood 1968; 32: 500–505. 7 Higuchi T, Okada S, Mori H, Nikura H, Omine M, Terada H. Leukemic transformation of polycythemia vera and essential thrombocythemia possibly associated with an alkylating agent. Cancer 1995; 75: 471–477. 8 Donovan PB, Kaplan ME, Goldberg JD, Tatarsky I, Najean Y, Silberstein EB et al. Treatment of polycythaemia vera with hydroxyurea. Am J Hematol 1984; 17: 329–334. 9 Lofvenberg E, Wahlin A. Management of polycythaemia vera, essential thrombocythaemia and myelofibrosis with hydroxyurea. Eur J Haematol 1988; 41: 375–381.

The Hematology Journal

10 Fruchtman SM, Mack K, Kaplan ME, Peterson P, Berk PD, Wasserman LR. From efficacy to safety: a Polycythemia Vera Study Group report on hydroxyurea in patients with polycythemia vera. Semin Hematol 1997; 34: 17–23. 11 Najean Y, Rain JD. Treatment of polycythemia vera: the use of hydroxyurea and pipobroman in 292 patients under the age of 65 years. Blood 1997; 90: 3370–3377. 12 Tefferi A, Solberg LA, Silverstein MN. A clinical update in polycythemia vera and essential thrombocythemia. Am J Med 2000; 109: 141–149. 13 Streiff M, Smith B, Spivak JL. The diagnosis and management of polycythemia vera in the era since the Polycythemia Vera Study Group: a survey of American Society of Hematology members’ practice patterns. Blood 2002; 99: 1144–1149. 14 Monto RW, Tenpas A, Battle JF, Rohn RJ, Louis J, Louis NB. A-8103 in polycythemia. JAMA 1964; 190: 833–836. 15 Bohnel J, Stacher A. On the effect of a cytostatic drug agent in chronic myelosis and polycythemia. Wien Med Wochenschr 1968; 118: 721–723. 16 Bilski-Pasquier G, Blanc C, Bousser J. Traitement de la polycythemie de Vaquez par le 1-4 bis 53 bromopropionyl piperazine. Etude de 36 cas. Presse Med 1968; 41: 1953–1956. 17 Petti MC, Spadea A, Avvisati G, Spadea T, Latagliata R, Montefusco E et al. Polycythemia vera treated with pipobroman as single agent: low incidence of secondary leukemia in a cohort of patients observed during 20 years (1971–1991). Leukemia 1998; 12: 869–874. 18 Passamonti F, Brusamolino E, Lazzarino M, Barate C, Klersy C, Orlandi E et al. Efficacy of pipobroman in the


207

19 20

21 22 23

24

25 26 27 28

29

30

treatment of polycythemia vera: long-term results in 163 patients. Haematologica 2000; 85: 1011–1018. Modan B, Lilienfeld AH. Polycythemia vera and leukemia. The role of radiation treatment. A study of 1222 patients. Medicine (Baltimore) 1965; 44: 305–344. Tubiana M, Flamant R, Attie E, Hayat M. A study of hematological complications occurring in patients with polycythemia vera treated with 32P (based on a series of 296 patients). Blood 1968; 32: 536–548. Chievitz E, Thiede T. Complications and causes of death in polycythemia vera. Acta Med Scand 1962; 44: 155–157. Brusamolino E, Salvaneschi L, Canevari A, Bernasconi C. Efficacy trial of pipobroman in polycythemia vera and incidence of acute leukemia. J Clin Oncol 1984; 2: 558–561. Bernard JF, Renoux M, Dhermy D, Lallemand R, Boivin P. Traitement de la polyglobulie primitive de Vaquez par le pipobroman. Re´sultats obtenus chez 63 malades traite´s entre 1969 et 1977. Nouv Press Med 1978; 7: 3527–3530. Bernard JF, Renoux M, Turlure P, Gardin C, Briere J, Boivin P. Efficacy trial (1969–1991) of pipobroman as first treatment of polycythemia vera. A long-term study of 165 patients [abstract]. Br J Haematol 1994; 87: 144. Wasserman LR. The management of polycythaemia vera. Br J Haematol 1971; 21: 371–376. Berlin NI. Diagnosis and classification of the polycythemias. Semin Hematol 1975; 12: 339–351. Pearson TC, Wetherley-Mein G. The course and complications of idiopathic erythrocytosis. Clin Lab Haematol 1979; 1: 189–196. Bennett JM, Catovsky D, Daniel MT, Flandrin G, Galton DA, Gralnick HR et al. Proposed revised criteria for the classification of acute myeloid leukemia. A report of the French–American–British Cooperative Group. Ann Intern Med 1985; 103: 620–625. Institut National de la Sante´ et de la Recherche Me´dicale: statistiques de mortalite´ en France pour les anneés 1979 a` 1998 – INSERM sc no. 8. Le Vesinet (78) – France (http:// sc8.vesinet.inserm.fr:1080/) Pearson TC, Messinezy M. The diagnostic criteria of polycythaemia rubra vera. Leukemia Lymphoma 1996; 22: 87–93.

31 Michiels JJ, Kutti J, Stark P, Bazzan M, Gugliotta L, Marchioli R et al. Diagnosis, pathogenesis and treatment of the myeloproliferative disorders essential thrombocythemia, polycythemia vera and essential megakaryocytic granulocytic metaplasia and myelofibrosis. Neth J Med 1999; 54: 46–62. 32 Kiladjian JJ, Dubernet E, Bruno F, Boivin P, Renoux M, Bernard JF. Idiopathic erythrocytosis: long-term outcomes of 39 patients treated with pipobroman in a prospective study [abstract]. Blood 2001; 98: 629a. 33 Pahl HL. Towards a molecular understanding of polycythemia rubra vera. Eur J Biochem 2000; 267: 3395–3401. 34 Parkin DM, Pisani P, Ferlay J. Estimates of the worldwide incidence of eighteen major cancers in 1985. Int J Cancer 1993; 54: 594–606. 35 Boivin P, Bernard JF. Polycythaemia and hyperparathyroidism: a fortuitous association?. Eur J Haematol 1992; 49: 153–155. 36 Michiels JJ, Thiele J. Clinical and pathological criteria for the diagnosis of essential thrombocythemia, polycythemia vera, and idiopathic myelofibrosis (agnogenic myeloid metaplasia). Int J Hematol 2002; 76: 133–145. 37 Berk PD, Wasserman LR, Fruchtman SM, Goldberg J. Treatment of polycythemia vera: a summary of clinical trials conducted by the polycythemia vera study group. In: Wasserman LR, Berk PD, Berlin NI (eds). Polycythemia Vera and The Myeloproliferative Disorders. Saunders: Philadelphia, PA, 1995, pp 166–194. 38 Ulich TR, del Castillo J, Senaldi G, Kinstler O, Yin S, Kaufman S et al. Systemic hematologic effects of PEGrHuMGDF-induced megakaryocyte hyperplasia in mice. Blood 1996; 87: 5006–5015. 39 Yan XQ, Lacey D, Hill D, Chen Y, Fletcher F, Hawley RG et al. A model of myelofibrosis and osteosclerosis in mice induced by over expressing thrombopoietin (mpl-ligand): reversal of disease by bone marrow transplantation. Blood 1996; 88: 402–409. 40 Tubiana M, Attie E, Parmentier C. Le devenir des polyglobulies essentielles traiteés par le phosphore radioactif: a` propos d’une se´rie de 303 malades suivis pendant 12 a` 24 ans. Nouv Presse Med 1975; 4: 1781–1786.

The Hematology Journal