Calreticulin mutation does not modify the IPSET score for predicting the risk of thrombosis among 1150 patients with essential thrombocythemia

To the editor:

An international prognostic score for the risk of thrombosis (IPSET-thrombosis) in essential thrombocythemia (ET) was developed.¹  Risk factors included the following: age >60 years (1 point), cardiovascular (CV) risk factors (1 point), previous thrombosis (2 points), and the presence of JAK2V617F mutation (2 points). Low-, intermediate-, and high-risk categories were identified by scores 0 to 1, 2, and ≥3, respectively. Mutations in the exon 9 of calreticulin (CALR) gene were recently identified in a large proportion of patients with JAK2V617F-negative ET and associated with a reduced thrombotic risk as compared with JAK2V617F-positive patients.^2-5  However, the utility of incorporating CALR mutation status into current risk stratification for thrombosis in ET is not yet tested. Answering this question was the purpose of the present study.

Under the auspices of the Associazione Italiana per la Ricerca sul Cancro Gruppo Italiano Malattie Mieloproliferative, 4 Italian centers convened to create a database of 1150 patients previously diagnosed with and treated for ET. The study was approved by each Institutional Review Board. Patients’ eligibility criteria included diagnosis according to the 2008 World Health Organization (WHO) criteria and full molecular characterization for JAK2V617F, MPLW515L/K, and CALR exon 9 mutations. The following methods were used: allele-specific polymerase chain reaction (PCR) or real-time quantitative PCR for JAK2V617F, and high-resolution melting analysis followed by bidirectional Sanger sequencing or next-generation sequencing for MPLW515L and CALR mutations.

Presenting features of the study population are shown in Table 1. JAK2V617F, MPLW515L/K, and CALR mutations were detected in 736 (64%), 44 (4%), and 164 (14%) patients, respectively. Eight patients (0.7%) showed a double positivity for JAK2V617F and MPL mutations and were excluded from further analysis. The remaining 198 patients (17%) were wild type for all 3 mutations. During a median follow-up of 4.1 years (range 0-29), 104 patients developed an arterial or venous thrombotic event, with a total incidence rate of 1.59% patients/year (pt-y). The IPSET-thrombosis ability to discriminate the thrombotic risk was confirmed. In fact, in the low-risk group (reference category), the rate was 0.57% pt-y; in the intermediate-risk group, 1.60% pt-y (hazard ratio [HR] 3.10; 95% confidence interval [CI], 1.55-6.18; P = .001); and in the high-risk group, 2.34% pt-y (HR 4.59; 95% CI, 2.41-8.77; P < .0001).

As to the impact of CALR mutation in the 3 categories of the IPSET-thrombosis score, we observed that CALR-mutated patients were more frequently distributed in the low- and intermediate-risk groups rather than in the high-risk IPSET group (Table 1). In patients carrying CALR mutation, a lower incidence of thrombosis during follow-up than in JAK2V617F-mutated patients was confirmed (1.30% vs 1.95% pt-y; HR 0.61; 95% CI, 0.34-1.09), although with a borderline statistical significance (P = .093). However, CALR-mutated patients were significantly younger (median age 53.5 vs 60.8 years, P = .001) and presented with less previous thrombosis (8% vs 17%, P = .005) than JAK2V617F-mutated patients (Table 1). In multivariable models, CALR mutation did not retain the association with the risk of thrombosis. This was demonstrated in the entire population (HR 0.81; 95% CI, 0.30-2.17; P = .674), as well as in the low-risk (HR 1.01; 95% CI, 0.27-3.81; P = .987) and intermediate-risk categories (HR 1.80; 95% CI, 0.57-5.72; P = .317); the high-risk category was not evaluable owing to the low proportion of CALR-mutated patients in this group. There are 2 possible reasons for this result. First, CALR mutation segregates with other factors associated with a lower thrombotic risk and already included in the score, such as younger age and less frequent previous history of thrombosis. A prognostic interaction between CALR mutations and common risk factors for thrombosis has been also reported by Gangat et al.⁶  Second, CALR mutation is virtually mutually exclusive with JAK2V617F mutation, which in turn was strongly associated with thrombosis in the score (2 points) and has been mechanistically associated with a prothrombotic state.⁷  None of the procoagulant features associated with JAK2V617F mutation have been hitherto studied in CALR-mutated patients.

In conclusion, CALR mutation status does not have a significant impact on the IPSET-thrombosis prognostic score. The score can be used as it is to predict the risk of thrombosis in molecularly annotated, WHO 2008–diagnosed ET patients.