γ-Glutamyl Transferase Is an Independent Biomarker of Splanchnic Thrombosis in Patients With Myeloproliferative Neoplasm

Abstract

Myeloproliferative neoplasms (MPNs) are associated with an increased risk of thrombotic events and constitute the major risk factor of splanchnic venous thrombosis (SVT) in Western countries. Although timely anticoagulation resolves SVT, unrecognized SVT frequently leads to portal hypertension and, potentially, variceal bleeding, which may render anticoagulation difficult. Thus, early identification of SVT development is clinically relevant in MPN patients.In this retrospective analysis, we included 126 patients with MPN and/or SVT referred to our hospital between 2009 and 2014. A total of 86 patients diagnosed with MPN formed the first cohort (PV n = 18, ET n = 16, and MF n = 40), whereas 40 patients who had SVT without adjunct MPN formed a control cohort. Median follow-up period was 960 days. Clinical and laboratory data were collected and analyzed for the identification of potential biomarkers applying descriptive statistics, nonparametric testing, Kaplan-Meier, and logistic regression analysis. The relevance of the identified biomarkers was evaluated in an independent 2nd cohort of 181 patients from the MPN registry of the Study Alliance of Leukemia (SAL-MPN).Thirty-three MPN patients (38%) in the 1st cohort had SVT. Elevated levels of aspartate aminotransferase, alanine aminotransferase, serum bilirubin, or γ-GT were significantly correlated to the presence of SVT. In multivariate testing, CRP and aspartate aminotransferase were predictors for survival and γ-GT remained the only significant variable associated with SVT in MPN patients (P < 0.05). These findings were confirmed in the 2nd cohort comprising 42% of patients with MPN suffering from SVT.Elevated γ-GT levels indicate SVT in MPN patients, whereas CRP levels are independent predictors of patient survival.

FIGURE 1

Laboratory parameters of patients with and without MPN. Serum levels of AST, ALT, and γ-GT were significantly higher in patients with non-MPN-associated SVT when compared to MPN patients of the control and 2nd cohort (A–C). Furthermore, MPN patients who had SVT showed significantly higher AST, ALT, and γ-GT levels than MPN patients without SVT in both the control and 2nd cohort. ALT = alanine transaminase, AST = aspartate transaminase, γ-GT = gamma-glutamyl transferase, MPN = myeloproliferative neoplasm, SVT = splanchnic vein thrombosis.

FIGURE 2

γ-GT as a risk factor for MPN patients. The AUROC was determined to illustrate the predictive values of γ-GT for the presence of SVT in MPN patients (A). In Kaplan–Meier analysis, patients with γ-GT levels higher than the upper limit of normal were significantly more likely to suffer from SVT (P < 0.01) (B). Upper limit of normal was defined as 38 U/L for female and 55 U/L for male patients. Furthermore, γ-GT levels were also associated with patient survival (C, D). AUROC = area under the receiver operating characteristic, γ-GT = gamma-glutamyl transferase, MPN = myeloproliferative neoplasm, SVT = splanchnic vein thrombosis.

FIGURE 3

Risk factors for survival in MPN patients. AUROC analysis suggests LDH and AST to be associated with survival in MPN patients (A, C). In Kaplan–Meier analysis, MPN patients with serum LDH higher than 245 U/L had significantly lower cumulative probability of survival (P < 0.01) (C). AST levels higher than the upper limit of normal also showed lower survival probability, especially when observed for more than 5 years (P < 0.05) (D). Upper limit of normal was defined as 35 U/L for female and 50 U/L for male patients. AST = aspartate transaminase, AUROC = area under the receiver operating characteristic, LDH = lactate dehydrogenase, MPN = myeloproliferative neoplasm.