N-acetyl-β-D-glucosaminidase as a prognostic marker depending on ejection fraction in ICD patients

Background: N-acetyl-β-D-glucosaminidase (NAG) is an innovative tubular marker, initially tested in cardiorenal syndrome. Prior studies have proven the prognostic value in heart failure and chronic kidney disease. However, a thorough analysis of NAG levels in relation to long time cardiac systolic function, especially in patients with an internal cardioverter-defibrillator (ICD), is still lacking.

Aim: The current study is aiming to determine the prognostic value of NAG regarding mortality and upcoming appropriate ICD therapies in patients depending on initial left ventricular ejection fraction (EF) and changes of EF.

Methods: A total of 289 ICD patients treated at the university hospital Regensburg were included in this study. Blood and urine samples were obtained and transthoracic echocardiography was performed during study inclusion and follow-up (median 8 echocardiographies per patient). Patients were assigned to subgroups depending on changes in EF throughout the study (preserved EF n = 75, reduced EF n = 40, recovered EF n = 32 and undulating EF n = 35). Deceased patients were assigned to a separate subgroup (n = 107).

Results: The median age of all included patients was 68 years (IQR 59-77 years). Patients were predominantly male (245, 84.8%). In 143 patients (49.5%), ICD implantation was performed due to primary prevention. 122 patients (42.2%) had a history of appropriate ICD therapies prior to study enrolment. Upon study inclusion, median EF was 40% (IQR 32-51%) and median levels of serum creatinine were 1.1mg/dl (IQR 0.9-1.5mg/dl). Median levels of NAG levels were 2.39 U/g creatinine (IQR 1.35-4.02 U/g creatinine).
Patients with preserved EF showed significantly lower levels of NAG in comparison to all other subgroups (each p < 0.05). NAG levels of deceased patients were significantly higher than in the other groups (p each < 0.05). Furthermore, patients who suffered from appropriate ICD-therapies during follow up showed significantly higher levels of NAG upon recruitment (p < 0.05). In ROC-analysis, NAG showed good predictive value regarding mortality (AUC 0.72; sensitivity 0.64; specificity 0.74) and acceptable predictive value regarding upcoming appropriate ICD therapies (AUC 0.67; sensitivity 0.66; specificity 0.67). In multivariate binary logistic regression analyses, higher NAG-levels were significantly associated with increased mortality and subsequent ICD therapies (each p < 0.05), independent from age, sex, GFR, EF and ICD indication.

Conclusion: In the current study, lower NAG levels were associated with preserved systolic heart function in ICD patients. Moreover, higher NAG levels were associated with higher all-cause mortality and an increased likelihood of subsequent appropriate ICD therapies. NAG was an independent, significant predictor for mortality in patients with ICD.