Do plasma neprilysin activity and plasma neprilysin concentration predict cardiac events in chronic kidney disease patients?

Do plasma neprilysin activity and plasma neprilysin concentration predict cardiac events in... ABSTRACT Background Since the introduction of sacubitril/valsartan in clinical cardiology, neprilysin has become a major target for heart failure treatment. Plasma neprilysin concentration has been discussed as a novel biomarker that predicts cardiac events. Natriuretic peptides may inhibit plasma neprilysin. As they accumulate in chronic kidney disease (CKD), we hypothesized that high plasma neprilysin loses its predictive role in CKD patients. Methods We measured plasma levels of neprilysin concentration, neprilysin activity and brain natriuretic peptide (BNP) in 542 CKD G2–G4 patients within the CARE FOR HOMe study. Patients were followed for predefined endpoints of hospitalization for acute decompensated heart failure and incident atherosclerotic cardiovascular events. Results During 5.1 ± 2.1 years, 63 patients had acute decompensated heart failure and 125 patients had incident atherosclerotic cardiovascular events. In both Kaplan–Meier and multivariate Cox regression analyses, high plasma BNP and low, rather than elevated, neprilysin activity predicted future hospitalization for acute decompensated heart failure; neprilysin concentration was not predictive. Furthermore, only BNP was an independent predictor of incident atherosclerotic cardiovascular events. Conclusions In line with experimental studies, high natriuretic peptides may inhibit neprilysin activity in CKD. Therefore, high neprilysin activity and concentrations are not predictors of adverse cardiovascular outcome in CKD patients. Thus neprilysin inhibitors should be implemented with caution in patients with advanced CKD. acute decompensated heart failure, cardiac biomarkers, chronic kidney disease, incident atherosclerotic cardiovascular events, natriuretic peptides INTRODUCTION Chronic kidney disease (CKD) is associated with a high cardiovascular risk burden that is particularly driven by the high prevalence of chronic heart failure [1, 2]. Consequently, the hospitalization rate of CKD patients is extremely high [3] and their treatment is particularly resource-consuming. Since cardiac alterations are nearly ubiquitous findings among CKD patients, identification of those CKD patients at highest risk for future cardiac events is of importance for individualized prevention and treatment pathways in clinical nephrology [4]. The implementation of disease-specific biomarkers may contribute to such an identification of high-risk patients who may specifically benefit from preventive and therapeutic interventions [5]. Among individuals without overt renal disease, plasma levels of natriuretic peptides—particularly brain natriuretic peptide (BNP)—have been successfully introduced into routine clinical care, since they allow tailoring of cardioprotective medication [6]. However, the use of BNP in CKD patients is more challenging. Even though elevated natriuretic peptides remain strong predictors for future cardiac events [7–10], BNP undergoes renal elimination, so their plasma levels also increase among CKD patients without any evidence of cardiac disease [11, 12]. In contrast to the general population, prospective studies among CKD patients that prove a beneficial effect of regular measurements of natriuretic peptides among CKD patients are outstanding. Therefore the identification of additional cardiac biomarkers is of specific interest in the field of nephrology. Based on recent epidemiologic work among ambulatory patients treated at a specific heart failure clinic, neprilysin plasma levels have been proposed as a promising emerging risk factor for prediction of future cardiac events [13]. Neprilysin degrades natriuretic peptides and numerous other vasoactive peptides, although BNP is fairly resistant to neprilysin degradation [14]. It is the target of the recently approved angiotensin II receptor blocker neprilysin inhibitor, which substantially reduces the risk of future cardiac decompensation and all-cause mortality among patients with prevalent heart failure [15]. We recently found that BNP may directly inhibit circulating neprilysin activity in heart failure [16]. Since most CKD patients have elevated natriuretic peptides, neprilysin activity may be paradoxically low in these patients, even if they have high plasma neprilysin levels. We therefore hypothesized that BNP levels and neprilysin activity are inversely associated in CKD patients and that unlike in the general population, neprilysin is not an independent cardiac biomarker in CKD patients. MATERIALS AND METHODS Study design The CARE FOR HOMe study was initiated in September 2008 to characterize novel predictors of cardiovascular and renal prognosis in non-dialysis CKD patients. To meet this goal, CARE FOR HOMe recruits clinically stable non-dialysis CKD patients [Kidney Disease: Improving Global Outcomes G2–G4; estimated glomerular filtration rate (eGFR) 15–89 mL/min/1.73 m2 according to the Modification of Diet in Renal Disease equation] from the German states of Saarland and Rhineland-Palatinate who were referred to the outpatient department of the Saarland University Medical Center. The study was approved by the local ethics committee and was conducted in concordance with the Helsinki Declaration. All included participants provided written informed consent. Major exclusion criteria were systemic immunosuppressive medication, infection with the human immunodeficiency virus, clinically apparent infections, active malignancy, acute kidney injury (defined as an increase in plasma creatinine >50% within the preceding 4 weeks), pregnancy or age <18 years. Prevalent diabetes mellitus was defined as self- or physician-reported diabetes mellitus, a fasting glucose >126 mg/dL or use of glucose-lowering medication. Active smoking status was defined as active smoking or having quit smoking <1 month before study enrolment. A total of 495 patients underwent echocardiography at study initiation (see Untersteller et al. [8] for details). For assessment of systolic ventricular function, we stratified patients into those with intact and those with impaired left ventricular systolic function; 482 patients had measurements of left ventricular mass indices and left ventricular hypertrophy was defined according to American Society of Echocardiography guidelines [8]. Prevalent cardiovascular disease, defined as previous myocardial infarction, coronary artery angioplasty/stenting/bypass grafting, major stroke, carotid endarterectomy/stenting, non-traumatic lower extremity amputation or lower limb artery bypass surgery/angioplasty/stenting, was documented by a standardized questionnaire and chart review. Blood pressure was measured after 5 min of rest with an automated blood pressure recording apparatus (Carescape DINAMAP V100, GE Healthcare, Chicago, IL, USA). Body mass index (BMI) was calculated as weight (kg)/height (m)2. Patients were invited annually for follow-up visits in which we assessed the occurrence of incident atherosclerotic cardiovascular events. If a patient was unwilling or unable to return to our study centre or if he/she became dialysis dependent, we contacted the patient, his/her next of kin and/or his/her family physician for a standardized telephone interview. The predefined endpoint was hospitalization for acute decompensated heart failure during follow-up and for incident atherosclerotic cardiovascular events, defined as acute myocardial infarction, surgical or interventional coronary/cerebrovascular/peripheral arterial revascularization, stroke, amputation above the ankle and death with a cardiovascular cause, whichever occurred first. As all study participants were also followed for renal outcome—with CKD progression defined as initiation of renal replacement therapy or halving of eGFR—we also assessed (in post hoc analyses) BNP, neprilysin activity and neprilysin concentration as predictors of CKD progression. For the present analyses, we followed all participants until January 2017. Biomarker measurements At baseline, blood samples were obtained after an overnight fast under standardized conditions after 5 min of rest. Plasma creatinine was traceable to isotope dilution mass spectrometry. Plasma was frozen at −80°C for further use. Standard laboratory parameters were directly measured at Saarland University Medical Center. BNP plasma levels were measured on an Abbott Architect system (Abbott Laboratories, Abbott Park, IL, USA). Circulating neprilysin concentration and activity were determined as previously described [16]: concentration was measured using the SEB785Hu enzyme-linked immunosorbent assay kit from USCN Life Science (Wuhan, China) and activity was measured by fluorometry. Statistical analyses Data management and statistical analyses were performed with SPSS statistics software (SPSS, Chicago, IL, USA). Two sided P-values <0.05 were considered significant. Categorical variables are presented as a percentage of patients and compared using the Fisher’s test. Continuous data are expressed as mean ± SD [or median (interquartile range) in case of a skewed distribution] and were compared using analysis of variance with linear contrast analyses. Correlation coefficients between continuous variables were calculated according to Spearman. We calculated survival probabilities using Kaplan–Meier analyses with the Breslow test and by univariate and multivariate Cox regression analyses. For Kaplan–Meier analyses, neprilysin activity, neprilysin concentration and BNP were divided into three equal groups (tertiles). For Cox regression analyses, we considered these variables both as categorized and as continuous variables. Different models were tested: Model 1 was univariate analysis; in Model 2 we adjust for eGFR; in Model 3 we adjusted for eGFR, age and sex and in Model 4 we adjusted for eGFR, age, sex, diabetes mellitus, systolic blood pressure, total cholesterol, current smoking, BMI and prevalent cardiovascular disease. Finally, we conducted receiver operating characteristics (ROC) curve analyses for hospitalization for acute decompensated heart failure and incident atherosclerotic cardiovascular events that occurred within 4 years after study initiation. RESULTS Patient characteristics Of 544 patients who were recruited into CARE FOR HOMe between September 2008 and April 2015, 542 patients had available blood samples for analysis. Their baseline characteristics are summarized in Table 1. A total of 114 participants had CKD GFR category 2, 187 patients had GFR category 3a, 146 patients had GFR category 3b and 97 patients had GFR category G4. The majority of the participants had a clinical diagnosis of vascular nephropathy (46%) or diabetic nephropathy (10%); few participants had undergone renal biopsy (12%). Table 1. Baseline characteristics     cNEP (pg/mL)   P-value  aNEP (nmol/mL/min)   P-value  BNP (pg/mL)   P-value    n=542  <308.5  >308.5    <0.204  >0.204    <46.8  >46.8    (n = 271)  (n = 271)  (n = 272)  (n = 270)  (n = 272)  (n = 270)  Age (years), mean ± SD  65.2 ± 12.3  65.8 ± 11.6  64.6 ± 12.9  0.251  66.0 ± 11.7  64.3 ± 12.8  0.106  60.1 ± 12.1  70.2 ± 10.2  <0.001  Gender (female), n (%)  223 (41.0)  110 (40.6)  112 (41.3)  0.930  110 (40.4)  112 (41.5)  0.861  97 (35.7)  125 (46.3)  0.014  Prevalent CVD (yes), n (%)  171 (31.4)  89 (32.8)  81 (29.9)  0.517  93 (34.2)  77 (28.5)  0.166  55 (20.2)  115 (42.6)  <0.001  Diabetes mellitus (yes), n (%)  207 (38.1)  98 (36.2)  108 (39.9)  0.426  113 (41.5)  93 (34.4)  0.093  95 (34.9)  111 (41.1)  0.157  Smoking (yes), n (%)  57 (10.5)  34 (12.5)  23 (8.5)  0.161  32 (11.8)  25 (9.3)  0.401  35 (12.9)  22 (8.1)  0.092  BMI (kg/m2), mean ± SD  30.4 ± 5.5  30.4 ± 5.3  30.4 ± 5.6  0.960  30.6 ± 5.7  30.2 ± 5.2  0.353  30.7 ± 5.6  30.1 ± 5.4  0.250  BPsys (mmHg), mean ± SD  152 ± 24  152 ± 22  153 ± 26  0.821  154 ± 25  151 ± 23  0.169  149 ± 21  156 ± 26  <0.001  BPdia (mmHg), mean ± SD  85 ± 13  86 ± 13  85 ± 12  0.722  86 ± 13  85 ± 12  0.374  88 ± 12  83 ± 13  <0.001  eGFR (mL/min/1.73 m2), mean ± SD  46 ± 16  46 ± 16  47 ± 16  0.343  46 ± 16  47 ± 16  0.305  52 ± 16  41 ± 14  <0.001  UAE (mg/g urea), median (IQR)  32.0 (7.4–196.4)  32.2 (6.9–202.6)  31.9 (8.1–193.3)  0.661  44.4 (9.9–214.3)  25.6 (6.9–153.2)  0.055  21.5 (6.0–153.0)  46.3 (14.3–218.7)  0.604  Haemoglobin (g/dL), mean ± SD  13.5 ± 1.6  13.5 ± 1.6  13.5 ± 1.6  0.658  13.4 ± 1.7  13.6 ± 1.5  0.220  13.9 ± 1.5  13.1 ± 1.6  <0.001  25-Hydroxy Vitamin D (ng/mL), mean ± SD  23 ± 13  24 ± 15  22 ± 12  0.040  22 ± 14  24 ± 13  0.132  23 ± 13  24 ± 14  0.346  proBNP (pg/mL), median (IQR)  206.7 (86.6–545.0)  221.3 (89.6–538.0)  193.8 (82.6–547.3)  0.161  209.4 (99.1–714.4)  201.4 (76.0–497.4)  0.013  89.3 (48.5–135.5)  547.3 (276.7–1354.8)  <0.001  BNP (pg/mL), median (IQR)  46.8 (19.7–109.9)  48.3 (20.2–106.0)  44.7 (19.5–112.7)  0.341  51.2 (23.3–123.2)  43.4 (16.2–94.9)  0.002  19.8 (11.5–29.7)  110.0 (68.8–203.9)  <0.001  aNEP (nmol/mL/min), median (IQR)  0.204 (0.156–0.285)  0.199 (0.155–0.274)  0.208 (0.158–0.293)  0.170  0.156 (0.126–0.180)  0.285 (0.233–0.360)  <0.001  0.208 (0.162–0.291)  0.199 (0.151–0.281)  0.216  cNEP (pg/mL), median (IQR)  308.5 (231.0–444.0)  231.0 (181.0–268.0)  444.0 (374.0–596.0)  <0.001  301.0 (234.3–425.0)  325.0 (226.8–454.3)  0.694  318.0 (232.2–445.5)  301.5 (227.0–444.0)  0.605  LV hypertrophy (yes), n (%)  119 (24.7)  64 (26.3)  55 (23.0)  0.401  67 (27.8)  52 (21.6)  0.139  39 (16.5)  80 (32.5)  <0.001  Impaired systolic LV function (yes), n (%)  65 (13.1)  23 (9.2)  42 (17.2)  0.011  35 (14.1)  30 (12.1)  0.595  11 (4.5)  54 (21.3)  <0.001      cNEP (pg/mL)   P-value  aNEP (nmol/mL/min)   P-value  BNP (pg/mL)   P-value    n=542  <308.5  >308.5    <0.204  >0.204    <46.8  >46.8    (n = 271)  (n = 271)  (n = 272)  (n = 270)  (n = 272)  (n = 270)  Age (years), mean ± SD  65.2 ± 12.3  65.8 ± 11.6  64.6 ± 12.9  0.251  66.0 ± 11.7  64.3 ± 12.8  0.106  60.1 ± 12.1  70.2 ± 10.2  <0.001  Gender (female), n (%)  223 (41.0)  110 (40.6)  112 (41.3)  0.930  110 (40.4)  112 (41.5)  0.861  97 (35.7)  125 (46.3)  0.014  Prevalent CVD (yes), n (%)  171 (31.4)  89 (32.8)  81 (29.9)  0.517  93 (34.2)  77 (28.5)  0.166  55 (20.2)  115 (42.6)  <0.001  Diabetes mellitus (yes), n (%)  207 (38.1)  98 (36.2)  108 (39.9)  0.426  113 (41.5)  93 (34.4)  0.093  95 (34.9)  111 (41.1)  0.157  Smoking (yes), n (%)  57 (10.5)  34 (12.5)  23 (8.5)  0.161  32 (11.8)  25 (9.3)  0.401  35 (12.9)  22 (8.1)  0.092  BMI (kg/m2), mean ± SD  30.4 ± 5.5  30.4 ± 5.3  30.4 ± 5.6  0.960  30.6 ± 5.7  30.2 ± 5.2  0.353  30.7 ± 5.6  30.1 ± 5.4  0.250  BPsys (mmHg), mean ± SD  152 ± 24  152 ± 22  153 ± 26  0.821  154 ± 25  151 ± 23  0.169  149 ± 21  156 ± 26  <0.001  BPdia (mmHg), mean ± SD  85 ± 13  86 ± 13  85 ± 12  0.722  86 ± 13  85 ± 12  0.374  88 ± 12  83 ± 13  <0.001  eGFR (mL/min/1.73 m2), mean ± SD  46 ± 16  46 ± 16  47 ± 16  0.343  46 ± 16  47 ± 16  0.305  52 ± 16  41 ± 14  <0.001  UAE (mg/g urea), median (IQR)  32.0 (7.4–196.4)  32.2 (6.9–202.6)  31.9 (8.1–193.3)  0.661  44.4 (9.9–214.3)  25.6 (6.9–153.2)  0.055  21.5 (6.0–153.0)  46.3 (14.3–218.7)  0.604  Haemoglobin (g/dL), mean ± SD  13.5 ± 1.6  13.5 ± 1.6  13.5 ± 1.6  0.658  13.4 ± 1.7  13.6 ± 1.5  0.220  13.9 ± 1.5  13.1 ± 1.6  <0.001  25-Hydroxy Vitamin D (ng/mL), mean ± SD  23 ± 13  24 ± 15  22 ± 12  0.040  22 ± 14  24 ± 13  0.132  23 ± 13  24 ± 14  0.346  proBNP (pg/mL), median (IQR)  206.7 (86.6–545.0)  221.3 (89.6–538.0)  193.8 (82.6–547.3)  0.161  209.4 (99.1–714.4)  201.4 (76.0–497.4)  0.013  89.3 (48.5–135.5)  547.3 (276.7–1354.8)  <0.001  BNP (pg/mL), median (IQR)  46.8 (19.7–109.9)  48.3 (20.2–106.0)  44.7 (19.5–112.7)  0.341  51.2 (23.3–123.2)  43.4 (16.2–94.9)  0.002  19.8 (11.5–29.7)  110.0 (68.8–203.9)  <0.001  aNEP (nmol/mL/min), median (IQR)  0.204 (0.156–0.285)  0.199 (0.155–0.274)  0.208 (0.158–0.293)  0.170  0.156 (0.126–0.180)  0.285 (0.233–0.360)  <0.001  0.208 (0.162–0.291)  0.199 (0.151–0.281)  0.216  cNEP (pg/mL), median (IQR)  308.5 (231.0–444.0)  231.0 (181.0–268.0)  444.0 (374.0–596.0)  <0.001  301.0 (234.3–425.0)  325.0 (226.8–454.3)  0.694  318.0 (232.2–445.5)  301.5 (227.0–444.0)  0.605  LV hypertrophy (yes), n (%)  119 (24.7)  64 (26.3)  55 (23.0)  0.401  67 (27.8)  52 (21.6)  0.139  39 (16.5)  80 (32.5)  <0.001  Impaired systolic LV function (yes), n (%)  65 (13.1)  23 (9.2)  42 (17.2)  0.011  35 (14.1)  30 (12.1)  0.595  11 (4.5)  54 (21.3)  <0.001  Echocardiographic data on left ventricular mass index and left ventricular function were available in 482 and 495 patients, respectively. BPsys, systolic blood pressure; BPdia, diastolic blood pressure; UAE, urinary albumin excretion; aNEP, neprilysin activity; cNEP, neprilysin concentration; IQR, interquartile range; LV, left ventricular. P-value below 0.05 are presented in bold letters. Table 1. Baseline characteristics     cNEP (pg/mL)   P-value  aNEP (nmol/mL/min)   P-value  BNP (pg/mL)   P-value    n=542  <308.5  >308.5    <0.204  >0.204    <46.8  >46.8    (n = 271)  (n = 271)  (n = 272)  (n = 270)  (n = 272)  (n = 270)  Age (years), mean ± SD  65.2 ± 12.3  65.8 ± 11.6  64.6 ± 12.9  0.251  66.0 ± 11.7  64.3 ± 12.8  0.106  60.1 ± 12.1  70.2 ± 10.2  <0.001  Gender (female), n (%)  223 (41.0)  110 (40.6)  112 (41.3)  0.930  110 (40.4)  112 (41.5)  0.861  97 (35.7)  125 (46.3)  0.014  Prevalent CVD (yes), n (%)  171 (31.4)  89 (32.8)  81 (29.9)  0.517  93 (34.2)  77 (28.5)  0.166  55 (20.2)  115 (42.6)  <0.001  Diabetes mellitus (yes), n (%)  207 (38.1)  98 (36.2)  108 (39.9)  0.426  113 (41.5)  93 (34.4)  0.093  95 (34.9)  111 (41.1)  0.157  Smoking (yes), n (%)  57 (10.5)  34 (12.5)  23 (8.5)  0.161  32 (11.8)  25 (9.3)  0.401  35 (12.9)  22 (8.1)  0.092  BMI (kg/m2), mean ± SD  30.4 ± 5.5  30.4 ± 5.3  30.4 ± 5.6  0.960  30.6 ± 5.7  30.2 ± 5.2  0.353  30.7 ± 5.6  30.1 ± 5.4  0.250  BPsys (mmHg), mean ± SD  152 ± 24  152 ± 22  153 ± 26  0.821  154 ± 25  151 ± 23  0.169  149 ± 21  156 ± 26  <0.001  BPdia (mmHg), mean ± SD  85 ± 13  86 ± 13  85 ± 12  0.722  86 ± 13  85 ± 12  0.374  88 ± 12  83 ± 13  <0.001  eGFR (mL/min/1.73 m2), mean ± SD  46 ± 16  46 ± 16  47 ± 16  0.343  46 ± 16  47 ± 16  0.305  52 ± 16  41 ± 14  <0.001  UAE (mg/g urea), median (IQR)  32.0 (7.4–196.4)  32.2 (6.9–202.6)  31.9 (8.1–193.3)  0.661  44.4 (9.9–214.3)  25.6 (6.9–153.2)  0.055  21.5 (6.0–153.0)  46.3 (14.3–218.7)  0.604  Haemoglobin (g/dL), mean ± SD  13.5 ± 1.6  13.5 ± 1.6  13.5 ± 1.6  0.658  13.4 ± 1.7  13.6 ± 1.5  0.220  13.9 ± 1.5  13.1 ± 1.6  <0.001  25-Hydroxy Vitamin D (ng/mL), mean ± SD  23 ± 13  24 ± 15  22 ± 12  0.040  22 ± 14  24 ± 13  0.132  23 ± 13  24 ± 14  0.346  proBNP (pg/mL), median (IQR)  206.7 (86.6–545.0)  221.3 (89.6–538.0)  193.8 (82.6–547.3)  0.161  209.4 (99.1–714.4)  201.4 (76.0–497.4)  0.013  89.3 (48.5–135.5)  547.3 (276.7–1354.8)  <0.001  BNP (pg/mL), median (IQR)  46.8 (19.7–109.9)  48.3 (20.2–106.0)  44.7 (19.5–112.7)  0.341  51.2 (23.3–123.2)  43.4 (16.2–94.9)  0.002  19.8 (11.5–29.7)  110.0 (68.8–203.9)  <0.001  aNEP (nmol/mL/min), median (IQR)  0.204 (0.156–0.285)  0.199 (0.155–0.274)  0.208 (0.158–0.293)  0.170  0.156 (0.126–0.180)  0.285 (0.233–0.360)  <0.001  0.208 (0.162–0.291)  0.199 (0.151–0.281)  0.216  cNEP (pg/mL), median (IQR)  308.5 (231.0–444.0)  231.0 (181.0–268.0)  444.0 (374.0–596.0)  <0.001  301.0 (234.3–425.0)  325.0 (226.8–454.3)  0.694  318.0 (232.2–445.5)  301.5 (227.0–444.0)  0.605  LV hypertrophy (yes), n (%)  119 (24.7)  64 (26.3)  55 (23.0)  0.401  67 (27.8)  52 (21.6)  0.139  39 (16.5)  80 (32.5)  <0.001  Impaired systolic LV function (yes), n (%)  65 (13.1)  23 (9.2)  42 (17.2)  0.011  35 (14.1)  30 (12.1)  0.595  11 (4.5)  54 (21.3)  <0.001      cNEP (pg/mL)   P-value  aNEP (nmol/mL/min)   P-value  BNP (pg/mL)   P-value    n=542  <308.5  >308.5    <0.204  >0.204    <46.8  >46.8    (n = 271)  (n = 271)  (n = 272)  (n = 270)  (n = 272)  (n = 270)  Age (years), mean ± SD  65.2 ± 12.3  65.8 ± 11.6  64.6 ± 12.9  0.251  66.0 ± 11.7  64.3 ± 12.8  0.106  60.1 ± 12.1  70.2 ± 10.2  <0.001  Gender (female), n (%)  223 (41.0)  110 (40.6)  112 (41.3)  0.930  110 (40.4)  112 (41.5)  0.861  97 (35.7)  125 (46.3)  0.014  Prevalent CVD (yes), n (%)  171 (31.4)  89 (32.8)  81 (29.9)  0.517  93 (34.2)  77 (28.5)  0.166  55 (20.2)  115 (42.6)  <0.001  Diabetes mellitus (yes), n (%)  207 (38.1)  98 (36.2)  108 (39.9)  0.426  113 (41.5)  93 (34.4)  0.093  95 (34.9)  111 (41.1)  0.157  Smoking (yes), n (%)  57 (10.5)  34 (12.5)  23 (8.5)  0.161  32 (11.8)  25 (9.3)  0.401  35 (12.9)  22 (8.1)  0.092  BMI (kg/m2), mean ± SD  30.4 ± 5.5  30.4 ± 5.3  30.4 ± 5.6  0.960  30.6 ± 5.7  30.2 ± 5.2  0.353  30.7 ± 5.6  30.1 ± 5.4  0.250  BPsys (mmHg), mean ± SD  152 ± 24  152 ± 22  153 ± 26  0.821  154 ± 25  151 ± 23  0.169  149 ± 21  156 ± 26  <0.001  BPdia (mmHg), mean ± SD  85 ± 13  86 ± 13  85 ± 12  0.722  86 ± 13  85 ± 12  0.374  88 ± 12  83 ± 13  <0.001  eGFR (mL/min/1.73 m2), mean ± SD  46 ± 16  46 ± 16  47 ± 16  0.343  46 ± 16  47 ± 16  0.305  52 ± 16  41 ± 14  <0.001  UAE (mg/g urea), median (IQR)  32.0 (7.4–196.4)  32.2 (6.9–202.6)  31.9 (8.1–193.3)  0.661  44.4 (9.9–214.3)  25.6 (6.9–153.2)  0.055  21.5 (6.0–153.0)  46.3 (14.3–218.7)  0.604  Haemoglobin (g/dL), mean ± SD  13.5 ± 1.6  13.5 ± 1.6  13.5 ± 1.6  0.658  13.4 ± 1.7  13.6 ± 1.5  0.220  13.9 ± 1.5  13.1 ± 1.6  <0.001  25-Hydroxy Vitamin D (ng/mL), mean ± SD  23 ± 13  24 ± 15  22 ± 12  0.040  22 ± 14  24 ± 13  0.132  23 ± 13  24 ± 14  0.346  proBNP (pg/mL), median (IQR)  206.7 (86.6–545.0)  221.3 (89.6–538.0)  193.8 (82.6–547.3)  0.161  209.4 (99.1–714.4)  201.4 (76.0–497.4)  0.013  89.3 (48.5–135.5)  547.3 (276.7–1354.8)  <0.001  BNP (pg/mL), median (IQR)  46.8 (19.7–109.9)  48.3 (20.2–106.0)  44.7 (19.5–112.7)  0.341  51.2 (23.3–123.2)  43.4 (16.2–94.9)  0.002  19.8 (11.5–29.7)  110.0 (68.8–203.9)  <0.001  aNEP (nmol/mL/min), median (IQR)  0.204 (0.156–0.285)  0.199 (0.155–0.274)  0.208 (0.158–0.293)  0.170  0.156 (0.126–0.180)  0.285 (0.233–0.360)  <0.001  0.208 (0.162–0.291)  0.199 (0.151–0.281)  0.216  cNEP (pg/mL), median (IQR)  308.5 (231.0–444.0)  231.0 (181.0–268.0)  444.0 (374.0–596.0)  <0.001  301.0 (234.3–425.0)  325.0 (226.8–454.3)  0.694  318.0 (232.2–445.5)  301.5 (227.0–444.0)  0.605  LV hypertrophy (yes), n (%)  119 (24.7)  64 (26.3)  55 (23.0)  0.401  67 (27.8)  52 (21.6)  0.139  39 (16.5)  80 (32.5)  <0.001  Impaired systolic LV function (yes), n (%)  65 (13.1)  23 (9.2)  42 (17.2)  0.011  35 (14.1)  30 (12.1)  0.595  11 (4.5)  54 (21.3)  <0.001  Echocardiographic data on left ventricular mass index and left ventricular function were available in 482 and 495 patients, respectively. BPsys, systolic blood pressure; BPdia, diastolic blood pressure; UAE, urinary albumin excretion; aNEP, neprilysin activity; cNEP, neprilysin concentration; IQR, interquartile range; LV, left ventricular. P-value below 0.05 are presented in bold letters. Biomarker analysis When stratifying these patients by their median plasma neprilysin concentration, neprilysin activity and BNP, patients with a neprilysin concentration above the median had lower levels of vitamin D and patients with neprilysin activity above the median had lower BNP. Patients with BNP above the median were older and had higher systolic and lower diastolic blood pressure, a lower glomerular filtration rate and lower haemoglobin. Additionally, they were more likely to be female and to have prevalent cardiovascular disease than patients with lower levels of BNP (Table 1). Higher levels of BNP were associated with both left ventricular hypertrophy and impaired left ventricular function, whereas no association was seen between neprilysin activity and echocardiographic parameters. Higher neprilysin concentration was associated with impaired left ventricular function but not with left ventricular hypertrophy (Table 1). Neprilysin activity was weakly correlated with lower eGFR (r = 0.090, P = 0.036) and higher BNP (r = −0.127, P = 0.003) (Figure 1); the latter was most pronounced when applying a BNP cut-off level of 916 pg/mL (Figure 1), as suggested in one of our earlier studies [16]. In contrast, neprilysin concentration was not correlated with BNP (r = 0.010, P = 0.823), eGFR (r = 0.036, P = 0.398) or neprilysin activity (r = 0.007, P = 0.874). FIGURE 1: View largeDownload slide Correlation between BNP levels and neprilysin activity. Both cut-off values of BNP at 916 pg/mL and neprilysin activity at 0.21 nmol/mL/min are shown. FIGURE 1: View largeDownload slide Correlation between BNP levels and neprilysin activity. Both cut-off values of BNP at 916 pg/mL and neprilysin activity at 0.21 nmol/mL/min are shown. Follow-up analysis During a mean follow-up of 5.1 ± 2.1 years, 63 patients were hospitalized for acute decompensated heart failure; incident atherosclerotic cardiovascular events occurred in 125 patients. When stratifying patients into tertiles for baseline plasma levels of BNP, neprilysin activity and neprilysin concentration, both BNP and neprilysin activity were significantly associated with hospitalization for acute decompensated heart failure in univariate Kaplan–Meier analyses, while neprilysin concentration was not. By visual inspection, patients with the lowest neprilysin activity and highest BNP had the lowest event-free survival (Figure 2). FIGURE 2: View largeDownload slide Kaplan–Meier analyses for hospitalization for acute decompensated heart failure. Higher plasma levels of BNP were significantly associated with lower event-free survival, whereas lower levels of plasma neprilysin activity were significantly associated with lower event-free survival. FIGURE 2: View largeDownload slide Kaplan–Meier analyses for hospitalization for acute decompensated heart failure. Higher plasma levels of BNP were significantly associated with lower event-free survival, whereas lower levels of plasma neprilysin activity were significantly associated with lower event-free survival. In multivariate Cox regression analyses, high BNP remained significantly associated with hospitalization for acute decompensated heart failure even after correction for renal function and traditional cardiovascular risk factors, whether considered as a continuous parameter (after log transform) or after categorization into tertiles. Similarly, low log-transformed neprilysin activity remained associated with hospitalization for acute decompensated heart failure in the fully adjusted model, while tertiles of neprilysin activity were not (Table 2). Table 2. Cox models (endpoint: hospitalization for acute decompensated heart failure) Exposure variable  Model 1   Model 2   Model 3   Model 4   HR (95% CI)  P-value  HR (95% CI)  P-value  HR (95% CI)  P-value  HR (95% CI)  P-value  Categorized predictors   BNP (pg/mL)    Second tertile α  3.219 (1.038–9.982)  0.043  2.859 (0.921–8.873)  0.069  2.390 (0.764–7.474)  0.134  2.346 (0.740–7.432)  0.147    Third tertile α  14.457 (5.201–40.180)  <0.001  9.943 (3.551–27.843)  <0.001  6.964 (2.442–19.860)  <0.001  7.630 (2.589–22.485)  <0.001   NEP activity (nmol/mL/min)    Second tertile α  0.498 (0.263–0.943)  0.032  0.545 (0.288–1.033)  0.063  0.551 (0.291–1.045)  0.068  0.607 (0.316–1.169)  0.135    Third tertile  α  0.783 (0.438–1.400)  0.410  0.895 (0.500–1.603)  0.710  0.934 (0.520–1.676)  0.818  0.990 (0.548–1.787)  0.973   NEP concentration (pg/mL)    Second tertile α  1.555 (0.830–2.912)  0.168  1.501 (0.800–2.815)  0.206  1.521 (0.808–2.862)  0.194  1.528 (0.810–2.884)  0.191    Third tertile α  1.439 (0.748–2.766)  0.275  1.471 (0.766–2.826)  0.246  1.551 (0.805–2.987)  0.189  1.789 (0.915–3.496)  0.089  Continuous predictors   logBNP (pg/mL)  10.144 (6.218–16.548)  <0.001  7.111 (4.250–11.899)  <0.001  5.527 (3.181–9.602)  <0.001  7.166 (3.860–13.303)  <0.001   logNEP activity (nmol/mL/min)  0.121 (0.033–0.443)  0.001  0.190 (0.053–0.680)  0.011  0.225 (0.067–0.758)  0.016  0.214 (0.059–0.777)  0.019   logNEP concentration (pg/mL)  1.406 (0.483–4.087)  0.532  1.614 (0.550–4.734)  0.383  1.537 (0.513–4.610)  0.443  1.673 (0.544–5.144)  0.369  Exposure variable  Model 1   Model 2   Model 3   Model 4   HR (95% CI)  P-value  HR (95% CI)  P-value  HR (95% CI)  P-value  HR (95% CI)  P-value  Categorized predictors   BNP (pg/mL)    Second tertile α  3.219 (1.038–9.982)  0.043  2.859 (0.921–8.873)  0.069  2.390 (0.764–7.474)  0.134  2.346 (0.740–7.432)  0.147    Third tertile α  14.457 (5.201–40.180)  <0.001  9.943 (3.551–27.843)  <0.001  6.964 (2.442–19.860)  <0.001  7.630 (2.589–22.485)  <0.001   NEP activity (nmol/mL/min)    Second tertile α  0.498 (0.263–0.943)  0.032  0.545 (0.288–1.033)  0.063  0.551 (0.291–1.045)  0.068  0.607 (0.316–1.169)  0.135    Third tertile  α  0.783 (0.438–1.400)  0.410  0.895 (0.500–1.603)  0.710  0.934 (0.520–1.676)  0.818  0.990 (0.548–1.787)  0.973   NEP concentration (pg/mL)    Second tertile α  1.555 (0.830–2.912)  0.168  1.501 (0.800–2.815)  0.206  1.521 (0.808–2.862)  0.194  1.528 (0.810–2.884)  0.191    Third tertile α  1.439 (0.748–2.766)  0.275  1.471 (0.766–2.826)  0.246  1.551 (0.805–2.987)  0.189  1.789 (0.915–3.496)  0.089  Continuous predictors   logBNP (pg/mL)  10.144 (6.218–16.548)  <0.001  7.111 (4.250–11.899)  <0.001  5.527 (3.181–9.602)  <0.001  7.166 (3.860–13.303)  <0.001   logNEP activity (nmol/mL/min)  0.121 (0.033–0.443)  0.001  0.190 (0.053–0.680)  0.011  0.225 (0.067–0.758)  0.016  0.214 (0.059–0.777)  0.019   logNEP concentration (pg/mL)  1.406 (0.483–4.087)  0.532  1.614 (0.550–4.734)  0.383  1.537 (0.513–4.610)  0.443  1.673 (0.544–5.144)  0.369  Model 1 is the univariate analysis. Model 2 is adjusted for eGFR. Model 3 is adjusted for eGFR, age and sex. Model 4 is adjusted for eGFR, age, sex, diabetes mellitus, current smoker, total cholesterol, prevalent cardiovascular events, BMI and BPsys. HR, hazard ratio; CI, confidence interval; NEP, neprilysin; BNP, brain natriuretic peptide; BPsys, systolic blood pressure. α reference is the first tertile. P-value below 0.05 are presented in bold letters. Table 2. Cox models (endpoint: hospitalization for acute decompensated heart failure) Exposure variable  Model 1   Model 2   Model 3   Model 4   HR (95% CI)  P-value  HR (95% CI)  P-value  HR (95% CI)  P-value  HR (95% CI)  P-value  Categorized predictors   BNP (pg/mL)    Second tertile α  3.219 (1.038–9.982)  0.043  2.859 (0.921–8.873)  0.069  2.390 (0.764–7.474)  0.134  2.346 (0.740–7.432)  0.147    Third tertile α  14.457 (5.201–40.180)  <0.001  9.943 (3.551–27.843)  <0.001  6.964 (2.442–19.860)  <0.001  7.630 (2.589–22.485)  <0.001   NEP activity (nmol/mL/min)    Second tertile α  0.498 (0.263–0.943)  0.032  0.545 (0.288–1.033)  0.063  0.551 (0.291–1.045)  0.068  0.607 (0.316–1.169)  0.135    Third tertile  α  0.783 (0.438–1.400)  0.410  0.895 (0.500–1.603)  0.710  0.934 (0.520–1.676)  0.818  0.990 (0.548–1.787)  0.973   NEP concentration (pg/mL)    Second tertile α  1.555 (0.830–2.912)  0.168  1.501 (0.800–2.815)  0.206  1.521 (0.808–2.862)  0.194  1.528 (0.810–2.884)  0.191    Third tertile α  1.439 (0.748–2.766)  0.275  1.471 (0.766–2.826)  0.246  1.551 (0.805–2.987)  0.189  1.789 (0.915–3.496)  0.089  Continuous predictors   logBNP (pg/mL)  10.144 (6.218–16.548)  <0.001  7.111 (4.250–11.899)  <0.001  5.527 (3.181–9.602)  <0.001  7.166 (3.860–13.303)  <0.001   logNEP activity (nmol/mL/min)  0.121 (0.033–0.443)  0.001  0.190 (0.053–0.680)  0.011  0.225 (0.067–0.758)  0.016  0.214 (0.059–0.777)  0.019   logNEP concentration (pg/mL)  1.406 (0.483–4.087)  0.532  1.614 (0.550–4.734)  0.383  1.537 (0.513–4.610)  0.443  1.673 (0.544–5.144)  0.369  Exposure variable  Model 1   Model 2   Model 3   Model 4   HR (95% CI)  P-value  HR (95% CI)  P-value  HR (95% CI)  P-value  HR (95% CI)  P-value  Categorized predictors   BNP (pg/mL)    Second tertile α  3.219 (1.038–9.982)  0.043  2.859 (0.921–8.873)  0.069  2.390 (0.764–7.474)  0.134  2.346 (0.740–7.432)  0.147    Third tertile α  14.457 (5.201–40.180)  <0.001  9.943 (3.551–27.843)  <0.001  6.964 (2.442–19.860)  <0.001  7.630 (2.589–22.485)  <0.001   NEP activity (nmol/mL/min)    Second tertile α  0.498 (0.263–0.943)  0.032  0.545 (0.288–1.033)  0.063  0.551 (0.291–1.045)  0.068  0.607 (0.316–1.169)  0.135    Third tertile  α  0.783 (0.438–1.400)  0.410  0.895 (0.500–1.603)  0.710  0.934 (0.520–1.676)  0.818  0.990 (0.548–1.787)  0.973   NEP concentration (pg/mL)    Second tertile α  1.555 (0.830–2.912)  0.168  1.501 (0.800–2.815)  0.206  1.521 (0.808–2.862)  0.194  1.528 (0.810–2.884)  0.191    Third tertile α  1.439 (0.748–2.766)  0.275  1.471 (0.766–2.826)  0.246  1.551 (0.805–2.987)  0.189  1.789 (0.915–3.496)  0.089  Continuous predictors   logBNP (pg/mL)  10.144 (6.218–16.548)  <0.001  7.111 (4.250–11.899)  <0.001  5.527 (3.181–9.602)  <0.001  7.166 (3.860–13.303)  <0.001   logNEP activity (nmol/mL/min)  0.121 (0.033–0.443)  0.001  0.190 (0.053–0.680)  0.011  0.225 (0.067–0.758)  0.016  0.214 (0.059–0.777)  0.019   logNEP concentration (pg/mL)  1.406 (0.483–4.087)  0.532  1.614 (0.550–4.734)  0.383  1.537 (0.513–4.610)  0.443  1.673 (0.544–5.144)  0.369  Model 1 is the univariate analysis. Model 2 is adjusted for eGFR. Model 3 is adjusted for eGFR, age and sex. Model 4 is adjusted for eGFR, age, sex, diabetes mellitus, current smoker, total cholesterol, prevalent cardiovascular events, BMI and BPsys. HR, hazard ratio; CI, confidence interval; NEP, neprilysin; BNP, brain natriuretic peptide; BPsys, systolic blood pressure. α reference is the first tertile. P-value below 0.05 are presented in bold letters. When considering incident atherosclerotic cardiovascular events, only tertiles of plasma BNP were significantly associated with event-free survival in Kaplan–Meier analyses, while tertiles of neprilysin activity and concentration were not predictive (Figure 3). FIGURE 3: View largeDownload slide Kaplan–Meier analyses for incident atherosclerotic cardiovascular events. Only higher plasma levels of BNP were significantly associated with a lower event-free survival. FIGURE 3: View largeDownload slide Kaplan–Meier analyses for incident atherosclerotic cardiovascular events. Only higher plasma levels of BNP were significantly associated with a lower event-free survival. Accordingly, in fully adjusted Cox regression analyses, incident atherosclerotic cardiovascular events was predicted by high BNP, whether considered as a continuous parameter (after log transformation) or after categorization into tertiles. Low log transformed neprilysin activity was associated with incident atherosclerotic cardiovascular events when adjusted for eGFR, age and sex, but not in the fully adjusted model (Table 3). Table 3. Cox models (endpoint: incident atherosclerotic cardiovascular events) Exposure variable  Model 1   Model 2   Model 3   Model 4     HR (95% CI)  P-value  HR (95% CI)  P-value  HR (95% CI)  P-value  HR (95% CI)  P-value  Categorized predictors   BNP (pg/mL)    Second tertile α  2.329 (1.265–4.290)  0.007  2.112 (1.146–3.892)  0.017  2.022 (1.086–3.763)  0.026  1.826 (0.978–3.407)  0.059    Third tertile α  6.566 (3.773–11.426)  <0.001  4.776 (2.718–8.394)  <0.001  4.025 (2.233–7.253)  <0.001  3.191 (1.753–5.811)  <0.001   NEP activity (nmol/mL/min)    Second tertile α  0.676 (0.468–0.976)  0.037  0.737 (0.510–1.067)  0.106  0.784 (0.542–1.134)  0.196  0.968 (0.661–1.418)  0.868    Third tertile α  0.642 (0.436–0.946)  0.025  0.694 (0.470–1.023)  0.065  0.723 (0.490–1.066)  0.101  0.889 (0.599–1.320)  0.560   NEP concentration (pg/mL)    Second tertile α  0.828 (0.564–1.214)  0.333  0.817 (0.557–1.199)  0.302  0.826 (0.563–1.212)  0.328  0.787 (0.533–1.164)  0.231    Third tertile α  1.010 (0.691–1.475)  0.960  1.054 (0.722–1.538)  0.787  1.069 (0.732–1.561)  0.731  1.037 (0.706–1.523)  0.852  Continuous predictors   logBNP (pg/mL)  4.900 (3.547–6.769)  <0.001  3.714 (2.639–5.226)  <0.001  3.188 (2.215–4.588)  <0.001  2.665 (1.818–3.907)  <0.001   logNEP activity (nmol/mL/min)  0.259 (0.109–0.617)  0.002  0.390 (0.167–0.909)  0.029  0.423 (0.184–0.970)  0.042  0.611 (0.262–1.424)  0.254   logNEP concentration (pg/mL)  0.736 (0.359–1.508)  0.402  0.845 (0.420–1.701)  0.638  0.787 (0.387–1.601)  0.509  0.652 (0.318–1.335)  0.242  Exposure variable  Model 1   Model 2   Model 3   Model 4     HR (95% CI)  P-value  HR (95% CI)  P-value  HR (95% CI)  P-value  HR (95% CI)  P-value  Categorized predictors   BNP (pg/mL)    Second tertile α  2.329 (1.265–4.290)  0.007  2.112 (1.146–3.892)  0.017  2.022 (1.086–3.763)  0.026  1.826 (0.978–3.407)  0.059    Third tertile α  6.566 (3.773–11.426)  <0.001  4.776 (2.718–8.394)  <0.001  4.025 (2.233–7.253)  <0.001  3.191 (1.753–5.811)  <0.001   NEP activity (nmol/mL/min)    Second tertile α  0.676 (0.468–0.976)  0.037  0.737 (0.510–1.067)  0.106  0.784 (0.542–1.134)  0.196  0.968 (0.661–1.418)  0.868    Third tertile α  0.642 (0.436–0.946)  0.025  0.694 (0.470–1.023)  0.065  0.723 (0.490–1.066)  0.101  0.889 (0.599–1.320)  0.560   NEP concentration (pg/mL)    Second tertile α  0.828 (0.564–1.214)  0.333  0.817 (0.557–1.199)  0.302  0.826 (0.563–1.212)  0.328  0.787 (0.533–1.164)  0.231    Third tertile α  1.010 (0.691–1.475)  0.960  1.054 (0.722–1.538)  0.787  1.069 (0.732–1.561)  0.731  1.037 (0.706–1.523)  0.852  Continuous predictors   logBNP (pg/mL)  4.900 (3.547–6.769)  <0.001  3.714 (2.639–5.226)  <0.001  3.188 (2.215–4.588)  <0.001  2.665 (1.818–3.907)  <0.001   logNEP activity (nmol/mL/min)  0.259 (0.109–0.617)  0.002  0.390 (0.167–0.909)  0.029  0.423 (0.184–0.970)  0.042  0.611 (0.262–1.424)  0.254   logNEP concentration (pg/mL)  0.736 (0.359–1.508)  0.402  0.845 (0.420–1.701)  0.638  0.787 (0.387–1.601)  0.509  0.652 (0.318–1.335)  0.242  Model 1 is the univariate analysis. Model 2 is adjusted for eGFR. Model 3 is adjusted for eGFR, age and sex. Model 4 is adjusted for eGFR, age, sex, diabetes mellitus, current smoker, total cholesterol, prevalent cardiovascular events, BMI and BPsys. HR, hazard ratio; CI, confidence interval; NEP, neprilysin; BNP, brain natriuretic peptide; BPsys, systolic blood pressure. α reference is the first tertile. P-value below 0.05 are presented in bold letters. Table 3. Cox models (endpoint: incident atherosclerotic cardiovascular events) Exposure variable  Model 1   Model 2   Model 3   Model 4     HR (95% CI)  P-value  HR (95% CI)  P-value  HR (95% CI)  P-value  HR (95% CI)  P-value  Categorized predictors   BNP (pg/mL)    Second tertile α  2.329 (1.265–4.290)  0.007  2.112 (1.146–3.892)  0.017  2.022 (1.086–3.763)  0.026  1.826 (0.978–3.407)  0.059    Third tertile α  6.566 (3.773–11.426)  <0.001  4.776 (2.718–8.394)  <0.001  4.025 (2.233–7.253)  <0.001  3.191 (1.753–5.811)  <0.001   NEP activity (nmol/mL/min)    Second tertile α  0.676 (0.468–0.976)  0.037  0.737 (0.510–1.067)  0.106  0.784 (0.542–1.134)  0.196  0.968 (0.661–1.418)  0.868    Third tertile α  0.642 (0.436–0.946)  0.025  0.694 (0.470–1.023)  0.065  0.723 (0.490–1.066)  0.101  0.889 (0.599–1.320)  0.560   NEP concentration (pg/mL)    Second tertile α  0.828 (0.564–1.214)  0.333  0.817 (0.557–1.199)  0.302  0.826 (0.563–1.212)  0.328  0.787 (0.533–1.164)  0.231    Third tertile α  1.010 (0.691–1.475)  0.960  1.054 (0.722–1.538)  0.787  1.069 (0.732–1.561)  0.731  1.037 (0.706–1.523)  0.852  Continuous predictors   logBNP (pg/mL)  4.900 (3.547–6.769)  <0.001  3.714 (2.639–5.226)  <0.001  3.188 (2.215–4.588)  <0.001  2.665 (1.818–3.907)  <0.001   logNEP activity (nmol/mL/min)  0.259 (0.109–0.617)  0.002  0.390 (0.167–0.909)  0.029  0.423 (0.184–0.970)  0.042  0.611 (0.262–1.424)  0.254   logNEP concentration (pg/mL)  0.736 (0.359–1.508)  0.402  0.845 (0.420–1.701)  0.638  0.787 (0.387–1.601)  0.509  0.652 (0.318–1.335)  0.242  Exposure variable  Model 1   Model 2   Model 3   Model 4     HR (95% CI)  P-value  HR (95% CI)  P-value  HR (95% CI)  P-value  HR (95% CI)  P-value  Categorized predictors   BNP (pg/mL)    Second tertile α  2.329 (1.265–4.290)  0.007  2.112 (1.146–3.892)  0.017  2.022 (1.086–3.763)  0.026  1.826 (0.978–3.407)  0.059    Third tertile α  6.566 (3.773–11.426)  <0.001  4.776 (2.718–8.394)  <0.001  4.025 (2.233–7.253)  <0.001  3.191 (1.753–5.811)  <0.001   NEP activity (nmol/mL/min)    Second tertile α  0.676 (0.468–0.976)  0.037  0.737 (0.510–1.067)  0.106  0.784 (0.542–1.134)  0.196  0.968 (0.661–1.418)  0.868    Third tertile α  0.642 (0.436–0.946)  0.025  0.694 (0.470–1.023)  0.065  0.723 (0.490–1.066)  0.101  0.889 (0.599–1.320)  0.560   NEP concentration (pg/mL)    Second tertile α  0.828 (0.564–1.214)  0.333  0.817 (0.557–1.199)  0.302  0.826 (0.563–1.212)  0.328  0.787 (0.533–1.164)  0.231    Third tertile α  1.010 (0.691–1.475)  0.960  1.054 (0.722–1.538)  0.787  1.069 (0.732–1.561)  0.731  1.037 (0.706–1.523)  0.852  Continuous predictors   logBNP (pg/mL)  4.900 (3.547–6.769)  <0.001  3.714 (2.639–5.226)  <0.001  3.188 (2.215–4.588)  <0.001  2.665 (1.818–3.907)  <0.001   logNEP activity (nmol/mL/min)  0.259 (0.109–0.617)  0.002  0.390 (0.167–0.909)  0.029  0.423 (0.184–0.970)  0.042  0.611 (0.262–1.424)  0.254   logNEP concentration (pg/mL)  0.736 (0.359–1.508)  0.402  0.845 (0.420–1.701)  0.638  0.787 (0.387–1.601)  0.509  0.652 (0.318–1.335)  0.242  Model 1 is the univariate analysis. Model 2 is adjusted for eGFR. Model 3 is adjusted for eGFR, age and sex. Model 4 is adjusted for eGFR, age, sex, diabetes mellitus, current smoker, total cholesterol, prevalent cardiovascular events, BMI and BPsys. HR, hazard ratio; CI, confidence interval; NEP, neprilysin; BNP, brain natriuretic peptide; BPsys, systolic blood pressure. α reference is the first tertile. P-value below 0.05 are presented in bold letters. In contrast, both tertiles of neprilysin concentration and log transformed neprilysin concentration did not predict hospital admission for acute decompensated heart failure or incident atherosclerotic cardiovascular events in Cox regression analyses (Tables 2 and 3). In explanatory post hoc analyses, we further adjusted for log transformed albuminuria, which did not affect the study results. Additionally we performed posthoc analyses among participants with eGFR ≤30 mL/min/1.73 m2. In these subgroup analyses, BNP—when considered as a continuous variable—was associated with hospitalization for acute decompensated heart failure and with incident atherosclerotic cardiovascular events in Cox regression analyses before and after adjustment for confounders (Supplementary data, Table S2). Similarly, BNP predicted a combined endpoint of hospitalization for acute decompensated heart failure and death of any cause. While neprilysin activity tended to be associated with all three cardiovascular endpoints, even after adjustment for renal function and traditional cardiovascular risk factors, neprilysin concentration was not associated with any of these cardiovascular endpoints. In a subsequent ROC curve analysis, BNP had the largest area under the curve for predicting hospitalization for acute decompensated heart failure (Figure 4) and incident atherosclerotic cardiovascular events (Figure 5) that occurred within 4 years after study initiation. FIGURE 4: View largeDownload slide ROC curve analyses for hospitalization for acute decompensated heart failure: AUC for BNP = 0.840; AUC for NEP activity = 0.410; AUC for NEP concentration = 0.546. FIGURE 4: View largeDownload slide ROC curve analyses for hospitalization for acute decompensated heart failure: AUC for BNP = 0.840; AUC for NEP activity = 0.410; AUC for NEP concentration = 0.546. FIGURE 5: View largeDownload slide ROC curve analyses for incident atherosclerotic cardiovascular events AUC for BNP = 0.773; AUC NEP activity = 0.480; AUC for NEP concentration = 0.460. FIGURE 5: View largeDownload slide ROC curve analyses for incident atherosclerotic cardiovascular events AUC for BNP = 0.773; AUC NEP activity = 0.480; AUC for NEP concentration = 0.460. In further post hoc analyses, we studied whether BNP, neprilysin activity or neprilysin concentration predict CKD progression. Seventy-nine participants had either initiation of renal replacement therapy or halving of eGFR during follow-up. After stratifying patients into tertiles for univariate Kaplan–Meier analyses, only BNP predicted renal events, while neprilysin activity or neprilysin concentration did not (Supplementary data, Figure S1). In further Cox regression analyses, BNP—both categorized and as a continuous variable—was significantly associated with CKD progression, even after adjustment for confounders. In contrast, neither activity nor concentration of neprilysin predicted renal events (Supplementary data, Table S1). In the final posthoc analyses, we divided participants by their neprilysin concentration (above or below median) cross-classified by their BNP level (above or below median; Supplementary data, Table S3 and Figure S2). In univariate Kaplan–Meier analyses and multivariate Cox regression analyses, participants with high BNP had worse cardiovascular outcome than patients with low BNP, whether or not their neprilysin concentration was high. Similarly, high BNP with either high or low neprilysin concentration predicted CKD progression in univariate Kaplan–Meier analyses but not in fully adjusted Cox regression analyses. DISCUSSION CKD patients have a high prevalence of chronic heart failure and require frequent hospital admission for cardiac decompensation [17]. Implementation of cardiac biomarker analyses into the clinical routine may allow us to identify CKD patients with a high risk for future heart failure decompensation and subsequently to tailor cardioprotective treatment [5]. Up until now, clinical biomarker research has focused upon measurements of natriuretic peptides, which were argued to have potential shortcomings among CKD patients [18, 19]. Therefore, alternative cardiac biomarkers are of interest for physicians who care for CKD patients. Based on recent epidemiological data from non-nephrological cohorts, plasma neprilysin was suggested as such an emerging biomarker, since its plasma concentration predicted adverse outcome in patients with stable heart failure with reduced ejection fraction [13] and with acute decompensated heart failure [16]. Compared with other innovative cardiac biomarkers, neprilysin may be of particular interest for risk stratification, since it recently became a target for innovative therapeutic strategies in heart failure patients [20]. In the landmark PARADIGM-HF study, patients who received the neprilysin inhibitor sacubitril in addition to the angiotensin II receptor blocker valsartan had a 20% reduction in death from cardiovascular causes and a 21% reduction of hospitalization for heart failure compared with standard therapy with the angiotensin-converting enzyme inhibitor enalapril [20]. However, both epidemiologic studies with plasma neprilysin as a cardiac biomarker and interventional trials with sacubitril in heart failure have focused upon patients with normal or mildly impaired renal function, while patients with advanced CKD were underrepresented [13] or excluded [20]. This is of particular concern since natriuretic peptides accumulate in advanced CKD in patients with and without overt cardiac disease and since elevated natriuretic peptides may inhibit neprilysin activity [16]. Therefore, in advanced CKD patients with a strong accumulation of natriuretic peptides, less active neprilysin might be available as a pharmacologic target for sacubitril. This may affect prognostic implications of plasma neprilysin concentration measurements and render sacubitril treatment less effective. Admittedly, the knowledge about the inhibitory effect of plasma natriuretic peptides on neprilysin activity largely derives from cross-sectional clinical studies, which cannot rule out the hypothesis that high plasma concentrations of BNP are the consequence rather than the cause of low neprilysin activity. Even though a direct inhibitory effect of BNP on neprilysin activity has been shown in vitro, this evidence is based on data from only 10 patients and awaits confirmation from independent groups. Furthermore, it was largely unknown how applicable these clinical and experimental findings may be to CKD patients. We therefore measured plasma neprilysin activity, neprilysin concentration and BNP among 542 patients with mild to severe CKD in order to analyse their association across the spectrum of non-dialysis CKD and to test how the interaction between BNP and neprilysin activity affects the use of neprilysin concentration as a cardiac outcome predictor in CKD patients. We first confirmed our earlier data that neprilysin activity and neprilysin concentration are not correlated at all [16]. Thus, measuring neprilysin concentration is neither informative for its activity in acute heart failure [16] nor in stable CKD patients, which questions the biological relevance of neprilysin concentration measurements at least in subgroups of patients with heart failure. Second, we confirmed that BNP levels are inversely associated with neprilysin activity and thereby externally validate our previous finding [16] that patients with BNP above a cut-off of 916 pg/mL have very low neprilysin activity. Third, despite the great increase in BNP with advanced CKD, the association of eGFR with neprilysin activity and concentration was negligible. This finding was partly unexpected, since the accumulation of natriuretic peptides in advanced CKD and the inhibition of neprilysin by high plasma levels of natriuretic peptides might have suggested a drop in neprilysin activity in advanced CKD. Apparently other factors may counteract the inhibition of neprilysin activity induced by natriuretic peptides. As expected, high plasma levels of BNP predicted CKD progression, hospitalization for acute decompensated heart failure and incident atherosclerotic cardiovascular events, in accordance with NT-proBNP data from CARE FOR HOMe [8], and with data from other CKD cohorts that analysed natriuretic peptides [21]. However, while earlier studies suggested high neprilysin concentration indicates future heart failure hospitalization [13], it was not predictive in our CKD patients and low rather than high neprilysin activity predicted adverse cardiac outcome. Interestingly, the prognostic role of plasma neprilysin concentration was not confirmed among 144 heart failure with preserved ejection fraction (HFpEF) patients, in whom neprilysin concentration did not predict future hospitalization for heart failure or cardiac death. Unlike in other studies [13, 16] and unlike in CARE FOR HOMe, neprilysin concentration was reported to correlate with NT-proBNP in HFpEF patients [22]. Thus our findings suggest against the implementation of neprilysin measurements into clinical nephrology for the time being. This follows numerous other studies that were unable to replicate epidemiologic data on classical and non-classical cardiovascular biomarkers from the general population among CKD patients. If BNP inhibits neprilysin substantially, then an accumulation of natriuretic peptides with low glomerular function may render neprilysin less of an attractive target for pharmacologic intervention in advanced CKD than among heart failure patients with intact renal function. Admittedly, subgroup analysis from PARADIGM-HF do not point towards less efficacy of sacubitril–valsartan in patients with moderate CKD compared with patients with normal GFR; however, patients with severe CKD were excluded. This is of major importance, as sacubitrilat—an active metabolite of sacubitril—may accumulate in severe CKD [23], along with several substrates of neprilysin, such as natriuretic peptides. Therefore we await final results from the United Kingdom Heart and Renal Protection III study that recruited CKD patients with an eGFR as low as 20 mL/min/1.73 m2 [24]. This study has some limitations. Different assays for measurements of neprilysin concentration are available, which yield strongly discrepant results [25]. We cannot fully rule out that measurements with an alternative neprilysin assay would have yielded better prognostic information; however, we are positive that the lack of standardization supports our critical attitude on measurements of neprilysin concentration in cardiorenal medicine for the present time. Next, neprilysin may degrade numerous other vasoactive peptides beyond natriuretic peptides [26], such as substance P, bradykinins, atrial natriuretic peptide or endothelin-1, which we did not measure in the current study. In our earlier paper we found that BNP-mediated inhibition of neprilysin may also prevent substance P breakdown [16]. Since our major study goal was to analyse neprilysin as a novel cardiac biomarker in CKD patients, we did not measure enzymes other than neprilysin that are involved in the degradation of NP precursors to active BNP. In particular, we did not measure corin, low plasma levels of which predicted cardiovascular events among 1382 patients [27]. We deliberately focused our analyses on stable patients from our outpatient department, which is in line with the seminal paper by Bayes-Genis et al. [13] that first introduced plasma neprilysin as a potential cardiac biomarker. However, Bayes-Genis et al. [28] subsequently expanded their epidemiological studies to patients with acute heart failure, who were recruited into our present study. Additionally, we did not assess New York Heart Association classes at study initiation. As a further limitation, the number of patients who had cardiac events was limited, which may decrease the statistical power of our Cox regression analyses. In conclusion, we first show CKD patients with high BNP levels to have particularly low neprilysin activity. Second, in contrast to earlier heart failure studies, we find that low levels of neprilysin activity rather than high neprilysin concentration predict poor cardiac outcome in CKD patients. These results question the usefulness of neprilysin as a biomarker in cardiorenal medicine and are a warning signal against the uncritical use of neprilysin inhibitors in heart failure patients with advanced CKD. SUPPLEMENTARY DATA Supplementary data are available at ndt online. FUNDING The present work was supported by a grant from Else Kröner-Fresenius-Stiftung. AUTHORS’ CONTRIBUTIONS I.E.E., G.H.H. and N.V. designed the research. I.E.E., L.F., S.S.-M., K.U. and J.-M.L. conducted the research. I.E.E., G.H.H., H.N. and N.V. analysed the data and performed the statistical analysis. I.E.E., G.H.H. and N.V. wrote the paper. CONFLICT OF INTEREST STATEMENT None declared. REFERENCES 1 Matsushita K, van der Velde M, Astor BC et al.   Association of estimated glomerular filtration rate and albuminuria with all-cause and cardiovascular mortality in general population cohorts: a collaborative meta-analysis. Lancet  2010; 375: 2073– 2081 Google Scholar CrossRef Search ADS PubMed  2 Hemmelgarn BR, Manns BJ, Lloyd A et al.   Relation between kidney function, proteinuria, and adverse outcomes. JAMA  2010; 303: 423– 429 Google Scholar CrossRef Search ADS PubMed  3 Go AS, Chertow GM, Fan D et al.   Chronic kidney disease and the risks of death, cardiovascular events, and hospitalization. N Engl J Med  2004; 351: 1296– 1305 Google Scholar CrossRef Search ADS PubMed  4 Herzog CA, Asinger RW, Berger AK et al.   Cardiovascular disease in chronic kidney disease. A clinical update from Kidney Disease: Improving Global Outcomes (KDIGO). Kidney Int  2011; 80: 572– 586 Google Scholar CrossRef Search ADS PubMed  5 Fassett RG, Venuthurupalli SK, Gobe GC et al.   Biomarkers in chronic kidney disease: a review. Kidney Int  2011; 80: 806– 821 Google Scholar CrossRef Search ADS PubMed  6 Ponikowski P, Voors AA, Anker SD et al.   2016 ESC guidelines for the diagnosis and treatment of acute and chronic heart failure: the Task Force for the diagnosis and treatment of acute and chronic heart failure of the European Society of Cardiology (ESC) developed with the special contribution of the Heart Failure Association (HFA) of the ESC. Eur Heart J  2016; 37: 2129– 2200 Google Scholar CrossRef Search ADS PubMed  7 Mishra RK, Li Y, Ricardo AC et al.   Association of N-terminal pro-B-type natriuretic peptide with left ventricular structure and function in chronic kidney disease (from the Chronic Renal Insufficiency Cohort [CRIC]). Am J Cardiol  2013; 111: 432– 438 Google Scholar CrossRef Search ADS PubMed  8 Untersteller K, Girerd N, Duarte K et al.   NT-proBNP and echocardiographic parameters for prediction of cardiovascular outcomes in patients with CKD stages G2–G4. Clin J Am Soc Nephrol  2016; 11: 1978– 1988 Google Scholar CrossRef Search ADS PubMed  9 DeFilippi CR, Fink JC, Nass CM et al.   N-terminal pro-B-type natriuretic peptide for predicting coronary disease and left ventricular hypertrophy in asymptomatic CKD not requiring dialysis. Am J Kidney Dis  2005; 46: 35– 44 Google Scholar CrossRef Search ADS PubMed  10 Scheven L, de Jong PE, Hillege HL et al.   High-sensitive troponin T and N-terminal pro-B type natriuretic peptide are associated with cardiovascular events despite the cross-sectional association with albuminuria and glomerular filtration rate. Eur Heart J  2012; 33: 2272– 2281 Google Scholar CrossRef Search ADS PubMed  11 Sakuma M, Nakamura M, Tanaka F et al.   Plasma B-type natriuretic peptide level and cardiovascular events in chronic kidney disease in a community-based population. Circ J  2010; 74: 792– 797 Google Scholar CrossRef Search ADS PubMed  12 Vickery S, Price CP, John RI et al.   B-type natriuretic peptide (BNP) and amino-terminal proBNP in patients with CKD: relationship to renal function and left ventricular hypertrophy. Am J Kidney Dis  2005; 46: 610– 620 Google Scholar CrossRef Search ADS PubMed  13 Bayes-Genis A, Barallat J, Galan A et al.   Soluble neprilysin is predictive of cardiovascular death and heart failure hospitalization in heart failure patients. J Am Coll Cardiol  2015; 65: 657– 665 Google Scholar CrossRef Search ADS PubMed  14 Vodovar N, Paquet C, Mebazaa A et al.   Neprilysin, cardiovascular, and Alzheimer’s diseases: the therapeutic split? Eur Heart J  2015; 36: 902– 905 Google Scholar CrossRef Search ADS PubMed  15 Bohm M, Young R, Jhund PS et al.   Systolic blood pressure, cardiovascular outcomes and efficacy and safety of sacubitril/valsartan (LCZ696) in patients with chronic heart failure and reduced ejection fraction: results from PARADIGM-HF. Eur Heart J  2017; 38: 1132– 1143 Google Scholar CrossRef Search ADS PubMed  16 Vodovar N, Seronde MF, Laribi S et al.   Elevated plasma B-type natriuretic peptide concentrations directly inhibit circulating neprilysin activity in heart failure. JACC Heart Fail  2015; 3: 629– 636 Google Scholar CrossRef Search ADS PubMed  17 He J, Shlipak M, Anderson A et al.   Risk factors for heart failure in patients with chronic kidney disease: the CRIC (Chronic Renal Insufficiency Cohort) study. J Am Heart Assoc  2017; 6: e005336 Google Scholar CrossRef Search ADS PubMed  18 Bansal N, Hyre Anderson A, Yang W et al.   High-sensitivity troponin T and N-terminal pro-B-type natriuretic peptide (NT-proBNP) and risk of incident heart failure in patients with CKD: the Chronic Renal Insufficiency Cohort (CRIC) Study. J Am Soc Nephrol  2015; 26: 946– 956 Google Scholar CrossRef Search ADS PubMed  19 Matsushita K, Sang Y, Ballew SH et al.   Cardiac and kidney markers for cardiovascular prediction in individuals with chronic kidney disease: the Atherosclerosis Risk in Communities study. Arterioscler Thromb Vasc Biol  2014; 34: 1770– 1777 Google Scholar CrossRef Search ADS PubMed  20 McMurray JJ, Packer M, Desai AS et al.   Angiotensin–neprilysin inhibition versus enalapril in heart failure. N Engl J Med  2014; 371: 993– 1004 Google Scholar CrossRef Search ADS PubMed  21 Colbert G, Jain N, de Lemos JA et al.   Utility of traditional circulating and imaging-based cardiac biomarkers in patients with predialysis CKD. Clin J Am Soc Nephrol  2015; 10: 515– 529 Google Scholar CrossRef Search ADS PubMed  22 Goliasch G, Pavo N, Zotter-Tufaro C et al.   Soluble neprilysin does not correlate with outcome in heart failure with preserved ejection fraction. Eur J Heart Fail  2016; 18: 89– 93 Google Scholar CrossRef Search ADS PubMed  23 Ayalasomayajula SP, Langenickel TH, Jordaan P et al.   Effect of renal function on the pharmacokinetics of LCZ696 (sacubitril/valsartan), an angiotensin receptor neprilysin inhibitor. Eur J Clin Pharmacol  2016; 72: 1065– 1073 Google Scholar CrossRef Search ADS PubMed  24 UK HARP-III Collaborative Group. Randomized multicentre pilot study of sacubitril/valsartan versus irbesartan in patients with chronic kidney disease: United Kingdom Heart and Renal Protection (HARP)-III—rationale, trial design and baseline data. Nephrol Dial Transplant  2017; 32: 2043– 2051 PubMed  25 Bayes-Genis A, Barallat J, Richards AM. A test in context: neprilysin: function, inhibition, and biomarker. J Am Coll Cardiol  2016; 68: 639– 653 Google Scholar CrossRef Search ADS PubMed  26 Mangiafico S, Costello-Boerrigter LC, Andersen IA et al.   Neutral endopeptidase inhibition and the natriuretic peptide system: an evolving strategy in cardiovascular therapeutics. Eur Heart J  2013; 34: 886– 893 Google Scholar CrossRef Search ADS PubMed  27 Zhou X, Chen J, Zhang Q et al.   Prognostic value of plasma soluble corin in patients with acute myocardial infarction. J Am Coll Cardiol  2016; 67: 2008– 2014 Google Scholar CrossRef Search ADS PubMed  28 Bayes-Genis A, Barallat J, Pascual-Figal D et al.   Prognostic value and kinetics of soluble neprilysin in acute heart failure: a pilot study. JACC Heart Fail  2015; 3: 641– 644 Google Scholar CrossRef Search ADS PubMed  © The Author(s) 2018. Published by Oxford University Press on behalf of ERA-EDTA. All rights reserved. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Nephrology Dialysis Transplantation Oxford University Press

Do plasma neprilysin activity and plasma neprilysin concentration predict cardiac events in chronic kidney disease patients?

Loading next page...
 
/lp/ou_press/do-plasma-neprilysin-activity-and-plasma-neprilysin-concentration-FLUafKpBM8
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of ERA-EDTA. All rights reserved.
ISSN
0931-0509
eISSN
1460-2385
D.O.I.
10.1093/ndt/gfy066
Publisher site
See Article on Publisher Site

Abstract

ABSTRACT Background Since the introduction of sacubitril/valsartan in clinical cardiology, neprilysin has become a major target for heart failure treatment. Plasma neprilysin concentration has been discussed as a novel biomarker that predicts cardiac events. Natriuretic peptides may inhibit plasma neprilysin. As they accumulate in chronic kidney disease (CKD), we hypothesized that high plasma neprilysin loses its predictive role in CKD patients. Methods We measured plasma levels of neprilysin concentration, neprilysin activity and brain natriuretic peptide (BNP) in 542 CKD G2–G4 patients within the CARE FOR HOMe study. Patients were followed for predefined endpoints of hospitalization for acute decompensated heart failure and incident atherosclerotic cardiovascular events. Results During 5.1 ± 2.1 years, 63 patients had acute decompensated heart failure and 125 patients had incident atherosclerotic cardiovascular events. In both Kaplan–Meier and multivariate Cox regression analyses, high plasma BNP and low, rather than elevated, neprilysin activity predicted future hospitalization for acute decompensated heart failure; neprilysin concentration was not predictive. Furthermore, only BNP was an independent predictor of incident atherosclerotic cardiovascular events. Conclusions In line with experimental studies, high natriuretic peptides may inhibit neprilysin activity in CKD. Therefore, high neprilysin activity and concentrations are not predictors of adverse cardiovascular outcome in CKD patients. Thus neprilysin inhibitors should be implemented with caution in patients with advanced CKD. acute decompensated heart failure, cardiac biomarkers, chronic kidney disease, incident atherosclerotic cardiovascular events, natriuretic peptides INTRODUCTION Chronic kidney disease (CKD) is associated with a high cardiovascular risk burden that is particularly driven by the high prevalence of chronic heart failure [1, 2]. Consequently, the hospitalization rate of CKD patients is extremely high [3] and their treatment is particularly resource-consuming. Since cardiac alterations are nearly ubiquitous findings among CKD patients, identification of those CKD patients at highest risk for future cardiac events is of importance for individualized prevention and treatment pathways in clinical nephrology [4]. The implementation of disease-specific biomarkers may contribute to such an identification of high-risk patients who may specifically benefit from preventive and therapeutic interventions [5]. Among individuals without overt renal disease, plasma levels of natriuretic peptides—particularly brain natriuretic peptide (BNP)—have been successfully introduced into routine clinical care, since they allow tailoring of cardioprotective medication [6]. However, the use of BNP in CKD patients is more challenging. Even though elevated natriuretic peptides remain strong predictors for future cardiac events [7–10], BNP undergoes renal elimination, so their plasma levels also increase among CKD patients without any evidence of cardiac disease [11, 12]. In contrast to the general population, prospective studies among CKD patients that prove a beneficial effect of regular measurements of natriuretic peptides among CKD patients are outstanding. Therefore the identification of additional cardiac biomarkers is of specific interest in the field of nephrology. Based on recent epidemiologic work among ambulatory patients treated at a specific heart failure clinic, neprilysin plasma levels have been proposed as a promising emerging risk factor for prediction of future cardiac events [13]. Neprilysin degrades natriuretic peptides and numerous other vasoactive peptides, although BNP is fairly resistant to neprilysin degradation [14]. It is the target of the recently approved angiotensin II receptor blocker neprilysin inhibitor, which substantially reduces the risk of future cardiac decompensation and all-cause mortality among patients with prevalent heart failure [15]. We recently found that BNP may directly inhibit circulating neprilysin activity in heart failure [16]. Since most CKD patients have elevated natriuretic peptides, neprilysin activity may be paradoxically low in these patients, even if they have high plasma neprilysin levels. We therefore hypothesized that BNP levels and neprilysin activity are inversely associated in CKD patients and that unlike in the general population, neprilysin is not an independent cardiac biomarker in CKD patients. MATERIALS AND METHODS Study design The CARE FOR HOMe study was initiated in September 2008 to characterize novel predictors of cardiovascular and renal prognosis in non-dialysis CKD patients. To meet this goal, CARE FOR HOMe recruits clinically stable non-dialysis CKD patients [Kidney Disease: Improving Global Outcomes G2–G4; estimated glomerular filtration rate (eGFR) 15–89 mL/min/1.73 m2 according to the Modification of Diet in Renal Disease equation] from the German states of Saarland and Rhineland-Palatinate who were referred to the outpatient department of the Saarland University Medical Center. The study was approved by the local ethics committee and was conducted in concordance with the Helsinki Declaration. All included participants provided written informed consent. Major exclusion criteria were systemic immunosuppressive medication, infection with the human immunodeficiency virus, clinically apparent infections, active malignancy, acute kidney injury (defined as an increase in plasma creatinine >50% within the preceding 4 weeks), pregnancy or age <18 years. Prevalent diabetes mellitus was defined as self- or physician-reported diabetes mellitus, a fasting glucose >126 mg/dL or use of glucose-lowering medication. Active smoking status was defined as active smoking or having quit smoking <1 month before study enrolment. A total of 495 patients underwent echocardiography at study initiation (see Untersteller et al. [8] for details). For assessment of systolic ventricular function, we stratified patients into those with intact and those with impaired left ventricular systolic function; 482 patients had measurements of left ventricular mass indices and left ventricular hypertrophy was defined according to American Society of Echocardiography guidelines [8]. Prevalent cardiovascular disease, defined as previous myocardial infarction, coronary artery angioplasty/stenting/bypass grafting, major stroke, carotid endarterectomy/stenting, non-traumatic lower extremity amputation or lower limb artery bypass surgery/angioplasty/stenting, was documented by a standardized questionnaire and chart review. Blood pressure was measured after 5 min of rest with an automated blood pressure recording apparatus (Carescape DINAMAP V100, GE Healthcare, Chicago, IL, USA). Body mass index (BMI) was calculated as weight (kg)/height (m)2. Patients were invited annually for follow-up visits in which we assessed the occurrence of incident atherosclerotic cardiovascular events. If a patient was unwilling or unable to return to our study centre or if he/she became dialysis dependent, we contacted the patient, his/her next of kin and/or his/her family physician for a standardized telephone interview. The predefined endpoint was hospitalization for acute decompensated heart failure during follow-up and for incident atherosclerotic cardiovascular events, defined as acute myocardial infarction, surgical or interventional coronary/cerebrovascular/peripheral arterial revascularization, stroke, amputation above the ankle and death with a cardiovascular cause, whichever occurred first. As all study participants were also followed for renal outcome—with CKD progression defined as initiation of renal replacement therapy or halving of eGFR—we also assessed (in post hoc analyses) BNP, neprilysin activity and neprilysin concentration as predictors of CKD progression. For the present analyses, we followed all participants until January 2017. Biomarker measurements At baseline, blood samples were obtained after an overnight fast under standardized conditions after 5 min of rest. Plasma creatinine was traceable to isotope dilution mass spectrometry. Plasma was frozen at −80°C for further use. Standard laboratory parameters were directly measured at Saarland University Medical Center. BNP plasma levels were measured on an Abbott Architect system (Abbott Laboratories, Abbott Park, IL, USA). Circulating neprilysin concentration and activity were determined as previously described [16]: concentration was measured using the SEB785Hu enzyme-linked immunosorbent assay kit from USCN Life Science (Wuhan, China) and activity was measured by fluorometry. Statistical analyses Data management and statistical analyses were performed with SPSS statistics software (SPSS, Chicago, IL, USA). Two sided P-values <0.05 were considered significant. Categorical variables are presented as a percentage of patients and compared using the Fisher’s test. Continuous data are expressed as mean ± SD [or median (interquartile range) in case of a skewed distribution] and were compared using analysis of variance with linear contrast analyses. Correlation coefficients between continuous variables were calculated according to Spearman. We calculated survival probabilities using Kaplan–Meier analyses with the Breslow test and by univariate and multivariate Cox regression analyses. For Kaplan–Meier analyses, neprilysin activity, neprilysin concentration and BNP were divided into three equal groups (tertiles). For Cox regression analyses, we considered these variables both as categorized and as continuous variables. Different models were tested: Model 1 was univariate analysis; in Model 2 we adjust for eGFR; in Model 3 we adjusted for eGFR, age and sex and in Model 4 we adjusted for eGFR, age, sex, diabetes mellitus, systolic blood pressure, total cholesterol, current smoking, BMI and prevalent cardiovascular disease. Finally, we conducted receiver operating characteristics (ROC) curve analyses for hospitalization for acute decompensated heart failure and incident atherosclerotic cardiovascular events that occurred within 4 years after study initiation. RESULTS Patient characteristics Of 544 patients who were recruited into CARE FOR HOMe between September 2008 and April 2015, 542 patients had available blood samples for analysis. Their baseline characteristics are summarized in Table 1. A total of 114 participants had CKD GFR category 2, 187 patients had GFR category 3a, 146 patients had GFR category 3b and 97 patients had GFR category G4. The majority of the participants had a clinical diagnosis of vascular nephropathy (46%) or diabetic nephropathy (10%); few participants had undergone renal biopsy (12%). Table 1. Baseline characteristics     cNEP (pg/mL)   P-value  aNEP (nmol/mL/min)   P-value  BNP (pg/mL)   P-value    n=542  <308.5  >308.5    <0.204  >0.204    <46.8  >46.8    (n = 271)  (n = 271)  (n = 272)  (n = 270)  (n = 272)  (n = 270)  Age (years), mean ± SD  65.2 ± 12.3  65.8 ± 11.6  64.6 ± 12.9  0.251  66.0 ± 11.7  64.3 ± 12.8  0.106  60.1 ± 12.1  70.2 ± 10.2  <0.001  Gender (female), n (%)  223 (41.0)  110 (40.6)  112 (41.3)  0.930  110 (40.4)  112 (41.5)  0.861  97 (35.7)  125 (46.3)  0.014  Prevalent CVD (yes), n (%)  171 (31.4)  89 (32.8)  81 (29.9)  0.517  93 (34.2)  77 (28.5)  0.166  55 (20.2)  115 (42.6)  <0.001  Diabetes mellitus (yes), n (%)  207 (38.1)  98 (36.2)  108 (39.9)  0.426  113 (41.5)  93 (34.4)  0.093  95 (34.9)  111 (41.1)  0.157  Smoking (yes), n (%)  57 (10.5)  34 (12.5)  23 (8.5)  0.161  32 (11.8)  25 (9.3)  0.401  35 (12.9)  22 (8.1)  0.092  BMI (kg/m2), mean ± SD  30.4 ± 5.5  30.4 ± 5.3  30.4 ± 5.6  0.960  30.6 ± 5.7  30.2 ± 5.2  0.353  30.7 ± 5.6  30.1 ± 5.4  0.250  BPsys (mmHg), mean ± SD  152 ± 24  152 ± 22  153 ± 26  0.821  154 ± 25  151 ± 23  0.169  149 ± 21  156 ± 26  <0.001  BPdia (mmHg), mean ± SD  85 ± 13  86 ± 13  85 ± 12  0.722  86 ± 13  85 ± 12  0.374  88 ± 12  83 ± 13  <0.001  eGFR (mL/min/1.73 m2), mean ± SD  46 ± 16  46 ± 16  47 ± 16  0.343  46 ± 16  47 ± 16  0.305  52 ± 16  41 ± 14  <0.001  UAE (mg/g urea), median (IQR)  32.0 (7.4–196.4)  32.2 (6.9–202.6)  31.9 (8.1–193.3)  0.661  44.4 (9.9–214.3)  25.6 (6.9–153.2)  0.055  21.5 (6.0–153.0)  46.3 (14.3–218.7)  0.604  Haemoglobin (g/dL), mean ± SD  13.5 ± 1.6  13.5 ± 1.6  13.5 ± 1.6  0.658  13.4 ± 1.7  13.6 ± 1.5  0.220  13.9 ± 1.5  13.1 ± 1.6  <0.001  25-Hydroxy Vitamin D (ng/mL), mean ± SD  23 ± 13  24 ± 15  22 ± 12  0.040  22 ± 14  24 ± 13  0.132  23 ± 13  24 ± 14  0.346  proBNP (pg/mL), median (IQR)  206.7 (86.6–545.0)  221.3 (89.6–538.0)  193.8 (82.6–547.3)  0.161  209.4 (99.1–714.4)  201.4 (76.0–497.4)  0.013  89.3 (48.5–135.5)  547.3 (276.7–1354.8)  <0.001  BNP (pg/mL), median (IQR)  46.8 (19.7–109.9)  48.3 (20.2–106.0)  44.7 (19.5–112.7)  0.341  51.2 (23.3–123.2)  43.4 (16.2–94.9)  0.002  19.8 (11.5–29.7)  110.0 (68.8–203.9)  <0.001  aNEP (nmol/mL/min), median (IQR)  0.204 (0.156–0.285)  0.199 (0.155–0.274)  0.208 (0.158–0.293)  0.170  0.156 (0.126–0.180)  0.285 (0.233–0.360)  <0.001  0.208 (0.162–0.291)  0.199 (0.151–0.281)  0.216  cNEP (pg/mL), median (IQR)  308.5 (231.0–444.0)  231.0 (181.0–268.0)  444.0 (374.0–596.0)  <0.001  301.0 (234.3–425.0)  325.0 (226.8–454.3)  0.694  318.0 (232.2–445.5)  301.5 (227.0–444.0)  0.605  LV hypertrophy (yes), n (%)  119 (24.7)  64 (26.3)  55 (23.0)  0.401  67 (27.8)  52 (21.6)  0.139  39 (16.5)  80 (32.5)  <0.001  Impaired systolic LV function (yes), n (%)  65 (13.1)  23 (9.2)  42 (17.2)  0.011  35 (14.1)  30 (12.1)  0.595  11 (4.5)  54 (21.3)  <0.001      cNEP (pg/mL)   P-value  aNEP (nmol/mL/min)   P-value  BNP (pg/mL)   P-value    n=542  <308.5  >308.5    <0.204  >0.204    <46.8  >46.8    (n = 271)  (n = 271)  (n = 272)  (n = 270)  (n = 272)  (n = 270)  Age (years), mean ± SD  65.2 ± 12.3  65.8 ± 11.6  64.6 ± 12.9  0.251  66.0 ± 11.7  64.3 ± 12.8  0.106  60.1 ± 12.1  70.2 ± 10.2  <0.001  Gender (female), n (%)  223 (41.0)  110 (40.6)  112 (41.3)  0.930  110 (40.4)  112 (41.5)  0.861  97 (35.7)  125 (46.3)  0.014  Prevalent CVD (yes), n (%)  171 (31.4)  89 (32.8)  81 (29.9)  0.517  93 (34.2)  77 (28.5)  0.166  55 (20.2)  115 (42.6)  <0.001  Diabetes mellitus (yes), n (%)  207 (38.1)  98 (36.2)  108 (39.9)  0.426  113 (41.5)  93 (34.4)  0.093  95 (34.9)  111 (41.1)  0.157  Smoking (yes), n (%)  57 (10.5)  34 (12.5)  23 (8.5)  0.161  32 (11.8)  25 (9.3)  0.401  35 (12.9)  22 (8.1)  0.092  BMI (kg/m2), mean ± SD  30.4 ± 5.5  30.4 ± 5.3  30.4 ± 5.6  0.960  30.6 ± 5.7  30.2 ± 5.2  0.353  30.7 ± 5.6  30.1 ± 5.4  0.250  BPsys (mmHg), mean ± SD  152 ± 24  152 ± 22  153 ± 26  0.821  154 ± 25  151 ± 23  0.169  149 ± 21  156 ± 26  <0.001  BPdia (mmHg), mean ± SD  85 ± 13  86 ± 13  85 ± 12  0.722  86 ± 13  85 ± 12  0.374  88 ± 12  83 ± 13  <0.001  eGFR (mL/min/1.73 m2), mean ± SD  46 ± 16  46 ± 16  47 ± 16  0.343  46 ± 16  47 ± 16  0.305  52 ± 16  41 ± 14  <0.001  UAE (mg/g urea), median (IQR)  32.0 (7.4–196.4)  32.2 (6.9–202.6)  31.9 (8.1–193.3)  0.661  44.4 (9.9–214.3)  25.6 (6.9–153.2)  0.055  21.5 (6.0–153.0)  46.3 (14.3–218.7)  0.604  Haemoglobin (g/dL), mean ± SD  13.5 ± 1.6  13.5 ± 1.6  13.5 ± 1.6  0.658  13.4 ± 1.7  13.6 ± 1.5  0.220  13.9 ± 1.5  13.1 ± 1.6  <0.001  25-Hydroxy Vitamin D (ng/mL), mean ± SD  23 ± 13  24 ± 15  22 ± 12  0.040  22 ± 14  24 ± 13  0.132  23 ± 13  24 ± 14  0.346  proBNP (pg/mL), median (IQR)  206.7 (86.6–545.0)  221.3 (89.6–538.0)  193.8 (82.6–547.3)  0.161  209.4 (99.1–714.4)  201.4 (76.0–497.4)  0.013  89.3 (48.5–135.5)  547.3 (276.7–1354.8)  <0.001  BNP (pg/mL), median (IQR)  46.8 (19.7–109.9)  48.3 (20.2–106.0)  44.7 (19.5–112.7)  0.341  51.2 (23.3–123.2)  43.4 (16.2–94.9)  0.002  19.8 (11.5–29.7)  110.0 (68.8–203.9)  <0.001  aNEP (nmol/mL/min), median (IQR)  0.204 (0.156–0.285)  0.199 (0.155–0.274)  0.208 (0.158–0.293)  0.170  0.156 (0.126–0.180)  0.285 (0.233–0.360)  <0.001  0.208 (0.162–0.291)  0.199 (0.151–0.281)  0.216  cNEP (pg/mL), median (IQR)  308.5 (231.0–444.0)  231.0 (181.0–268.0)  444.0 (374.0–596.0)  <0.001  301.0 (234.3–425.0)  325.0 (226.8–454.3)  0.694  318.0 (232.2–445.5)  301.5 (227.0–444.0)  0.605  LV hypertrophy (yes), n (%)  119 (24.7)  64 (26.3)  55 (23.0)  0.401  67 (27.8)  52 (21.6)  0.139  39 (16.5)  80 (32.5)  <0.001  Impaired systolic LV function (yes), n (%)  65 (13.1)  23 (9.2)  42 (17.2)  0.011  35 (14.1)  30 (12.1)  0.595  11 (4.5)  54 (21.3)  <0.001  Echocardiographic data on left ventricular mass index and left ventricular function were available in 482 and 495 patients, respectively. BPsys, systolic blood pressure; BPdia, diastolic blood pressure; UAE, urinary albumin excretion; aNEP, neprilysin activity; cNEP, neprilysin concentration; IQR, interquartile range; LV, left ventricular. P-value below 0.05 are presented in bold letters. Table 1. Baseline characteristics     cNEP (pg/mL)   P-value  aNEP (nmol/mL/min)   P-value  BNP (pg/mL)   P-value    n=542  <308.5  >308.5    <0.204  >0.204    <46.8  >46.8    (n = 271)  (n = 271)  (n = 272)  (n = 270)  (n = 272)  (n = 270)  Age (years), mean ± SD  65.2 ± 12.3  65.8 ± 11.6  64.6 ± 12.9  0.251  66.0 ± 11.7  64.3 ± 12.8  0.106  60.1 ± 12.1  70.2 ± 10.2  <0.001  Gender (female), n (%)  223 (41.0)  110 (40.6)  112 (41.3)  0.930  110 (40.4)  112 (41.5)  0.861  97 (35.7)  125 (46.3)  0.014  Prevalent CVD (yes), n (%)  171 (31.4)  89 (32.8)  81 (29.9)  0.517  93 (34.2)  77 (28.5)  0.166  55 (20.2)  115 (42.6)  <0.001  Diabetes mellitus (yes), n (%)  207 (38.1)  98 (36.2)  108 (39.9)  0.426  113 (41.5)  93 (34.4)  0.093  95 (34.9)  111 (41.1)  0.157  Smoking (yes), n (%)  57 (10.5)  34 (12.5)  23 (8.5)  0.161  32 (11.8)  25 (9.3)  0.401  35 (12.9)  22 (8.1)  0.092  BMI (kg/m2), mean ± SD  30.4 ± 5.5  30.4 ± 5.3  30.4 ± 5.6  0.960  30.6 ± 5.7  30.2 ± 5.2  0.353  30.7 ± 5.6  30.1 ± 5.4  0.250  BPsys (mmHg), mean ± SD  152 ± 24  152 ± 22  153 ± 26  0.821  154 ± 25  151 ± 23  0.169  149 ± 21  156 ± 26  <0.001  BPdia (mmHg), mean ± SD  85 ± 13  86 ± 13  85 ± 12  0.722  86 ± 13  85 ± 12  0.374  88 ± 12  83 ± 13  <0.001  eGFR (mL/min/1.73 m2), mean ± SD  46 ± 16  46 ± 16  47 ± 16  0.343  46 ± 16  47 ± 16  0.305  52 ± 16  41 ± 14  <0.001  UAE (mg/g urea), median (IQR)  32.0 (7.4–196.4)  32.2 (6.9–202.6)  31.9 (8.1–193.3)  0.661  44.4 (9.9–214.3)  25.6 (6.9–153.2)  0.055  21.5 (6.0–153.0)  46.3 (14.3–218.7)  0.604  Haemoglobin (g/dL), mean ± SD  13.5 ± 1.6  13.5 ± 1.6  13.5 ± 1.6  0.658  13.4 ± 1.7  13.6 ± 1.5  0.220  13.9 ± 1.5  13.1 ± 1.6  <0.001  25-Hydroxy Vitamin D (ng/mL), mean ± SD  23 ± 13  24 ± 15  22 ± 12  0.040  22 ± 14  24 ± 13  0.132  23 ± 13  24 ± 14  0.346  proBNP (pg/mL), median (IQR)  206.7 (86.6–545.0)  221.3 (89.6–538.0)  193.8 (82.6–547.3)  0.161  209.4 (99.1–714.4)  201.4 (76.0–497.4)  0.013  89.3 (48.5–135.5)  547.3 (276.7–1354.8)  <0.001  BNP (pg/mL), median (IQR)  46.8 (19.7–109.9)  48.3 (20.2–106.0)  44.7 (19.5–112.7)  0.341  51.2 (23.3–123.2)  43.4 (16.2–94.9)  0.002  19.8 (11.5–29.7)  110.0 (68.8–203.9)  <0.001  aNEP (nmol/mL/min), median (IQR)  0.204 (0.156–0.285)  0.199 (0.155–0.274)  0.208 (0.158–0.293)  0.170  0.156 (0.126–0.180)  0.285 (0.233–0.360)  <0.001  0.208 (0.162–0.291)  0.199 (0.151–0.281)  0.216  cNEP (pg/mL), median (IQR)  308.5 (231.0–444.0)  231.0 (181.0–268.0)  444.0 (374.0–596.0)  <0.001  301.0 (234.3–425.0)  325.0 (226.8–454.3)  0.694  318.0 (232.2–445.5)  301.5 (227.0–444.0)  0.605  LV hypertrophy (yes), n (%)  119 (24.7)  64 (26.3)  55 (23.0)  0.401  67 (27.8)  52 (21.6)  0.139  39 (16.5)  80 (32.5)  <0.001  Impaired systolic LV function (yes), n (%)  65 (13.1)  23 (9.2)  42 (17.2)  0.011  35 (14.1)  30 (12.1)  0.595  11 (4.5)  54 (21.3)  <0.001      cNEP (pg/mL)   P-value  aNEP (nmol/mL/min)   P-value  BNP (pg/mL)   P-value    n=542  <308.5  >308.5    <0.204  >0.204    <46.8  >46.8    (n = 271)  (n = 271)  (n = 272)  (n = 270)  (n = 272)  (n = 270)  Age (years), mean ± SD  65.2 ± 12.3  65.8 ± 11.6  64.6 ± 12.9  0.251  66.0 ± 11.7  64.3 ± 12.8  0.106  60.1 ± 12.1  70.2 ± 10.2  <0.001  Gender (female), n (%)  223 (41.0)  110 (40.6)  112 (41.3)  0.930  110 (40.4)  112 (41.5)  0.861  97 (35.7)  125 (46.3)  0.014  Prevalent CVD (yes), n (%)  171 (31.4)  89 (32.8)  81 (29.9)  0.517  93 (34.2)  77 (28.5)  0.166  55 (20.2)  115 (42.6)  <0.001  Diabetes mellitus (yes), n (%)  207 (38.1)  98 (36.2)  108 (39.9)  0.426  113 (41.5)  93 (34.4)  0.093  95 (34.9)  111 (41.1)  0.157  Smoking (yes), n (%)  57 (10.5)  34 (12.5)  23 (8.5)  0.161  32 (11.8)  25 (9.3)  0.401  35 (12.9)  22 (8.1)  0.092  BMI (kg/m2), mean ± SD  30.4 ± 5.5  30.4 ± 5.3  30.4 ± 5.6  0.960  30.6 ± 5.7  30.2 ± 5.2  0.353  30.7 ± 5.6  30.1 ± 5.4  0.250  BPsys (mmHg), mean ± SD  152 ± 24  152 ± 22  153 ± 26  0.821  154 ± 25  151 ± 23  0.169  149 ± 21  156 ± 26  <0.001  BPdia (mmHg), mean ± SD  85 ± 13  86 ± 13  85 ± 12  0.722  86 ± 13  85 ± 12  0.374  88 ± 12  83 ± 13  <0.001  eGFR (mL/min/1.73 m2), mean ± SD  46 ± 16  46 ± 16  47 ± 16  0.343  46 ± 16  47 ± 16  0.305  52 ± 16  41 ± 14  <0.001  UAE (mg/g urea), median (IQR)  32.0 (7.4–196.4)  32.2 (6.9–202.6)  31.9 (8.1–193.3)  0.661  44.4 (9.9–214.3)  25.6 (6.9–153.2)  0.055  21.5 (6.0–153.0)  46.3 (14.3–218.7)  0.604  Haemoglobin (g/dL), mean ± SD  13.5 ± 1.6  13.5 ± 1.6  13.5 ± 1.6  0.658  13.4 ± 1.7  13.6 ± 1.5  0.220  13.9 ± 1.5  13.1 ± 1.6  <0.001  25-Hydroxy Vitamin D (ng/mL), mean ± SD  23 ± 13  24 ± 15  22 ± 12  0.040  22 ± 14  24 ± 13  0.132  23 ± 13  24 ± 14  0.346  proBNP (pg/mL), median (IQR)  206.7 (86.6–545.0)  221.3 (89.6–538.0)  193.8 (82.6–547.3)  0.161  209.4 (99.1–714.4)  201.4 (76.0–497.4)  0.013  89.3 (48.5–135.5)  547.3 (276.7–1354.8)  <0.001  BNP (pg/mL), median (IQR)  46.8 (19.7–109.9)  48.3 (20.2–106.0)  44.7 (19.5–112.7)  0.341  51.2 (23.3–123.2)  43.4 (16.2–94.9)  0.002  19.8 (11.5–29.7)  110.0 (68.8–203.9)  <0.001  aNEP (nmol/mL/min), median (IQR)  0.204 (0.156–0.285)  0.199 (0.155–0.274)  0.208 (0.158–0.293)  0.170  0.156 (0.126–0.180)  0.285 (0.233–0.360)  <0.001  0.208 (0.162–0.291)  0.199 (0.151–0.281)  0.216  cNEP (pg/mL), median (IQR)  308.5 (231.0–444.0)  231.0 (181.0–268.0)  444.0 (374.0–596.0)  <0.001  301.0 (234.3–425.0)  325.0 (226.8–454.3)  0.694  318.0 (232.2–445.5)  301.5 (227.0–444.0)  0.605  LV hypertrophy (yes), n (%)  119 (24.7)  64 (26.3)  55 (23.0)  0.401  67 (27.8)  52 (21.6)  0.139  39 (16.5)  80 (32.5)  <0.001  Impaired systolic LV function (yes), n (%)  65 (13.1)  23 (9.2)  42 (17.2)  0.011  35 (14.1)  30 (12.1)  0.595  11 (4.5)  54 (21.3)  <0.001  Echocardiographic data on left ventricular mass index and left ventricular function were available in 482 and 495 patients, respectively. BPsys, systolic blood pressure; BPdia, diastolic blood pressure; UAE, urinary albumin excretion; aNEP, neprilysin activity; cNEP, neprilysin concentration; IQR, interquartile range; LV, left ventricular. P-value below 0.05 are presented in bold letters. Biomarker analysis When stratifying these patients by their median plasma neprilysin concentration, neprilysin activity and BNP, patients with a neprilysin concentration above the median had lower levels of vitamin D and patients with neprilysin activity above the median had lower BNP. Patients with BNP above the median were older and had higher systolic and lower diastolic blood pressure, a lower glomerular filtration rate and lower haemoglobin. Additionally, they were more likely to be female and to have prevalent cardiovascular disease than patients with lower levels of BNP (Table 1). Higher levels of BNP were associated with both left ventricular hypertrophy and impaired left ventricular function, whereas no association was seen between neprilysin activity and echocardiographic parameters. Higher neprilysin concentration was associated with impaired left ventricular function but not with left ventricular hypertrophy (Table 1). Neprilysin activity was weakly correlated with lower eGFR (r = 0.090, P = 0.036) and higher BNP (r = −0.127, P = 0.003) (Figure 1); the latter was most pronounced when applying a BNP cut-off level of 916 pg/mL (Figure 1), as suggested in one of our earlier studies [16]. In contrast, neprilysin concentration was not correlated with BNP (r = 0.010, P = 0.823), eGFR (r = 0.036, P = 0.398) or neprilysin activity (r = 0.007, P = 0.874). FIGURE 1: View largeDownload slide Correlation between BNP levels and neprilysin activity. Both cut-off values of BNP at 916 pg/mL and neprilysin activity at 0.21 nmol/mL/min are shown. FIGURE 1: View largeDownload slide Correlation between BNP levels and neprilysin activity. Both cut-off values of BNP at 916 pg/mL and neprilysin activity at 0.21 nmol/mL/min are shown. Follow-up analysis During a mean follow-up of 5.1 ± 2.1 years, 63 patients were hospitalized for acute decompensated heart failure; incident atherosclerotic cardiovascular events occurred in 125 patients. When stratifying patients into tertiles for baseline plasma levels of BNP, neprilysin activity and neprilysin concentration, both BNP and neprilysin activity were significantly associated with hospitalization for acute decompensated heart failure in univariate Kaplan–Meier analyses, while neprilysin concentration was not. By visual inspection, patients with the lowest neprilysin activity and highest BNP had the lowest event-free survival (Figure 2). FIGURE 2: View largeDownload slide Kaplan–Meier analyses for hospitalization for acute decompensated heart failure. Higher plasma levels of BNP were significantly associated with lower event-free survival, whereas lower levels of plasma neprilysin activity were significantly associated with lower event-free survival. FIGURE 2: View largeDownload slide Kaplan–Meier analyses for hospitalization for acute decompensated heart failure. Higher plasma levels of BNP were significantly associated with lower event-free survival, whereas lower levels of plasma neprilysin activity were significantly associated with lower event-free survival. In multivariate Cox regression analyses, high BNP remained significantly associated with hospitalization for acute decompensated heart failure even after correction for renal function and traditional cardiovascular risk factors, whether considered as a continuous parameter (after log transform) or after categorization into tertiles. Similarly, low log-transformed neprilysin activity remained associated with hospitalization for acute decompensated heart failure in the fully adjusted model, while tertiles of neprilysin activity were not (Table 2). Table 2. Cox models (endpoint: hospitalization for acute decompensated heart failure) Exposure variable  Model 1   Model 2   Model 3   Model 4   HR (95% CI)  P-value  HR (95% CI)  P-value  HR (95% CI)  P-value  HR (95% CI)  P-value  Categorized predictors   BNP (pg/mL)    Second tertile α  3.219 (1.038–9.982)  0.043  2.859 (0.921–8.873)  0.069  2.390 (0.764–7.474)  0.134  2.346 (0.740–7.432)  0.147    Third tertile α  14.457 (5.201–40.180)  <0.001  9.943 (3.551–27.843)  <0.001  6.964 (2.442–19.860)  <0.001  7.630 (2.589–22.485)  <0.001   NEP activity (nmol/mL/min)    Second tertile α  0.498 (0.263–0.943)  0.032  0.545 (0.288–1.033)  0.063  0.551 (0.291–1.045)  0.068  0.607 (0.316–1.169)  0.135    Third tertile  α  0.783 (0.438–1.400)  0.410  0.895 (0.500–1.603)  0.710  0.934 (0.520–1.676)  0.818  0.990 (0.548–1.787)  0.973   NEP concentration (pg/mL)    Second tertile α  1.555 (0.830–2.912)  0.168  1.501 (0.800–2.815)  0.206  1.521 (0.808–2.862)  0.194  1.528 (0.810–2.884)  0.191    Third tertile α  1.439 (0.748–2.766)  0.275  1.471 (0.766–2.826)  0.246  1.551 (0.805–2.987)  0.189  1.789 (0.915–3.496)  0.089  Continuous predictors   logBNP (pg/mL)  10.144 (6.218–16.548)  <0.001  7.111 (4.250–11.899)  <0.001  5.527 (3.181–9.602)  <0.001  7.166 (3.860–13.303)  <0.001   logNEP activity (nmol/mL/min)  0.121 (0.033–0.443)  0.001  0.190 (0.053–0.680)  0.011  0.225 (0.067–0.758)  0.016  0.214 (0.059–0.777)  0.019   logNEP concentration (pg/mL)  1.406 (0.483–4.087)  0.532  1.614 (0.550–4.734)  0.383  1.537 (0.513–4.610)  0.443  1.673 (0.544–5.144)  0.369  Exposure variable  Model 1   Model 2   Model 3   Model 4   HR (95% CI)  P-value  HR (95% CI)  P-value  HR (95% CI)  P-value  HR (95% CI)  P-value  Categorized predictors   BNP (pg/mL)    Second tertile α  3.219 (1.038–9.982)  0.043  2.859 (0.921–8.873)  0.069  2.390 (0.764–7.474)  0.134  2.346 (0.740–7.432)  0.147    Third tertile α  14.457 (5.201–40.180)  <0.001  9.943 (3.551–27.843)  <0.001  6.964 (2.442–19.860)  <0.001  7.630 (2.589–22.485)  <0.001   NEP activity (nmol/mL/min)    Second tertile α  0.498 (0.263–0.943)  0.032  0.545 (0.288–1.033)  0.063  0.551 (0.291–1.045)  0.068  0.607 (0.316–1.169)  0.135    Third tertile  α  0.783 (0.438–1.400)  0.410  0.895 (0.500–1.603)  0.710  0.934 (0.520–1.676)  0.818  0.990 (0.548–1.787)  0.973   NEP concentration (pg/mL)    Second tertile α  1.555 (0.830–2.912)  0.168  1.501 (0.800–2.815)  0.206  1.521 (0.808–2.862)  0.194  1.528 (0.810–2.884)  0.191    Third tertile α  1.439 (0.748–2.766)  0.275  1.471 (0.766–2.826)  0.246  1.551 (0.805–2.987)  0.189  1.789 (0.915–3.496)  0.089  Continuous predictors   logBNP (pg/mL)  10.144 (6.218–16.548)  <0.001  7.111 (4.250–11.899)  <0.001  5.527 (3.181–9.602)  <0.001  7.166 (3.860–13.303)  <0.001   logNEP activity (nmol/mL/min)  0.121 (0.033–0.443)  0.001  0.190 (0.053–0.680)  0.011  0.225 (0.067–0.758)  0.016  0.214 (0.059–0.777)  0.019   logNEP concentration (pg/mL)  1.406 (0.483–4.087)  0.532  1.614 (0.550–4.734)  0.383  1.537 (0.513–4.610)  0.443  1.673 (0.544–5.144)  0.369  Model 1 is the univariate analysis. Model 2 is adjusted for eGFR. Model 3 is adjusted for eGFR, age and sex. Model 4 is adjusted for eGFR, age, sex, diabetes mellitus, current smoker, total cholesterol, prevalent cardiovascular events, BMI and BPsys. HR, hazard ratio; CI, confidence interval; NEP, neprilysin; BNP, brain natriuretic peptide; BPsys, systolic blood pressure. α reference is the first tertile. P-value below 0.05 are presented in bold letters. Table 2. Cox models (endpoint: hospitalization for acute decompensated heart failure) Exposure variable  Model 1   Model 2   Model 3   Model 4   HR (95% CI)  P-value  HR (95% CI)  P-value  HR (95% CI)  P-value  HR (95% CI)  P-value  Categorized predictors   BNP (pg/mL)    Second tertile α  3.219 (1.038–9.982)  0.043  2.859 (0.921–8.873)  0.069  2.390 (0.764–7.474)  0.134  2.346 (0.740–7.432)  0.147    Third tertile α  14.457 (5.201–40.180)  <0.001  9.943 (3.551–27.843)  <0.001  6.964 (2.442–19.860)  <0.001  7.630 (2.589–22.485)  <0.001   NEP activity (nmol/mL/min)    Second tertile α  0.498 (0.263–0.943)  0.032  0.545 (0.288–1.033)  0.063  0.551 (0.291–1.045)  0.068  0.607 (0.316–1.169)  0.135    Third tertile  α  0.783 (0.438–1.400)  0.410  0.895 (0.500–1.603)  0.710  0.934 (0.520–1.676)  0.818  0.990 (0.548–1.787)  0.973   NEP concentration (pg/mL)    Second tertile α  1.555 (0.830–2.912)  0.168  1.501 (0.800–2.815)  0.206  1.521 (0.808–2.862)  0.194  1.528 (0.810–2.884)  0.191    Third tertile α  1.439 (0.748–2.766)  0.275  1.471 (0.766–2.826)  0.246  1.551 (0.805–2.987)  0.189  1.789 (0.915–3.496)  0.089  Continuous predictors   logBNP (pg/mL)  10.144 (6.218–16.548)  <0.001  7.111 (4.250–11.899)  <0.001  5.527 (3.181–9.602)  <0.001  7.166 (3.860–13.303)  <0.001   logNEP activity (nmol/mL/min)  0.121 (0.033–0.443)  0.001  0.190 (0.053–0.680)  0.011  0.225 (0.067–0.758)  0.016  0.214 (0.059–0.777)  0.019   logNEP concentration (pg/mL)  1.406 (0.483–4.087)  0.532  1.614 (0.550–4.734)  0.383  1.537 (0.513–4.610)  0.443  1.673 (0.544–5.144)  0.369  Exposure variable  Model 1   Model 2   Model 3   Model 4   HR (95% CI)  P-value  HR (95% CI)  P-value  HR (95% CI)  P-value  HR (95% CI)  P-value  Categorized predictors   BNP (pg/mL)    Second tertile α  3.219 (1.038–9.982)  0.043  2.859 (0.921–8.873)  0.069  2.390 (0.764–7.474)  0.134  2.346 (0.740–7.432)  0.147    Third tertile α  14.457 (5.201–40.180)  <0.001  9.943 (3.551–27.843)  <0.001  6.964 (2.442–19.860)  <0.001  7.630 (2.589–22.485)  <0.001   NEP activity (nmol/mL/min)    Second tertile α  0.498 (0.263–0.943)  0.032  0.545 (0.288–1.033)  0.063  0.551 (0.291–1.045)  0.068  0.607 (0.316–1.169)  0.135    Third tertile  α  0.783 (0.438–1.400)  0.410  0.895 (0.500–1.603)  0.710  0.934 (0.520–1.676)  0.818  0.990 (0.548–1.787)  0.973   NEP concentration (pg/mL)    Second tertile α  1.555 (0.830–2.912)  0.168  1.501 (0.800–2.815)  0.206  1.521 (0.808–2.862)  0.194  1.528 (0.810–2.884)  0.191    Third tertile α  1.439 (0.748–2.766)  0.275  1.471 (0.766–2.826)  0.246  1.551 (0.805–2.987)  0.189  1.789 (0.915–3.496)  0.089  Continuous predictors   logBNP (pg/mL)  10.144 (6.218–16.548)  <0.001  7.111 (4.250–11.899)  <0.001  5.527 (3.181–9.602)  <0.001  7.166 (3.860–13.303)  <0.001   logNEP activity (nmol/mL/min)  0.121 (0.033–0.443)  0.001  0.190 (0.053–0.680)  0.011  0.225 (0.067–0.758)  0.016  0.214 (0.059–0.777)  0.019   logNEP concentration (pg/mL)  1.406 (0.483–4.087)  0.532  1.614 (0.550–4.734)  0.383  1.537 (0.513–4.610)  0.443  1.673 (0.544–5.144)  0.369  Model 1 is the univariate analysis. Model 2 is adjusted for eGFR. Model 3 is adjusted for eGFR, age and sex. Model 4 is adjusted for eGFR, age, sex, diabetes mellitus, current smoker, total cholesterol, prevalent cardiovascular events, BMI and BPsys. HR, hazard ratio; CI, confidence interval; NEP, neprilysin; BNP, brain natriuretic peptide; BPsys, systolic blood pressure. α reference is the first tertile. P-value below 0.05 are presented in bold letters. When considering incident atherosclerotic cardiovascular events, only tertiles of plasma BNP were significantly associated with event-free survival in Kaplan–Meier analyses, while tertiles of neprilysin activity and concentration were not predictive (Figure 3). FIGURE 3: View largeDownload slide Kaplan–Meier analyses for incident atherosclerotic cardiovascular events. Only higher plasma levels of BNP were significantly associated with a lower event-free survival. FIGURE 3: View largeDownload slide Kaplan–Meier analyses for incident atherosclerotic cardiovascular events. Only higher plasma levels of BNP were significantly associated with a lower event-free survival. Accordingly, in fully adjusted Cox regression analyses, incident atherosclerotic cardiovascular events was predicted by high BNP, whether considered as a continuous parameter (after log transformation) or after categorization into tertiles. Low log transformed neprilysin activity was associated with incident atherosclerotic cardiovascular events when adjusted for eGFR, age and sex, but not in the fully adjusted model (Table 3). Table 3. Cox models (endpoint: incident atherosclerotic cardiovascular events) Exposure variable  Model 1   Model 2   Model 3   Model 4     HR (95% CI)  P-value  HR (95% CI)  P-value  HR (95% CI)  P-value  HR (95% CI)  P-value  Categorized predictors   BNP (pg/mL)    Second tertile α  2.329 (1.265–4.290)  0.007  2.112 (1.146–3.892)  0.017  2.022 (1.086–3.763)  0.026  1.826 (0.978–3.407)  0.059    Third tertile α  6.566 (3.773–11.426)  <0.001  4.776 (2.718–8.394)  <0.001  4.025 (2.233–7.253)  <0.001  3.191 (1.753–5.811)  <0.001   NEP activity (nmol/mL/min)    Second tertile α  0.676 (0.468–0.976)  0.037  0.737 (0.510–1.067)  0.106  0.784 (0.542–1.134)  0.196  0.968 (0.661–1.418)  0.868    Third tertile α  0.642 (0.436–0.946)  0.025  0.694 (0.470–1.023)  0.065  0.723 (0.490–1.066)  0.101  0.889 (0.599–1.320)  0.560   NEP concentration (pg/mL)    Second tertile α  0.828 (0.564–1.214)  0.333  0.817 (0.557–1.199)  0.302  0.826 (0.563–1.212)  0.328  0.787 (0.533–1.164)  0.231    Third tertile α  1.010 (0.691–1.475)  0.960  1.054 (0.722–1.538)  0.787  1.069 (0.732–1.561)  0.731  1.037 (0.706–1.523)  0.852  Continuous predictors   logBNP (pg/mL)  4.900 (3.547–6.769)  <0.001  3.714 (2.639–5.226)  <0.001  3.188 (2.215–4.588)  <0.001  2.665 (1.818–3.907)  <0.001   logNEP activity (nmol/mL/min)  0.259 (0.109–0.617)  0.002  0.390 (0.167–0.909)  0.029  0.423 (0.184–0.970)  0.042  0.611 (0.262–1.424)  0.254   logNEP concentration (pg/mL)  0.736 (0.359–1.508)  0.402  0.845 (0.420–1.701)  0.638  0.787 (0.387–1.601)  0.509  0.652 (0.318–1.335)  0.242  Exposure variable  Model 1   Model 2   Model 3   Model 4     HR (95% CI)  P-value  HR (95% CI)  P-value  HR (95% CI)  P-value  HR (95% CI)  P-value  Categorized predictors   BNP (pg/mL)    Second tertile α  2.329 (1.265–4.290)  0.007  2.112 (1.146–3.892)  0.017  2.022 (1.086–3.763)  0.026  1.826 (0.978–3.407)  0.059    Third tertile α  6.566 (3.773–11.426)  <0.001  4.776 (2.718–8.394)  <0.001  4.025 (2.233–7.253)  <0.001  3.191 (1.753–5.811)  <0.001   NEP activity (nmol/mL/min)    Second tertile α  0.676 (0.468–0.976)  0.037  0.737 (0.510–1.067)  0.106  0.784 (0.542–1.134)  0.196  0.968 (0.661–1.418)  0.868    Third tertile α  0.642 (0.436–0.946)  0.025  0.694 (0.470–1.023)  0.065  0.723 (0.490–1.066)  0.101  0.889 (0.599–1.320)  0.560   NEP concentration (pg/mL)    Second tertile α  0.828 (0.564–1.214)  0.333  0.817 (0.557–1.199)  0.302  0.826 (0.563–1.212)  0.328  0.787 (0.533–1.164)  0.231    Third tertile α  1.010 (0.691–1.475)  0.960  1.054 (0.722–1.538)  0.787  1.069 (0.732–1.561)  0.731  1.037 (0.706–1.523)  0.852  Continuous predictors   logBNP (pg/mL)  4.900 (3.547–6.769)  <0.001  3.714 (2.639–5.226)  <0.001  3.188 (2.215–4.588)  <0.001  2.665 (1.818–3.907)  <0.001   logNEP activity (nmol/mL/min)  0.259 (0.109–0.617)  0.002  0.390 (0.167–0.909)  0.029  0.423 (0.184–0.970)  0.042  0.611 (0.262–1.424)  0.254   logNEP concentration (pg/mL)  0.736 (0.359–1.508)  0.402  0.845 (0.420–1.701)  0.638  0.787 (0.387–1.601)  0.509  0.652 (0.318–1.335)  0.242  Model 1 is the univariate analysis. Model 2 is adjusted for eGFR. Model 3 is adjusted for eGFR, age and sex. Model 4 is adjusted for eGFR, age, sex, diabetes mellitus, current smoker, total cholesterol, prevalent cardiovascular events, BMI and BPsys. HR, hazard ratio; CI, confidence interval; NEP, neprilysin; BNP, brain natriuretic peptide; BPsys, systolic blood pressure. α reference is the first tertile. P-value below 0.05 are presented in bold letters. Table 3. Cox models (endpoint: incident atherosclerotic cardiovascular events) Exposure variable  Model 1   Model 2   Model 3   Model 4     HR (95% CI)  P-value  HR (95% CI)  P-value  HR (95% CI)  P-value  HR (95% CI)  P-value  Categorized predictors   BNP (pg/mL)    Second tertile α  2.329 (1.265–4.290)  0.007  2.112 (1.146–3.892)  0.017  2.022 (1.086–3.763)  0.026  1.826 (0.978–3.407)  0.059    Third tertile α  6.566 (3.773–11.426)  <0.001  4.776 (2.718–8.394)  <0.001  4.025 (2.233–7.253)  <0.001  3.191 (1.753–5.811)  <0.001   NEP activity (nmol/mL/min)    Second tertile α  0.676 (0.468–0.976)  0.037  0.737 (0.510–1.067)  0.106  0.784 (0.542–1.134)  0.196  0.968 (0.661–1.418)  0.868    Third tertile α  0.642 (0.436–0.946)  0.025  0.694 (0.470–1.023)  0.065  0.723 (0.490–1.066)  0.101  0.889 (0.599–1.320)  0.560   NEP concentration (pg/mL)    Second tertile α  0.828 (0.564–1.214)  0.333  0.817 (0.557–1.199)  0.302  0.826 (0.563–1.212)  0.328  0.787 (0.533–1.164)  0.231    Third tertile α  1.010 (0.691–1.475)  0.960  1.054 (0.722–1.538)  0.787  1.069 (0.732–1.561)  0.731  1.037 (0.706–1.523)  0.852  Continuous predictors   logBNP (pg/mL)  4.900 (3.547–6.769)  <0.001  3.714 (2.639–5.226)  <0.001  3.188 (2.215–4.588)  <0.001  2.665 (1.818–3.907)  <0.001   logNEP activity (nmol/mL/min)  0.259 (0.109–0.617)  0.002  0.390 (0.167–0.909)  0.029  0.423 (0.184–0.970)  0.042  0.611 (0.262–1.424)  0.254   logNEP concentration (pg/mL)  0.736 (0.359–1.508)  0.402  0.845 (0.420–1.701)  0.638  0.787 (0.387–1.601)  0.509  0.652 (0.318–1.335)  0.242  Exposure variable  Model 1   Model 2   Model 3   Model 4     HR (95% CI)  P-value  HR (95% CI)  P-value  HR (95% CI)  P-value  HR (95% CI)  P-value  Categorized predictors   BNP (pg/mL)    Second tertile α  2.329 (1.265–4.290)  0.007  2.112 (1.146–3.892)  0.017  2.022 (1.086–3.763)  0.026  1.826 (0.978–3.407)  0.059    Third tertile α  6.566 (3.773–11.426)  <0.001  4.776 (2.718–8.394)  <0.001  4.025 (2.233–7.253)  <0.001  3.191 (1.753–5.811)  <0.001   NEP activity (nmol/mL/min)    Second tertile α  0.676 (0.468–0.976)  0.037  0.737 (0.510–1.067)  0.106  0.784 (0.542–1.134)  0.196  0.968 (0.661–1.418)  0.868    Third tertile α  0.642 (0.436–0.946)  0.025  0.694 (0.470–1.023)  0.065  0.723 (0.490–1.066)  0.101  0.889 (0.599–1.320)  0.560   NEP concentration (pg/mL)    Second tertile α  0.828 (0.564–1.214)  0.333  0.817 (0.557–1.199)  0.302  0.826 (0.563–1.212)  0.328  0.787 (0.533–1.164)  0.231    Third tertile α  1.010 (0.691–1.475)  0.960  1.054 (0.722–1.538)  0.787  1.069 (0.732–1.561)  0.731  1.037 (0.706–1.523)  0.852  Continuous predictors   logBNP (pg/mL)  4.900 (3.547–6.769)  <0.001  3.714 (2.639–5.226)  <0.001  3.188 (2.215–4.588)  <0.001  2.665 (1.818–3.907)  <0.001   logNEP activity (nmol/mL/min)  0.259 (0.109–0.617)  0.002  0.390 (0.167–0.909)  0.029  0.423 (0.184–0.970)  0.042  0.611 (0.262–1.424)  0.254   logNEP concentration (pg/mL)  0.736 (0.359–1.508)  0.402  0.845 (0.420–1.701)  0.638  0.787 (0.387–1.601)  0.509  0.652 (0.318–1.335)  0.242  Model 1 is the univariate analysis. Model 2 is adjusted for eGFR. Model 3 is adjusted for eGFR, age and sex. Model 4 is adjusted for eGFR, age, sex, diabetes mellitus, current smoker, total cholesterol, prevalent cardiovascular events, BMI and BPsys. HR, hazard ratio; CI, confidence interval; NEP, neprilysin; BNP, brain natriuretic peptide; BPsys, systolic blood pressure. α reference is the first tertile. P-value below 0.05 are presented in bold letters. In contrast, both tertiles of neprilysin concentration and log transformed neprilysin concentration did not predict hospital admission for acute decompensated heart failure or incident atherosclerotic cardiovascular events in Cox regression analyses (Tables 2 and 3). In explanatory post hoc analyses, we further adjusted for log transformed albuminuria, which did not affect the study results. Additionally we performed posthoc analyses among participants with eGFR ≤30 mL/min/1.73 m2. In these subgroup analyses, BNP—when considered as a continuous variable—was associated with hospitalization for acute decompensated heart failure and with incident atherosclerotic cardiovascular events in Cox regression analyses before and after adjustment for confounders (Supplementary data, Table S2). Similarly, BNP predicted a combined endpoint of hospitalization for acute decompensated heart failure and death of any cause. While neprilysin activity tended to be associated with all three cardiovascular endpoints, even after adjustment for renal function and traditional cardiovascular risk factors, neprilysin concentration was not associated with any of these cardiovascular endpoints. In a subsequent ROC curve analysis, BNP had the largest area under the curve for predicting hospitalization for acute decompensated heart failure (Figure 4) and incident atherosclerotic cardiovascular events (Figure 5) that occurred within 4 years after study initiation. FIGURE 4: View largeDownload slide ROC curve analyses for hospitalization for acute decompensated heart failure: AUC for BNP = 0.840; AUC for NEP activity = 0.410; AUC for NEP concentration = 0.546. FIGURE 4: View largeDownload slide ROC curve analyses for hospitalization for acute decompensated heart failure: AUC for BNP = 0.840; AUC for NEP activity = 0.410; AUC for NEP concentration = 0.546. FIGURE 5: View largeDownload slide ROC curve analyses for incident atherosclerotic cardiovascular events AUC for BNP = 0.773; AUC NEP activity = 0.480; AUC for NEP concentration = 0.460. FIGURE 5: View largeDownload slide ROC curve analyses for incident atherosclerotic cardiovascular events AUC for BNP = 0.773; AUC NEP activity = 0.480; AUC for NEP concentration = 0.460. In further post hoc analyses, we studied whether BNP, neprilysin activity or neprilysin concentration predict CKD progression. Seventy-nine participants had either initiation of renal replacement therapy or halving of eGFR during follow-up. After stratifying patients into tertiles for univariate Kaplan–Meier analyses, only BNP predicted renal events, while neprilysin activity or neprilysin concentration did not (Supplementary data, Figure S1). In further Cox regression analyses, BNP—both categorized and as a continuous variable—was significantly associated with CKD progression, even after adjustment for confounders. In contrast, neither activity nor concentration of neprilysin predicted renal events (Supplementary data, Table S1). In the final posthoc analyses, we divided participants by their neprilysin concentration (above or below median) cross-classified by their BNP level (above or below median; Supplementary data, Table S3 and Figure S2). In univariate Kaplan–Meier analyses and multivariate Cox regression analyses, participants with high BNP had worse cardiovascular outcome than patients with low BNP, whether or not their neprilysin concentration was high. Similarly, high BNP with either high or low neprilysin concentration predicted CKD progression in univariate Kaplan–Meier analyses but not in fully adjusted Cox regression analyses. DISCUSSION CKD patients have a high prevalence of chronic heart failure and require frequent hospital admission for cardiac decompensation [17]. Implementation of cardiac biomarker analyses into the clinical routine may allow us to identify CKD patients with a high risk for future heart failure decompensation and subsequently to tailor cardioprotective treatment [5]. Up until now, clinical biomarker research has focused upon measurements of natriuretic peptides, which were argued to have potential shortcomings among CKD patients [18, 19]. Therefore, alternative cardiac biomarkers are of interest for physicians who care for CKD patients. Based on recent epidemiological data from non-nephrological cohorts, plasma neprilysin was suggested as such an emerging biomarker, since its plasma concentration predicted adverse outcome in patients with stable heart failure with reduced ejection fraction [13] and with acute decompensated heart failure [16]. Compared with other innovative cardiac biomarkers, neprilysin may be of particular interest for risk stratification, since it recently became a target for innovative therapeutic strategies in heart failure patients [20]. In the landmark PARADIGM-HF study, patients who received the neprilysin inhibitor sacubitril in addition to the angiotensin II receptor blocker valsartan had a 20% reduction in death from cardiovascular causes and a 21% reduction of hospitalization for heart failure compared with standard therapy with the angiotensin-converting enzyme inhibitor enalapril [20]. However, both epidemiologic studies with plasma neprilysin as a cardiac biomarker and interventional trials with sacubitril in heart failure have focused upon patients with normal or mildly impaired renal function, while patients with advanced CKD were underrepresented [13] or excluded [20]. This is of particular concern since natriuretic peptides accumulate in advanced CKD in patients with and without overt cardiac disease and since elevated natriuretic peptides may inhibit neprilysin activity [16]. Therefore, in advanced CKD patients with a strong accumulation of natriuretic peptides, less active neprilysin might be available as a pharmacologic target for sacubitril. This may affect prognostic implications of plasma neprilysin concentration measurements and render sacubitril treatment less effective. Admittedly, the knowledge about the inhibitory effect of plasma natriuretic peptides on neprilysin activity largely derives from cross-sectional clinical studies, which cannot rule out the hypothesis that high plasma concentrations of BNP are the consequence rather than the cause of low neprilysin activity. Even though a direct inhibitory effect of BNP on neprilysin activity has been shown in vitro, this evidence is based on data from only 10 patients and awaits confirmation from independent groups. Furthermore, it was largely unknown how applicable these clinical and experimental findings may be to CKD patients. We therefore measured plasma neprilysin activity, neprilysin concentration and BNP among 542 patients with mild to severe CKD in order to analyse their association across the spectrum of non-dialysis CKD and to test how the interaction between BNP and neprilysin activity affects the use of neprilysin concentration as a cardiac outcome predictor in CKD patients. We first confirmed our earlier data that neprilysin activity and neprilysin concentration are not correlated at all [16]. Thus, measuring neprilysin concentration is neither informative for its activity in acute heart failure [16] nor in stable CKD patients, which questions the biological relevance of neprilysin concentration measurements at least in subgroups of patients with heart failure. Second, we confirmed that BNP levels are inversely associated with neprilysin activity and thereby externally validate our previous finding [16] that patients with BNP above a cut-off of 916 pg/mL have very low neprilysin activity. Third, despite the great increase in BNP with advanced CKD, the association of eGFR with neprilysin activity and concentration was negligible. This finding was partly unexpected, since the accumulation of natriuretic peptides in advanced CKD and the inhibition of neprilysin by high plasma levels of natriuretic peptides might have suggested a drop in neprilysin activity in advanced CKD. Apparently other factors may counteract the inhibition of neprilysin activity induced by natriuretic peptides. As expected, high plasma levels of BNP predicted CKD progression, hospitalization for acute decompensated heart failure and incident atherosclerotic cardiovascular events, in accordance with NT-proBNP data from CARE FOR HOMe [8], and with data from other CKD cohorts that analysed natriuretic peptides [21]. However, while earlier studies suggested high neprilysin concentration indicates future heart failure hospitalization [13], it was not predictive in our CKD patients and low rather than high neprilysin activity predicted adverse cardiac outcome. Interestingly, the prognostic role of plasma neprilysin concentration was not confirmed among 144 heart failure with preserved ejection fraction (HFpEF) patients, in whom neprilysin concentration did not predict future hospitalization for heart failure or cardiac death. Unlike in other studies [13, 16] and unlike in CARE FOR HOMe, neprilysin concentration was reported to correlate with NT-proBNP in HFpEF patients [22]. Thus our findings suggest against the implementation of neprilysin measurements into clinical nephrology for the time being. This follows numerous other studies that were unable to replicate epidemiologic data on classical and non-classical cardiovascular biomarkers from the general population among CKD patients. If BNP inhibits neprilysin substantially, then an accumulation of natriuretic peptides with low glomerular function may render neprilysin less of an attractive target for pharmacologic intervention in advanced CKD than among heart failure patients with intact renal function. Admittedly, subgroup analysis from PARADIGM-HF do not point towards less efficacy of sacubitril–valsartan in patients with moderate CKD compared with patients with normal GFR; however, patients with severe CKD were excluded. This is of major importance, as sacubitrilat—an active metabolite of sacubitril—may accumulate in severe CKD [23], along with several substrates of neprilysin, such as natriuretic peptides. Therefore we await final results from the United Kingdom Heart and Renal Protection III study that recruited CKD patients with an eGFR as low as 20 mL/min/1.73 m2 [24]. This study has some limitations. Different assays for measurements of neprilysin concentration are available, which yield strongly discrepant results [25]. We cannot fully rule out that measurements with an alternative neprilysin assay would have yielded better prognostic information; however, we are positive that the lack of standardization supports our critical attitude on measurements of neprilysin concentration in cardiorenal medicine for the present time. Next, neprilysin may degrade numerous other vasoactive peptides beyond natriuretic peptides [26], such as substance P, bradykinins, atrial natriuretic peptide or endothelin-1, which we did not measure in the current study. In our earlier paper we found that BNP-mediated inhibition of neprilysin may also prevent substance P breakdown [16]. Since our major study goal was to analyse neprilysin as a novel cardiac biomarker in CKD patients, we did not measure enzymes other than neprilysin that are involved in the degradation of NP precursors to active BNP. In particular, we did not measure corin, low plasma levels of which predicted cardiovascular events among 1382 patients [27]. We deliberately focused our analyses on stable patients from our outpatient department, which is in line with the seminal paper by Bayes-Genis et al. [13] that first introduced plasma neprilysin as a potential cardiac biomarker. However, Bayes-Genis et al. [28] subsequently expanded their epidemiological studies to patients with acute heart failure, who were recruited into our present study. Additionally, we did not assess New York Heart Association classes at study initiation. As a further limitation, the number of patients who had cardiac events was limited, which may decrease the statistical power of our Cox regression analyses. In conclusion, we first show CKD patients with high BNP levels to have particularly low neprilysin activity. Second, in contrast to earlier heart failure studies, we find that low levels of neprilysin activity rather than high neprilysin concentration predict poor cardiac outcome in CKD patients. These results question the usefulness of neprilysin as a biomarker in cardiorenal medicine and are a warning signal against the uncritical use of neprilysin inhibitors in heart failure patients with advanced CKD. SUPPLEMENTARY DATA Supplementary data are available at ndt online. FUNDING The present work was supported by a grant from Else Kröner-Fresenius-Stiftung. AUTHORS’ CONTRIBUTIONS I.E.E., G.H.H. and N.V. designed the research. I.E.E., L.F., S.S.-M., K.U. and J.-M.L. conducted the research. I.E.E., G.H.H., H.N. and N.V. analysed the data and performed the statistical analysis. I.E.E., G.H.H. and N.V. wrote the paper. CONFLICT OF INTEREST STATEMENT None declared. REFERENCES 1 Matsushita K, van der Velde M, Astor BC et al.   Association of estimated glomerular filtration rate and albuminuria with all-cause and cardiovascular mortality in general population cohorts: a collaborative meta-analysis. Lancet  2010; 375: 2073– 2081 Google Scholar CrossRef Search ADS PubMed  2 Hemmelgarn BR, Manns BJ, Lloyd A et al.   Relation between kidney function, proteinuria, and adverse outcomes. JAMA  2010; 303: 423– 429 Google Scholar CrossRef Search ADS PubMed  3 Go AS, Chertow GM, Fan D et al.   Chronic kidney disease and the risks of death, cardiovascular events, and hospitalization. N Engl J Med  2004; 351: 1296– 1305 Google Scholar CrossRef Search ADS PubMed  4 Herzog CA, Asinger RW, Berger AK et al.   Cardiovascular disease in chronic kidney disease. A clinical update from Kidney Disease: Improving Global Outcomes (KDIGO). Kidney Int  2011; 80: 572– 586 Google Scholar CrossRef Search ADS PubMed  5 Fassett RG, Venuthurupalli SK, Gobe GC et al.   Biomarkers in chronic kidney disease: a review. Kidney Int  2011; 80: 806– 821 Google Scholar CrossRef Search ADS PubMed  6 Ponikowski P, Voors AA, Anker SD et al.   2016 ESC guidelines for the diagnosis and treatment of acute and chronic heart failure: the Task Force for the diagnosis and treatment of acute and chronic heart failure of the European Society of Cardiology (ESC) developed with the special contribution of the Heart Failure Association (HFA) of the ESC. Eur Heart J  2016; 37: 2129– 2200 Google Scholar CrossRef Search ADS PubMed  7 Mishra RK, Li Y, Ricardo AC et al.   Association of N-terminal pro-B-type natriuretic peptide with left ventricular structure and function in chronic kidney disease (from the Chronic Renal Insufficiency Cohort [CRIC]). Am J Cardiol  2013; 111: 432– 438 Google Scholar CrossRef Search ADS PubMed  8 Untersteller K, Girerd N, Duarte K et al.   NT-proBNP and echocardiographic parameters for prediction of cardiovascular outcomes in patients with CKD stages G2–G4. Clin J Am Soc Nephrol  2016; 11: 1978– 1988 Google Scholar CrossRef Search ADS PubMed  9 DeFilippi CR, Fink JC, Nass CM et al.   N-terminal pro-B-type natriuretic peptide for predicting coronary disease and left ventricular hypertrophy in asymptomatic CKD not requiring dialysis. Am J Kidney Dis  2005; 46: 35– 44 Google Scholar CrossRef Search ADS PubMed  10 Scheven L, de Jong PE, Hillege HL et al.   High-sensitive troponin T and N-terminal pro-B type natriuretic peptide are associated with cardiovascular events despite the cross-sectional association with albuminuria and glomerular filtration rate. Eur Heart J  2012; 33: 2272– 2281 Google Scholar CrossRef Search ADS PubMed  11 Sakuma M, Nakamura M, Tanaka F et al.   Plasma B-type natriuretic peptide level and cardiovascular events in chronic kidney disease in a community-based population. Circ J  2010; 74: 792– 797 Google Scholar CrossRef Search ADS PubMed  12 Vickery S, Price CP, John RI et al.   B-type natriuretic peptide (BNP) and amino-terminal proBNP in patients with CKD: relationship to renal function and left ventricular hypertrophy. Am J Kidney Dis  2005; 46: 610– 620 Google Scholar CrossRef Search ADS PubMed  13 Bayes-Genis A, Barallat J, Galan A et al.   Soluble neprilysin is predictive of cardiovascular death and heart failure hospitalization in heart failure patients. J Am Coll Cardiol  2015; 65: 657– 665 Google Scholar CrossRef Search ADS PubMed  14 Vodovar N, Paquet C, Mebazaa A et al.   Neprilysin, cardiovascular, and Alzheimer’s diseases: the therapeutic split? Eur Heart J  2015; 36: 902– 905 Google Scholar CrossRef Search ADS PubMed  15 Bohm M, Young R, Jhund PS et al.   Systolic blood pressure, cardiovascular outcomes and efficacy and safety of sacubitril/valsartan (LCZ696) in patients with chronic heart failure and reduced ejection fraction: results from PARADIGM-HF. Eur Heart J  2017; 38: 1132– 1143 Google Scholar CrossRef Search ADS PubMed  16 Vodovar N, Seronde MF, Laribi S et al.   Elevated plasma B-type natriuretic peptide concentrations directly inhibit circulating neprilysin activity in heart failure. JACC Heart Fail  2015; 3: 629– 636 Google Scholar CrossRef Search ADS PubMed  17 He J, Shlipak M, Anderson A et al.   Risk factors for heart failure in patients with chronic kidney disease: the CRIC (Chronic Renal Insufficiency Cohort) study. J Am Heart Assoc  2017; 6: e005336 Google Scholar CrossRef Search ADS PubMed  18 Bansal N, Hyre Anderson A, Yang W et al.   High-sensitivity troponin T and N-terminal pro-B-type natriuretic peptide (NT-proBNP) and risk of incident heart failure in patients with CKD: the Chronic Renal Insufficiency Cohort (CRIC) Study. J Am Soc Nephrol  2015; 26: 946– 956 Google Scholar CrossRef Search ADS PubMed  19 Matsushita K, Sang Y, Ballew SH et al.   Cardiac and kidney markers for cardiovascular prediction in individuals with chronic kidney disease: the Atherosclerosis Risk in Communities study. Arterioscler Thromb Vasc Biol  2014; 34: 1770– 1777 Google Scholar CrossRef Search ADS PubMed  20 McMurray JJ, Packer M, Desai AS et al.   Angiotensin–neprilysin inhibition versus enalapril in heart failure. N Engl J Med  2014; 371: 993– 1004 Google Scholar CrossRef Search ADS PubMed  21 Colbert G, Jain N, de Lemos JA et al.   Utility of traditional circulating and imaging-based cardiac biomarkers in patients with predialysis CKD. Clin J Am Soc Nephrol  2015; 10: 515– 529 Google Scholar CrossRef Search ADS PubMed  22 Goliasch G, Pavo N, Zotter-Tufaro C et al.   Soluble neprilysin does not correlate with outcome in heart failure with preserved ejection fraction. Eur J Heart Fail  2016; 18: 89– 93 Google Scholar CrossRef Search ADS PubMed  23 Ayalasomayajula SP, Langenickel TH, Jordaan P et al.   Effect of renal function on the pharmacokinetics of LCZ696 (sacubitril/valsartan), an angiotensin receptor neprilysin inhibitor. Eur J Clin Pharmacol  2016; 72: 1065– 1073 Google Scholar CrossRef Search ADS PubMed  24 UK HARP-III Collaborative Group. Randomized multicentre pilot study of sacubitril/valsartan versus irbesartan in patients with chronic kidney disease: United Kingdom Heart and Renal Protection (HARP)-III—rationale, trial design and baseline data. Nephrol Dial Transplant  2017; 32: 2043– 2051 PubMed  25 Bayes-Genis A, Barallat J, Richards AM. A test in context: neprilysin: function, inhibition, and biomarker. J Am Coll Cardiol  2016; 68: 639– 653 Google Scholar CrossRef Search ADS PubMed  26 Mangiafico S, Costello-Boerrigter LC, Andersen IA et al.   Neutral endopeptidase inhibition and the natriuretic peptide system: an evolving strategy in cardiovascular therapeutics. Eur Heart J  2013; 34: 886– 893 Google Scholar CrossRef Search ADS PubMed  27 Zhou X, Chen J, Zhang Q et al.   Prognostic value of plasma soluble corin in patients with acute myocardial infarction. J Am Coll Cardiol  2016; 67: 2008– 2014 Google Scholar CrossRef Search ADS PubMed  28 Bayes-Genis A, Barallat J, Pascual-Figal D et al.   Prognostic value and kinetics of soluble neprilysin in acute heart failure: a pilot study. JACC Heart Fail  2015; 3: 641– 644 Google Scholar CrossRef Search ADS PubMed  © The Author(s) 2018. Published by Oxford University Press on behalf of ERA-EDTA. All rights reserved. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

Nephrology Dialysis TransplantationOxford University Press

Published: Apr 9, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off