Pregnancy-associated plasma protein-A predicts survival in end-stage renal disease—confounding and modifying effects of cardiovascular disease, body composition and inflammation

Pregnancy-associated plasma protein-A predicts survival in end-stage renal disease—confounding... ABSTRACT Background High pregnancy-associated plasma protein-A (PAPP-A) levels are linked to atherosclerosis and associate with increased mortality in prevalent dialysis patients. We investigated associations of PAPP-A, measured at dialysis initiation, with cardiovascular disease (CVD), CVD risk factors and mortality in incident dialysis patients, and explored if body composition and inflammation modulated these associations. Methods Baseline plasma PAPP-A levels, inflammation biomarkers and body composition, using dual-energy X-ray absorptiometry, were measured in 286 incident dialysis patients. Primary outcome was survival during 60 months follow-up. Quantile (median) regression was used for cross-sectional analysis and Kaplan–Meier diagrams and Cox proportional hazards regression for survival analysis. Results In cross-sectional analysis adjusted for age and sex, PAPP-A levels were associated with lean tissue index (LTI) and high-sensitivity C-reactive protein (hsCRP) but not with fat tissue index (FTI) or history of CVD. In a model also including diabetes mellitus (DM), the association with LTI did not remain statistically significant. When adjusted for cardiovascular risk factors and body composition, higher PAPP-A levels showed a moderate but significant association [hazard ratio (HR) = 1.2, 95% confidence interval (CI): 1–1.4, P = 0.04] with mortality. When also including hsCRP the association was attenuated (HR = 1.2, 95% CI: 0.99–1.4, P = 0.06). In survival analysis, interactions with PAPP-A on the multiplicative scale were found for hsCRP (HR = 1.6, 95% CI: 1.2–2.2, P = 0.004) and DM (HR = 1.6, 95% CI: 1.1–2.2, P = 0.01) and with DM and FTI on the additive scale. Conclusions Higher PAPP-A levels are associated with worse survival in incident dialysis patients following adjustment for established cardiovascular risk factors and body composition indices, but not clearly so when adjusted for hsCRP. Inflammation, body composition (FTI) and DM were found to be potential effect modifiers for the observed moderate association of PAPP-A with survival. biomarkers, cardiovascular disease, dialysis, end-stage renal disease, pregnancy-associated plasma protein-A INTRODUCTION Pregnancy-associated plasma protein-A (PAPP-A) plays a role in the development and ageing processes by modulating growth hormone (GH) effects on lipid, glucose and protein metabolism, as well as on cardiovascular function in adults [1, 2]. Both excess and deficiency of GH are associated with cardiovascular disease (CVD) and GH and its main mediator, insulin-like growth factor 1 (IGF-1), have been implicated as having both pro- and antiatherogenic effects in experimental models [2, 3]. Both high and low circulating concentrations of IGF-1 are associated with increased mortality risk in the general population [4] and in patients with end-stage renal disease (ESRD) low IGF-1 predicts worse survival [5]. Availability of IGF-1 at receptor level is influenced by concentrations of its multiple binding proteins (insulin-like growth factor binding proteins, IGFBPs) in plasma and by membrane-bound metalloproteinase PAPP-A [6], which cleaves IGFBP-4 bound to IGF-1, making IGF-1 available to its receptor. PAPP-A is therefore considered an important regulator of IGF-1 action [7]. GH levels are associated with body composition [8] and PAPP-A is expressed relatively more in visceral than in subcutaneous fat. Conversely, in PAPP-A knock-out mice mesenteric fat depots are reduced [9]. This indicates an interdependency between fat tissue and PAPP-A. A negative association between PAPP-A and body mass index (BMI) has been described in ESRD and a link to protein-energy wasting (PEW) suggested [10]. Elevated plasma PAPP-A concentration is associated with the extent of coronary artery disease and has been evaluated as a prognostic marker in chronic stable angina pectoris and acute coronary syndromes, where elevated plasma levels predict increased risk of death [11]. Interestingly, an interaction between PAPP-A and the anti-inflammatory cytokine IL-10 levels has been observed in this context, indicating that elevated PAPP-A is detrimental only when IL-10 levels are low. Positive associations between PAPP-A levels and carotid and peripheral artery disease have also been reported [11, 12]. In prevalent haemodialysis (HD) patients, PAPP-A levels are elevated compared with controls, predict all-cause and cardiovascular mortality [7, 10, 13, 14] and are associated with cardiac troponins [15]. Notably, in adjusted models an association with sudden death, stroke and death due to infection was found, but there was no statistically significant association with risk of myocardial infarction [7, 10]. PAPP-A has not been evaluated as a prognostic marker in incident dialysis patients. Due to the high early mortality after initiation of dialysis [16], studies on prevalent patients are not representative for the incident dialysis population. The aim of the present study was to investigate PAPP-A as a risk marker for mortality and CVD in incident dialysis patients, highlighting body composition as a potential confounder and effect modifier. We hypothesized that PAPP-A levels are associated with survival in incident dialysis patients, that this association is confounded by body composition as represented by lean tissue index (LTI) and fat tissue index (FTI), and further, that there would be an interaction between PAPP-A and fat tissue and inflammation in relation to survival. We also report on the associations between PAPP-A and CVD risk factors in this population. MATERIALS AND METHODS Study design We measured PAPP-A levels in an ongoing prospective cohort originally designed to investigate malnutrition, PEW, inflammation and atherosclerosis in incident dialysis patients, described previously [17]. Exclusion criteria were age  <18 or  >75 years, signs of infection and unwillingness to participate. During 1993 to 2014, 552 patients were recruited to the cohort. Among these, 286 patients had plasma samples available in 2015 and were included in the present study, having started dialysis treatment in years 2000–14. Participants were followed until death, renal transplantation or until 1 March 2016. The Ethics Committee of Karolinska Institutet, Stockholm, Sweden, approved the study protocol. Informed consent was obtained from all participants. The study comprised a cross-sectional analysis of PAPP-A levels in relation to cardiovascular risk factors, inflammation and body composition and a prospective analysis of mortality, where follow-up was truncated at 60 months. Blood sampling and biochemical analysis Blood samples were drawn after overnight fast and analysed immediately or centrifuged and stored at −70°C. PAPP-A was assayed using ELISA (R&D Systems, Minneapolis, MN, USA). High-sensitivity C-reactive protein (hsCRP) was measured using nephelometry and tumour necrosis factor (TNF) was analysed at an Immulite system (Diagnostic Products Corp., Los Angeles, CA, USA). Routine biochemical analyses were performed at the Department of Clinical Chemistry at Karolinska University Hospital, Huddinge. Glomerular filtration rate (GFR) was measured as the mean of renal urea and creatinine clearances during a 24-h urine collection, which is thought to provide a reliable estimate of GFR in patients starting on dialysis. Body composition, anthropometry and nutritional status Body composition was measured using dual-energy X-ray absorptiometry (DEXA) (Lunar Corp., Madison, WI, USA), within 7 days of inclusion and after HD if already started. LTI and FTI were calculated in kilograms per meters squared [18, 19]. Handgrip strength (HGS) was measured in the nonfistula or dominant hand using a Harpenden Handgrip Dynamometer (Yamar, Jackson, MI, USA), repeated three times and the highest value noted. HGS was then normalized within each sex category by dividing the value with mean values from individuals randomly selected by Statistics Sweden (a government agency) from the Stockholm region to constitute a control population. Nutritional status was estimated by subjective global assessment (SGA) and categorized into normal versus not normal nutritional status, the latter encompassing signs of mild to severe malnutrition [20]. Definition of CVD Presence of CVD at baseline was determined through review of medical history and included any previous occurrence of coronary artery disease (myocardial infarction, angina pectoris, coronary artery bypass graft or percutaneous transluminal coronary angioplasty), cerebrovascular disease (ischaemic or haemorrhagic stroke or transitory ischaemic attack), peripheral vascular disease (in extremities, carotides or renal arteries), heart failure or aortic aneurysm. Statistical methods For descriptive statistics median and interquartile range or percentages were used as appropriate. Follow-up time was truncated at 60 months and interval variables were standardized in cross-sectional and survival analysis. Numbers were rounded at two significant digits for ratios and three for descriptive data, trailing zeroes omitted. Quantile (median) regression was utilized for cross-sectional associations and Cox regression was used for survival analysis with censoring at renal transplantation and end of follow-up. Confounders included in multivariable models were selected a priori based on known associations to survival or (for cross-sectional analysis) CVD and PAPP-A levels. Variables were then included based on the specific hypotheses presented. Kaplan–Meier curves were used to visualize potential (unadjusted) interactions between PAPP-A and body composition indices (LTI, FBI) and inflammation (hsCRP). Variables were dichotomized at median for this analysis. These interactions were then tested in Cox models. Cox model assumptions were tested using Schoenfeld residuals plot for proportional hazards and influence diagnostics for identifying influential data points. Testing of normal distribution was not done, since assumption of normality of distributions does not apply for the methods used in the study. All tests were two-tailed and a P < 0.05 considered statistically significant. The statistical analyses were performed in R v 3.3.1 [21] using ‘survival’ [22] and ‘quantreg’ [23] packages. RESULTS Baseline characteristics and outcomes Baseline characteristics are presented in Table 1. DEXA was performed in 213 patients. Compared with a similar study in prevalent HD patients [7], participants in the current study were somewhat younger (mean age 55 versus 63 years) and with a higher proportion of males (62% versus 54%). The frequency of diabetes mellitus (DM) was similar (30% versus 33%) but history of CVD less common in the present study (37% versus 61%). Causes of primary renal disease included diabetic nephropathy (30%), hypertensive renal disease (7%), systemic inflammatory disease (2%), primary glomerular disease (18%), hereditary disease (15%), and other (4%) or unknown (21%) aetiology. Medications used included erythropoesis-stimulating agents (90%), diuretics (86%), phosphate binders (84%), active vitamin D (81%), angiotensin-converting enzyme inhibitors/angiotensin receptor blockers (69%), beta blockers (68%), vitamin B1 (63%), calcium channel blockers (52%), intravenous iron (43%), statins (36%), insulin (26%), amino acid substitution (24%), allopurinol (17%) and peroral iron (15%). Table 1 Demographic, clinical and laboratory characteristics of 286 incident dialysis patients Variable Median (IQR) or % Age, years 57 (47–65) Male 62 DM 30 CVD 37 Smoker, current or former 48 Malnutrition, SGA >1 26 eGFR, mL/min/1.73 m2 6.4 (5–8) BMI, kg/m2 24.6 (22–28.3) Lean body mass, % 69.1 (63.4–77) Fat body mass, % 30.9 (23–36.6) FTI, kg/m2 7.16 (5.02–9.81) LTI, kg/m2 16.3 (14.8–18.2) Height, cm 172 (165–179) Cholesterol, mmol/L 4.4 (3.7–5.2) Triglycerid, mmol/L 1.6 (1.1–2.2) High-density lipoprotein, mmol/L 1.3 (1–1.6) Low-density lipoprotein, mmol/L 2.25 (1.6–2.97) Creatinine, µmol/L 733 (576–910) Phosphate, mmol/L 1.9 (1.6–2.32) Calcium, mmol/L 2.39 (2.21–2.56) Parathyroid hormone, ng/L 232 (120–370) hsCRP, mg/L 3.9 (1.3–10.8) s-Albumin, g/L 34 (30–36.2) TNF, pg/mL 14.4 (11.3–17.7) PAPP-A, ng/mL 0.139 (0.035–0.275) Triceps skinfold, mm 14 (10–18) HGS, % 29 (22.2–37.5) Variable Median (IQR) or % Age, years 57 (47–65) Male 62 DM 30 CVD 37 Smoker, current or former 48 Malnutrition, SGA >1 26 eGFR, mL/min/1.73 m2 6.4 (5–8) BMI, kg/m2 24.6 (22–28.3) Lean body mass, % 69.1 (63.4–77) Fat body mass, % 30.9 (23–36.6) FTI, kg/m2 7.16 (5.02–9.81) LTI, kg/m2 16.3 (14.8–18.2) Height, cm 172 (165–179) Cholesterol, mmol/L 4.4 (3.7–5.2) Triglycerid, mmol/L 1.6 (1.1–2.2) High-density lipoprotein, mmol/L 1.3 (1–1.6) Low-density lipoprotein, mmol/L 2.25 (1.6–2.97) Creatinine, µmol/L 733 (576–910) Phosphate, mmol/L 1.9 (1.6–2.32) Calcium, mmol/L 2.39 (2.21–2.56) Parathyroid hormone, ng/L 232 (120–370) hsCRP, mg/L 3.9 (1.3–10.8) s-Albumin, g/L 34 (30–36.2) TNF, pg/mL 14.4 (11.3–17.7) PAPP-A, ng/mL 0.139 (0.035–0.275) Triceps skinfold, mm 14 (10–18) HGS, % 29 (22.2–37.5) Table 1 Demographic, clinical and laboratory characteristics of 286 incident dialysis patients Variable Median (IQR) or % Age, years 57 (47–65) Male 62 DM 30 CVD 37 Smoker, current or former 48 Malnutrition, SGA >1 26 eGFR, mL/min/1.73 m2 6.4 (5–8) BMI, kg/m2 24.6 (22–28.3) Lean body mass, % 69.1 (63.4–77) Fat body mass, % 30.9 (23–36.6) FTI, kg/m2 7.16 (5.02–9.81) LTI, kg/m2 16.3 (14.8–18.2) Height, cm 172 (165–179) Cholesterol, mmol/L 4.4 (3.7–5.2) Triglycerid, mmol/L 1.6 (1.1–2.2) High-density lipoprotein, mmol/L 1.3 (1–1.6) Low-density lipoprotein, mmol/L 2.25 (1.6–2.97) Creatinine, µmol/L 733 (576–910) Phosphate, mmol/L 1.9 (1.6–2.32) Calcium, mmol/L 2.39 (2.21–2.56) Parathyroid hormone, ng/L 232 (120–370) hsCRP, mg/L 3.9 (1.3–10.8) s-Albumin, g/L 34 (30–36.2) TNF, pg/mL 14.4 (11.3–17.7) PAPP-A, ng/mL 0.139 (0.035–0.275) Triceps skinfold, mm 14 (10–18) HGS, % 29 (22.2–37.5) Variable Median (IQR) or % Age, years 57 (47–65) Male 62 DM 30 CVD 37 Smoker, current or former 48 Malnutrition, SGA >1 26 eGFR, mL/min/1.73 m2 6.4 (5–8) BMI, kg/m2 24.6 (22–28.3) Lean body mass, % 69.1 (63.4–77) Fat body mass, % 30.9 (23–36.6) FTI, kg/m2 7.16 (5.02–9.81) LTI, kg/m2 16.3 (14.8–18.2) Height, cm 172 (165–179) Cholesterol, mmol/L 4.4 (3.7–5.2) Triglycerid, mmol/L 1.6 (1.1–2.2) High-density lipoprotein, mmol/L 1.3 (1–1.6) Low-density lipoprotein, mmol/L 2.25 (1.6–2.97) Creatinine, µmol/L 733 (576–910) Phosphate, mmol/L 1.9 (1.6–2.32) Calcium, mmol/L 2.39 (2.21–2.56) Parathyroid hormone, ng/L 232 (120–370) hsCRP, mg/L 3.9 (1.3–10.8) s-Albumin, g/L 34 (30–36.2) TNF, pg/mL 14.4 (11.3–17.7) PAPP-A, ng/mL 0.139 (0.035–0.275) Triceps skinfold, mm 14 (10–18) HGS, % 29 (22.2–37.5) At 60 months follow-up 86 patients had died, 144 were transplanted and 14 censored due to end of follow-up. Causes of death included cardiovascular causes (37%), infection (15%), malignancy (5%), other (10%) and unknown (35%). Factors associated with PAPP-A in cross-sectional analysis Univariable and multivariable associations between PAPP-A and body composition, nutritional parameters, cardiovascular risk factors and inflammatory markers are presented in Table 2. Among body composition variables, PAPP-A was associated only with LTI. This association was positive and remained when adjusted for age and sex but was not statistically significant when DM was also included. Conversely, HGS had a negative association with PAPP-A when adjusted for age and sex but the association did not remain when further adjusted for DM. Median PAPP-A was also predicted to be higher in DM, although this association was less clear when age and sex were considered. Notably, hsCRP and calcium were positively associated with PAPP-A while associations with s-albumin and TNF were negative in all models. Table 2 Associations between cardiovascular risk markers and plasma PAPP-A levels in 286 incident dialysis patients Variable Model 1 P-value Model 2 P-value Model 3 P-value Age 0.08 (0.073) 0.28 0.14 (0.062)a 0.03 0.095 (0.065)a 0.15 Male 0.078 (0.13) 0.56 0.22 (0.13)b 0.08 0.088 (0.14)b 0.53 DM 0.44 (0.13) 0.001 0.3 (0.16) 0.06 CVD 0.14 (0.14) 0.33 0.025 (0.13) 0.85 0.00073 (0.14) 1.00 Smoker, current or former 0.28 (0.17) 0.09 0.097 (0.16) 0.53 0.16 (0.15) 0.26 eGFR −0.052 (0.066) 0.44 −0.074 (0.06) 0.22 −0.097 (0.057) 0.09 BMI 0.018 (0.057) 0.75 0.02 (0.056) 0.71 −0.015 (0.063) 0.82 Lean body mass 0.12 (0.091) 0.18 0.11 (0.11) 0.33 0.13 (0.093) 0.16 Fat body mass −0.12 (0.091) 0.19 −0.11 (0.11) 0.34 −0.13 (0.094) 0.16 FTI −0.0029 (0.081) 0.97 −0.014 (0.085) 0.87 −0.034 (0.082) 0.68 LTI 0.19 (0.074) 0.010 0.18 (0.083) 0.03 0.15 (0.079) 0.06 Height 0.0000 (0.053) 1.00 −0.059 (0.067) 0.38 −0.014 (0.069) 0.84 Cholesterol 0.11 (0.058) 0.06 0.13 (0.066) 0.05 0.14 (0.062) 0.03 Triglycerides 0.016 (0.062) 0.80 0.017 (0.071) 0.82 0.024 (0.064) 0.71 High-density lipoprotein −0.025 (0.058) 0.66 −0.02 (0.058) 0.74 −0.016 (0.053) 0.75 Low-density lipoprotein 0.1 (0.063) 0.11 0.098 (0.056) 0.08 0.11 (0.063) 0.09 Creatinine 0.062 (0.054) 0.25 0.073 (0.061) 0.23 0.097 (0.065) 0.13 Phosphate 0.012 (0.066) 0.86 0.0037 (0.066) 0.96 0.015 (0.072) 0.84 Calcium 0.15 (0.069) 0.03 0.14 (0.057) 0.01 0.21 (0.062) <0.001 Parathyroid hormone −0.013 (0.084) 0.88 0.0073 (0.072) 0.92 0.033 (0.084) 0.70 hsCRP 0.21 (0.077) 0.006 0.2 (0.075) 0.007 0.2 (0.078) 0.010 s-Albumin −0.2 (0.059) <0.001 −0.22 (0.063) <0.001 −0.2 (0.055) <0.001 TNF −0.22 (0.089) 0.02 −0.25 (0.079) 0.002 −0.27 (0.085) 0.002 HGS −0.055 (0.075) 0.47 −0.19 (0.09) 0.04 −0.13 (0.093) 0.16 Malnutrition, SGA >1 −0.074 (0.16) 0.64 −0.13 (0.14) 0.35 −0.13 (0.15) 0.40 Triceps skinfold −0.025 (0.079) 0.75 0.035 (0.079) 0.65 0.026 (0.078) 0.74 Variable Model 1 P-value Model 2 P-value Model 3 P-value Age 0.08 (0.073) 0.28 0.14 (0.062)a 0.03 0.095 (0.065)a 0.15 Male 0.078 (0.13) 0.56 0.22 (0.13)b 0.08 0.088 (0.14)b 0.53 DM 0.44 (0.13) 0.001 0.3 (0.16) 0.06 CVD 0.14 (0.14) 0.33 0.025 (0.13) 0.85 0.00073 (0.14) 1.00 Smoker, current or former 0.28 (0.17) 0.09 0.097 (0.16) 0.53 0.16 (0.15) 0.26 eGFR −0.052 (0.066) 0.44 −0.074 (0.06) 0.22 −0.097 (0.057) 0.09 BMI 0.018 (0.057) 0.75 0.02 (0.056) 0.71 −0.015 (0.063) 0.82 Lean body mass 0.12 (0.091) 0.18 0.11 (0.11) 0.33 0.13 (0.093) 0.16 Fat body mass −0.12 (0.091) 0.19 −0.11 (0.11) 0.34 −0.13 (0.094) 0.16 FTI −0.0029 (0.081) 0.97 −0.014 (0.085) 0.87 −0.034 (0.082) 0.68 LTI 0.19 (0.074) 0.010 0.18 (0.083) 0.03 0.15 (0.079) 0.06 Height 0.0000 (0.053) 1.00 −0.059 (0.067) 0.38 −0.014 (0.069) 0.84 Cholesterol 0.11 (0.058) 0.06 0.13 (0.066) 0.05 0.14 (0.062) 0.03 Triglycerides 0.016 (0.062) 0.80 0.017 (0.071) 0.82 0.024 (0.064) 0.71 High-density lipoprotein −0.025 (0.058) 0.66 −0.02 (0.058) 0.74 −0.016 (0.053) 0.75 Low-density lipoprotein 0.1 (0.063) 0.11 0.098 (0.056) 0.08 0.11 (0.063) 0.09 Creatinine 0.062 (0.054) 0.25 0.073 (0.061) 0.23 0.097 (0.065) 0.13 Phosphate 0.012 (0.066) 0.86 0.0037 (0.066) 0.96 0.015 (0.072) 0.84 Calcium 0.15 (0.069) 0.03 0.14 (0.057) 0.01 0.21 (0.062) <0.001 Parathyroid hormone −0.013 (0.084) 0.88 0.0073 (0.072) 0.92 0.033 (0.084) 0.70 hsCRP 0.21 (0.077) 0.006 0.2 (0.075) 0.007 0.2 (0.078) 0.010 s-Albumin −0.2 (0.059) <0.001 −0.22 (0.063) <0.001 −0.2 (0.055) <0.001 TNF −0.22 (0.089) 0.02 −0.25 (0.079) 0.002 −0.27 (0.085) 0.002 HGS −0.055 (0.075) 0.47 −0.19 (0.09) 0.04 −0.13 (0.093) 0.16 Malnutrition, SGA >1 −0.074 (0.16) 0.64 −0.13 (0.14) 0.35 −0.13 (0.15) 0.40 Triceps skinfold −0.025 (0.079) 0.75 0.035 (0.079) 0.65 0.026 (0.078) 0.74 Median regression coefficients with standard errors for cardiovascular risk markers on plasma PAPP-A levels. Model 1 is univariable; Model 2 includes age and sex; Model 3 also includes diabetes status. In models marked a and b no adjustment was made for age and sex, respectively. Interval variables are standardized and centered to the mean. Table 2 Associations between cardiovascular risk markers and plasma PAPP-A levels in 286 incident dialysis patients Variable Model 1 P-value Model 2 P-value Model 3 P-value Age 0.08 (0.073) 0.28 0.14 (0.062)a 0.03 0.095 (0.065)a 0.15 Male 0.078 (0.13) 0.56 0.22 (0.13)b 0.08 0.088 (0.14)b 0.53 DM 0.44 (0.13) 0.001 0.3 (0.16) 0.06 CVD 0.14 (0.14) 0.33 0.025 (0.13) 0.85 0.00073 (0.14) 1.00 Smoker, current or former 0.28 (0.17) 0.09 0.097 (0.16) 0.53 0.16 (0.15) 0.26 eGFR −0.052 (0.066) 0.44 −0.074 (0.06) 0.22 −0.097 (0.057) 0.09 BMI 0.018 (0.057) 0.75 0.02 (0.056) 0.71 −0.015 (0.063) 0.82 Lean body mass 0.12 (0.091) 0.18 0.11 (0.11) 0.33 0.13 (0.093) 0.16 Fat body mass −0.12 (0.091) 0.19 −0.11 (0.11) 0.34 −0.13 (0.094) 0.16 FTI −0.0029 (0.081) 0.97 −0.014 (0.085) 0.87 −0.034 (0.082) 0.68 LTI 0.19 (0.074) 0.010 0.18 (0.083) 0.03 0.15 (0.079) 0.06 Height 0.0000 (0.053) 1.00 −0.059 (0.067) 0.38 −0.014 (0.069) 0.84 Cholesterol 0.11 (0.058) 0.06 0.13 (0.066) 0.05 0.14 (0.062) 0.03 Triglycerides 0.016 (0.062) 0.80 0.017 (0.071) 0.82 0.024 (0.064) 0.71 High-density lipoprotein −0.025 (0.058) 0.66 −0.02 (0.058) 0.74 −0.016 (0.053) 0.75 Low-density lipoprotein 0.1 (0.063) 0.11 0.098 (0.056) 0.08 0.11 (0.063) 0.09 Creatinine 0.062 (0.054) 0.25 0.073 (0.061) 0.23 0.097 (0.065) 0.13 Phosphate 0.012 (0.066) 0.86 0.0037 (0.066) 0.96 0.015 (0.072) 0.84 Calcium 0.15 (0.069) 0.03 0.14 (0.057) 0.01 0.21 (0.062) <0.001 Parathyroid hormone −0.013 (0.084) 0.88 0.0073 (0.072) 0.92 0.033 (0.084) 0.70 hsCRP 0.21 (0.077) 0.006 0.2 (0.075) 0.007 0.2 (0.078) 0.010 s-Albumin −0.2 (0.059) <0.001 −0.22 (0.063) <0.001 −0.2 (0.055) <0.001 TNF −0.22 (0.089) 0.02 −0.25 (0.079) 0.002 −0.27 (0.085) 0.002 HGS −0.055 (0.075) 0.47 −0.19 (0.09) 0.04 −0.13 (0.093) 0.16 Malnutrition, SGA >1 −0.074 (0.16) 0.64 −0.13 (0.14) 0.35 −0.13 (0.15) 0.40 Triceps skinfold −0.025 (0.079) 0.75 0.035 (0.079) 0.65 0.026 (0.078) 0.74 Variable Model 1 P-value Model 2 P-value Model 3 P-value Age 0.08 (0.073) 0.28 0.14 (0.062)a 0.03 0.095 (0.065)a 0.15 Male 0.078 (0.13) 0.56 0.22 (0.13)b 0.08 0.088 (0.14)b 0.53 DM 0.44 (0.13) 0.001 0.3 (0.16) 0.06 CVD 0.14 (0.14) 0.33 0.025 (0.13) 0.85 0.00073 (0.14) 1.00 Smoker, current or former 0.28 (0.17) 0.09 0.097 (0.16) 0.53 0.16 (0.15) 0.26 eGFR −0.052 (0.066) 0.44 −0.074 (0.06) 0.22 −0.097 (0.057) 0.09 BMI 0.018 (0.057) 0.75 0.02 (0.056) 0.71 −0.015 (0.063) 0.82 Lean body mass 0.12 (0.091) 0.18 0.11 (0.11) 0.33 0.13 (0.093) 0.16 Fat body mass −0.12 (0.091) 0.19 −0.11 (0.11) 0.34 −0.13 (0.094) 0.16 FTI −0.0029 (0.081) 0.97 −0.014 (0.085) 0.87 −0.034 (0.082) 0.68 LTI 0.19 (0.074) 0.010 0.18 (0.083) 0.03 0.15 (0.079) 0.06 Height 0.0000 (0.053) 1.00 −0.059 (0.067) 0.38 −0.014 (0.069) 0.84 Cholesterol 0.11 (0.058) 0.06 0.13 (0.066) 0.05 0.14 (0.062) 0.03 Triglycerides 0.016 (0.062) 0.80 0.017 (0.071) 0.82 0.024 (0.064) 0.71 High-density lipoprotein −0.025 (0.058) 0.66 −0.02 (0.058) 0.74 −0.016 (0.053) 0.75 Low-density lipoprotein 0.1 (0.063) 0.11 0.098 (0.056) 0.08 0.11 (0.063) 0.09 Creatinine 0.062 (0.054) 0.25 0.073 (0.061) 0.23 0.097 (0.065) 0.13 Phosphate 0.012 (0.066) 0.86 0.0037 (0.066) 0.96 0.015 (0.072) 0.84 Calcium 0.15 (0.069) 0.03 0.14 (0.057) 0.01 0.21 (0.062) <0.001 Parathyroid hormone −0.013 (0.084) 0.88 0.0073 (0.072) 0.92 0.033 (0.084) 0.70 hsCRP 0.21 (0.077) 0.006 0.2 (0.075) 0.007 0.2 (0.078) 0.010 s-Albumin −0.2 (0.059) <0.001 −0.22 (0.063) <0.001 −0.2 (0.055) <0.001 TNF −0.22 (0.089) 0.02 −0.25 (0.079) 0.002 −0.27 (0.085) 0.002 HGS −0.055 (0.075) 0.47 −0.19 (0.09) 0.04 −0.13 (0.093) 0.16 Malnutrition, SGA >1 −0.074 (0.16) 0.64 −0.13 (0.14) 0.35 −0.13 (0.15) 0.40 Triceps skinfold −0.025 (0.079) 0.75 0.035 (0.079) 0.65 0.026 (0.078) 0.74 Median regression coefficients with standard errors for cardiovascular risk markers on plasma PAPP-A levels. Model 1 is univariable; Model 2 includes age and sex; Model 3 also includes diabetes status. In models marked a and b no adjustment was made for age and sex, respectively. Interval variables are standardized and centered to the mean. Survival analysis Survival by PAPP-A quartiles is presented in Figure 1, showing that survival was worse with increasing quartile. Kaplan–Meier plots with log rank test of potential interactions between dichotomic PAPP-A and hsCRP, LTI, FTI and DM, respectively, showed that median survival time was shorter among patients with high levels of both hsCRP and PAPP-A (Figure 2) and that low PAPP-A was protective only in the high FTI category, while there was no apparent effect of PAPP-A in the low FTI category. Worse survival was seen in the group with simultaneously low LTI and high PAPP-A. Altogether these findings suggest interactions between PAPP-A and hsCRP, FTI and LTI, respectively. A possible interaction with DM was also seen, with high PAPP-A appearing detrimental only in the presence of DM (Figure 2). FIGURE 1 View largeDownload slide Kaplan–Meier diagram of survival in 286 dialysis patients by PAPP-A quartiles. FIGURE 1 View largeDownload slide Kaplan–Meier diagram of survival in 286 dialysis patients by PAPP-A quartiles. FIGURE 2 View largeDownload slide Kaplan–Meier diagram of survival in 286 dialysis patients. Interactions between PAPP-A and LTI (A), FTI (B), hsCRP (C) and DM (D) are shown. Continuous variables were categorized as ‘high’ if above median value and ‘low’ if not. FIGURE 2 View largeDownload slide Kaplan–Meier diagram of survival in 286 dialysis patients. Interactions between PAPP-A and LTI (A), FTI (B), hsCRP (C) and DM (D) are shown. Continuous variables were categorized as ‘high’ if above median value and ‘low’ if not. Multivariable Cox regression models on survival are presented in Table 3. The first model includes PAPP-A in combination with Framingham cardiovascular risk factors. PAPP-A, DM, age, smoking and low-density lipoprotein (LDL) were associated with worse survival. For BMI, the association was negative and CVD appeared detrimental only in the first stratum, i.e. before day 400. PAPP-A levels had a relatively small but robust effect on mortality risk. In Model 2, BMI was replaced by FTI and LTI, which improved model performance and modulated mainly the effects of DM, CVD and LDL. Of the body composition indices only LTI showed a statistically significant association with survival. When hsCRP was also included (Model 3), the association of PAPP-A with survival did not reach statistical significance, although no major changes to the size or direction of coefficients were observed. In a post hoc analysis where the cohort was stratified by diabetes status, PAPP-A was predictive of survival only in patients with diabetes and this association remained when adjusted for Framingham cardiovascular risk factors, LTI, FTI and CRP (Supplementary Tables S7 and S8). Further, when adjusted for GFR in a separate analysis, PAPP-A was independently associated with survival [hazard ratio (HR) = 1.27, 95% confidence interval (CI): 1.09–1.47, P = 0.002]. Table 3 PAPP-A, body composition and survival in 286 incident dialysis patients Variable Model 1 Model 2 Model 3 HR (95% CI) P-value HR (95% CI) P-value HR (95% CI) P-value Age 1.06 (1.03–1.10) <0.001 1.07 (1.04–1.11) <0.001 1.07 (1.04–1.11) <0.001 Male 0.80 (0.48–1.35) 0.41 1.05 (0.53–2.10) 0.88 1.06 (0.53–2.12) 0.87 DM 2.27 (1.35–3.82) 0.002 1.86 (1.02–3.38) 0.04 1.76 (0.96–3.22) 0.07 CVD, strata 1 = day 0–400 3.78 (1.48–9.64) 0.005 3.61 (1.28–10.23) 0.02 3.30 (1.15–9.48) 0.03 CVD, strata 2 = day 400+ 0.62 (0.32–1.18) 0.15 0.54 (0.25–1.17) 0.12 0.53 (0.24–1.15) 0.11 Smoker, current or former 1.74 (0.98–3.11) 0.06 1.83 (0.93–3.60) 0.08 1.79 (0.91–3.52) 0.09 Low-density lipoprotein 1.30 (1.01–1.69) 0.04 1.27 (0.95–1.69) 0.10 1.24 (0.94–1.64) 0.13 PAPP-A 1.21 (1.02–1.42) 0.03 1.19 (1.00–1.41) 0.04 1.19 (0.99–1.41) 0.06 BMI 0.66 (0.51–0.86) 0.002 FTI 0.73 (0.52–1.02) 0.07 0.75 (0.53–1.05) 0.09 LTI 0.66 (0.48–0.91) 0.010 0.68 (0.49–0.94) 0.02 hsCRP 1.20 (0.95–1.52) 0.14 −2 Loglik 635 449 447 Variable Model 1 Model 2 Model 3 HR (95% CI) P-value HR (95% CI) P-value HR (95% CI) P-value Age 1.06 (1.03–1.10) <0.001 1.07 (1.04–1.11) <0.001 1.07 (1.04–1.11) <0.001 Male 0.80 (0.48–1.35) 0.41 1.05 (0.53–2.10) 0.88 1.06 (0.53–2.12) 0.87 DM 2.27 (1.35–3.82) 0.002 1.86 (1.02–3.38) 0.04 1.76 (0.96–3.22) 0.07 CVD, strata 1 = day 0–400 3.78 (1.48–9.64) 0.005 3.61 (1.28–10.23) 0.02 3.30 (1.15–9.48) 0.03 CVD, strata 2 = day 400+ 0.62 (0.32–1.18) 0.15 0.54 (0.25–1.17) 0.12 0.53 (0.24–1.15) 0.11 Smoker, current or former 1.74 (0.98–3.11) 0.06 1.83 (0.93–3.60) 0.08 1.79 (0.91–3.52) 0.09 Low-density lipoprotein 1.30 (1.01–1.69) 0.04 1.27 (0.95–1.69) 0.10 1.24 (0.94–1.64) 0.13 PAPP-A 1.21 (1.02–1.42) 0.03 1.19 (1.00–1.41) 0.04 1.19 (0.99–1.41) 0.06 BMI 0.66 (0.51–0.86) 0.002 FTI 0.73 (0.52–1.02) 0.07 0.75 (0.53–1.05) 0.09 LTI 0.66 (0.48–0.91) 0.010 0.68 (0.49–0.94) 0.02 hsCRP 1.20 (0.95–1.52) 0.14 −2 Loglik 635 449 447 Cox proportional hazards regression models on survival. Interval variables are standardized and centered to the mean. Loglik, log-likelihood. Table 3 PAPP-A, body composition and survival in 286 incident dialysis patients Variable Model 1 Model 2 Model 3 HR (95% CI) P-value HR (95% CI) P-value HR (95% CI) P-value Age 1.06 (1.03–1.10) <0.001 1.07 (1.04–1.11) <0.001 1.07 (1.04–1.11) <0.001 Male 0.80 (0.48–1.35) 0.41 1.05 (0.53–2.10) 0.88 1.06 (0.53–2.12) 0.87 DM 2.27 (1.35–3.82) 0.002 1.86 (1.02–3.38) 0.04 1.76 (0.96–3.22) 0.07 CVD, strata 1 = day 0–400 3.78 (1.48–9.64) 0.005 3.61 (1.28–10.23) 0.02 3.30 (1.15–9.48) 0.03 CVD, strata 2 = day 400+ 0.62 (0.32–1.18) 0.15 0.54 (0.25–1.17) 0.12 0.53 (0.24–1.15) 0.11 Smoker, current or former 1.74 (0.98–3.11) 0.06 1.83 (0.93–3.60) 0.08 1.79 (0.91–3.52) 0.09 Low-density lipoprotein 1.30 (1.01–1.69) 0.04 1.27 (0.95–1.69) 0.10 1.24 (0.94–1.64) 0.13 PAPP-A 1.21 (1.02–1.42) 0.03 1.19 (1.00–1.41) 0.04 1.19 (0.99–1.41) 0.06 BMI 0.66 (0.51–0.86) 0.002 FTI 0.73 (0.52–1.02) 0.07 0.75 (0.53–1.05) 0.09 LTI 0.66 (0.48–0.91) 0.010 0.68 (0.49–0.94) 0.02 hsCRP 1.20 (0.95–1.52) 0.14 −2 Loglik 635 449 447 Variable Model 1 Model 2 Model 3 HR (95% CI) P-value HR (95% CI) P-value HR (95% CI) P-value Age 1.06 (1.03–1.10) <0.001 1.07 (1.04–1.11) <0.001 1.07 (1.04–1.11) <0.001 Male 0.80 (0.48–1.35) 0.41 1.05 (0.53–2.10) 0.88 1.06 (0.53–2.12) 0.87 DM 2.27 (1.35–3.82) 0.002 1.86 (1.02–3.38) 0.04 1.76 (0.96–3.22) 0.07 CVD, strata 1 = day 0–400 3.78 (1.48–9.64) 0.005 3.61 (1.28–10.23) 0.02 3.30 (1.15–9.48) 0.03 CVD, strata 2 = day 400+ 0.62 (0.32–1.18) 0.15 0.54 (0.25–1.17) 0.12 0.53 (0.24–1.15) 0.11 Smoker, current or former 1.74 (0.98–3.11) 0.06 1.83 (0.93–3.60) 0.08 1.79 (0.91–3.52) 0.09 Low-density lipoprotein 1.30 (1.01–1.69) 0.04 1.27 (0.95–1.69) 0.10 1.24 (0.94–1.64) 0.13 PAPP-A 1.21 (1.02–1.42) 0.03 1.19 (1.00–1.41) 0.04 1.19 (0.99–1.41) 0.06 BMI 0.66 (0.51–0.86) 0.002 FTI 0.73 (0.52–1.02) 0.07 0.75 (0.53–1.05) 0.09 LTI 0.66 (0.48–0.91) 0.010 0.68 (0.49–0.94) 0.02 hsCRP 1.20 (0.95–1.52) 0.14 −2 Loglik 635 449 447 Cox proportional hazards regression models on survival. Interval variables are standardized and centered to the mean. Loglik, log-likelihood. When modelling the aforementioned interactions using Cox regression, including one interaction term between PAPP-A and the parameter of interest each in turn, there were significant interactions on the multiplicative scale for hsCRP (HR = 1.6, 95% CI: 1.2–2.2, P = 0.004), DM (HR = 1.6, 95% CI: 1.1–2.2, P = 0.01) and (borderline) for FTI (HR = 1.3, 95% CI: 0.99–1.8, P = 0.06), but not for LTI (HR = 0.93, 95% CI: 0.7–1.2, P = 0.59). Multivariable analysis of these interactions is available in the Supplementary Material and shows that interactions with PAPP-A remained statistically significant for hsCRP and DM (Supplementary Tables S1–S4). Analysis of interactions on the additive scale using relative excess risk due to interaction (RERI) and attributable proportion (AP), showed that FTI and DM interacted with PAPP-A levels on survival (Supplementary Tables S5 and S6). When the cohort was stratified by diabetes status, higher PAPP-A predicted worse survival in diabetic patients, while in nondiabetic patients this association was not seen (Supplementary Tables S7 and S8). Statistical models testing and modifications The proportional hazards assumption was violated for CVD. Thus, a time varying coefficient was included for survival time splitting at 400 days. When removing three influential observations, LDL (HR = 1.35, P = 0.05) and the second strata of CVD (HR = 0.42, P < 0.05) became somewhat more strongly associated with survival. These influential cases were kept in the models presented. DISCUSSION Our study shows that higher PAPP-A levels are associated with worse survival in incident dialysis patients and that this association remains when adjusted for established cardiovascular risk factors and body composition but is attenuated by hsCRP. This accords with previous results in prevalent HD patients [7, 10, 13, 14]. However, in contradiction to hypotheses presented in a previous study [10] we did not find markers of PEW, such as low LTI, to be associated with elevated PAPP-A. On the contrary, LTI was positively associated with PAPP-A levels. DM appears to be a confounder in this context and in the study by Kalousová et al. [10] only diabetic patients were included. In apparent conflict with these results, HGS was negatively associated with PAPP-A when adjusted for age and sex, perhaps reflecting age-related dissociation between muscle mass and muscle strength [24]. It is worth pointing out that neither BMI nor fat body mass or lean body mass percentages predicted PAPP-A levels. BMI may not reflect opposing changes in fat and lean tissue mass [25] while FBM and LBM indices are interdependent when expressed as percentages of body weight, making interpretations of separate effects of lean and fat tissue components difficult. FTI and LTI could therefore be more appropriate indicators when investigating the relative effects of lean and fat mass. In addition to different inclusion criteria, changes in body composition during the course of dialysis treatment [26], along with selection due to high mortality rates, could explain discrepancies between our results and those from studies on prevalent HD patients. Furthermore, analysis of interactions on the additive scale showed a statistically significant positive interaction between PAPP-A and FTI, suggesting that PAPP-A has a stronger association with survival in persons with higher FTI. Whereas PAPP-A was positively associated with hsCRP, a negative correlation to TNF was observed. This is surprising since TNF stimulates PAPP-A expression in cell models [27]. Although it could reflect a complex interplay between PAPP-A and different inflammatory pathways, confounding or spurious findings could not be ruled out. In line with previous findings [11], we observed an interaction between inflammation and PAPP-A in that the mortality predictive impact of PAPP-A was enhanced when inflammation was present. This observation supports a catalytic effect of inflammation for the risk factor profile [28]. We also report that elevated PAPP-A might be especially detrimental in diabetic dialysis patients or possibly prognostic only in this group. However, this finding should be regarded with caution since confounding from comorbid conditions may be present and due to the stratified analysis being post hoc. GH axis components, including IGF-1 and PAPP-A, have been linked to atherosclerosis but we found no association between PAPP-A levels and history of CVD. Due to differences in case-mix along with GH axis disturbances in ESRD, it may not be appropriate to extrapolate from findings in other populations or experimental models. Furthermore, systemic levels of PAPP-A may not reflect local PAPP-A activity associated with atherosclerotic pathways in animal models and it is unclear if plasma levels of PAPP-A represent local up-regulation or increased release due to tissue damage. Neither do plasma levels reveal from what organ PAPP-A is released. In a previous study on HD patients, PAPP-A predicted death by infection, stroke or sudden cardiac death rather than myocardial infarction [10]. Despite pathophysiological links between local PAPP-A activity and CVD, plasma PAPP-A in ESRD could hypothetically be a biomarker for conditions associated with increased risk for cerebrovascular or infectious complications rather than coronary artery disease. A number of caveats should be considered when interpreting the results of the present study. Although the mean of renal urea and creatinine clearances during 24-h was used as a measure of GFR, and we cannot exclude association of GFR with muscle mass, adjustment for residual renal function would be inappropriate since associations to lean tissue was part of the research question. Furthermore, as residual renal function is positively associated with both survival and PAPP-A, including it in the statistical models is unlikely to weaken the associations between PAPP-A and mortality. Also, in the present study we did not find any association between PAPP-A and GFR (Table 2). Further, the relative small sample size limits the number of confounders and interactions that can be investigated in multivariable analysis. Although we have adjusted for major CVD risk factors, the broad inclusion criteria increase the possibility of residual confounding from comorbid conditions. Ideally, studies with larger sample size would allow for more adjustment for the significant case-mix present in the dialysis population. In summary, a higher circulating concentration of PAPP-A is associated with inflammatory markers and predicts mortality risk in incident dialysis. Body composition, inflammation and DM are effect modifiers when modelling the effects of PAPP-A on survival. SUPPLEMENTARY DATA Supplementary data are available online at http://ndt.oxfordjournals.org. ACKNOWLEDGEMENTS We are grateful to the patients and control subjects participating in the study. We thank Åsa Lindé and Annika Nilsson, Anki Emmoth and Ulrika Jensen for collection of samples, and Ann-Christin Bragfors-Helin and Monica Eriksson for laboratory analyses. Baxter Novum is the result of a grant from Baxter Healthcare Corporation to the Karolinska Institutet. FUNDING This study was supported by grants from The Swedish Research Council (grant 521-2013-2764), Westman's Foundation (no grant number) and The Swedish Kidney Foundation (no grant number). CONFLICT OF INTEREST STATEMENT B.L. is employed by and has received grants, lecturing fees, consultancy fees and travel funding from Baxter Healthcare. P.S. has received grants from Bayer, lecturing fees from AbbVie, Shire, Bayer, Pfizer and Asahi, and has been a member of Scientific Advisory Boards of ARO group (Amgen), Vifor, Keryx and Astellas. REFERENCES 1 Root A. Growth hormone . Pediatrics 1965 ; 36 : 940 – 950 Google Scholar PubMed 2 Palmeiro CR , Anand R , Dardi IK et al. . Growth hormone and the cardiovascular system . Cardiol Rev 2012 ; 20 : 197 – 207 Google Scholar CrossRef Search ADS PubMed 3 Higashi Y , Sukhanov S , Anwar A et al. . Aging, atherosclerosis, and IGF-1 . J Gerontol A Biol Sci Med Sci 2012 ; 67A : 626 – 639 Google Scholar CrossRef Search ADS 4 Burgers AMG , Biermasz NR , Schoones JW et al. . Meta-analysis and dose-response metaregression: circulating insulin-like growth factor 1 (IGF-1) and mortality . J Clin Endocrinol Metab 2011 ; 96 : 2912 – 2920 Google Scholar CrossRef Search ADS PubMed 5 Jia T , Gama Axelsson T , Heimbürger O et al. . IGF-1 and survival in ESRD . Clin J Am Soc Nephrol 2014 ; 9 : 120 – 127 Google Scholar CrossRef Search ADS PubMed 6 Maile LA , Holly JMP. Insulin-like growth factor binding protein (IGFBP) proteolysis: Occurrence, identification, role and regulation . Growth Horm IGF Res 1999 ; 9 : 85 – 95 Google Scholar CrossRef Search ADS PubMed 7 Kalousova M , Benakova H , Kubena AA et al. . Pregnancy-associated plasma protein A as an independent mortality predictor in long-term hemodialysis patients . Kidney Blood Press Res 2012 ; 35 : 192 – 201 Google Scholar CrossRef Search ADS PubMed 8 Murray RD , Adams JE , Shalet SM. Adults with partial growth hormone deficiency have an adverse body composition . J Clin Endocrinol Metab 2004 ; 89 : 1586 – 1591 Google Scholar CrossRef Search ADS PubMed 9 Conover CA , Bale LK , Marler RJ. Pregnancy-associated plasma protein-A deficiency improves survival of mice on a high fat diet . Exp Gerontol 2015 ; 70 : 131 – 134 Google Scholar CrossRef Search ADS PubMed 10 Kalousová M , Zima T , Krane V et al. . Pregnancy-associated plasma protein A associates with cardiovascular events in diabetic hemodialysis patients . Atherosclerosis 2014 ; 236 : 263 – 269 Google Scholar CrossRef Search ADS PubMed 11 Consuegra-Sanchez L , Fredericks S , Kaski JC. Pregnancy-associated plasma protein-A (PAPP-A) and cardiovascular risk . Atherosclerosis 2009 ; 203 : 346 – 352 Google Scholar CrossRef Search ADS PubMed 12 Mueller T , Dieplinger B , Poelz W et al. . Increased pregnancy-associated plasma protein-A as a marker for peripheral atherosclerosis: results from the Linz peripheral arterial disease study . Clin Chem 2006 ; 52 : 1096 – 1103 Google Scholar CrossRef Search ADS PubMed 13 Kalousova M , Horejsi M , Fialova L et al. . Increased levels of pregnancy-associated plasma protein A are associated with mortality in hemodialysis patients: preliminary results . Blood Purif 2004 ; 22 : 298 – 300 Google Scholar CrossRef Search ADS PubMed 14 Etter C , Straub Y , Hersberger M et al. . Pregnancy-associated plasma protein-A is an independent short-time predictor of mortality in patients on maintenance haemodialysis . Eur Heart J 2010 ; 31 : 354 – 359 Google Scholar CrossRef Search ADS PubMed 15 Jefferies HJ , Tertti R , Wittfooth S et al. . Elevated serum free pregnancy-associated plasma protein-A independently predicts mortality in haemodialysis patients but is not associated with recurrent haemodialysis-induced ischaemic myocardial injury . Nephron 2015 ; 129 : 171 – 178 Google Scholar CrossRef Search ADS PubMed 16 Eckardt K-U , Gillespie IA , Kronenberg F et al. . High cardiovascular event rates occur within the first weeks of starting hemodialysis . Kidney Int 2015 ; 88 : 1117 – 1125 Google Scholar CrossRef Search ADS PubMed 17 Stenvinkel P , Heimbürger O , Paultre F et al. . Strong association between malnutrition, inflammation, and atherosclerosis in chronic renal failure . Kidney Int 1999 ; 55 : 1899 – 1911 Google Scholar CrossRef Search ADS PubMed 18 Kyle UG , Schutz Y , Dupertuis YM , Pichard C. Body composition interpretation: contributions of the fat-free mass index and the body fat mass index . Nutrition 2003 ; 19 : 597 – 604 Google Scholar CrossRef Search ADS PubMed 19 VanItallie TB , Yang MU , Heymsfield SB et al. . Height-normalized indices of the body’s fat-free mass and fat mass: potentially useful indicators of nutritional status . Am J Clin Nutr 1990 ; 52 : 953 – 959 Google Scholar CrossRef Search ADS PubMed 20 Qureshi AR , Alvestrand A , Danielsson A et al. . Factors predicting malnutrition in hemodialysis patients: a cross-sectional study . Kidney Int 1998 ; 53 : 773 – 782 Google Scholar CrossRef Search ADS PubMed 21 R Development Core Team . R: A Language and Environment for Statistical Computing . Vienna, Austria : R Foundation for Statistical Computing , 2016 22 Therneau TM. Survival: Survival analysis. 2015 . https://CRAN.R-project.org/package=survival (30 May 2017, date last accessed) 23 Koenker R. Quantreg: Quantile regression. 2016 . https://CRAN.R-project.org/package=quantreg (30 May 2017, date last accessed) 24 Clark BC , Manini TM. Sarcopenia =/= dynapenia . J Gerontol Ser A 2008 ; 63 : 829 – 834 Google Scholar CrossRef Search ADS 25 Marcelli D , Brand K , Ponce P et al. . Longitudinal changes in body composition in patients after initiation of hemodialysis therapy: Results from an international cohort . J Ren Nutr 2016 ; 26 : 72 – 80 Google Scholar CrossRef Search ADS PubMed 26 Chumlea WC. Anthropometric and body composition assessment in dialysis patients . Semin Dial 2004 ; 17 : 466 – 470 Google Scholar CrossRef Search ADS PubMed 27 Conover CA , Bale LK , Harrington SC et al. . Cytokine stimulation of pregnancy-associated plasma protein A expression in human coronary artery smooth muscle cells: inhibition by resveratrol . Am J Physiol Cell Physiol 2006 ; 290 : C183 – C188 Google Scholar CrossRef Search ADS PubMed 28 Carrero JJ , Stenvinkel P. Persistent inflammation as a catalyst for other risk factors in chronic kidney disease: a hypothesis proposal . Clin J Am Soc Nephrol 2009 ; 4 : S49 – S55 Google Scholar CrossRef Search ADS PubMed © The Author 2017. 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

Pregnancy-associated plasma protein-A predicts survival in end-stage renal disease—confounding and modifying effects of cardiovascular disease, body composition and inflammation

Loading next page...
 
/lp/ou_press/pregnancy-associated-plasma-protein-a-predicts-survival-in-end-stage-wEQwirX8ZB
Publisher
Oxford University Press
Copyright
© The Author 2017. 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/gfx215
Publisher site
See Article on Publisher Site

Abstract

ABSTRACT Background High pregnancy-associated plasma protein-A (PAPP-A) levels are linked to atherosclerosis and associate with increased mortality in prevalent dialysis patients. We investigated associations of PAPP-A, measured at dialysis initiation, with cardiovascular disease (CVD), CVD risk factors and mortality in incident dialysis patients, and explored if body composition and inflammation modulated these associations. Methods Baseline plasma PAPP-A levels, inflammation biomarkers and body composition, using dual-energy X-ray absorptiometry, were measured in 286 incident dialysis patients. Primary outcome was survival during 60 months follow-up. Quantile (median) regression was used for cross-sectional analysis and Kaplan–Meier diagrams and Cox proportional hazards regression for survival analysis. Results In cross-sectional analysis adjusted for age and sex, PAPP-A levels were associated with lean tissue index (LTI) and high-sensitivity C-reactive protein (hsCRP) but not with fat tissue index (FTI) or history of CVD. In a model also including diabetes mellitus (DM), the association with LTI did not remain statistically significant. When adjusted for cardiovascular risk factors and body composition, higher PAPP-A levels showed a moderate but significant association [hazard ratio (HR) = 1.2, 95% confidence interval (CI): 1–1.4, P = 0.04] with mortality. When also including hsCRP the association was attenuated (HR = 1.2, 95% CI: 0.99–1.4, P = 0.06). In survival analysis, interactions with PAPP-A on the multiplicative scale were found for hsCRP (HR = 1.6, 95% CI: 1.2–2.2, P = 0.004) and DM (HR = 1.6, 95% CI: 1.1–2.2, P = 0.01) and with DM and FTI on the additive scale. Conclusions Higher PAPP-A levels are associated with worse survival in incident dialysis patients following adjustment for established cardiovascular risk factors and body composition indices, but not clearly so when adjusted for hsCRP. Inflammation, body composition (FTI) and DM were found to be potential effect modifiers for the observed moderate association of PAPP-A with survival. biomarkers, cardiovascular disease, dialysis, end-stage renal disease, pregnancy-associated plasma protein-A INTRODUCTION Pregnancy-associated plasma protein-A (PAPP-A) plays a role in the development and ageing processes by modulating growth hormone (GH) effects on lipid, glucose and protein metabolism, as well as on cardiovascular function in adults [1, 2]. Both excess and deficiency of GH are associated with cardiovascular disease (CVD) and GH and its main mediator, insulin-like growth factor 1 (IGF-1), have been implicated as having both pro- and antiatherogenic effects in experimental models [2, 3]. Both high and low circulating concentrations of IGF-1 are associated with increased mortality risk in the general population [4] and in patients with end-stage renal disease (ESRD) low IGF-1 predicts worse survival [5]. Availability of IGF-1 at receptor level is influenced by concentrations of its multiple binding proteins (insulin-like growth factor binding proteins, IGFBPs) in plasma and by membrane-bound metalloproteinase PAPP-A [6], which cleaves IGFBP-4 bound to IGF-1, making IGF-1 available to its receptor. PAPP-A is therefore considered an important regulator of IGF-1 action [7]. GH levels are associated with body composition [8] and PAPP-A is expressed relatively more in visceral than in subcutaneous fat. Conversely, in PAPP-A knock-out mice mesenteric fat depots are reduced [9]. This indicates an interdependency between fat tissue and PAPP-A. A negative association between PAPP-A and body mass index (BMI) has been described in ESRD and a link to protein-energy wasting (PEW) suggested [10]. Elevated plasma PAPP-A concentration is associated with the extent of coronary artery disease and has been evaluated as a prognostic marker in chronic stable angina pectoris and acute coronary syndromes, where elevated plasma levels predict increased risk of death [11]. Interestingly, an interaction between PAPP-A and the anti-inflammatory cytokine IL-10 levels has been observed in this context, indicating that elevated PAPP-A is detrimental only when IL-10 levels are low. Positive associations between PAPP-A levels and carotid and peripheral artery disease have also been reported [11, 12]. In prevalent haemodialysis (HD) patients, PAPP-A levels are elevated compared with controls, predict all-cause and cardiovascular mortality [7, 10, 13, 14] and are associated with cardiac troponins [15]. Notably, in adjusted models an association with sudden death, stroke and death due to infection was found, but there was no statistically significant association with risk of myocardial infarction [7, 10]. PAPP-A has not been evaluated as a prognostic marker in incident dialysis patients. Due to the high early mortality after initiation of dialysis [16], studies on prevalent patients are not representative for the incident dialysis population. The aim of the present study was to investigate PAPP-A as a risk marker for mortality and CVD in incident dialysis patients, highlighting body composition as a potential confounder and effect modifier. We hypothesized that PAPP-A levels are associated with survival in incident dialysis patients, that this association is confounded by body composition as represented by lean tissue index (LTI) and fat tissue index (FTI), and further, that there would be an interaction between PAPP-A and fat tissue and inflammation in relation to survival. We also report on the associations between PAPP-A and CVD risk factors in this population. MATERIALS AND METHODS Study design We measured PAPP-A levels in an ongoing prospective cohort originally designed to investigate malnutrition, PEW, inflammation and atherosclerosis in incident dialysis patients, described previously [17]. Exclusion criteria were age  <18 or  >75 years, signs of infection and unwillingness to participate. During 1993 to 2014, 552 patients were recruited to the cohort. Among these, 286 patients had plasma samples available in 2015 and were included in the present study, having started dialysis treatment in years 2000–14. Participants were followed until death, renal transplantation or until 1 March 2016. The Ethics Committee of Karolinska Institutet, Stockholm, Sweden, approved the study protocol. Informed consent was obtained from all participants. The study comprised a cross-sectional analysis of PAPP-A levels in relation to cardiovascular risk factors, inflammation and body composition and a prospective analysis of mortality, where follow-up was truncated at 60 months. Blood sampling and biochemical analysis Blood samples were drawn after overnight fast and analysed immediately or centrifuged and stored at −70°C. PAPP-A was assayed using ELISA (R&D Systems, Minneapolis, MN, USA). High-sensitivity C-reactive protein (hsCRP) was measured using nephelometry and tumour necrosis factor (TNF) was analysed at an Immulite system (Diagnostic Products Corp., Los Angeles, CA, USA). Routine biochemical analyses were performed at the Department of Clinical Chemistry at Karolinska University Hospital, Huddinge. Glomerular filtration rate (GFR) was measured as the mean of renal urea and creatinine clearances during a 24-h urine collection, which is thought to provide a reliable estimate of GFR in patients starting on dialysis. Body composition, anthropometry and nutritional status Body composition was measured using dual-energy X-ray absorptiometry (DEXA) (Lunar Corp., Madison, WI, USA), within 7 days of inclusion and after HD if already started. LTI and FTI were calculated in kilograms per meters squared [18, 19]. Handgrip strength (HGS) was measured in the nonfistula or dominant hand using a Harpenden Handgrip Dynamometer (Yamar, Jackson, MI, USA), repeated three times and the highest value noted. HGS was then normalized within each sex category by dividing the value with mean values from individuals randomly selected by Statistics Sweden (a government agency) from the Stockholm region to constitute a control population. Nutritional status was estimated by subjective global assessment (SGA) and categorized into normal versus not normal nutritional status, the latter encompassing signs of mild to severe malnutrition [20]. Definition of CVD Presence of CVD at baseline was determined through review of medical history and included any previous occurrence of coronary artery disease (myocardial infarction, angina pectoris, coronary artery bypass graft or percutaneous transluminal coronary angioplasty), cerebrovascular disease (ischaemic or haemorrhagic stroke or transitory ischaemic attack), peripheral vascular disease (in extremities, carotides or renal arteries), heart failure or aortic aneurysm. Statistical methods For descriptive statistics median and interquartile range or percentages were used as appropriate. Follow-up time was truncated at 60 months and interval variables were standardized in cross-sectional and survival analysis. Numbers were rounded at two significant digits for ratios and three for descriptive data, trailing zeroes omitted. Quantile (median) regression was utilized for cross-sectional associations and Cox regression was used for survival analysis with censoring at renal transplantation and end of follow-up. Confounders included in multivariable models were selected a priori based on known associations to survival or (for cross-sectional analysis) CVD and PAPP-A levels. Variables were then included based on the specific hypotheses presented. Kaplan–Meier curves were used to visualize potential (unadjusted) interactions between PAPP-A and body composition indices (LTI, FBI) and inflammation (hsCRP). Variables were dichotomized at median for this analysis. These interactions were then tested in Cox models. Cox model assumptions were tested using Schoenfeld residuals plot for proportional hazards and influence diagnostics for identifying influential data points. Testing of normal distribution was not done, since assumption of normality of distributions does not apply for the methods used in the study. All tests were two-tailed and a P < 0.05 considered statistically significant. The statistical analyses were performed in R v 3.3.1 [21] using ‘survival’ [22] and ‘quantreg’ [23] packages. RESULTS Baseline characteristics and outcomes Baseline characteristics are presented in Table 1. DEXA was performed in 213 patients. Compared with a similar study in prevalent HD patients [7], participants in the current study were somewhat younger (mean age 55 versus 63 years) and with a higher proportion of males (62% versus 54%). The frequency of diabetes mellitus (DM) was similar (30% versus 33%) but history of CVD less common in the present study (37% versus 61%). Causes of primary renal disease included diabetic nephropathy (30%), hypertensive renal disease (7%), systemic inflammatory disease (2%), primary glomerular disease (18%), hereditary disease (15%), and other (4%) or unknown (21%) aetiology. Medications used included erythropoesis-stimulating agents (90%), diuretics (86%), phosphate binders (84%), active vitamin D (81%), angiotensin-converting enzyme inhibitors/angiotensin receptor blockers (69%), beta blockers (68%), vitamin B1 (63%), calcium channel blockers (52%), intravenous iron (43%), statins (36%), insulin (26%), amino acid substitution (24%), allopurinol (17%) and peroral iron (15%). Table 1 Demographic, clinical and laboratory characteristics of 286 incident dialysis patients Variable Median (IQR) or % Age, years 57 (47–65) Male 62 DM 30 CVD 37 Smoker, current or former 48 Malnutrition, SGA >1 26 eGFR, mL/min/1.73 m2 6.4 (5–8) BMI, kg/m2 24.6 (22–28.3) Lean body mass, % 69.1 (63.4–77) Fat body mass, % 30.9 (23–36.6) FTI, kg/m2 7.16 (5.02–9.81) LTI, kg/m2 16.3 (14.8–18.2) Height, cm 172 (165–179) Cholesterol, mmol/L 4.4 (3.7–5.2) Triglycerid, mmol/L 1.6 (1.1–2.2) High-density lipoprotein, mmol/L 1.3 (1–1.6) Low-density lipoprotein, mmol/L 2.25 (1.6–2.97) Creatinine, µmol/L 733 (576–910) Phosphate, mmol/L 1.9 (1.6–2.32) Calcium, mmol/L 2.39 (2.21–2.56) Parathyroid hormone, ng/L 232 (120–370) hsCRP, mg/L 3.9 (1.3–10.8) s-Albumin, g/L 34 (30–36.2) TNF, pg/mL 14.4 (11.3–17.7) PAPP-A, ng/mL 0.139 (0.035–0.275) Triceps skinfold, mm 14 (10–18) HGS, % 29 (22.2–37.5) Variable Median (IQR) or % Age, years 57 (47–65) Male 62 DM 30 CVD 37 Smoker, current or former 48 Malnutrition, SGA >1 26 eGFR, mL/min/1.73 m2 6.4 (5–8) BMI, kg/m2 24.6 (22–28.3) Lean body mass, % 69.1 (63.4–77) Fat body mass, % 30.9 (23–36.6) FTI, kg/m2 7.16 (5.02–9.81) LTI, kg/m2 16.3 (14.8–18.2) Height, cm 172 (165–179) Cholesterol, mmol/L 4.4 (3.7–5.2) Triglycerid, mmol/L 1.6 (1.1–2.2) High-density lipoprotein, mmol/L 1.3 (1–1.6) Low-density lipoprotein, mmol/L 2.25 (1.6–2.97) Creatinine, µmol/L 733 (576–910) Phosphate, mmol/L 1.9 (1.6–2.32) Calcium, mmol/L 2.39 (2.21–2.56) Parathyroid hormone, ng/L 232 (120–370) hsCRP, mg/L 3.9 (1.3–10.8) s-Albumin, g/L 34 (30–36.2) TNF, pg/mL 14.4 (11.3–17.7) PAPP-A, ng/mL 0.139 (0.035–0.275) Triceps skinfold, mm 14 (10–18) HGS, % 29 (22.2–37.5) Table 1 Demographic, clinical and laboratory characteristics of 286 incident dialysis patients Variable Median (IQR) or % Age, years 57 (47–65) Male 62 DM 30 CVD 37 Smoker, current or former 48 Malnutrition, SGA >1 26 eGFR, mL/min/1.73 m2 6.4 (5–8) BMI, kg/m2 24.6 (22–28.3) Lean body mass, % 69.1 (63.4–77) Fat body mass, % 30.9 (23–36.6) FTI, kg/m2 7.16 (5.02–9.81) LTI, kg/m2 16.3 (14.8–18.2) Height, cm 172 (165–179) Cholesterol, mmol/L 4.4 (3.7–5.2) Triglycerid, mmol/L 1.6 (1.1–2.2) High-density lipoprotein, mmol/L 1.3 (1–1.6) Low-density lipoprotein, mmol/L 2.25 (1.6–2.97) Creatinine, µmol/L 733 (576–910) Phosphate, mmol/L 1.9 (1.6–2.32) Calcium, mmol/L 2.39 (2.21–2.56) Parathyroid hormone, ng/L 232 (120–370) hsCRP, mg/L 3.9 (1.3–10.8) s-Albumin, g/L 34 (30–36.2) TNF, pg/mL 14.4 (11.3–17.7) PAPP-A, ng/mL 0.139 (0.035–0.275) Triceps skinfold, mm 14 (10–18) HGS, % 29 (22.2–37.5) Variable Median (IQR) or % Age, years 57 (47–65) Male 62 DM 30 CVD 37 Smoker, current or former 48 Malnutrition, SGA >1 26 eGFR, mL/min/1.73 m2 6.4 (5–8) BMI, kg/m2 24.6 (22–28.3) Lean body mass, % 69.1 (63.4–77) Fat body mass, % 30.9 (23–36.6) FTI, kg/m2 7.16 (5.02–9.81) LTI, kg/m2 16.3 (14.8–18.2) Height, cm 172 (165–179) Cholesterol, mmol/L 4.4 (3.7–5.2) Triglycerid, mmol/L 1.6 (1.1–2.2) High-density lipoprotein, mmol/L 1.3 (1–1.6) Low-density lipoprotein, mmol/L 2.25 (1.6–2.97) Creatinine, µmol/L 733 (576–910) Phosphate, mmol/L 1.9 (1.6–2.32) Calcium, mmol/L 2.39 (2.21–2.56) Parathyroid hormone, ng/L 232 (120–370) hsCRP, mg/L 3.9 (1.3–10.8) s-Albumin, g/L 34 (30–36.2) TNF, pg/mL 14.4 (11.3–17.7) PAPP-A, ng/mL 0.139 (0.035–0.275) Triceps skinfold, mm 14 (10–18) HGS, % 29 (22.2–37.5) At 60 months follow-up 86 patients had died, 144 were transplanted and 14 censored due to end of follow-up. Causes of death included cardiovascular causes (37%), infection (15%), malignancy (5%), other (10%) and unknown (35%). Factors associated with PAPP-A in cross-sectional analysis Univariable and multivariable associations between PAPP-A and body composition, nutritional parameters, cardiovascular risk factors and inflammatory markers are presented in Table 2. Among body composition variables, PAPP-A was associated only with LTI. This association was positive and remained when adjusted for age and sex but was not statistically significant when DM was also included. Conversely, HGS had a negative association with PAPP-A when adjusted for age and sex but the association did not remain when further adjusted for DM. Median PAPP-A was also predicted to be higher in DM, although this association was less clear when age and sex were considered. Notably, hsCRP and calcium were positively associated with PAPP-A while associations with s-albumin and TNF were negative in all models. Table 2 Associations between cardiovascular risk markers and plasma PAPP-A levels in 286 incident dialysis patients Variable Model 1 P-value Model 2 P-value Model 3 P-value Age 0.08 (0.073) 0.28 0.14 (0.062)a 0.03 0.095 (0.065)a 0.15 Male 0.078 (0.13) 0.56 0.22 (0.13)b 0.08 0.088 (0.14)b 0.53 DM 0.44 (0.13) 0.001 0.3 (0.16) 0.06 CVD 0.14 (0.14) 0.33 0.025 (0.13) 0.85 0.00073 (0.14) 1.00 Smoker, current or former 0.28 (0.17) 0.09 0.097 (0.16) 0.53 0.16 (0.15) 0.26 eGFR −0.052 (0.066) 0.44 −0.074 (0.06) 0.22 −0.097 (0.057) 0.09 BMI 0.018 (0.057) 0.75 0.02 (0.056) 0.71 −0.015 (0.063) 0.82 Lean body mass 0.12 (0.091) 0.18 0.11 (0.11) 0.33 0.13 (0.093) 0.16 Fat body mass −0.12 (0.091) 0.19 −0.11 (0.11) 0.34 −0.13 (0.094) 0.16 FTI −0.0029 (0.081) 0.97 −0.014 (0.085) 0.87 −0.034 (0.082) 0.68 LTI 0.19 (0.074) 0.010 0.18 (0.083) 0.03 0.15 (0.079) 0.06 Height 0.0000 (0.053) 1.00 −0.059 (0.067) 0.38 −0.014 (0.069) 0.84 Cholesterol 0.11 (0.058) 0.06 0.13 (0.066) 0.05 0.14 (0.062) 0.03 Triglycerides 0.016 (0.062) 0.80 0.017 (0.071) 0.82 0.024 (0.064) 0.71 High-density lipoprotein −0.025 (0.058) 0.66 −0.02 (0.058) 0.74 −0.016 (0.053) 0.75 Low-density lipoprotein 0.1 (0.063) 0.11 0.098 (0.056) 0.08 0.11 (0.063) 0.09 Creatinine 0.062 (0.054) 0.25 0.073 (0.061) 0.23 0.097 (0.065) 0.13 Phosphate 0.012 (0.066) 0.86 0.0037 (0.066) 0.96 0.015 (0.072) 0.84 Calcium 0.15 (0.069) 0.03 0.14 (0.057) 0.01 0.21 (0.062) <0.001 Parathyroid hormone −0.013 (0.084) 0.88 0.0073 (0.072) 0.92 0.033 (0.084) 0.70 hsCRP 0.21 (0.077) 0.006 0.2 (0.075) 0.007 0.2 (0.078) 0.010 s-Albumin −0.2 (0.059) <0.001 −0.22 (0.063) <0.001 −0.2 (0.055) <0.001 TNF −0.22 (0.089) 0.02 −0.25 (0.079) 0.002 −0.27 (0.085) 0.002 HGS −0.055 (0.075) 0.47 −0.19 (0.09) 0.04 −0.13 (0.093) 0.16 Malnutrition, SGA >1 −0.074 (0.16) 0.64 −0.13 (0.14) 0.35 −0.13 (0.15) 0.40 Triceps skinfold −0.025 (0.079) 0.75 0.035 (0.079) 0.65 0.026 (0.078) 0.74 Variable Model 1 P-value Model 2 P-value Model 3 P-value Age 0.08 (0.073) 0.28 0.14 (0.062)a 0.03 0.095 (0.065)a 0.15 Male 0.078 (0.13) 0.56 0.22 (0.13)b 0.08 0.088 (0.14)b 0.53 DM 0.44 (0.13) 0.001 0.3 (0.16) 0.06 CVD 0.14 (0.14) 0.33 0.025 (0.13) 0.85 0.00073 (0.14) 1.00 Smoker, current or former 0.28 (0.17) 0.09 0.097 (0.16) 0.53 0.16 (0.15) 0.26 eGFR −0.052 (0.066) 0.44 −0.074 (0.06) 0.22 −0.097 (0.057) 0.09 BMI 0.018 (0.057) 0.75 0.02 (0.056) 0.71 −0.015 (0.063) 0.82 Lean body mass 0.12 (0.091) 0.18 0.11 (0.11) 0.33 0.13 (0.093) 0.16 Fat body mass −0.12 (0.091) 0.19 −0.11 (0.11) 0.34 −0.13 (0.094) 0.16 FTI −0.0029 (0.081) 0.97 −0.014 (0.085) 0.87 −0.034 (0.082) 0.68 LTI 0.19 (0.074) 0.010 0.18 (0.083) 0.03 0.15 (0.079) 0.06 Height 0.0000 (0.053) 1.00 −0.059 (0.067) 0.38 −0.014 (0.069) 0.84 Cholesterol 0.11 (0.058) 0.06 0.13 (0.066) 0.05 0.14 (0.062) 0.03 Triglycerides 0.016 (0.062) 0.80 0.017 (0.071) 0.82 0.024 (0.064) 0.71 High-density lipoprotein −0.025 (0.058) 0.66 −0.02 (0.058) 0.74 −0.016 (0.053) 0.75 Low-density lipoprotein 0.1 (0.063) 0.11 0.098 (0.056) 0.08 0.11 (0.063) 0.09 Creatinine 0.062 (0.054) 0.25 0.073 (0.061) 0.23 0.097 (0.065) 0.13 Phosphate 0.012 (0.066) 0.86 0.0037 (0.066) 0.96 0.015 (0.072) 0.84 Calcium 0.15 (0.069) 0.03 0.14 (0.057) 0.01 0.21 (0.062) <0.001 Parathyroid hormone −0.013 (0.084) 0.88 0.0073 (0.072) 0.92 0.033 (0.084) 0.70 hsCRP 0.21 (0.077) 0.006 0.2 (0.075) 0.007 0.2 (0.078) 0.010 s-Albumin −0.2 (0.059) <0.001 −0.22 (0.063) <0.001 −0.2 (0.055) <0.001 TNF −0.22 (0.089) 0.02 −0.25 (0.079) 0.002 −0.27 (0.085) 0.002 HGS −0.055 (0.075) 0.47 −0.19 (0.09) 0.04 −0.13 (0.093) 0.16 Malnutrition, SGA >1 −0.074 (0.16) 0.64 −0.13 (0.14) 0.35 −0.13 (0.15) 0.40 Triceps skinfold −0.025 (0.079) 0.75 0.035 (0.079) 0.65 0.026 (0.078) 0.74 Median regression coefficients with standard errors for cardiovascular risk markers on plasma PAPP-A levels. Model 1 is univariable; Model 2 includes age and sex; Model 3 also includes diabetes status. In models marked a and b no adjustment was made for age and sex, respectively. Interval variables are standardized and centered to the mean. Table 2 Associations between cardiovascular risk markers and plasma PAPP-A levels in 286 incident dialysis patients Variable Model 1 P-value Model 2 P-value Model 3 P-value Age 0.08 (0.073) 0.28 0.14 (0.062)a 0.03 0.095 (0.065)a 0.15 Male 0.078 (0.13) 0.56 0.22 (0.13)b 0.08 0.088 (0.14)b 0.53 DM 0.44 (0.13) 0.001 0.3 (0.16) 0.06 CVD 0.14 (0.14) 0.33 0.025 (0.13) 0.85 0.00073 (0.14) 1.00 Smoker, current or former 0.28 (0.17) 0.09 0.097 (0.16) 0.53 0.16 (0.15) 0.26 eGFR −0.052 (0.066) 0.44 −0.074 (0.06) 0.22 −0.097 (0.057) 0.09 BMI 0.018 (0.057) 0.75 0.02 (0.056) 0.71 −0.015 (0.063) 0.82 Lean body mass 0.12 (0.091) 0.18 0.11 (0.11) 0.33 0.13 (0.093) 0.16 Fat body mass −0.12 (0.091) 0.19 −0.11 (0.11) 0.34 −0.13 (0.094) 0.16 FTI −0.0029 (0.081) 0.97 −0.014 (0.085) 0.87 −0.034 (0.082) 0.68 LTI 0.19 (0.074) 0.010 0.18 (0.083) 0.03 0.15 (0.079) 0.06 Height 0.0000 (0.053) 1.00 −0.059 (0.067) 0.38 −0.014 (0.069) 0.84 Cholesterol 0.11 (0.058) 0.06 0.13 (0.066) 0.05 0.14 (0.062) 0.03 Triglycerides 0.016 (0.062) 0.80 0.017 (0.071) 0.82 0.024 (0.064) 0.71 High-density lipoprotein −0.025 (0.058) 0.66 −0.02 (0.058) 0.74 −0.016 (0.053) 0.75 Low-density lipoprotein 0.1 (0.063) 0.11 0.098 (0.056) 0.08 0.11 (0.063) 0.09 Creatinine 0.062 (0.054) 0.25 0.073 (0.061) 0.23 0.097 (0.065) 0.13 Phosphate 0.012 (0.066) 0.86 0.0037 (0.066) 0.96 0.015 (0.072) 0.84 Calcium 0.15 (0.069) 0.03 0.14 (0.057) 0.01 0.21 (0.062) <0.001 Parathyroid hormone −0.013 (0.084) 0.88 0.0073 (0.072) 0.92 0.033 (0.084) 0.70 hsCRP 0.21 (0.077) 0.006 0.2 (0.075) 0.007 0.2 (0.078) 0.010 s-Albumin −0.2 (0.059) <0.001 −0.22 (0.063) <0.001 −0.2 (0.055) <0.001 TNF −0.22 (0.089) 0.02 −0.25 (0.079) 0.002 −0.27 (0.085) 0.002 HGS −0.055 (0.075) 0.47 −0.19 (0.09) 0.04 −0.13 (0.093) 0.16 Malnutrition, SGA >1 −0.074 (0.16) 0.64 −0.13 (0.14) 0.35 −0.13 (0.15) 0.40 Triceps skinfold −0.025 (0.079) 0.75 0.035 (0.079) 0.65 0.026 (0.078) 0.74 Variable Model 1 P-value Model 2 P-value Model 3 P-value Age 0.08 (0.073) 0.28 0.14 (0.062)a 0.03 0.095 (0.065)a 0.15 Male 0.078 (0.13) 0.56 0.22 (0.13)b 0.08 0.088 (0.14)b 0.53 DM 0.44 (0.13) 0.001 0.3 (0.16) 0.06 CVD 0.14 (0.14) 0.33 0.025 (0.13) 0.85 0.00073 (0.14) 1.00 Smoker, current or former 0.28 (0.17) 0.09 0.097 (0.16) 0.53 0.16 (0.15) 0.26 eGFR −0.052 (0.066) 0.44 −0.074 (0.06) 0.22 −0.097 (0.057) 0.09 BMI 0.018 (0.057) 0.75 0.02 (0.056) 0.71 −0.015 (0.063) 0.82 Lean body mass 0.12 (0.091) 0.18 0.11 (0.11) 0.33 0.13 (0.093) 0.16 Fat body mass −0.12 (0.091) 0.19 −0.11 (0.11) 0.34 −0.13 (0.094) 0.16 FTI −0.0029 (0.081) 0.97 −0.014 (0.085) 0.87 −0.034 (0.082) 0.68 LTI 0.19 (0.074) 0.010 0.18 (0.083) 0.03 0.15 (0.079) 0.06 Height 0.0000 (0.053) 1.00 −0.059 (0.067) 0.38 −0.014 (0.069) 0.84 Cholesterol 0.11 (0.058) 0.06 0.13 (0.066) 0.05 0.14 (0.062) 0.03 Triglycerides 0.016 (0.062) 0.80 0.017 (0.071) 0.82 0.024 (0.064) 0.71 High-density lipoprotein −0.025 (0.058) 0.66 −0.02 (0.058) 0.74 −0.016 (0.053) 0.75 Low-density lipoprotein 0.1 (0.063) 0.11 0.098 (0.056) 0.08 0.11 (0.063) 0.09 Creatinine 0.062 (0.054) 0.25 0.073 (0.061) 0.23 0.097 (0.065) 0.13 Phosphate 0.012 (0.066) 0.86 0.0037 (0.066) 0.96 0.015 (0.072) 0.84 Calcium 0.15 (0.069) 0.03 0.14 (0.057) 0.01 0.21 (0.062) <0.001 Parathyroid hormone −0.013 (0.084) 0.88 0.0073 (0.072) 0.92 0.033 (0.084) 0.70 hsCRP 0.21 (0.077) 0.006 0.2 (0.075) 0.007 0.2 (0.078) 0.010 s-Albumin −0.2 (0.059) <0.001 −0.22 (0.063) <0.001 −0.2 (0.055) <0.001 TNF −0.22 (0.089) 0.02 −0.25 (0.079) 0.002 −0.27 (0.085) 0.002 HGS −0.055 (0.075) 0.47 −0.19 (0.09) 0.04 −0.13 (0.093) 0.16 Malnutrition, SGA >1 −0.074 (0.16) 0.64 −0.13 (0.14) 0.35 −0.13 (0.15) 0.40 Triceps skinfold −0.025 (0.079) 0.75 0.035 (0.079) 0.65 0.026 (0.078) 0.74 Median regression coefficients with standard errors for cardiovascular risk markers on plasma PAPP-A levels. Model 1 is univariable; Model 2 includes age and sex; Model 3 also includes diabetes status. In models marked a and b no adjustment was made for age and sex, respectively. Interval variables are standardized and centered to the mean. Survival analysis Survival by PAPP-A quartiles is presented in Figure 1, showing that survival was worse with increasing quartile. Kaplan–Meier plots with log rank test of potential interactions between dichotomic PAPP-A and hsCRP, LTI, FTI and DM, respectively, showed that median survival time was shorter among patients with high levels of both hsCRP and PAPP-A (Figure 2) and that low PAPP-A was protective only in the high FTI category, while there was no apparent effect of PAPP-A in the low FTI category. Worse survival was seen in the group with simultaneously low LTI and high PAPP-A. Altogether these findings suggest interactions between PAPP-A and hsCRP, FTI and LTI, respectively. A possible interaction with DM was also seen, with high PAPP-A appearing detrimental only in the presence of DM (Figure 2). FIGURE 1 View largeDownload slide Kaplan–Meier diagram of survival in 286 dialysis patients by PAPP-A quartiles. FIGURE 1 View largeDownload slide Kaplan–Meier diagram of survival in 286 dialysis patients by PAPP-A quartiles. FIGURE 2 View largeDownload slide Kaplan–Meier diagram of survival in 286 dialysis patients. Interactions between PAPP-A and LTI (A), FTI (B), hsCRP (C) and DM (D) are shown. Continuous variables were categorized as ‘high’ if above median value and ‘low’ if not. FIGURE 2 View largeDownload slide Kaplan–Meier diagram of survival in 286 dialysis patients. Interactions between PAPP-A and LTI (A), FTI (B), hsCRP (C) and DM (D) are shown. Continuous variables were categorized as ‘high’ if above median value and ‘low’ if not. Multivariable Cox regression models on survival are presented in Table 3. The first model includes PAPP-A in combination with Framingham cardiovascular risk factors. PAPP-A, DM, age, smoking and low-density lipoprotein (LDL) were associated with worse survival. For BMI, the association was negative and CVD appeared detrimental only in the first stratum, i.e. before day 400. PAPP-A levels had a relatively small but robust effect on mortality risk. In Model 2, BMI was replaced by FTI and LTI, which improved model performance and modulated mainly the effects of DM, CVD and LDL. Of the body composition indices only LTI showed a statistically significant association with survival. When hsCRP was also included (Model 3), the association of PAPP-A with survival did not reach statistical significance, although no major changes to the size or direction of coefficients were observed. In a post hoc analysis where the cohort was stratified by diabetes status, PAPP-A was predictive of survival only in patients with diabetes and this association remained when adjusted for Framingham cardiovascular risk factors, LTI, FTI and CRP (Supplementary Tables S7 and S8). Further, when adjusted for GFR in a separate analysis, PAPP-A was independently associated with survival [hazard ratio (HR) = 1.27, 95% confidence interval (CI): 1.09–1.47, P = 0.002]. Table 3 PAPP-A, body composition and survival in 286 incident dialysis patients Variable Model 1 Model 2 Model 3 HR (95% CI) P-value HR (95% CI) P-value HR (95% CI) P-value Age 1.06 (1.03–1.10) <0.001 1.07 (1.04–1.11) <0.001 1.07 (1.04–1.11) <0.001 Male 0.80 (0.48–1.35) 0.41 1.05 (0.53–2.10) 0.88 1.06 (0.53–2.12) 0.87 DM 2.27 (1.35–3.82) 0.002 1.86 (1.02–3.38) 0.04 1.76 (0.96–3.22) 0.07 CVD, strata 1 = day 0–400 3.78 (1.48–9.64) 0.005 3.61 (1.28–10.23) 0.02 3.30 (1.15–9.48) 0.03 CVD, strata 2 = day 400+ 0.62 (0.32–1.18) 0.15 0.54 (0.25–1.17) 0.12 0.53 (0.24–1.15) 0.11 Smoker, current or former 1.74 (0.98–3.11) 0.06 1.83 (0.93–3.60) 0.08 1.79 (0.91–3.52) 0.09 Low-density lipoprotein 1.30 (1.01–1.69) 0.04 1.27 (0.95–1.69) 0.10 1.24 (0.94–1.64) 0.13 PAPP-A 1.21 (1.02–1.42) 0.03 1.19 (1.00–1.41) 0.04 1.19 (0.99–1.41) 0.06 BMI 0.66 (0.51–0.86) 0.002 FTI 0.73 (0.52–1.02) 0.07 0.75 (0.53–1.05) 0.09 LTI 0.66 (0.48–0.91) 0.010 0.68 (0.49–0.94) 0.02 hsCRP 1.20 (0.95–1.52) 0.14 −2 Loglik 635 449 447 Variable Model 1 Model 2 Model 3 HR (95% CI) P-value HR (95% CI) P-value HR (95% CI) P-value Age 1.06 (1.03–1.10) <0.001 1.07 (1.04–1.11) <0.001 1.07 (1.04–1.11) <0.001 Male 0.80 (0.48–1.35) 0.41 1.05 (0.53–2.10) 0.88 1.06 (0.53–2.12) 0.87 DM 2.27 (1.35–3.82) 0.002 1.86 (1.02–3.38) 0.04 1.76 (0.96–3.22) 0.07 CVD, strata 1 = day 0–400 3.78 (1.48–9.64) 0.005 3.61 (1.28–10.23) 0.02 3.30 (1.15–9.48) 0.03 CVD, strata 2 = day 400+ 0.62 (0.32–1.18) 0.15 0.54 (0.25–1.17) 0.12 0.53 (0.24–1.15) 0.11 Smoker, current or former 1.74 (0.98–3.11) 0.06 1.83 (0.93–3.60) 0.08 1.79 (0.91–3.52) 0.09 Low-density lipoprotein 1.30 (1.01–1.69) 0.04 1.27 (0.95–1.69) 0.10 1.24 (0.94–1.64) 0.13 PAPP-A 1.21 (1.02–1.42) 0.03 1.19 (1.00–1.41) 0.04 1.19 (0.99–1.41) 0.06 BMI 0.66 (0.51–0.86) 0.002 FTI 0.73 (0.52–1.02) 0.07 0.75 (0.53–1.05) 0.09 LTI 0.66 (0.48–0.91) 0.010 0.68 (0.49–0.94) 0.02 hsCRP 1.20 (0.95–1.52) 0.14 −2 Loglik 635 449 447 Cox proportional hazards regression models on survival. Interval variables are standardized and centered to the mean. Loglik, log-likelihood. Table 3 PAPP-A, body composition and survival in 286 incident dialysis patients Variable Model 1 Model 2 Model 3 HR (95% CI) P-value HR (95% CI) P-value HR (95% CI) P-value Age 1.06 (1.03–1.10) <0.001 1.07 (1.04–1.11) <0.001 1.07 (1.04–1.11) <0.001 Male 0.80 (0.48–1.35) 0.41 1.05 (0.53–2.10) 0.88 1.06 (0.53–2.12) 0.87 DM 2.27 (1.35–3.82) 0.002 1.86 (1.02–3.38) 0.04 1.76 (0.96–3.22) 0.07 CVD, strata 1 = day 0–400 3.78 (1.48–9.64) 0.005 3.61 (1.28–10.23) 0.02 3.30 (1.15–9.48) 0.03 CVD, strata 2 = day 400+ 0.62 (0.32–1.18) 0.15 0.54 (0.25–1.17) 0.12 0.53 (0.24–1.15) 0.11 Smoker, current or former 1.74 (0.98–3.11) 0.06 1.83 (0.93–3.60) 0.08 1.79 (0.91–3.52) 0.09 Low-density lipoprotein 1.30 (1.01–1.69) 0.04 1.27 (0.95–1.69) 0.10 1.24 (0.94–1.64) 0.13 PAPP-A 1.21 (1.02–1.42) 0.03 1.19 (1.00–1.41) 0.04 1.19 (0.99–1.41) 0.06 BMI 0.66 (0.51–0.86) 0.002 FTI 0.73 (0.52–1.02) 0.07 0.75 (0.53–1.05) 0.09 LTI 0.66 (0.48–0.91) 0.010 0.68 (0.49–0.94) 0.02 hsCRP 1.20 (0.95–1.52) 0.14 −2 Loglik 635 449 447 Variable Model 1 Model 2 Model 3 HR (95% CI) P-value HR (95% CI) P-value HR (95% CI) P-value Age 1.06 (1.03–1.10) <0.001 1.07 (1.04–1.11) <0.001 1.07 (1.04–1.11) <0.001 Male 0.80 (0.48–1.35) 0.41 1.05 (0.53–2.10) 0.88 1.06 (0.53–2.12) 0.87 DM 2.27 (1.35–3.82) 0.002 1.86 (1.02–3.38) 0.04 1.76 (0.96–3.22) 0.07 CVD, strata 1 = day 0–400 3.78 (1.48–9.64) 0.005 3.61 (1.28–10.23) 0.02 3.30 (1.15–9.48) 0.03 CVD, strata 2 = day 400+ 0.62 (0.32–1.18) 0.15 0.54 (0.25–1.17) 0.12 0.53 (0.24–1.15) 0.11 Smoker, current or former 1.74 (0.98–3.11) 0.06 1.83 (0.93–3.60) 0.08 1.79 (0.91–3.52) 0.09 Low-density lipoprotein 1.30 (1.01–1.69) 0.04 1.27 (0.95–1.69) 0.10 1.24 (0.94–1.64) 0.13 PAPP-A 1.21 (1.02–1.42) 0.03 1.19 (1.00–1.41) 0.04 1.19 (0.99–1.41) 0.06 BMI 0.66 (0.51–0.86) 0.002 FTI 0.73 (0.52–1.02) 0.07 0.75 (0.53–1.05) 0.09 LTI 0.66 (0.48–0.91) 0.010 0.68 (0.49–0.94) 0.02 hsCRP 1.20 (0.95–1.52) 0.14 −2 Loglik 635 449 447 Cox proportional hazards regression models on survival. Interval variables are standardized and centered to the mean. Loglik, log-likelihood. When modelling the aforementioned interactions using Cox regression, including one interaction term between PAPP-A and the parameter of interest each in turn, there were significant interactions on the multiplicative scale for hsCRP (HR = 1.6, 95% CI: 1.2–2.2, P = 0.004), DM (HR = 1.6, 95% CI: 1.1–2.2, P = 0.01) and (borderline) for FTI (HR = 1.3, 95% CI: 0.99–1.8, P = 0.06), but not for LTI (HR = 0.93, 95% CI: 0.7–1.2, P = 0.59). Multivariable analysis of these interactions is available in the Supplementary Material and shows that interactions with PAPP-A remained statistically significant for hsCRP and DM (Supplementary Tables S1–S4). Analysis of interactions on the additive scale using relative excess risk due to interaction (RERI) and attributable proportion (AP), showed that FTI and DM interacted with PAPP-A levels on survival (Supplementary Tables S5 and S6). When the cohort was stratified by diabetes status, higher PAPP-A predicted worse survival in diabetic patients, while in nondiabetic patients this association was not seen (Supplementary Tables S7 and S8). Statistical models testing and modifications The proportional hazards assumption was violated for CVD. Thus, a time varying coefficient was included for survival time splitting at 400 days. When removing three influential observations, LDL (HR = 1.35, P = 0.05) and the second strata of CVD (HR = 0.42, P < 0.05) became somewhat more strongly associated with survival. These influential cases were kept in the models presented. DISCUSSION Our study shows that higher PAPP-A levels are associated with worse survival in incident dialysis patients and that this association remains when adjusted for established cardiovascular risk factors and body composition but is attenuated by hsCRP. This accords with previous results in prevalent HD patients [7, 10, 13, 14]. However, in contradiction to hypotheses presented in a previous study [10] we did not find markers of PEW, such as low LTI, to be associated with elevated PAPP-A. On the contrary, LTI was positively associated with PAPP-A levels. DM appears to be a confounder in this context and in the study by Kalousová et al. [10] only diabetic patients were included. In apparent conflict with these results, HGS was negatively associated with PAPP-A when adjusted for age and sex, perhaps reflecting age-related dissociation between muscle mass and muscle strength [24]. It is worth pointing out that neither BMI nor fat body mass or lean body mass percentages predicted PAPP-A levels. BMI may not reflect opposing changes in fat and lean tissue mass [25] while FBM and LBM indices are interdependent when expressed as percentages of body weight, making interpretations of separate effects of lean and fat tissue components difficult. FTI and LTI could therefore be more appropriate indicators when investigating the relative effects of lean and fat mass. In addition to different inclusion criteria, changes in body composition during the course of dialysis treatment [26], along with selection due to high mortality rates, could explain discrepancies between our results and those from studies on prevalent HD patients. Furthermore, analysis of interactions on the additive scale showed a statistically significant positive interaction between PAPP-A and FTI, suggesting that PAPP-A has a stronger association with survival in persons with higher FTI. Whereas PAPP-A was positively associated with hsCRP, a negative correlation to TNF was observed. This is surprising since TNF stimulates PAPP-A expression in cell models [27]. Although it could reflect a complex interplay between PAPP-A and different inflammatory pathways, confounding or spurious findings could not be ruled out. In line with previous findings [11], we observed an interaction between inflammation and PAPP-A in that the mortality predictive impact of PAPP-A was enhanced when inflammation was present. This observation supports a catalytic effect of inflammation for the risk factor profile [28]. We also report that elevated PAPP-A might be especially detrimental in diabetic dialysis patients or possibly prognostic only in this group. However, this finding should be regarded with caution since confounding from comorbid conditions may be present and due to the stratified analysis being post hoc. GH axis components, including IGF-1 and PAPP-A, have been linked to atherosclerosis but we found no association between PAPP-A levels and history of CVD. Due to differences in case-mix along with GH axis disturbances in ESRD, it may not be appropriate to extrapolate from findings in other populations or experimental models. Furthermore, systemic levels of PAPP-A may not reflect local PAPP-A activity associated with atherosclerotic pathways in animal models and it is unclear if plasma levels of PAPP-A represent local up-regulation or increased release due to tissue damage. Neither do plasma levels reveal from what organ PAPP-A is released. In a previous study on HD patients, PAPP-A predicted death by infection, stroke or sudden cardiac death rather than myocardial infarction [10]. Despite pathophysiological links between local PAPP-A activity and CVD, plasma PAPP-A in ESRD could hypothetically be a biomarker for conditions associated with increased risk for cerebrovascular or infectious complications rather than coronary artery disease. A number of caveats should be considered when interpreting the results of the present study. Although the mean of renal urea and creatinine clearances during 24-h was used as a measure of GFR, and we cannot exclude association of GFR with muscle mass, adjustment for residual renal function would be inappropriate since associations to lean tissue was part of the research question. Furthermore, as residual renal function is positively associated with both survival and PAPP-A, including it in the statistical models is unlikely to weaken the associations between PAPP-A and mortality. Also, in the present study we did not find any association between PAPP-A and GFR (Table 2). Further, the relative small sample size limits the number of confounders and interactions that can be investigated in multivariable analysis. Although we have adjusted for major CVD risk factors, the broad inclusion criteria increase the possibility of residual confounding from comorbid conditions. Ideally, studies with larger sample size would allow for more adjustment for the significant case-mix present in the dialysis population. In summary, a higher circulating concentration of PAPP-A is associated with inflammatory markers and predicts mortality risk in incident dialysis. Body composition, inflammation and DM are effect modifiers when modelling the effects of PAPP-A on survival. SUPPLEMENTARY DATA Supplementary data are available online at http://ndt.oxfordjournals.org. ACKNOWLEDGEMENTS We are grateful to the patients and control subjects participating in the study. We thank Åsa Lindé and Annika Nilsson, Anki Emmoth and Ulrika Jensen for collection of samples, and Ann-Christin Bragfors-Helin and Monica Eriksson for laboratory analyses. Baxter Novum is the result of a grant from Baxter Healthcare Corporation to the Karolinska Institutet. FUNDING This study was supported by grants from The Swedish Research Council (grant 521-2013-2764), Westman's Foundation (no grant number) and The Swedish Kidney Foundation (no grant number). CONFLICT OF INTEREST STATEMENT B.L. is employed by and has received grants, lecturing fees, consultancy fees and travel funding from Baxter Healthcare. P.S. has received grants from Bayer, lecturing fees from AbbVie, Shire, Bayer, Pfizer and Asahi, and has been a member of Scientific Advisory Boards of ARO group (Amgen), Vifor, Keryx and Astellas. REFERENCES 1 Root A. Growth hormone . Pediatrics 1965 ; 36 : 940 – 950 Google Scholar PubMed 2 Palmeiro CR , Anand R , Dardi IK et al. . Growth hormone and the cardiovascular system . Cardiol Rev 2012 ; 20 : 197 – 207 Google Scholar CrossRef Search ADS PubMed 3 Higashi Y , Sukhanov S , Anwar A et al. . Aging, atherosclerosis, and IGF-1 . J Gerontol A Biol Sci Med Sci 2012 ; 67A : 626 – 639 Google Scholar CrossRef Search ADS 4 Burgers AMG , Biermasz NR , Schoones JW et al. . Meta-analysis and dose-response metaregression: circulating insulin-like growth factor 1 (IGF-1) and mortality . J Clin Endocrinol Metab 2011 ; 96 : 2912 – 2920 Google Scholar CrossRef Search ADS PubMed 5 Jia T , Gama Axelsson T , Heimbürger O et al. . IGF-1 and survival in ESRD . Clin J Am Soc Nephrol 2014 ; 9 : 120 – 127 Google Scholar CrossRef Search ADS PubMed 6 Maile LA , Holly JMP. Insulin-like growth factor binding protein (IGFBP) proteolysis: Occurrence, identification, role and regulation . Growth Horm IGF Res 1999 ; 9 : 85 – 95 Google Scholar CrossRef Search ADS PubMed 7 Kalousova M , Benakova H , Kubena AA et al. . Pregnancy-associated plasma protein A as an independent mortality predictor in long-term hemodialysis patients . Kidney Blood Press Res 2012 ; 35 : 192 – 201 Google Scholar CrossRef Search ADS PubMed 8 Murray RD , Adams JE , Shalet SM. Adults with partial growth hormone deficiency have an adverse body composition . J Clin Endocrinol Metab 2004 ; 89 : 1586 – 1591 Google Scholar CrossRef Search ADS PubMed 9 Conover CA , Bale LK , Marler RJ. Pregnancy-associated plasma protein-A deficiency improves survival of mice on a high fat diet . Exp Gerontol 2015 ; 70 : 131 – 134 Google Scholar CrossRef Search ADS PubMed 10 Kalousová M , Zima T , Krane V et al. . Pregnancy-associated plasma protein A associates with cardiovascular events in diabetic hemodialysis patients . Atherosclerosis 2014 ; 236 : 263 – 269 Google Scholar CrossRef Search ADS PubMed 11 Consuegra-Sanchez L , Fredericks S , Kaski JC. Pregnancy-associated plasma protein-A (PAPP-A) and cardiovascular risk . Atherosclerosis 2009 ; 203 : 346 – 352 Google Scholar CrossRef Search ADS PubMed 12 Mueller T , Dieplinger B , Poelz W et al. . Increased pregnancy-associated plasma protein-A as a marker for peripheral atherosclerosis: results from the Linz peripheral arterial disease study . Clin Chem 2006 ; 52 : 1096 – 1103 Google Scholar CrossRef Search ADS PubMed 13 Kalousova M , Horejsi M , Fialova L et al. . Increased levels of pregnancy-associated plasma protein A are associated with mortality in hemodialysis patients: preliminary results . Blood Purif 2004 ; 22 : 298 – 300 Google Scholar CrossRef Search ADS PubMed 14 Etter C , Straub Y , Hersberger M et al. . Pregnancy-associated plasma protein-A is an independent short-time predictor of mortality in patients on maintenance haemodialysis . Eur Heart J 2010 ; 31 : 354 – 359 Google Scholar CrossRef Search ADS PubMed 15 Jefferies HJ , Tertti R , Wittfooth S et al. . Elevated serum free pregnancy-associated plasma protein-A independently predicts mortality in haemodialysis patients but is not associated with recurrent haemodialysis-induced ischaemic myocardial injury . Nephron 2015 ; 129 : 171 – 178 Google Scholar CrossRef Search ADS PubMed 16 Eckardt K-U , Gillespie IA , Kronenberg F et al. . High cardiovascular event rates occur within the first weeks of starting hemodialysis . Kidney Int 2015 ; 88 : 1117 – 1125 Google Scholar CrossRef Search ADS PubMed 17 Stenvinkel P , Heimbürger O , Paultre F et al. . Strong association between malnutrition, inflammation, and atherosclerosis in chronic renal failure . Kidney Int 1999 ; 55 : 1899 – 1911 Google Scholar CrossRef Search ADS PubMed 18 Kyle UG , Schutz Y , Dupertuis YM , Pichard C. Body composition interpretation: contributions of the fat-free mass index and the body fat mass index . Nutrition 2003 ; 19 : 597 – 604 Google Scholar CrossRef Search ADS PubMed 19 VanItallie TB , Yang MU , Heymsfield SB et al. . Height-normalized indices of the body’s fat-free mass and fat mass: potentially useful indicators of nutritional status . Am J Clin Nutr 1990 ; 52 : 953 – 959 Google Scholar CrossRef Search ADS PubMed 20 Qureshi AR , Alvestrand A , Danielsson A et al. . Factors predicting malnutrition in hemodialysis patients: a cross-sectional study . Kidney Int 1998 ; 53 : 773 – 782 Google Scholar CrossRef Search ADS PubMed 21 R Development Core Team . R: A Language and Environment for Statistical Computing . Vienna, Austria : R Foundation for Statistical Computing , 2016 22 Therneau TM. Survival: Survival analysis. 2015 . https://CRAN.R-project.org/package=survival (30 May 2017, date last accessed) 23 Koenker R. Quantreg: Quantile regression. 2016 . https://CRAN.R-project.org/package=quantreg (30 May 2017, date last accessed) 24 Clark BC , Manini TM. Sarcopenia =/= dynapenia . J Gerontol Ser A 2008 ; 63 : 829 – 834 Google Scholar CrossRef Search ADS 25 Marcelli D , Brand K , Ponce P et al. . Longitudinal changes in body composition in patients after initiation of hemodialysis therapy: Results from an international cohort . J Ren Nutr 2016 ; 26 : 72 – 80 Google Scholar CrossRef Search ADS PubMed 26 Chumlea WC. Anthropometric and body composition assessment in dialysis patients . Semin Dial 2004 ; 17 : 466 – 470 Google Scholar CrossRef Search ADS PubMed 27 Conover CA , Bale LK , Harrington SC et al. . Cytokine stimulation of pregnancy-associated plasma protein A expression in human coronary artery smooth muscle cells: inhibition by resveratrol . Am J Physiol Cell Physiol 2006 ; 290 : C183 – C188 Google Scholar CrossRef Search ADS PubMed 28 Carrero JJ , Stenvinkel P. Persistent inflammation as a catalyst for other risk factors in chronic kidney disease: a hypothesis proposal . Clin J Am Soc Nephrol 2009 ; 4 : S49 – S55 Google Scholar CrossRef Search ADS PubMed © The Author 2017. 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: Jul 24, 2017

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