Advertisement

Advanced glycation end products and mortality in hemodialysis patients

      Advanced glycation end products and mortality in hemodialysis patients.

      Background

      Advanced glycation and oxidation end products (AGEs) cause oxidative stress and trigger cytokine driven inflammatory reactions in vitro. The net effects on markers of inflammation and acute phase proteins in vivo as well as their influence on survival in hemodialysis patients are unknown.

      Methods

      We conducted a cross-sectional study in 312 stable hemodialysis patients and analyzed the interrelationships of AGEs and C-reactive protein (CRP) and their predictive effect on all-cause as well as cardiovascular mortality. Mortality was monitored prospectively over a period of 32 months. AGEs were determined by measuring total serum fluorescent AGEs (AGE-fl) and Ne-(carboxymethyl)-lysine (CML).

      Results

      The levels of AGE-fl, CML and CRP were 3.2-, 3.8- and 10-fold higher as compared to healthy controls. AGE-fl and CML levels correlated significantly with each other but not with CRP or serum albumin. Patients with high (above median values) AGE-fl or CML levels (109 × 103 AU and 1.4 μg/mL, respectively) had a significant better survival than those with low (below median values) AGE-fl or CML levels. Patients with high CRP levels (above 7.7 mg/L = median value) had a better survival than those with low CRP (below median value) when AGE-fl or CML levels were high in parallel.

      Conclusions

      In contrast to in vitro data and to current hypotheses, the presence of high serum AGEs, as measured by AGE-fl and CML, were not linked to increased mortality. Statistically, high serum AGEs partly overcame the negative impact of the acute phase response on mortality in hemodialysis patients. Whether the benefit of high serum AGEs is an epiphenomenon or reflects a better nutritional support needs further studies.

      Keywords

      The causes for the high frequency of cardiovascular disease and high incidence of cardiovascular mortality in dialysis patients are multifactorial in origin. Disturbances in the carbohydrate and lipid metabolism, the balance between oxidants and antioxidants and the immuno-inflammatory system are all thought to play a role. More specifically, chronic uremia is characterized by the accumulation of advanced glycation end products (AGEs) as well as an activation of the acute phase response, for which C-reactive protein (CRP) is the prototype in humans. AGEs are diverse fluorescent and non-fluorescent compounds that are formed in complex not yet fully understood reactions starting with the non-enzymatic reaction of reducing sugars or other carbonyl compounds with free amino groups of proteins. Only a few physiologically occurring AGEs have been characterized chemically, such as Ne-(carboxymethyl)-lysine (CML), a colorless and nonfluorescent compound
      • Westwood M.E.
      • Thornalley P.J.
      The glycation hypothesis,.
      . AGEs accumulate during aging and in the course of many degenerative diseases
      • Brownlee M.
      Advanced protein glycosylation in diabetes and aging.
      . Makita and co-workers first noted in 1991 that end-stage renal disease patients had dramatically increased serum levels of unspecific AGE-modified peptides
      • Makita Z.
      • Radoff S.
      • Rayfield E.J.
      • et al.
      Advanced glycosylation end products in patients with diabetic nephropathy.
      . Meanwhile, numerous studies measuring levels of total unspecific AGEs, AGE-dependent fluorescence, CML and in particular pentosidine have confirmed these results
      • Dawnay A.
      • Millar D.J.
      The pathogenesis and consequences of AGE formation in uraemia and its treatment.
      ,
      • Schinzel R.
      • Münch G.
      • Heidland A.
      Advanced glycation endproducts in end-stage renal disease and their removal.
      . Since the kidney has been postulated to be responsible for the removal of AGEs
      • Miyata T.
      • Ueda Y.
      • Horie K.
      • et al.
      Renal catabolism of advanced glycation end products: The fate of pentosidine.
      ,
      • Gugliucci A.
      • Bendayan M.
      Renal fate of circulating advanced glycated end products (AGE): Evidence for reabsorption and catabolism of AGE-peptides by renal proximal tubular cells.
      , the accumulation of AGEs in chronic renal insufficiency has been partly attributed to the progressive loss in renal function. On the other hand, increased de novo generation due to enhanced oxidative stress and accumulation of reactive precursors such as dicarbonyls also seem to be implicated
      • Miyata T.
      • Van Ypersele de Strihou C.
      • Kurokowa K.
      • Baynes J.W.
      Alterations in nonenzymatic biochemistry in uremia: Origin and significance of “carbonyl stress” in long-term uremic complications.
      .
      Advanced glycation end-products may lead to progression of atherosclerosis by interaction with receptors that induce production of free oxygen radicals followed by a release of inflammatory cytokines in different cell types
      • Raj D.S.
      • Choudhury D.
      • Welbourne T.C.
      • Levi M.
      Advanced glycation end products: A Nephrologist's perspective.
      ,
      • Bierhaus A.
      • Hofmann M.A.
      • Ziegler R.
      • Nawroth P.P.
      AGEs and their interaction with AGE-receptors in vascular disease and diabetes mellitus. I. The AGE concept.
      such as interleukin-6 (IL-6), ultimately leading to the production of CRP by the liver. A recent observation suggested that glycoaldehyde generated by the myeloperoxidase-H2O2-chloride system involving activated phagocytes might lead to CML production, and thereby play an important role in tissue damage at sites of inflammation as encountered in atherosclerosis. Besides the classical pathway, CML may be formed by glyoxal generated during lipid peroxidation or glucose autoxidation
      • Fu M.X.
      • Requena J.R.
      • Jenkins A.J.
      • et al.
      The advanced glycation end product, Nepsilon-(carboxymethyl)lysine, is a product of both lipid peroxidation and glycoxidation reactions.
      ,
      • Thornalley P.J.
      Pharmacology of methylglyoxal: Formation, modification of proteins and nucleic acids, and enzymatic detoxification—A role in pathogenesis and antiproliferative chemotherapy.
      . Moreover, co-staining of CML and macrophages was found in atherosclerotic lesions where CRP is deposited as well
      • Schleicher E.D.
      • Wagner E.
      • Nerlich A.G.
      Increased accumulation of the glycoxidation product N(epsilon)-(carboxymethyl)lysine in human tissues in diabetes and aging.
      . It also has been proposed that CRP may play a part in the recruitment of monocytes during atherosclerosis
      • Torzewski M.
      • Rist C.
      • Mortensen R.F.
      • et al.
      C-reactive protein in the arterial intima: Role of C-reactive protein receptor-dependent monocyte recruitment in atherogenesis.
      . While the finding of CRP in the atherosclerotic lesion is of yet unclear pathobiological meaning, the predictive association between serological CRP and myocardial events, cardiovascular mortality, ischemic stroke and peripheral vascular disease has been extensively described in the general population
      • Danesh J.
      • Whincup P.
      • Walker M.
      • et al.
      Low grade inflammation and coronary heart disease: prospective study and updated meta-analyses.
      . We demonstrated that elevated CRP was the most powerful predictor of all-cause and cardiovascular mortality in hemodialysis patients
      • Zimmermann J.
      • Herrlinger S.
      • Pruy A.
      • et al.
      Inflammation enhances cardiovascular risk and mortality in hemodialysis patients.
      .
      In our current study, we hypothesized that circulating AGEs may promote the inflammatory status seen in hemodialysis patients
      • Schwedler S.
      • Schinzel R.
      • Vaith P.
      • Wanner C.
      Inflammation and advanced glycation end products in uremia: Simple coexistence, potentiation or causal relationship?.
      , subsequently leading to high concentrations of CRP in serum and to cardiovascular mortality. We therefore investigated the relationship between AGEs and CRP, and analyzed whether total serum fluorescent AGEs (AGE-fl) and CML were predictors of overall and cardiovascular mortality in hemodialysis patients.

      METHODS

       Study population

      A total of 312 Caucasian patients on chronic hemodialysis treatment from the dialysis unit of the University Clinic of Würzburg and three other outpatient dialysis centers were prospectively observed for 32 months, from August 1997 until April 2000. The study was carried out according to the principles of the Declaration of Helsinki. All patients gave their informed consent before entering the study, and the study protocol was approved by the appropriate institutional review board. Data on age, sex, body mass index (BMI; weight in kilograms divided by the square of the height in meters), primary renal disease (diabetes, hypertension, other), time since initiation of hemodialysis (years) and weekly dialysis time (hours/week) were recorded. Sixty middle-aged healthy subjects served as controls.

       Baseline examination

      Patients were examined at baseline and no clinical signs of overt infection were present. The presence of coronary artery disease (CAD), peripheral vascular disease (PVD) and cerebral vascular disease (CVD) was assessed by interviews and by reviewing clinical charts. Patients with a history of myocardial infarction or angina pectoris were classified as having coronary artery disease. A previous myocardial infarction was defined by the presence of chest pain, a confirmatory electrocardiogram, enzyme courses or pathological findings from coronary angiography. Angina pectoris was diagnosed when either a stress test was positive and/or a coronary stenosis>75% of luminal diameter was found by coronary angiography. Peripheral vascular disease was diagnosed by the presence of claudicatio and diminished pulses on clinical examination combined with duplex sonography or angiography. Cerebral vascular disease was defined as a stroke or history of transient ischemic attacks and positive findings in computed tomography or magnetic resonance imaging, or an acute functional deficit lasting for 24 hours or less. A patient was considered to have vascular disease (VD) if at least one of these three defined diseases was present.
      Smoking was defined on the basis of current smoking or previous habitual smoking. Hypertension was defined as a predialysis systolic blood pressure>140 mm Hg or a diastolic pressure>90 mm Hg or the use of antihypertensive drugs. Diabetes was defined by glucose values of 200 mg/dL and higher at any time or fasting glucose values ≥126 mg/dL or the use of insulin or oral antidiabetic drugs.

       Follow-up study

      Data of all hemodialysis patients from the original cohort were followed for analysis of mortality. Date and cause of death were evaluated from records and from interviews with the physicians in the outpatient dialysis centers. Causes of death were classified as mortality due to cardiovascular events (myocardial infarction, congestive heart failure, sudden death and stroke) and non-cardiovascular events (infection, neoplasma, unknown causes).

       Sampling procedure and laboratory analysis

      Blood samples were collected on ice (4°C) prior to initiation of the hemodialysis session and prior to heparin administration and were centrifuged within 60 minutes after collection. Measurements of CRP, albumin, fibrinogen, apolipoproteins and lipids were performed on fresh samples. Serum samples for measurement of AGE were stored at -70°C. The serum concentration of CRP was measured by a nephelometric immunoassay (NA Latex CRP Reagent; Behring Institute, Marburg, Germany). Creatinine and blood urea nitrogen (BUN) were determined by enzymatic methods. Measurement of serum albumin was carried out using the bromocresol green method. Fibrinogen was measured by the thrombin time method (Clauss) in a blood sample anticoagulated with 3.8% sodium citrate. Apolipoproteins A-1 (apo A-1), A-2 (apo A-2) and B (apo B) were measured by nephelometric methods. Total cholesterol and triglycerides were measured by enzymatic-colorimetric methods (cholesterol CHOD-PAP, triglycerides PAP; Roche Diagnostics Mannheim, Mannheim, Germany). High-density lipoprotein cholesterol (HDL-C) was measured after precipitation with MgCl2 and low-density lipoprotein cholesterol (LDL-C) was calculated using the Friedewald formula (in patients with triglycerides <400 mg/dL).

       AGE measurements

      In thawed serum samples we assessed CML concentrations and AGE-fl. CML was measured in triplicate with competitive ELISA using monoclonal antibody against CML (Alteon Inc., New York, NY, USA) developed by Roche Diagnostics as described earlier
      • Wagner Z.
      • Wittmann I.
      • Mazák I.
      • et al.
      Serum Ne-(carboxymethyl)lysine levels in type 2 diabetic patients—Impact on renal function.
      . Prior to determination of CML content, serum was digested with proteinase K (1 mg/mL) for three hours at 37°C. Reaction was stopped by adding 2 mmol/L phenylmethylsulfonyl fluoride (PMSF; Roche Diagnostics). To estimate the absolute CML concentration of digested and diluted serum (20-fold) N-(carboxymethyl)amino-caproic acid was used as the standard (Alteon, Inc.). Microtiter enzyme-linked immunosorbent assay (ELISA) plates were incubated with AGE-bovine serum albumin (AGE-BSA; 100 ng/well) for one hour followed by three washing steps. Thereafter, serum and CML standards were incubated with peroxidase-conjugated monoclonal antibody against CML for two hours. After three washing steps color reaction was induced with 2,2′-azino-di-3-ethylbenzthiazoline-sulfonic acid (Roche Diagnostics). Absorbance was read in a microtiter ELISA plate reader (Multiscan Ascent; Labsystems, Helsinki, Finland) at 405 nm (reference 603 nm). All steps were carried out at room temperature (RT). Results were expressed as μg CML/mL.
      Fluorescent AGE intensity of 50-fold diluted serum samples was measured in triplicate at 460 nm after excitation at 355 nm, using a fluorescence reader (Victor 1; Wallace, Turku, Finland)
      • Münch G.
      • Keis R.
      • Wessels A.
      • et al.
      Determination of advanced glycation end products in serum by fluorescence spectroscopy and competitive ELISA.
      . Serum AGE-fl was expressed as relative fluorescence intensity in arbitrary units (AU). Hemolytic sera were excluded from AGE measurements (N = 10). For five other patients serum was not sufficient to perform both CML and AGE-fl measurements in triplicates, and so for these patients only CML measurements were done.

       Statistical analysis

      Comparisons between all pairs were performed using the Mann-Whitney U test for non-normally distributed variables, comparisons between more than two groups were performed using the Kruskal-Wallis H test. Bivariate regression and correlation analyses were performed between all available data by using the Spearman rank method. The cumulative incidence of death during follow-up was estimated by the Kaplan-Meier method. Relative risk of mortality for different parameters was estimated using the Cox proportional hazards model. Relative risks and their 95% confidence intervals (95% CI) were calculated with the use of the estimated regression coefficients and their standard errors in the Cox regression analysis. In addition, chi-square statistics were given. Data are expressed as mean ± SD. P values less than 0.05 were considered significant. Data analysis was performed using the SPSS/PC+ package (SPSS Inc., Chicago, IL, USA).

      RESULTS

       Clinical and biochemical characteristics

      Twenty-eight percent of the 312 patients suffered from diabetes mellitus type 1 or 2. The primary renal diagnosis was split among diabetes (23%), hypertension (33%) and other causes. The mean time on dialysis was 6.1 ± 4.8 years (range 3 to 299 months). Weekly dialysis time was 12.5 ± 1.4 hours. Hemodialysis treatment was performed using conventional bicarbonate-buffered dialysate in all patients. Dialyzer membranes used were made of polysulfone (47%), polyamide (14%), cellulose (38%) and other materials (1%). Dialyzer membranes with an ultrafiltration coefficient>20 mL/(mm Hg × h) were considered as high flux, all those below this value as low flux membranes. Based on this definition 72% of all patients were dialyzed with low flux and 28% with high flux membranes. Medication of patients included antihypertensive drugs (64%), phosphate binders (calcium acetate or carbonate and/or aluminum hydroxide), vitamin B, C and D supplementation, epoetin, iron gluconate, aspirin (37%), and lipid lowering agents (19%). Aspirin was given in a dose of 100 to 300 mg/day.
      Table 1 depicts clinical and biochemical characteristics of patients and controls. Hemodialysis patients were significantly older and had a lower BMI than controls. In both groups gender was equally distributed. Forty percent of patients suffered from vascular disease (25% CAD, 14% PVD, 11% CVD) and 65% had hypertension. Smoking habits were similar in patients and controls (21 vs. 32%). Patients had considerably higher CRP (10.2-fold) and CML (3.8-fold) levels and higher AGE-fl (3.2-fold) than controls. In addition, they had higher levels of fibrinogen, cholesterol, triglycerides, LDL-C, but lower levels of albumin, apo B, apo A-1, apo A-2 and HDL-C. Smokers (N = 65) had significantly higher AGE-fl (119.0 ± 32.9 vs. 105.8 ± 31.6 AU × 103, P < 0.01) but similar CML (1.51 ± 0.65 vs. 1.43 ± 0.60 μg/mL) levels compared to non-smokers (N = 247). There was a trend toward higher CRP levels in smokers (21.4 ± 31.6 vs. 14.9 ± 23.5 mg/L) but this was not statistically significant. In controls no differences were found between smokers and non-smokers concerning AGE-fl and CML levels.
      Table 1Clinical and laboratory characteristics of hemodialysis patients and controls
      CharacteristicsAll patients (N = 312)Controls (N = 60)
      Age years range62.7±12.842.9±11
      P < 0.0001
      27–8630–67
      Gender (M/F) %54/4653/47
      BMI kg/m224.1±4.625.2±3.3
      P < 0.0001
      CRP mg/L16.3±25.11.6±2.4
      P < 0.0001
      Albumin g/L40.5±4.147.3±2.4
      P < 0.0001
      Fibrinogen g/L4.1±1.32.8±0.5
      P < 0.0001
      AGE-fl AU (×103)108.7±32.133.8±5.2
      P < 0.0001
      CML μg/mL1.44±0.610.38±0.1
      P < 0.0001
      Creatinine mg/dL9.5±2.71.0±0.1
      P < 0.0001
      BUN mg/dL63±1716±4
      P < 0.0001
      Apo B mg/dL100±28101±21
      P < 0.0001
      Apo A-1 mg/dL127±23166±27
      P < 0.0001
      Apo A-2 mg/dL33±1438±12
      P < 0.0001
      Cholesterol mg/dL194±47187±33
      P < 0.0001
      Triglycerides mg/dL227±181166±121
      P < 0.0001
      LDL-C mg/dL116±3698±28
      P < 0.0001
      HDL-C mg/dL35±1358±15
      P < 0.0001
      Data are given as mean ± SD. Abbreviations are: BMI, body mass index; CRP, C-reactive protein; AGE-fl, fluorescent advanced glycation end-products; CML, Ne-(carboxymethyl)-lysine; BUN, blood urea nitrogen; Apo, apolipoprotein; LDL-C, low-density lipoprotein cholesterol; HDL-C, high-density lipoprotein cholesterol.
      a P < 0.0001
      Diabetic patients (N = 87) had a higher BMI (25.0 ± 4.7 vs. 23.7 ± 4.4 kg/m2, P < 0.01) and a higher incidence of hypertension (74 vs. 61%, P < 0.05), CAD (47 vs. 17%, P < 0.001), PVD (34 vs. 6%, P < 0.001) and CVD (17 vs. 8%, P < 0.05) as compared to non-diabetic patients (N = 225). Taken together, 69% of diabetic and 28% of non-diabetic patients had vascular disease (P < 0.001). While smoking habits, CRP (16.3 ± 22.2 vs. 16.3 ± 26.3 mg/L), AGE-fl (107.7 ± 27.6 vs. 108.8 ± 33.8 AU × 103) and fibrinogen levels were similar, albumin and CML (1.24 ± 0.55 vs. 1.53 ± 0.62 μg/mL, P < 0.001) levels were lower in diabetic patients as compared to non-diabetic patients.

       Clinical and biochemical characteristics of patients according to AGE-fl and CML levels

      Table 2 shows the data of patients divided according to high and low AGE-fl and CML levels. Patients with high AGE-fl were significantly younger and had higher albumin and CML levels than patients with low AGE-fl. The same incidences of vascular disease, diabetes and smoking were observed.
      Table 2Clinical and laboratory characteristics of patients according to serum levels of AGE-fl and CML (below and above the median)
      AGE-flCML
      Characteristics≤median (N = 148)>median (N = 149)≤median (N = 151)>median (N = 151)
      Age years64.4±12.760.9±13.1b63.7±11.361.6±14.3
      BMI kg/m224.5±4.824.0±4.124.7±4.423.5±4.5b
      CRP mg/L17.6±25.414.3±20.618.1±25.913.4±19.5
      P < 0.05
      Albumin g/L39.7±4.541.3±3.4
      P < 0.01
      39.8±4.441.2±3.6
      P < 0.05
      Fibrinogen g/L4.1±1.24.1±1.34.2±1.24.0±1.3
      P < 0.05
      AGE-fl AU (×103)102.5±30.6115.3±32.5
      P < 0.01
      CML μg/mL1.31±0.571.56±0.62
      P < 0.01
      Creatinine mg/dL9.0±2.910.0±2.48.8±2.510.3±2.7
      P < 0.001 between AGE-fl and CML above and below the median
      BUN mg/dL62±1864±1760±1666±18
      P < 0.01
      Vascular disease42374534
      P < 0.05
      Diabetes29283819
      P < 0.001 between AGE-fl and CML above and below the median
      Smoking16251922
      Aspirin medication37404235
      High flux membrane30233025
      Data are given as mean ± SD. Data of AGE-fl are given in arbitrary units (AU) × 103 with a median of 109; data of CML are given in μg/mL with a median of 1.4. Vascular disease, diabetes, smoking, aspirin medication, high flux membrane are given in %. Abbreviations are in Table 1.
      a P < 0.05
      b P < 0.01
      c P < 0.001 between AGE-fl and CML above and below the median
      In the group with high CML levels a lower BMI, lower CRP and fibrinogen levels, but higher albumin, creatinine and BUN levels and AGE-fl were found. The incidence of vascular disease and diabetes was significantly lower. No differences in smoking habits were seen. In all four subgroups aspirin medication and high flux membrane use were similar.
      Patients with very low serum albumin (≤35 g/L, N = 28%) also had very low AGE-fl (97.8 ± 28.4 vs. 109.7 ± 32.3 × 103 AU, P = 0.055), CML (1.07 ± 0.56 vs. 1.48 ± 0.61 μg/mL, P < 0.001) in comparison to those patients with albumin>35 g/L.

       Analysis of mortality rates according to high and low AGE-fl, CML and CRP

      During the follow-up period of 32 months 34% of patients (N = 105) had died, which corresponds to an annual mortality rate of 13%. Forty-seven percent were cardiovascular and 53% were non-cardiovascular deaths. At the beginning of the observation period the calculated mean survival time per patient was 25.1 months. Presence of vascular disease (VD) and diabetes decreased survival significantly [22.3 months vs. 27 months for VD (P < 0.01), 22.6 months vs. 26 months for diabetes (P < 0.001)].
      The all-cause as well as cardiovascular mortality rates were significantly higher in patients with low AGE-fl as compared to those with high AGE-fl (44 vs. 23%, P < 0.0001, and 20 vs. 12%, P < 0.05, respectively). Thus, the expected lifespan for patients with high AGE-fl at time zero increased from 22.8 months (for patients with low AGE-fl) up to 27.8 months Figure 1a.
      Figure thumbnail gr1
      Figure 1Kaplan-Meier estimates of survival during follow-up with regard to all-cause mortality in relation to high (↑, above median values) or low (↓, below median values) fluorescent advanced glycation end-products (AGE-fl; A) and Ne-(carboxymethyl)-lysine (CML; B).
      A higher all-cause mortality rate also was observed in patients with low CML levels as compared to those with high CML levels (38 vs. 27%, P < 0.05; Figure 1b). The mean survival time was 24.1 and 26.5 months, accordingly. There was a tendency toward higher cardiovascular mortality (19 vs. 12%, P = 0.08) in the former group, as compared to the latter.
      For CRP, the opposite was observed. All-cause and cardiovascular mortality rates were significantly higher in patients with high CRP levels as compared to those with low CRP levels (46 vs. 21%, P < 0.0001, and 21 vs. 11%, P < 0.01, respectively). This observation was confirmed when comparing the relative risk of all-cause and cardiovascular mortality of patients divided in quartiles Table 3. For CRP the relative risk of death in general as well as death caused by cardiovascular events increased with each quartile of baseline concentration of CRP, while for AGE-fl it decreased. For CML a similar trend was present for all-cause mortality. Thus, patients in the highest CRP quartile had a 6.1-fold higher risk for all-cause and a 5.2-fold higher risk for cardiovascular mortality during follow-up than those in the lowest quartile. By contrast, patients in the highest AGE-fl quartile had a 0.36-fold lower risk for all-cause and a 0.28-fold lower risk for cardiovascular mortality than those in the lowest AGE quartile.
      Table 3Relative risks of mortality after 32 months according to quartiles of baseline serum concentrations of CRP, AGE-fl and CML
      All-cause mortalityCardiovascular mortality
      Risk factorNRelative risk (95% CI)P valueRelative risk (95% CI)P value
      CRP mg/L
       <3.47911
       3.4–7.7761.96 (0.97–3.99)0.061.61 (0.61–4.22)0.34
       7.8–17.2772.18 (1.09–4.39)0.0281.31 (0.48–3.62)0.60
       >17.2786.09 (3.23–11.46)<0.0015.20 (2.24–12.08)<0.0001
      AGE-fl AU (×103)
       <83.77411
       83.7–109.0740.97 (0.60–1.58)0.910.67 (0.32–1.38)0.28
       109.1–131.0740.53 (0.30–0.92)0.0240.53 (0.25–1.11)0.09
       >131.0730.36 (0.19–0.67)0.0010.28 (0.11–0.70)0.007
      CML μg/mL
       <0.997411
       0.99–1.40760.71 (0.42–1.19)0.191.16 (0.55–2.44)0.71
       1.41–1.86760.51 (0.29–0.89)0.0190.50 (0.20–1.26)0.14
       >1.86740.61 (0.36–1.05)0.080.81 (0.36–1.84)0.62
      By univariate Cox regression analysis, various factors such as age, pre-existing vascular disease, presence of diabetes, CRP, albumin, fibrinogen, apolipoprotein A-1 (apo A-1), apo A-2 and AGE-fl were significantly associated with all-cause mortality while sex, BMI, smoking, hypertension, CML, lipids and apo B were not. The multivariate stepwise regression analysis demonstrated that age, pre-existing vascular disease, CRP and albumin were independent predictors of all-cause mortality Table 4. Diabetes, fibrinogen, apo A-1, apo A-2 and AGE-fl lost their significance as survival factors in the multivariate Cox model.
      Table 4Predictiors of overall mortality after 32 months in hemodialysis patients by the univariate and multivariate Cox regression analysis
      Univariate Cox regression analysisMultivariate Cox regression analysis
      VariableUnitsκ2Relative riskP valueκ2Relative riskP value
      Ageyears29.301.05<0.0018.821.03=0.004
      Pre-existing CVDRef = no CVD12.391.97=0.0019.241.93=0.002
      DiabetesRef = nondiabetic9.081.81=0.003NS
      CRPmg/L36.601.13<0.0017.451.10=0.003
      Albuming/L77.420.15<0.00136.480.19<0.001
      Fibrinogeng/L5.991.20=0.014NS
      Apo A-1mg/dL13.040.98<0.001NS
      Apo A-2mg/dL9.060.97=0.001NS
      AGE-flAU8.731.00=0.003NS

       Analysis of mortality rates according to various combinations of high and low CRP and AGE-fl/CML levels

      Since AGEs and CRP showed the opposite influence on overall and cardiovascular mortality, we analyzed groups of various combinations of both parameters. Patients with high CRP and low AGE-fl had the worst survival Figure 2a. If in parallel to high CRP levels AGE-fl was high, a dramatic increase in the survival rate (from 18.5 to 26.3 months mean survival) was noticed. Patients with low CRP and high AGE-fl values had the best survival (29.2 months mean survival). No significant difference in survival was present between the high and low AGE-fl groups if CRP was low.
      Figure thumbnail gr2
      Figure 2Kaplan-Meier estimates of survival during follow-up with regard to all-cause mortality in relation to groups of high (↑, above median values) or low (↓, below median values) C-reactive protein (CRP) and AGE-fl (A) or CML (B) values.
      Comparing the characteristics of the two AGE-fl/CRP groups with best and worst survival rates, the incidence of diabetes, smoking habits, aspirin medication and high flux membrane use were similar Table 5. The group with the worst survival was older and had lower serum albumin, creatinine and BUN levels. With decreasing survival the incidence of vascular disease increased consistently from 30 to 46%, but no significant level was reached. Comparing the two groups with low CRP and low or high AGE-fl albumin and creatinine increased with increasing AGE-fl Table 5.
      Table 5Clinical and laboratory characteristics of patients divided in four groups according to AGE-fl and CRP above or below the median with decreasing survival from left to right
      CharacteristicsAGE-fl> median + CRP ≤ median (N = 74)AGE-fl ≤ median + CRP ≤ median (N = 72)AGE-fl >median + CRP> median (N = 73)AGE-fl > median + CRP> median (N = 76)
      Age years58.6±13.763.0±13.162.9±12.266.0±12.2
      P < 0.01
      BMI kg/m223.2±3.524.7±4.824.7±4.524.2±4.7
      CRP mg/L3.4±1.93.9±2.125.3±24.830.6±30.2
      Albumin g/L42.1±3.041.5±2.840.5±3.637.9±5.0
      P < 0.01
      ,
      P < 0.05 for AGE-fl ≤+ CRP> median (last column) vs. AGE-fl> + CRP> median (third column)
      Fibrinogen g/L3.5±0.83.6±0.84.7±1.44.6±0.1
      AGE-fl AU (×103)137.1±18.481.7±16.1133.4±19.482.8±18.5
      CML μg/mL1.63±0.551.37±0.561.50±0.681.24±0.59
      Creatinine mg/dL9.9±2.49.4±3.010.1±2.38.7±2.8
      P < 0.01
      ,
      P < 0.001
      BUN mg/dL66±1765±1561±1658±19
      P < 0.05 for AGE-fl ≤+ CRP> median (last column) vs. AGE-fl> + CRP ≤ median (first column)
      Vascular disease30374546
      Diabetes27272832
      Smoking23122720
      Aspirin medication32324742
      High flux membrane27321929
      Data are given as mean ± SD. Data of AGE-fl are given in arbitrary units (AU) × 103 with a median of 109; data of CML are given in μg/ml with a median of 1.4. Vascular disease, diabetes, smoking, aspirin medication, high flux membrane are given in %.
      a P < 0.01
      b P < 0.05 for AGE-fl ≤+ CRP> median (last column) vs. AGE-fl> + CRP ≤ median (first column)
      c P < 0.001
      d P < 0.05 for AGE-fl ≤+ CRP> median (last column) vs. AGE-fl> + CRP> median (third column)
      The same trend of increased survival was seen if comparing the low CML/high CRP-group with the high CML/high CRP-group, but it did not reach a significant level Figure 2b. Again, patients with low CRP had the best survival no matter how elevated the CML levels.
      Significant differences were observed between the best and worst survival rates in terms of higher serum albumin, BUN and creatinine levels and lower incidence of diabetes and vascular disease in the group with best survival Table 6. Again, in the two groups with both low CRP but different CML levels, albumin and creatinine increased when in parallel CML levels increased. By contrast, a significantly higher percentage of diabetic patients was present in the group with the worst survival rate in comparison to the high CML/high CRP-group. No differences were noticed between all groups concerning smoking habits, aspirin treatment and use of high flux membranes. There was an overall trend toward higher incidence of vascular disease with decreasing survival.
      Table 6Clinical and laboratory characteristics of patients divided into groups according to CML and CRP above or below the median with decreasing survival from left to right
      CharacteristicsCML < median + CRP ≤ median (N = 84)CML ≤ median + CRP ≤ median (N = 66)CML> median + CRP> median (N = 66)CML ≤ median + CRP> median (N = 84)
      Age years59.7±14.762.4±11.764.0±13.564.7±11.0
      BMI kg/m222.9±4.025.0±4.324.3±5.024.4±4.5
      P < 0.01 for CML ≤+ CRP> median (last column) vs. CML> + CRP ≤ median (first column)
      CRP mg/L3.6±2.13.7±2.125.8±24.129.6±30.2
      Albumin g/L42.0±3.041.4±3.040.1±3.938.6±4.9
      P < 0.001
      ,
      P < 0.05 for CML ≤+ CRP> median (last column) vs. CML> + CRP> median (third column)
      Fibrinogen g/L3.4±0.83.6±0.94.6±1.54.6±1.2
      AGE-fl AU (×103)115.8±33.1102.0±30.7114.7±31.8102.8±30.6
      CML μg/mL1.92±0.411.00±0.251.97±0.450.90±0.27
      Creatinine mg/dL10.5±2.78.6±2.510.1±2.68.9±2.6
      P < 0.001
      ,
      P < 0.05 for CML ≤+ CRP> median (last column) vs. CML> + CRP> median (third column)
      BUN mg/dL68±1863±1463±1857±17
      P < 0.01 for CML ≤+ CRP> median (last column) vs. CML> + CRP ≤ median (first column)
      Vascular disease27424248
      P < 0.01 for CML ≤+ CRP> median (last column) vs. CML> + CRP ≤ median (first column)
      Diabetes19371838
      P < 0.01 for CML ≤+ CRP> median (last column) vs. CML> + CRP ≤ median (first column)
      ,
      P < 0.01
      Smoking21132224
      Aspirin medication27394544
      High flux membrane27332127
      Data are given as mean ± SD. Data of AGE-fl are given in arbitrary units (AU) × 103 with a median of 109; data of CML are given in μg/ml with a median of 1.4. Vascular disease, diabetes, smoking, aspirin medication, high flux membrane are given in %.
      a P < 0.001
      b P < 0.01 for CML ≤+ CRP> median (last column) vs. CML> + CRP ≤ median (first column)
      c P < 0.01
      d P < 0.05 for CML ≤+ CRP> median (last column) vs. CML> + CRP> median (third column)

       Aspirin, inflammation and mortality

      Interestingly, patients on aspirin medication had higher CRP levels (18.2 ± 26.0 vs. 15.5 ± 24.9 mg/L, P < 0.05). CRP levels were higher in patients with vascular disease as compared to those without (19.5 ± 29.9 vs. 14.1 ± 21.1 mg/L, P < 0.05). Fifty percent of patients under aspirin treatment had vascular disease compared to 33% without (P < 0.01). AGE-fl (107.8 ± 31.9 vs. 109.4 ± 32.3 AU × 103) and CML levels (1.42 ± 0.57 vs. 1.45 ± 0.64 μg/mL) were similar in those with or without aspirin treatment. When comparing patients with and without vascular disease no differences in AGE-fl and CML levels were observed. Aspirin treatment had no benefit on survival (25.0 vs. 24.9 months without aspirin; Figure 3).
      Figure thumbnail gr3
      Figure 3Kaplan-Meier estimates of survival during follow-up with regard to all-cause mortality in relation to aspirin intake.

       Correlation between parameters

      A significant correlation was found between AGE-fl and CML levels (R = 0.25, P < 0.001), while a negative correlation was present between albumin and CRP (R = -0.37, P < 0.001) as well as fibrinogen (R = -0.2, P < 0.001). Creatinine and BUN correlated positively with each other (R = 0.46, P < 0.001) and with albumin (R = 0.38, P < 0.001, and R = 0.29, P < 0.001, respectively), while they correlated negatively with age (R = -0.45, P < 0.001, and R = -0.26, P < 0.001, respectively). Between creatinine and CML a positive correlation (R = 0.31, P < 0.001) was found. A weak but significant correlation also was demonstrated between creatinine and AGE-fl (R = 0.16, P < 0.01) and age and AGE-fl (R = -0.15, P < 0.05). No correlation was found between AGE-fl and age in controls.

      DISCUSSION

      The present study tested the hypothesis that AGEs are linked to mortality and cardiovascular disease by activating the acute phase response in uremia. Serum fluorescent AGEs (AGE-fl) and carboxymethyllysine (CML), one specific AGE component, as well as CRP, a prominent acute phase protein, were determined and their effect on all-cause as well as cardiovascular mortality was studied over a period of more than two years in a cohort of patients on maintenance hemodialysis treatment. The results of the present study demonstrate high levels of serum AGEs in HD patients, thus confirming earlier reports
      • Dolhofer-Bliesener R.
      • Lechner B.
      • Deppisch R.
      • et al.
      Immunological determination of advanced glycosylation end-products in human blood and urine.
      • Papanastasiou P.
      • Grass L.
      • Rodela H.
      • et al.
      Immunological quantification of advanced glycosylation end-products in the serum of patients on hemodialysis or CAPD.
      • Makita Z.
      • Bucala R.
      • Rayfield E.J.
      • et al.
      Reactive glycosylation endproducts in diabetic uraemia and treatment of renal failure.
      • Odetti P.
      • Fogarty J.
      • Sell D.R.
      • Monnier V.M.
      Chromatographic quantitation of plasma and erythrocyte pentosidine in diabetic and uremic subjects.
      . The main findings of the present study were that diabetes and vascular disease were linked to increased mortality, while high (above the median) serum levels of AGE-fl and CML were not. In parallel, diabetics or patients with vascular disease didn't have increased serum AGE levels. Statistically, high levels of AGE-fl and CML were even linked to a better survival. Since high CRP levels were associated with increased mortality, we addressed the question of how the combination of high levels of AGE-fl/CML and CRP influenced mortality. For both parameters, AGE-fl and CML, survival was worst when the combination of low AGE and high CRP levels coincided. Strikingly, survival increased significantly when AGE-fl increased in the identical inflammatory setting. A similar, but not statistical significant trend was observed for CML.
      To explore these unexpected findings we analyzed different factors that may have interfered such as diabetes, smoking, dialysis membrane use, and aspirin medication. The incidence of diabetes was identical in all AGE-fl/CRP groups, while more diabetics were found in the low CML/high CRP (highest mortality) as compared to the high CML/high CRP (lowest mortality) and the high CML/high CRP group. Therefore, diabetes may have interfered in the CML/CRP subgroups. The similarity of AGE levels in diabetic and non-diabetic hemodialysis patients indicates that uremia-associated factors such as oxidative or α-dicarbonyl stress are probably of greater importance for the late stage of the Maillard reaction than glycemic control. In our study diabetic patients had even lower CML levels than non-diabetics, a yet unexplained finding already described previously
      • Weiss M.F.
      • Erhard P.
      • Kader-Attia F.A.
      • et al.
      Mechanisms for the formation of glycoxidation products in end-stage renal disease.
      . AGE-fl correlated significantly with CML levels showing common origins of formation
      • Reddy S.
      • Bichler J.
      • Wells-Knecht K.J.
      • et al.
      N epsilon-(carboxymethyl)lysine is a dominant advanced glycation end product (AGE) antigen in tissue proteins.
      . Differences in smoking habits and membrane use (percentage of high flux membrane use) did not contribute to differences in survival.
      When comparing clinical parameters in relation to increasing survival in all groups, consistently higher levels of serum creatinine, BUN and albumin were found. The significance of protein caloric malnutrition in chronic hemodialysis patients has been emphasized often. Lowrie, Huang and Lew showed that four of the six most important predictors of outcome were nutritionally related
      • Lowrie E.G.
      • Huang W.H.
      • Lew N.L.
      Death risk predictors among peritoneal dialysis and hemodialysis patients: A preliminary comparison.
      . Serum albumin has been considered to be a key index of nutritional status, and is commonly used to assess malnutrition. However, data show that serum albumin is dependent on several other factors such as age, fluid overload, capillary leakage, and most importantly inflammation
      • Zimmermann J.
      • Herrlinger S.
      • Pruy A.
      • et al.
      Inflammation enhances cardiovascular risk and mortality in hemodialysis patients.
      ,
      • Kaysen G.A.
      • Rathore V.
      • Shearer G.C.
      • Depner T.A.
      Mechanisms of hypoalbuminemia in hemodialysis patients.
      . In our study albumin was related to inflammation, as assessed by the inverse correlation with CRP, but also was positively related to creatinine, which itself was independent of CRP. Creatinine and CML correlated significantly. Although we were not able to reliably determine the actual nutritional status of our patients retrospectively, we speculate that the higher creatinine, albumin and BUN values found in those patients with higher AGE levels all may be indicators of a better nutritional status. Whether high AGE levels in serum are due to AGE absorption from food
      • Koschinsky T.
      • He C.J.
      • Mitsuhashi T.
      • et al.
      Orally absorbed reactive glycation products (glycotoxins): An environmental risk factor in diabetic nephropathy.
      is disputed by others
      • Henle T.
      • Deppisch R.
      • Ritz E.
      The Maillard reaction–from food chemistry to uraemia research.
      and remains controversial. Thus, high AGE levels in serum may be an epiphenomenon of a better nutritional state.
      Certainly, fluorescence is not specific for individual compounds and interference with other fluorescent compounds may occur
      • Münch G.
      • Keis R.
      • Wessels A.
      • et al.
      Determination of advanced glycation end products in serum by fluorescence spectroscopy and competitive ELISA.
      . However, it may comprise unknown AGE compounds with physiological impact. The physiological activity of the well characterized AGEs such as CML, pentosidine or imidazolone is still not unequivocally proven. We also aimed for second line evidence by measuring CML, and the results confirmed those obtained with AGE-fl. Clearly, the findings were not in accordance with our expectation and the hypothesis of the study
      • Schwedler S.
      • Schinzel R.
      • Vaith P.
      • Wanner C.
      Inflammation and advanced glycation end products in uremia: Simple coexistence, potentiation or causal relationship?.
      and cannot be extrapolated to other AGE compounds. For example, pentosidine shows a much higher degree of accumulation (approximately 10-fold) and different pathways of AGE-formation may exist in vivo. However, the concentration of pentosidine is 100-fold lower than the concentration of CML. Nevertheless, similar data were obtained in a cohort of patients from southern Italy when pentosidine levels were correlated with mortality (C. Zoccali, personal communication). A similar and initially conflicting finding was recently reported by Suliman et al, who showed that patients with serum homocysteine levels>24 μmol/L (above median values) had a significantly better four-year survival rate than those with homocysteine levels below the median
      • Suliman M.E.
      • Qureshi A.R.
      • Barany P.
      • et al.
      Hyperhomocysteinemia, nutritional status, and cardiovascular disease in hemodialysis patients.
      . The results demonstrated that global nutritional assessment and protein intake as well as total plasma cysteine levels influenced total homocysteine levels in hemodialysis patients. Until these results were published homocysteine was undoubtedly considered to be a strong cardiovascular risk factor in hemodialysis patients, a finding that still holds true in the general population. The mechanism by which homocysteine exerts its initial cardiovascular damage was recently proposed to be similar to that of AGEs
      • De C.R.
      Endothelial dysfunctions: Common denominators in vascular disease.
      . Both activators of endothelial dysfunction—homocysteine and AGEs—may signal their action through the transcription factor nuclear factor-κB via generation of oxygen free radicals. As a result of the initial oxidative stress, multiple genes are turned on that induce the generation of cytokines (IL-6, IL-8), chemokines (MCP-1, M-CSF) and adhesion molecules (VCAM-1, ICAM-1). These in vitro findings are strongly supported by a number of experiments. The cascade finally may result in an inflammatory response that can be determined by CRP in serum
      • Rader D.J.
      Inflammatory markers of coronary risk.
      .
      Aspirin has been shown to prevent future cardiovascular events in previously healthy men with elevated levels of CRP
      • Ridker P.M.
      • Cushman M.
      • Stampfer M.J.
      • et al.
      Inflammation, aspirin and the risk of cardiovascular disease in apparently healthy men.
      . Further analysis of our data in relation to aspirin intake revealed that a considerable number of patients were on aspirin medication but apparently had no benefit in terms of a better survival. Patients on aspirin treatment also had higher CRP levels, another unexpected finding in this study. One explanation is that more patients suffered from vascular disease in the aspirin group, which therefore represented a selection of patients at a high risk for cardiovascular events. In general, vascular disease patients do have higher CRP levels. In this context aspirin treatment might be promising since the survival was similar in groups of patients with different cardiovascular risk profiles and CRP levels. We suggest that the aspirin results presented here are typical for a morbid population in which some “expected” results are confounded by diseases. It appears that these data represent another example of reverse causality often seen in such patient populations. Only prospective studies may clarify this issue.
      The present findings demonstrate that the mechanisms proposed are not translated into an in vivo clinical situation. However, it may still be possible that the damaging effects of AGEs on the cardiovascular system may be seen locally in tissues or cells
      • Nerlich A.G.
      • Schleicher E.D.
      N(epsilon)-(carboxymethyl)lysine in atherosclerotic vascular lesions as a marker for local oxidative stress.
      or with other types of AGEs
      • Takeuchi M.
      • Makita Z.
      • Yanagisawa K.
      • et al.
      Detection of noncarboxymethyllysine and carboxymethyllysine advanced glycation end products (AGE) in serum of diabetic patients.
      . On the other hand, a recent study found that pentosidine, a reliable marker of “carbonyl stress,” was unrelated to intima media thickness and to the number of atherosclerotic plaques
      • Zoccali C.
      • Mallamaci F.
      • Asahia K.
      • et al.
      Pentosidine, carotid atherosclerosis and alterations in left ventricular geometry in hemodialysis patients.
      . Furthermore, it is likely that the blood load of AGEs resembles only a small fraction of total body AGE content and that the serum levels reflect—besides kidney function—particular changes in the overall AGE pool (Faist and Erbersdobler, personal communication). Thus, circulating AGEs may not be an adequate parameter for demonstrating effects on outcome. Most likely the focus should be on intracellular AGEs
      • Hammes H.P.
      • Brownlee M.
      • Lin J.
      • et al.
      Diabetic retinopathy risk correlates with intracellular concentrations of the glycoxidation product Nepsilon-(carboxymethyl) lysine independently of glycohaemoglobin concentrations.
      • Shinohara M.
      • Thornalley P.J.
      • Giardino I.
      • et al.
      Overexpression of glyoxalase-I in bovine endothelial cells inhibits intracellular advanced glycation endproduct formation and prevents hyperglycemia-induced increases in macromolecular endocytosis.
      • Giardino I.
      • Edelstein D.
      • Brownlee M.
      BCL-2 expression or antioxidants prevent hyperglycemia-induced formation of intracellular advanced glycation endproducts in bovine endothelial cells.
      .
      In summary, our study indicates that low levels of AGEs, as determined by AGE-fl and CML, were not linked to a better survival rate in hemodialysis patients when observed over a period of 32 months. A correlation to the acute phase response (as determined by CRP) could not be demonstrated, and the hypothesis that high levels of CML and AGE-fl may activate an acute phase response was not demonstrated by serum markers. Whether high AGE levels are reflective of a better nutritional support cannot be clarified at this point. We emphasize that our results don't mean that AGEs are “beneficial” per se, thus contradicting previously published studies about the role of AGEs in vascular disease. Rather, our results suggest that serum AGE levels in hemodialysis patients are probably not very conclusive when determining cardiovascular outcome.

      ACKNOWLEDGMENTS

      The research was supported from the “Interdisziplinäres Zentrum für Klinische Forschung” (Z 4/3) of the University of Würzburg. The authors express appreciation to Drs. Udo Bahner, Gisela Buschmann, Gabriele Eckert, Barbara Hammerl, Margit Brandl, and Helga Bettger from the Kuratorium für Dialyse und Transplantation (KfH) Würzburg; Drs. Kurt Bausewein, Ursula Dittmann, Verena Felder-Piepenbrink, and Heiko Ehrich from the KfH Kitzingen; and to Drs. Christoph Blaser and Ulf Grunewald from the KfH Lohr for collaborative assistance of the multicenter study.

      REFERENCES

        • Westwood M.E.
        • Thornalley P.J.
        The glycation hypothesis,.
        Glycation and Advanced Glycation Endproducts. edited by Colaco C. Landes Bioscience, Georgetown, 1997: 59
        • Brownlee M.
        Advanced protein glycosylation in diabetes and aging.
        Annu Rev Med. 1995; 46: 223-234
        • Makita Z.
        • Radoff S.
        • Rayfield E.J.
        • et al.
        Advanced glycosylation end products in patients with diabetic nephropathy.
        N Engl J Med. 1991; 325: 836-842
        • Dawnay A.
        • Millar D.J.
        The pathogenesis and consequences of AGE formation in uraemia and its treatment.
        Cell Mol Biol. 1998; 44: 1081-1094
        • Schinzel R.
        • Münch G.
        • Heidland A.
        Advanced glycation endproducts in end-stage renal disease and their removal.
        Nephron. 2001; 87: 295-303
        • Miyata T.
        • Ueda Y.
        • Horie K.
        • et al.
        Renal catabolism of advanced glycation end products: The fate of pentosidine.
        Kidney Int. 1998; 53: 416-422
        • Gugliucci A.
        • Bendayan M.
        Renal fate of circulating advanced glycated end products (AGE): Evidence for reabsorption and catabolism of AGE-peptides by renal proximal tubular cells.
        Diabetologia. 1996; 39: 149-160
        • Miyata T.
        • Van Ypersele de Strihou C.
        • Kurokowa K.
        • Baynes J.W.
        Alterations in nonenzymatic biochemistry in uremia: Origin and significance of “carbonyl stress” in long-term uremic complications.
        Kidney Int. 1999; 55: 389-399
        • Raj D.S.
        • Choudhury D.
        • Welbourne T.C.
        • Levi M.
        Advanced glycation end products: A Nephrologist's perspective.
        Am J Kidney Dis. 2000; 35: 365-380
        • Bierhaus A.
        • Hofmann M.A.
        • Ziegler R.
        • Nawroth P.P.
        AGEs and their interaction with AGE-receptors in vascular disease and diabetes mellitus. I. The AGE concept.
        Cardiovasc Res. 1998; 37: 586-600
        • Fu M.X.
        • Requena J.R.
        • Jenkins A.J.
        • et al.
        The advanced glycation end product, Nepsilon-(carboxymethyl)lysine, is a product of both lipid peroxidation and glycoxidation reactions.
        J Biol Chem. 1996; 271: 9982-9986
        • Thornalley P.J.
        Pharmacology of methylglyoxal: Formation, modification of proteins and nucleic acids, and enzymatic detoxification—A role in pathogenesis and antiproliferative chemotherapy.
        Gen Pharmacol. 1996; 27: 565-573
        • Schleicher E.D.
        • Wagner E.
        • Nerlich A.G.
        Increased accumulation of the glycoxidation product N(epsilon)-(carboxymethyl)lysine in human tissues in diabetes and aging.
        J Clin Invest. 1997; 99: 457-468
        • Torzewski M.
        • Rist C.
        • Mortensen R.F.
        • et al.
        C-reactive protein in the arterial intima: Role of C-reactive protein receptor-dependent monocyte recruitment in atherogenesis.
        Arterioscler Thromb Vasc Biol. 2000; 20: 2094-2099
        • Danesh J.
        • Whincup P.
        • Walker M.
        • et al.
        Low grade inflammation and coronary heart disease: prospective study and updated meta-analyses.
        BMJ. 2000; 321: 199-204
        • Zimmermann J.
        • Herrlinger S.
        • Pruy A.
        • et al.
        Inflammation enhances cardiovascular risk and mortality in hemodialysis patients.
        Kidney Int. 1999; 55: 648-658
        • Schwedler S.
        • Schinzel R.
        • Vaith P.
        • Wanner C.
        Inflammation and advanced glycation end products in uremia: Simple coexistence, potentiation or causal relationship?.
        Kidney Int. 2001; 59: S32-S36
        • Wagner Z.
        • Wittmann I.
        • Mazák I.
        • et al.
        Serum Ne-(carboxymethyl)lysine levels in type 2 diabetic patients—Impact on renal function.
        Am J Kidney Dis. 2001; 38: 785-791
        • Münch G.
        • Keis R.
        • Wessels A.
        • et al.
        Determination of advanced glycation end products in serum by fluorescence spectroscopy and competitive ELISA.
        Eur J Clin Chem Clin Biochem. 1997; 35: 669-677
        • Dolhofer-Bliesener R.
        • Lechner B.
        • Deppisch R.
        • et al.
        Immunological determination of advanced glycosylation end-products in human blood and urine.
        Nephrol Dial Transplant. 1995; 10: 657-664
        • Papanastasiou P.
        • Grass L.
        • Rodela H.
        • et al.
        Immunological quantification of advanced glycosylation end-products in the serum of patients on hemodialysis or CAPD.
        Kidney Int. 1994; 46: 216-222
        • Makita Z.
        • Bucala R.
        • Rayfield E.J.
        • et al.
        Reactive glycosylation endproducts in diabetic uraemia and treatment of renal failure.
        Lancet. 1994; 343: 1519-1522
        • Odetti P.
        • Fogarty J.
        • Sell D.R.
        • Monnier V.M.
        Chromatographic quantitation of plasma and erythrocyte pentosidine in diabetic and uremic subjects.
        Diabetes 1992. 1992; 41: 153-159
        • Weiss M.F.
        • Erhard P.
        • Kader-Attia F.A.
        • et al.
        Mechanisms for the formation of glycoxidation products in end-stage renal disease.
        Kidney Int. 2000; 57: 2571-2585
        • Reddy S.
        • Bichler J.
        • Wells-Knecht K.J.
        • et al.
        N epsilon-(carboxymethyl)lysine is a dominant advanced glycation end product (AGE) antigen in tissue proteins.
        Biochemistry. 1995; 34: 10872-10878
        • Lowrie E.G.
        • Huang W.H.
        • Lew N.L.
        Death risk predictors among peritoneal dialysis and hemodialysis patients: A preliminary comparison.
        Am J Kidney Dis. 1995; 26: 220-228
        • Kaysen G.A.
        • Rathore V.
        • Shearer G.C.
        • Depner T.A.
        Mechanisms of hypoalbuminemia in hemodialysis patients.
        Kidney Int. 1995; 48: 510-516
        • Koschinsky T.
        • He C.J.
        • Mitsuhashi T.
        • et al.
        Orally absorbed reactive glycation products (glycotoxins): An environmental risk factor in diabetic nephropathy.
        Proc Natl Acad Sci USA. 1997; 94: 6474-6479
        • Henle T.
        • Deppisch R.
        • Ritz E.
        The Maillard reaction–from food chemistry to uraemia research.
        Nephrol Dial Transplant. 1996; 11: 1718-1722
        • Suliman M.E.
        • Qureshi A.R.
        • Barany P.
        • et al.
        Hyperhomocysteinemia, nutritional status, and cardiovascular disease in hemodialysis patients.
        Kidney Int. 2000; 57: 1727-1735
        • De C.R.
        Endothelial dysfunctions: Common denominators in vascular disease.
        Curr Opin Lipidol. 2000; 11: 9-23
        • Rader D.J.
        Inflammatory markers of coronary risk.
        N Engl J Med. 2000; 343: 1179-1182
        • Ridker P.M.
        • Cushman M.
        • Stampfer M.J.
        • et al.
        Inflammation, aspirin and the risk of cardiovascular disease in apparently healthy men.
        N Engl J Med. 1997; 336: 973-979
        • Nerlich A.G.
        • Schleicher E.D.
        N(epsilon)-(carboxymethyl)lysine in atherosclerotic vascular lesions as a marker for local oxidative stress.
        Atherosclerosis. 1999; 144: 41-47
        • Takeuchi M.
        • Makita Z.
        • Yanagisawa K.
        • et al.
        Detection of noncarboxymethyllysine and carboxymethyllysine advanced glycation end products (AGE) in serum of diabetic patients.
        Mol Med. 1999; 5: 393-405
        • Zoccali C.
        • Mallamaci F.
        • Asahia K.
        • et al.
        Pentosidine, carotid atherosclerosis and alterations in left ventricular geometry in hemodialysis patients.
        J Nephrol. 2001; 14: 293-298
        • Hammes H.P.
        • Brownlee M.
        • Lin J.
        • et al.
        Diabetic retinopathy risk correlates with intracellular concentrations of the glycoxidation product Nepsilon-(carboxymethyl) lysine independently of glycohaemoglobin concentrations.
        Diabetologia. 1999; 42: 603-607
        • Shinohara M.
        • Thornalley P.J.
        • Giardino I.
        • et al.
        Overexpression of glyoxalase-I in bovine endothelial cells inhibits intracellular advanced glycation endproduct formation and prevents hyperglycemia-induced increases in macromolecular endocytosis.
        J Clin Invest. 1998; 101: 1142-1147
        • Giardino I.
        • Edelstein D.
        • Brownlee M.
        BCL-2 expression or antioxidants prevent hyperglycemia-induced formation of intracellular advanced glycation endproducts in bovine endothelial cells.
        J Clin Invest. 1996; 97: 1422-1428