amadori-albumin correlates with microvascular complications and precedes nephropathy in type 1...

European Journal of Clinical Investigation (2002) 32, 500–506

© 2002 Blackwell Science Ltd

Blackwell Science, LtdAmadori-albumin correlates with microvascular complications and precedes nephropathy in type 1 diabetic patients

C. G. Schalkwijk*, N. Chaturvedi†, H. Twaafhoven*, V. W. M. van Hinsbergh*,‡, C. D. A. Stehouwer* and the EUCLID Study Group*Academic Hospital Vrije Universiteit, Amsterdam, the Netherlands, †Imperial College of Medicine at St Mary’s, London,

UK, ‡Gaubius Laboratory TNO-PG, Leiden, the Netherlands

Abstract Background Amadori-albumin, a major glycated protein, is involved in experimentalhyperglycaemia-induced microvascular complications, and is associated with advancednephropathy in Type I diabetic patients in humans. Our aim was to assess the associationof Amadori-albumin with early nephropathy and with retinopathy in Type I diabetic patientsand the involvement of chronic low-degree inflammation therein.

Design and Methods Amadori-albumin, the Amadori product of haemoglobin (HbA1c),C-reactive protein, and fibrinogen levels were measured in the EUCLID study, a2-year randomised, double-blind, placebo-controlled trial of lisinopril in 447 Type Idiabetic patients. Retinal photographs were taken in 341 patients at baseline and 294 atfollow up.

Results Amadori-albumin was positively associated with albumin the excretion rate andretinopathy status (P = 0·0001 and P = 0·02 for trend, respectively) and with the progressionfrom normoalbuminuria to (micro)albuminuria (38·6 U mL−1 in nonprogressors,44·3 U mL−1 in progressors; P = 0·02), but not with the development or progression ofretinopathy during a 2-year follow up. Amadori-albumin levels at baseline were associatedwith C-reactive protein and fibrinogen (P = 0·0007 and P = 0·0001, respectively). C-reactiveprotein and fibrinogen were also associated with albumin excretion rates (P = 0·03 andP = 0·01, respectively) and retinopathy status (P = 0·02 and P = 0·0006, respectively).Adjustment for these inflammatory markers did not markedly attenuate the associationbetween Amadori-albumin and the albumin excretion rate, while adjustment for fibrinogen,but not C-reactive protein, abolished the association between Amadori-albumin andretinopathy. Lisinopril had no impact on the association between the levels of Amadori-albumin and albumin excretion rates or retinopathy.

Conclusions Amadori-albumin was associated with early nephropathy and with retinopathyin Type I diabetic patients and preceded an increase in albumin excretion rate, but notretinopathy. A chronic low-degree inflammation does not appear to be involved in Amadori-albumin-associated microvascular complications in Type I diabetes.

Keywords Amadori-albumin, diabetic nephropathy, diabetic retinopathy.Eur J Clin Invest 2002; 32 (7): 500–506

Department of Clinical Chemistry (C. G. Schalkwijk, H. Twaafhoven), and Internal Medicine (C. D. A. Stehouwer), Academic Hospital, and Institute for Cardiovascular Research (C. G. Schalkwijk, V. W. M. van Hinsbergh, C. D. A. Stehouwer), Vrije Universiteit, Amsterdam, the Netherlands; EURODIAB, Department of Epidemiology and Public Health, Imperial College of Medicine at St Mary’s, London, UK (N. Chaturvedi); Gaubius Laboratory TNO-PG, Leiden, the Netherlands (V. W. M. van Hinsbergh).

Correspondence to: Dr C. G. Schalkwijk, Department of Clinical Chemistry, Academic Hospital Vrije Universiteit, PO Box 7057, 1007 MB Amsterdam, the Netherlands. Tel.: +31 20–4443870; fax: +31 20–4443895; e-mail: [email protected]

Received 22 October 2001; accepted 15 March 2002

Amadori-albumin in Type I diabetic patients 501

© 2002 Blackwell Science Ltd, European Journal of Clinical Investigation, 32, 500–506

Introduction

Type I diabetes mellitus is complicated by the developmentof severe microvascular disease, the aetiology of whichremains incompletely understood. Various glucose-inducedmechanisms may induce microvascular disease in diabetes,and recent information suggests that the effect of glucoseis mediated, at least in part, via nonenzymatic glycation [1].In addition, several observations support the theory thathyperglycaemia-associated chronic low-degree inflamma-tion and endothelial dysfunction are involved in the patho-genesis and progression of micro- and macrovasculardiseases [2–4]. In Type I diabetic patients, features ofinflammatory activity and markers of endothelial dysfunc-tion are increased and are correlated [4], suggesting a roleof low-degree inflammation in vascular complications.

Nonenzymatic glycation involves the reaction of thecarbonyl group of sugar aldehydes with the N-terminus offree-amino groups of proteins, resulting in the formation ofstable Amadori products. Further irreversible chemical reac-tions of these Amadori products may lead to the formationof advanced glycation endproducts (AGEs) [1]. These havebeen shown to activate macrophages [5], increase oxidativestress [6], and to induce production of cytokines [6–8], i.e.features of inflammation. Therefore, many of the possiblemechanisms leading to chronic low-degree inflammationmay be related to an increase in the formation of AGEs [9].The role of Amadori-albumin, which is by far the mostprominent glycated product, has recently come under con-sideration [10–12]. It has been demonstrated that Amadori-albumin is able to induce biological effects, such as theexpression of cytokines in macrophages [6] and, in vivo, inmice [7]; effects believed to be involved in vascular compli-cations. In contrast to the circulating Amadori-adduct ofalbumin, the Amadori product of haemoglobin (HbA1c) isbelieved to have little if any biological activity on accountof its storage in erythrocytes. We have demonstrated thatplasma Amadori-albumin concentration is increasedapproximately twofold in Type I diabetic patients, is corre-lated with markers of endothelial dysfunction, and is inde-pendently associated with advanced diabetic nephropathy[10]. In addition, Amadori-albumin is localized in thecapillaries of the retinas of diabetic patients [10]. Furthermore,previous studies demonstrating that antiglycated albuminadministered to diabetic mice prevented diabetic nephro-pathy [13] and diabetic retinopathy [14] indicate thatAmadori-albumin is causally involved in the pathogenesisof diabetic microvascular complications.

As a next step, we have investigated, in Type I diabeticpatients recruited from the EUCLID study [15], the asso-ciation of Amadori-albumin with early nephropathy andwith retinopathy, and, in a 2-year follow up, whetherelevated plasma levels of Amadori-albumin precede theprogression of these microvascular complications. Wehypothesized that Amadori-albumin induces a chroniclow-degree inflammation, contributing to microvascularcomplications. We therefore investigated the role of chroniclow-degree inflammation, as estimated by plasma C-reactive protein (CRP) and fibrinogen, in the association of

Amadori-albumin with nephropathy and retinopathy. As theEUCLID study was a placebo-controlled investigation ofthe effect of the ACE-inhibitor, lisinopril, on the develop-ment of microangiopathy [15], we also had the opportunityto investigate whether ACE-inhibition modified the relation-ship between Amadori-albumin and microvascular disease,as recently suggested [16].

Methods

Subjects

The EUCLID study was a double-blind, randomised,parallel-designed, 24-month clinical trial of lisinopril andplacebo in 18 European centres in 530 Type I diabeticpatients aged 20–59 years. For the present study, sampleswere available on 447 out of 530 patients for Amadori-albumin. Analysis could not be performed on some samplesbecause of insufficient material or spoiling of the sample.Statistical analyses have been restricted to patients withan Amadori-albumin measurement. Patient selection andmethods have been described in detail elsewhere [15,17].Patients were normotensive according to guidelines inforce at the time of the study design. Most of the patientswere normoalbuminuric [albumin excretion rate (AER)< 20 µg min−1] and 16% were microalbuminuric (AER 20–200 µg min−1). Retinopathy was classified from photographson a five-level scale (none to proliferative). Photographswere available on 341 out of 447 patients at baseline andin 294 out of 341 at follow up.

Laboratory data

The Amadori product of haemoglobin was measured at theRoyal London Hospital, UK, by an enzyme immunoassayEIA with monoclonal antibodies raised against HbA1c(Dako Ltd, Ely, UK). The normal range of this assay is 2·9–4·8%. C-reactive protein was measured with a sensitivein-house ELISA with rabbit anti-CRP (Dako, Copenhagen,Denmark) as a catching and tagging antibody, as describedpreviously [3], with intra- and interassay coefficients ofvariation of 3·8% and 4·7%, respectively. Plasma fibrinogenwas also measured centrally on citrated plasma. Fibrinogenwas assayed with a clotting assay [18].

Urine albumin was measured by an immunoturbidi-metric assay, and AER was assessed as the mean of twoovernight urine collections [19].

Amadori-albumin was determined in a competitiveELISA, as recently described [10]. Briefly, each well wascoated with 100 L of 0·1 mg mL−1 glycated human serumalbumin (HSA), as prepared by the incubation of HSA with0·5 M glucose for 4 weeks. This preparation was also usedas standard in the competitive ELISA. To each well wereadded 50 L of the antibody conjugated with biotin(1 : 2000) and 50 L of standard or a sample to be testeddiluted in phosphate-buffered saline (PBS) including

502 C. G. Schalkwijk et al.


30 mg mL−1 HSA and incubated for 2 h. After threewashes with PBS-Tween, the wells were incubated withstreptavidine-peroxidase (CLB, Amsterdam, the Netherlands)for 1 h followed by substrate development with 100 µL oftetramethylbenzidine. The reaction was stopped with 100 µL0·5 M H2SO4. The extinction at 450 nm was measured witha multichannel spectrophotometer (SLT Microplate reader,Wilten Bioteknika, Etten-Leur, the Netherlands). Plasmalevels were expressed as Amadori-albumin units, and 1 Uwas defined as the antibody-reactive material equivalentto 1 g of the glycated-HSA standard. Plasma Amadori-albumin in control subjects was 20·9 ± 4·0 U mL−1 [10].The intra- and interassay coefficients of variation were 5·4%and 11·6%, respectively.

Statistical analysis

Development or progression of retinopathy was defined asa worsening by at least one level of retinopathy from baselineto follow up. The denominator for progression of albuminu-ria was all those who were normoalbuminuric at baseline(n = 374). Progression was defined as an AER in the micro-albuminuric range (> 20 µg min−1) at two successive visits orat the final visit for that patient.

The distributions of CRP and AER were positivelyskewed, and were log-transformed for analysis. Pearsoncorrelation coefficients were first calculated. Analysis ofvariance, with adjustments for covariates, was used to estimatemean values of continuous variables. Mean levels ofAmadori-albumin and CRP at 2 years were estimated by thetreatment group, with adjustment for the baseline value.Multivariate regression models are often used to understandthe relative importance of several predictors for a given out-come. However, the direct comparison of the size of the betacoefficients of these explanatory variables is problematic, as,for example, a 1-year increase in age cannot be said to beequivalent to a 1% increase in HbA1c. Standardised regres-sion effects were therefore used to overcome this limitation.These were calculated for continuous variables by multiply-ing the beta estimate from the regression models by thestandard deviation of that variable; log-transformed vari-ables were not converted back. This allows the directcomparison of the degree of importance of each variableby standardising for population variance.

Results

The mean age and diabetes duration of the patients at base-line were 33·8 and 14·1 years, respectively (Table 1). Theprevalence of microalbuminuria was 16% and of retino-pathy, 64%. Amadori-albumin was significantly correlatedwith HbA1c, body mass index, cholesterol, and the inflam-matory markers CRP and fibrinogen, but not with ageor duration of Type I diabetes (Table 2). There was norelationship between either Amadori-albumin or CRP andsmoking status; however, fibrinogen was significantly

elevated in ex-(3·2 g L−1) and current smokers (3·0 g L−1)compared with nonsmokers (2·9 g l−1, P = 0·04).

Associations of Amadori-albumin and inflammatory markers with albumin excretion rates

Amadori-albumin was strongly positively associated withAER (Table 3). The Amadori product of haemoglobin(HbA1c) was also positively associated with AER (P = 0·0001).The association between Amadori-albumin and AER wasmarkedly attenuated after adjustment for HbA1c (fromP = 0·0001 to P = 0·2). Both CRP and fibrinogen wereassociated with AER (Table 3). Adjustment for CRP alone,fibrinogen alone, or these inflammatory markers in combi-nation (data not shown) did not attenuate the associationbetween Amadori-albumin and AER. When both thesefactors were added to a model containing HbA1c, onlyHbA1c remained significantly related to AER (P = 0·0001).

The univariate standardised-regression effect for associ-ation between Amadori-albumin and AER was 0·16 (95%CI 0·07–0·26, P = 0·0003); and for HbA1c it was 0·23(95% CI 0·14–0·31, P = 0·0001). In the bivariate modelwith both Amadori-albumin and HbA1c, Amadori-albuminwas still associated with AER after adjustment for HbA1c,although the β was halved (0·09, P = 0·05). In contrast, theβ for HbA1c in the bivariate model was 0·20 (P = 0·0001).Addition of CRP and fibrinogen did not markedly alter theβ for Amadori-albumin, but the association was no longerstatistically significant.

We next determined whether Amadori-albumin levelsat baseline were associated with the development of(micro)albuminuria. Baseline Amadori-albumin wasassociated with progression of normalbuminuria to (micro)albuminuria in a 2-year follow up, which was attenuatedafter adjustment for HbA1c (from P = 0·02 to P = 0·2).

Table 1 Baseline patient characteristics

Males: n (%) 267 (60)Age (years) 33·8 ± 8·6Diabetes duration (years) 14·1 ± 8·0Blood pressure

Systolic (mmHg) 122 ± 11Diastolic (mmHg) 80 ± 4

HbA1c (%) 7·2 ± 2·0Current smokers: n (%) 135 (30)Body mass index (kg/m2) 24·6 ± 2·8Waist-to-hip ratio 0·84 ± 0·09Total cholesterol (mmol L−1) 5·38 ± 1·10Any retinopathy: n (%)* 218 (64)Microalbuminuria: n (%) 71 (16)AER (µg min−1)† 7·96 (4·44, 13·60)Amadori-albumin (U mL−1) 39·6 ± 12·1C-reactive protein (mg L−1)† 0·83 (0·33, 1·97)Fibrinogen (g L−1) 3·0 ± 0·9

Data are shown as a percentage or mean ± SD.*Photographs available on 341 patients at baseline.†Geometric mean (25th, 75th percentile).



In addition, baseline HbA1c was strongly associated withprogression to microalbuminuria at follow up (8·09% inprogressors and 6·91% in nonprogressors, P = 0·0007).This strong association remained after adjustment forAmadori-albumin (data not shown). Baseline CRP andfibrinogen were not associated with the progression normalalbuminuria to (micro)albuminuria (P = 0·9 and 0·1,respectively).

Associations of Amadori-albumin and inflammatory markers with retinopathy

Amadori-albumin and the markers of inflammation weresignificantly associated with retinopathy (Table 4). The

Amadori product of haemoglobin (HbA1c) was also posit-ively associated with retinopathy (P = 0·01). Patients withproliferative retinopathy or photocoagulation had lowerAmadori-albumin, and correspondingly lower HbA1c levelsthan those with moderate and severe nonproliferativeretinopathy. The association between Amadori-albuminand retinopathy status was attenuated after adjustmentfor HbA1c (from P = 0·02 to P = 0·08). Again, HbA1cremained independently associated with retinopathy afteradjustment for Amadori-albumin (P = 0·0001). Adjust-ment for CRP had little impact on the association betweenAmadori-albumin and retinopathy, but adjustment forfibrinogen abolished the association with retinopathy. Whenall these factors were entered into the model, only fibrinogenremained significantly associated with retinopathy.

Table 2 Correlation coefficients at baseline of Amadori-albumin, C-reactive protein and fibrinogen with variables associated with diabetic complications

Table 3 Mean baseline levels of Amadori-albumin, C-reactive protein and fibrinogen by albumin excretion rate

Table 4 Mean baseline levels of Amadori-albumin, C-reactive protein and fibrinogen by retinopathy status

Amadori-albumin

Markers of chronic inflammation

C-reactive protein Fibrinogen

Age (years) −0·07 (0·1) 0·13 (0·008) 0·08 (0·1)Diabetes duration (years) 0·02 (0·7) 0·11 (0·02) 0·04 (0·4)Systolic blood pressure (mmHg) −0·05 (0·3) 0·02 (0·07) −0·01 (0·9)Diastolic blood pressure (mmHg) 0·03 (0·5) 0 (1·0) 0·05 (0·3)HbA1c (%) 0·38 (0·001) 0·22 (0·0001) 0·23 (0·0001)Albumin excretion rate (µg min−1) 0·20 (0·0001) 0·05 (0·3) 0·15 (0·02)Body mass index (kg/m2) 0·13 (0·01) 0·24 (0·0001) 0·08 (0·1)Waist-to-hip ratio 0·00 (0·90) 0·07 (0·1) −0·10 (0·03)Total cholesterol (mmol L−1) 0·17 (0·0004) 0·15 (0·002) 0·22 (0·0001)C-reactive protein (mg L−1) 0·16 (0·0007) – –Fibrinogen (g L−1) 0·28 (0·0001) 0·33 (0·0001) –

P-values are shown in parentheses.

Albumin excretion rate (µg min−1) n

Amadori-albumin (U mL−1)

Markers of inflammation Amadori-albumin adjusted for

CRP (mg L−1) Fibrinogen (g L−1) C-reactive protein Fibrinogen

≤ 5 133 37·1 (35·0–39·0) 0·7 (0·57–0·88) 2·8 (2·75–3·03) 37·3 37·9> 5 ≤ 10 161 38·3 (36·5–40·1) 0·79 (0·65–0·97) 2·90 (2·76–3·04) 38·4 39·2> 10 ≤ 20 80 43·4 (40·9–45·9) 1·07 (0·82–1·51) 3·01 (2·83–3·19) 43·0 43·6> 20 ≤ 70 52 42·1 (38·9–45·3) 1·13 (0·82–1·56) 3·21 (2·99–3·43) 41·5 40·7> 70 19 46·2 (40·9–51·5) 0·67 (0·40–1·14) 3·43 (3·06–3·80) 46·7 46·3P for trend – P = 0·0001 P = 0·03 P = 0·01 P = 0·0003 P = 0·002

Data are given as mean and (95% confidence limits).

Albumin excretion rate (µg min−1) nAmadori-albumin (U mL−1)

Markers of inflammation Amadori-albumin adjusted for

CRP (mg L−1) Fibrinogen (g L−1) C-reactive protein Fibrinogen

None 123 39·8 (37·6–42·0) 0·65 (0·54–0·79) 2·81 (2·65–2·97) 40·2 41·0Minimal non proliferative 149 41·2 (39·2–43·2) 0·83 (0·69–0·99) 3·21 (3·05–3·37) 41·0 41·4Moderate + severe non proliferative 45 45·8 (42·3–49·3) 1·04 (0·75–1·45) 3·27 (3·03–3·51) 45·4 45·2Proliferative + photocoagulated 24 37·3 (32·4–42·2) 1·21 (0·77–1·90) 3·13 (2·82–3·44) 36·7 37·1P for trend – P = 0·02 P = 0·02 P = 0·0006 0·04 0·9

Data are given as mean and as 95% confidence limits.



We next determined whether Amadori-albumin levels atbaseline were associated with the development or progres-sion of retinopathy. Amadori-albumin was not associatedwith development or progression of retinopathy (Table 5).Baseline HbA1c was strongly predictive for the develop-ment or progression of retinopathy (7·85% in progressorsand 6·90% in nonprogressors; P = 0·0005). This associationwas not attenuated by the addition of Amadori-albumin.Fibrinogen, but not CRP (P = 0·1), was associated with thedevelopment or progression of retinopathy (P = 0·03).

Effects of lisinopril

Amadori-albumin at baseline did not differ by treatmentgroup (lisinopril vs. placebo). Levels in those on lisinopriland placebo were 42·6 U mL−1 and 41·8 U mL−1, respect-ively (P = 0·5) at 2-year follow up. In addition, lisinopril didnot affect the associations between baseline Amadori-albumin levels and the development of (micro)albuminuriaand retinopathy (data not shown).

Additional analyses

None of the above analyses were altered materially byadjusting for age, diabetes duration, sex, smoking, systolicblood pressure, body mass index, waist-to-hip ratio, or totalcholesterol (data not shown), except that the relationshipsbetween CRP on the one hand and AER and retinopathyon the other were no longer statistically significant, largelybecause of body mass index and waist-to-hip ratio.

Discussion

We have previously shown that Amadori-albumin levels arestrongly associated with the presence of advanced diabeticnephropathy in Type I diabetes [10]. The present studyshows that Amadori-albumin is associated with: (I)increased AER in the normo- to (micro)albuminuricrange (i.e. very early nephropathy) and with retinopathy status(from none to proliferative); (ii) the development of(micro)albuminuria but not the development or progressionof retinopathy in a 2-year follow up; and (iii) markersof inflammation. The association of Amadori-albumin with

nephropathy and retinopathy could largely be accounted forby ‘confounding’ by HbA1c (see later text). Markers ofinflammation were also associated with the presence of albu-minuria and retinopathy at baseline. In general, however,adjustment for the inflammatory markers did not attenuatethe association between Amadori-albumin and the presenceof albuminuria and retinopathy at baseline. The only excep-tion to this was fibrinogen, adjustment for which stronglyattenuated the association of Amadori-albumin andretinopathy.

The ACE-inhibitor, lisinopril, had no impact on the levelsof Amadori-albumin, and, in addition, did not affect theassociation between Amadori-albumin and the developmentof early diabetic nephropathy.

Amadori-albumin and nephropathy

Amadori-albumin, rather than the Schiff base form orAGEs, is a major form of glycated proteins [20]. The con-centrations of Amadori products typically constitute 2–10%of serum proteins, whereas the concentration of AGEs innormal and diabetic serum is less than 0·01% [21]. Wepreviously reported that Amadori-albumin is independentlyassociated with advanced nephropathy [10]. Furthermore,the presence of Amadori-albumin in the glomeruli ofpatients with diabetic nephropathy correlates with the sever-ity of tissue damage [22]. In the present study we extendedthese findings by showing a strong association of Amadori-albumin with elevated AER, and also demonstrated thatAmadori-albumin precedes the development of(micro)albuminuria in a 2-year follow up, indicating apathophysiological role of Amadori-albumin quite early inthe development of diabetic nephropathy. In accordancewith these findings, antiglycated albumin therapy reducednephropathy in genetically diabetic mice [13]. Possiblemechanisms for Amadori-albumin-induced nephropathy arean interaction of Amadori-albumin with specific receptorsthat can induce receptor-mediated responses [23,24] and,more indirectly, the induction of chronic low-degree inflam-mation [6,7,25]. However, as the association of Amadori-albumin with AER could not be explained by markers ofinflammation, our data may suggest that an interaction ofAmadori-albumin with specific receptors is involved inAmadori-albumin-induced nephropathy. This is consistentwith in vitro data showing that Amadori-albumin modulatesendothelial and glomerular mesangial cell biology, which

Table 5 Amadori-albumin at baseline in those who did and did not progress to (micro)albuminuria and in those who developed retinopathy or in whom retinopathy progressed by at least one level

n Amadori-albumin (U mL−1) (95% CI) P

Progression to (micro)albuminuria 0·02No 350 38·6 (37·4–39·8)Yes 24 44·3 (39·6–49·0)

Developing or progression to retinopathy 0.09No 226 41·1 (39·5–42·7)Yes 47 40·9 (37·4–44·4)



has been implicated in the development of diabetic nephro-pathy via interaction with specific receptors [25–28]

Amadori-albumin and retinopathy

Amadori-albumin, CRP, and fibrinogen were associatedwith the presence of retinopathy at baseline. The associationof Amadori-albumin with retinopathy remained largelyunchanged after adjustment for CRP, but could largelybe explained by fibrinogen. Indeed the association betweenfibrinogen and retinopathy appeared largely independent ofboth Amadori-albumin and HbA1c. As adjusting Amadori-albumin for CRP did not affect association with retinopathy,these data suggest that Amadori-albumin and CRP workthrough different mechanisms. Therefore, a direct interac-tion of Amadori-albumin with its receptor appears to havean effect on retinopathy. In agreement, we recently demon-strated Amadori-albumin in the capillaries of the retinas ofdiabetic patients [10]. However, Amadori-albumin was notassociated with progression of retinopathy in a 2-year followup, which may be because of a lack of power (retinopathydeveloped or progressed in only 60 patients). In addition,17 out of 47 ‘progressors’ had minimal nonproliferativeretinopathy, whereas raised Amadori-albumin is morestrongly associated with moderate/severe nonproliferativeretinopathy (Table 4).

Adjustment for the Amadori product of haemoglobin

When an adjustment was made for baseline HbA1c, theassociations between Amadori-albumin and early nephro-pathy and retinopathy were attenuated. However, it isquestionable whether this should be done becauseAmadori-products of haemoglobin and albumin are bothformed as a function of hyperglycaemia and are thus part ofthe same pathway. Therefore, such an attenuation will resultin overadjustment and may lead to an underestimation of aputative pathophysiological role of Amadori-albumin. In thisrespect it is important to note that we recently demonstratedthat after adjustment for several factors including HbA1c,Amadori-albumin continued to be significantly associatedwith advanced nephropathy [10]. Furthermore, in thatstudy, Amadori-albumin, but not HbA1c, correlated withvascular cell adhesion molecule-1, while HbA1c, but notAmadori-albumin, was highly correlated with E-selectin,demonstrating the difference between biological activeglycated products, i.e. Amadori-albumin, and an estimatedglycaemic control, i.e. HbA1c.

In summary, we found that Amadori-albumin is associ-ated with increased AER and with retinopathy status,and that Amadori-albumin precedes the development of(micro)albuminuria but not that of retinopathy in a 2-yearfollow up. Association of Amadori-albumin with nephro-pathy and retinopathy could not be explained by an inflam-matory reaction. These data are consistent with theproposed pathophysiological role of Amadori-albumin inmicrovascular complications in Type I diabetes.

Acknowledgements

Casper G. Schalkwijk is supported by a USF grant of theVrije Universiteit and a Clinical Research Fellowship fromthe Diabetes Fonds, the Netherlands. Victor W. M. vanHinsbergh is supported by a USF grant from Vrije Univer-siteit, the Netherlands. The EUCLID study was supportedby AstraZeneca pharmaceuticals.

References

1 Vlassara H, Bucala R, Striker L. Pathogenic effects of advanced glycosylation. Biochemical, biological and clinical implications for diabetes and aging. Lab Invest 1994;70:138–51.

2 Ross R. The pathogenesis of atherosclerosis: a perspective for the 1990s. Nature 1993;362:801–9.

3 Stehouwer CDA, Lambert J, Donker AJM, Van Hinsbergh VWM. Endothelial dysfunction and pathogenesis of diabetic angiopathy. Cardiovasc Res 1997;34:55–68.

4 Schalkwijk CG, Poland DCW, Van Dijk W, Kok A, Emeis JJ, Drager AM et al. Plasma concentration of C-reactive protein is increased in Type I diabetic patients without clinical macroangiopathy and correlates with markers of endothelial dysfunction: evidence for chronic inflammation. Diabetologia 1999;41:351–7.

5 Kirstein M, Brett J, Radoff S, Ogawa S, Stern D, Vlassara H. Advanced protein glycation induces transendothelial human monocyte chemotaxis and secretion of PDGF. role in vascular disease of diabetes and aging. Proc Natl Acad Sci USA 1990;87:9010–4.

6 Yan SD, Schmidt AM, Anderson GM, Zhang J, Brett J et al. Enhanced cellular oxidant stress by the interaction of advanced glycation endproducts with their receptors/binding proteins. J Biol Chem 1994;268:9889–97.

7 Vlassara H, Brownlee M, Manogue HR, Dinarello C, Pasagian A. Cachetin/TNF and IL-1 induced by glucose modified proteins: role in normal tissue remodeling. Science 1988;240:1546–1548.

8 Schmidt AM, Hasu M, Popov M, Zhang JH, Chen J et al. Receptor for advanced glycation endproducts (AGEs) has a central role in vessel wall interaction and gene activation in response to circulating AGE proteins. Proc Natl Acad Sci USA 1994;91:8807–8811.

9 Smulders RA, Stehouwer CDA, Schalkwijk CG, Donker AJ, Van Hinsbergh VW, Tekoppele JM. Distinct association of hba1c and the urinary excretion of pentosidine, an advanced glycosylation end–product, with endothelial function in insulin-dependent diabetes mellitus. Thromb Haemost 1998;80:52–7.

10 Schalkwijk CG, Ligtvoet N, Twaalfhoven H, Jager A, Blaauwgeers HG et al. Amadori-albumin in Type I diabetic patients: correlation with markers of endothelial function, association with diabetic nephropathy and localization in retinal capilaries. Diabetes 1999;48:2446–53.

11 Cohen MP, Ziyadeh FN. Role of Amadori-modified nonenzymatically glycated serum proteins in the pathogenesis of diabetic nephropathy. J Am Soc Nephrol 1996;7:183–90.

12 Chen S, Cohen MP, Ziyadeh FN. Amadori-glycated albumin in diabetic nephropathy: pathophysiologic connections. Kidney Int 2000;58 (Suppl. 77):S40–S44.

13 Cohen MP, Hud E, Wu V-Y. Amelioration of diabetic nephropathy by treatment with monoclonal antibodies against glycated albumin. Kidney Int 1994;45:1673–9.



14 Clements RS, Robison WG, Cohen MP. Anti-glycated albumin therapy ameliorates early retinal microvascular pathology in db/db mice. J Diab Comp 1998;12:28–33.

15 The EUCLID. Study Group: Randomised placebo-controlled trial of lisinopril in normotensive patients with insulin-dependent diabetes and normoalbuminuria or microalbuminuria. Lancet 1997;349:1787–92.

16 Osicha TM, Yu Y, Panagiotopoulos S, Clavant SP, Kiriazis Z, Pike RN et al. Prevention of albuminuria by aminoguanidine or ramipril in streptozotocin-induced diabetic rats is associated with the normalization of glomerular protein kinase C. Diabetes 2000;49:87–93.

17 Chaturvedi N, Sjolie A-K, Stephanson JM, Abrahamian H, Keipes M et al. and the EUCLID Study Group. Effect of lisinopril on progression of retinopathy in normotensive people with Type I diabetes. Lancet 1998;351:28–31.

18 Greaves M, Malia RG, Goodfellow K, Mattock M, Stevens LK et al. and the EURODIAB IDDM Complications Study Group. Fibrinogen and von Willebrand factor in IDDM relationships to lipid vascular risk factors, blood pressure, glycaemic control and urinary albumin excretion rate: the EURODIAB IDDM Complications study. Diabetologia 1997;40:698–705.

19 Medcalf E, Newman DJ, Gorman EG, Price CP. Rapid, robust method for measuring low concentrations of albumin in urine. Clin Chem 1990;36:446–9.

20 Curtiss LK, Witztum JL. A novel method for generating region-specific monoclonal antibodies to modified proteins. J Clin Invest 1983;72:1427–38.

21 Makita Z, Vlassara H, Cerami A, Bucala R. Immunochemical

detection of advanced glycosylation end products in vivo. J Biol Chem 1992;267:5133–8.

22 Sakai H, Jinde K, Suzuki D, Yagame M, Nomoto Y. Localization of glycated proteins in the glomeruli of patiens with diabetic nephropathy. Nephrol Dial Transplant 1996;11 (Suppl. 5):66–71.

23 Predescu D, Simionescu M, Simionescu N, Palade G. Binding and transcytosis of glycoalbumin by the microvascular endothelium of the myocardium: Evidence that glycoalbumin behaves as a bifunctional ligand. J Cell Biol 1988;107:1729–38.

24 Brandt R, Landmesser C, Vogt L, Hehmke B, Hanschke R et al. Differential expression of fructosyllysine-specific receptors on monocytes and macropahages and possible pathophysiological significance. Diabetologia 1996;39:1140–7.

25 Amore A, Cirina P, Mitola S, Peruzzi L, Gianoglio B et al. Nonenzymatically glycated albumin (Amadori adducts) enhances nitric oxide synthase activity and gene expression in endothelial cells. Kidney Int 1997;51:27–35.

26 Chen S, Cohen MP, Lautenslager GT, Shearman CW, Ziyadeh FN. Glycated albumin stimulates TGF-1 production and protein kinase C activity in glomerular endothelial cells. Kidney Int 2001;59:673–81.

27 Ziyadeh FN, Han D-C, Cohen JA, Guo J, Cohen MP. Glycated albumin stimulates fibronectin gene expression in glomerular mesangial cells: Involvement of the transforming growth factor-β system. Kidney Int 1998;53:631–8.

28 Cohen M, Ziyadeh FN. Amadori glucose adducts modulate mesangial cell growth and collagen gene expression. Kidney Int 1994;45:475–84.

amadori-albumin correlates with microvascular complications and precedes nephropathy in type 1...

Documents