Limits...
Multicentric validation of proteomic biomarkers in urine specific for diabetic nephropathy.

Alkhalaf A, Zürbig P, Bakker SJ, Bilo HJ, Cerna M, Fischer C, Fuchs S, Janssen B, Medek K, Mischak H, Roob JM, Rossing K, Rossing P, Rychlík I, Sourij H, Tiran B, Winklhofer-Roob BM, Navis GJ, PREDICTIONS Gro - PLoS ONE (2010)

Bottom Line: In <10% of cases and controls classification by proteome analysis not entirely resulted in the expected clinical outcome.These data provide the first independent confirmation that profiling of the urinary proteome by CE-MS can adequately identify subjects with DN, supporting the generalizability of this approach.The data further establish urinary collagen fragments as biomarkers for diabetes-induced renal damage that may serve as earlier and more specific biomarkers than the currently used urinary albumin.

View Article: PubMed Central - PubMed

Affiliation: Department of Internal Medicine, University Medical Center Groningen, University of Groningen, Groningen, The Netherlands.

ABSTRACT

Background: Urine proteome analysis is rapidly emerging as a tool for diagnosis and prognosis in disease states. For diagnosis of diabetic nephropathy (DN), urinary proteome analysis was successfully applied in a pilot study. The validity of the previously established proteomic biomarkers with respect to the diagnostic and prognostic potential was assessed on a separate set of patients recruited at three different European centers. In this case-control study of 148 Caucasian patients with diabetes mellitus type 2 and duration ≥5 years, cases of DN were defined as albuminuria >300 mg/d and diabetic retinopathy (n = 66). Controls were matched for gender and diabetes duration (n = 82).

Methodology/principal findings: Proteome analysis was performed blinded using high-resolution capillary electrophoresis coupled with mass spectrometry (CE-MS). Data were evaluated employing the previously developed model for DN. Upon unblinding, the model for DN showed 93.8% sensitivity and 91.4% specificity, with an AUC of 0.948 (95% CI 0.898-0.978). Of 65 previously identified peptides, 60 were significantly different between cases and controls of this study. In <10% of cases and controls classification by proteome analysis not entirely resulted in the expected clinical outcome. Analysis of patient's subsequent clinical course revealed later progression to DN in some of the false positive classified DN control patients.

Conclusions: These data provide the first independent confirmation that profiling of the urinary proteome by CE-MS can adequately identify subjects with DN, supporting the generalizability of this approach. The data further establish urinary collagen fragments as biomarkers for diabetes-induced renal damage that may serve as earlier and more specific biomarkers than the currently used urinary albumin.

Show MeSH

Related in: MedlinePlus

ROC curves for classification of the patient collectives with the ‘DN type 2’ pattern.ROC analysis for CKD diagnosis of the training set (A) and the test set (B).
© Copyright Policy
Related In: Results  -  Collection


getmorefigures.php?uid=PMC2958112&req=5

pone-0013421-g007: ROC curves for classification of the patient collectives with the ‘DN type 2’ pattern.ROC analysis for CKD diagnosis of the training set (A) and the test set (B).

Mentions: For the generation of a new model for DN in diabetic type 2 patients (using the PREDICTIONS cohort), urinary polypeptides of the control group were compared with those of patients with diabetic nephropathy. This analysis identified 103 peptides of statistical significance using multivariate statistic analysis like Benjamini-Hochberg [40] (p = 0.05; see table S3b). A support vector machine-based model with these biomarkers discriminated controls from cases with 98% sensitivity and 99% specificity. The distribution of the polypeptides in the two groups is shown in figure 7A. The validity of the ‘DN type 2’ biomarkers was further evaluated in a diabetes type 1 test-set cohort (trainingset of Rossing et al. [1]) and resulted in 86% sensitivity and 100% specificity with an AUC value of 0.948 (see figure 7B). Of the 103 defined differentially expressed peptides, 65% (67 markers) could be confirmed as also being significantly different in the ‘Rossing’ cohort between diabetic patients with DN and diabetes controls. The results of the statistical analyses are listed in table S3b. Of note, most biomarkers with high significance were found to be significantly different in both cohorts.


Multicentric validation of proteomic biomarkers in urine specific for diabetic nephropathy.

Alkhalaf A, Zürbig P, Bakker SJ, Bilo HJ, Cerna M, Fischer C, Fuchs S, Janssen B, Medek K, Mischak H, Roob JM, Rossing K, Rossing P, Rychlík I, Sourij H, Tiran B, Winklhofer-Roob BM, Navis GJ, PREDICTIONS Gro - PLoS ONE (2010)

ROC curves for classification of the patient collectives with the ‘DN type 2’ pattern.ROC analysis for CKD diagnosis of the training set (A) and the test set (B).
© Copyright Policy
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC2958112&req=5

pone-0013421-g007: ROC curves for classification of the patient collectives with the ‘DN type 2’ pattern.ROC analysis for CKD diagnosis of the training set (A) and the test set (B).
Mentions: For the generation of a new model for DN in diabetic type 2 patients (using the PREDICTIONS cohort), urinary polypeptides of the control group were compared with those of patients with diabetic nephropathy. This analysis identified 103 peptides of statistical significance using multivariate statistic analysis like Benjamini-Hochberg [40] (p = 0.05; see table S3b). A support vector machine-based model with these biomarkers discriminated controls from cases with 98% sensitivity and 99% specificity. The distribution of the polypeptides in the two groups is shown in figure 7A. The validity of the ‘DN type 2’ biomarkers was further evaluated in a diabetes type 1 test-set cohort (trainingset of Rossing et al. [1]) and resulted in 86% sensitivity and 100% specificity with an AUC value of 0.948 (see figure 7B). Of the 103 defined differentially expressed peptides, 65% (67 markers) could be confirmed as also being significantly different in the ‘Rossing’ cohort between diabetic patients with DN and diabetes controls. The results of the statistical analyses are listed in table S3b. Of note, most biomarkers with high significance were found to be significantly different in both cohorts.

Bottom Line: In <10% of cases and controls classification by proteome analysis not entirely resulted in the expected clinical outcome.These data provide the first independent confirmation that profiling of the urinary proteome by CE-MS can adequately identify subjects with DN, supporting the generalizability of this approach.The data further establish urinary collagen fragments as biomarkers for diabetes-induced renal damage that may serve as earlier and more specific biomarkers than the currently used urinary albumin.

View Article: PubMed Central - PubMed

Affiliation: Department of Internal Medicine, University Medical Center Groningen, University of Groningen, Groningen, The Netherlands.

ABSTRACT

Background: Urine proteome analysis is rapidly emerging as a tool for diagnosis and prognosis in disease states. For diagnosis of diabetic nephropathy (DN), urinary proteome analysis was successfully applied in a pilot study. The validity of the previously established proteomic biomarkers with respect to the diagnostic and prognostic potential was assessed on a separate set of patients recruited at three different European centers. In this case-control study of 148 Caucasian patients with diabetes mellitus type 2 and duration ≥5 years, cases of DN were defined as albuminuria >300 mg/d and diabetic retinopathy (n = 66). Controls were matched for gender and diabetes duration (n = 82).

Methodology/principal findings: Proteome analysis was performed blinded using high-resolution capillary electrophoresis coupled with mass spectrometry (CE-MS). Data were evaluated employing the previously developed model for DN. Upon unblinding, the model for DN showed 93.8% sensitivity and 91.4% specificity, with an AUC of 0.948 (95% CI 0.898-0.978). Of 65 previously identified peptides, 60 were significantly different between cases and controls of this study. In <10% of cases and controls classification by proteome analysis not entirely resulted in the expected clinical outcome. Analysis of patient's subsequent clinical course revealed later progression to DN in some of the false positive classified DN control patients.

Conclusions: These data provide the first independent confirmation that profiling of the urinary proteome by CE-MS can adequately identify subjects with DN, supporting the generalizability of this approach. The data further establish urinary collagen fragments as biomarkers for diabetes-induced renal damage that may serve as earlier and more specific biomarkers than the currently used urinary albumin.

Show MeSH
Related in: MedlinePlus