Comparison of Methods for Renal Risk Prediction in Patients with Type 2 Diabetes (ZODIAC-36)

Content Provider	Semantic Scholar
Author	Riphagen, Ineke J. Kleefstra, Nanne Drion, Iefke Alkhalaf, Alaa Diepen, Merel Van Groenier, Klaas H. Landman, Gijs Wilhelmus Diederik Navis, Gerjan J. Bilo, Henk J. G. Bakker, Stephan J. L.
Copyright Year	2015
Abstract	BACKGROUND Patients with diabetes are at high risk of death prior to reaching end-stage renal disease, but most models predicting the risk of kidney disease do not take this competing risk into account. We aimed to compare the performance of Cox regression and competing risk models for prediction of early- and late-stage renal complications in type 2 diabetes. METHODS Patients with type 2 diabetes participating in the observational ZODIAC study were included. Prediction models for (micro)albuminuria and 50% increase in serum creatinine (SCr) were developed using Cox regression and competing risk analyses. Model performance was assessed by discrimination and calibration. RESULTS During a total follow-up period of 10 years, 183 out of 640 patients (28.6%) with normoalbuminuria developed (micro)albuminuria, and 22 patients (3.4%) died without developing (micro)albuminuria (i.e. experienced the competing event). Seventy-nine out of 1,143 patients (6.9%) reached the renal end point of 50% increase in SCr, while 219 (19.2%) died without developing the renal end point. Performance of the Cox and competing risk models predicting (micro)albuminuria was similar and differences in predicted risks were small. However, the Cox model increasingly overestimated the risk of increase in SCr in presence of a substantial number of competing events, while the performance of the competing risk model was quite good. CONCLUSIONS In this study, we demonstrated that, in case of substantial numbers of competing events, it is important to account for the competing risk of death in renal risk prediction in patients with type 2 diabetes.
File Format	PDF HTM / HTML
DOI	10.1371/journal.pone.0120477
PubMed reference number	25775414
Journal	Medline
Volume Number	10
Alternate Webpage(s)	https://www.rug.nl/research/portal/files/20158302/journal.pone.0120477.pdf
Alternate Webpage(s)	http://ftp.ncbi.nlm.nih.gov/pub/pmc/f8/30/pone.0120477.PMC4361549.pdf
Alternate Webpage(s)	https://doi.org/10.1371/journal.pone.0120477
Journal	PloS one
Language	English
Access Restriction	Open
Content Type	Text
Resource Type	Article