UZH-Logo

Development and validation of a prediction score for postoperative acute renal failure following liver resection


Slankamenac, K; Breitenstein, S; Held, U; Beck-Schimmer, B; Puhan, M A; Clavien, P A (2009). Development and validation of a prediction score for postoperative acute renal failure following liver resection. Annals of Surgery, 250(5):720-728.

Abstract

OBJECTIVE: To develop and validate a score to predict postoperative acute renal failure (ARF) after liver resection. BACKGROUND: Postoperative ARF after major surgery is associated with morbidity and mortality. Early identification of patients at risk of ARF is important in order to provide protective kidney treatment. METHODS: Postoperative ARF was prospectively assessed in consecutive patients undergoing liver resection. In randomly selected two-third of the total number of patients, multivariate logistic regression analysis was used to develop a new prediction score (including a full and a reduced model), based on the preoperative parameters of age, gender, preexisting chronic renal dysfunction, cardiovascular disease, diabetes, bilirubin, and alanine aminotransferase (ALT) levels. In the remaining last third of the patients, the new score was validated by calibrating the accuracy of the score (ClinicalTrials.gov NCT 00743132). RESULTS: Postoperative ARF occurred in 15.1% (86 of 569 consecutive patients) from 2002 to 2007 and was highly associated with mortality (22.5% vs. 0.8% without ARF, P < 0.001). In the 380 (two-third of the population) patients selected for the development of the prediction score, preoperatively elevated ALT, preexisting cardiovascular disease, chronic renal failure, and diabetes were the strongest predictors of ARF. Validating the full prediction model (0-22 points) to the remaining 189 patients (one-third of the population), the risk could be predicted accurately (mean predicted risk of 11.5% vs. an observed risk of 14.8%) without significant differences between predicted and observed risks across different risk categories (P = 0.98). Prediction with the reduced model including the 4 strongest predictors (0-7 points) was almost as accurate as with the full model (11.4% predicted vs. 14.8% observed) and also without significant differences across different risk categories (P = 0.75). CONCLUSIONS: The new prediction score (the full as well as the reduced model) accurately predicted postoperative ARF after liver resection. The use of these scores allows early identification of patients at high risk of ARF, and may support decision making for protective kidney interventions perioperatively.

OBJECTIVE: To develop and validate a score to predict postoperative acute renal failure (ARF) after liver resection. BACKGROUND: Postoperative ARF after major surgery is associated with morbidity and mortality. Early identification of patients at risk of ARF is important in order to provide protective kidney treatment. METHODS: Postoperative ARF was prospectively assessed in consecutive patients undergoing liver resection. In randomly selected two-third of the total number of patients, multivariate logistic regression analysis was used to develop a new prediction score (including a full and a reduced model), based on the preoperative parameters of age, gender, preexisting chronic renal dysfunction, cardiovascular disease, diabetes, bilirubin, and alanine aminotransferase (ALT) levels. In the remaining last third of the patients, the new score was validated by calibrating the accuracy of the score (ClinicalTrials.gov NCT 00743132). RESULTS: Postoperative ARF occurred in 15.1% (86 of 569 consecutive patients) from 2002 to 2007 and was highly associated with mortality (22.5% vs. 0.8% without ARF, P < 0.001). In the 380 (two-third of the population) patients selected for the development of the prediction score, preoperatively elevated ALT, preexisting cardiovascular disease, chronic renal failure, and diabetes were the strongest predictors of ARF. Validating the full prediction model (0-22 points) to the remaining 189 patients (one-third of the population), the risk could be predicted accurately (mean predicted risk of 11.5% vs. an observed risk of 14.8%) without significant differences between predicted and observed risks across different risk categories (P = 0.98). Prediction with the reduced model including the 4 strongest predictors (0-7 points) was almost as accurate as with the full model (11.4% predicted vs. 14.8% observed) and also without significant differences across different risk categories (P = 0.75). CONCLUSIONS: The new prediction score (the full as well as the reduced model) accurately predicted postoperative ARF after liver resection. The use of these scores allows early identification of patients at high risk of ARF, and may support decision making for protective kidney interventions perioperatively.

Citations

23 citations in Web of Science®
29 citations in Scopus®
Google Scholar™

Altmetrics

Downloads

69 downloads since deposited on 13 Jan 2010
19 downloads since 12 months
Detailed statistics

Additional indexing

Item Type:Journal Article, refereed, original work
Communities & Collections:04 Faculty of Medicine > University Hospital Zurich > Clinic and Policlinic for Internal Medicine
04 Faculty of Medicine > University Hospital Zurich > Clinic for Visceral and Transplantation Surgery
04 Faculty of Medicine > University Hospital Zurich > Institute of Anesthesiology
Dewey Decimal Classification:610 Medicine & health
Language:English
Date:2009
Deposited On:13 Jan 2010 11:48
Last Modified:05 Apr 2016 13:39
Publisher:Lippincott Wiliams & Wilkins
ISSN:0003-4932
Publisher DOI:10.1097/SLA.0b013e3181bdd840
PubMed ID:19809295
Permanent URL: http://doi.org/10.5167/uzh-26074

Download

[img]
Preview
Content: Accepted Version
Filetype: PDF
Size: 1MB
View at publisher

TrendTerms

TrendTerms displays relevant terms of the abstract of this publication and related documents on a map. The terms and their relations were extracted from ZORA using word statistics. Their timelines are taken from ZORA as well. The bubble size of a term is proportional to the number of documents where the term occurs. Red, orange, yellow and green colors are used for terms that occur in the current document; red indicates high interlinkedness of a term with other terms, orange, yellow and green decreasing interlinkedness. Blue is used for terms that have a relation with the terms in this document, but occur in other documents.
You can navigate and zoom the map. Mouse-hovering a term displays its timeline, clicking it yields the associated documents.

Author Collaborations