Detailed Assessment of Augmented Renal Clearance in a Large Mixed Intensive Care Unit Population

• Conditions: Critical Illness; Augmented Renal Clearance
• Interventions: Other: no intervention

• Registration Number: NCT03954275
• Lead Sponsor: Universitaire Ziekenhuizen KU Leuven

Brief Summary

This multi-center retrospective cohort study presents a detailed assessment of augmented renal clearance (ARC) in a mixed population of adult critically ill patients. Epidemiology of ARC will be studied in detail in a very heterogeneous population. Risk factors for ARC will be identified and a predictive scoring system for ARC ready to use in clinical practice will be constructed and validated. Performance of estimators of kidney function will be measured and a cutoff for ARC will be determined for the best estimator. Finally clinical impact of ARC will be explored using vancomycine and aminoglycosides levels as surrogate marker.

Detailed Description

Augmented renal clearance will be assessed in detail in a very large and heterogeneous adult critically ill population. Analysis will be conducted retrospectively on a multi-center database collected by the M@tric research group. M@tric collects data from all intensive care units (surgical, medical, cardiac) in 3 Belgian University Hospitals (Leuven, Ghent, Antwerp).

Anonymised admission, demographic, clinical and laboratory data collected from 2013 until the present will be retrieved from the M@tric database. These data will then be coded and analysed in R statistical software. ARC will be defined based on a 24h creatinine clearance (CrCl24h) \>=130ml/min/1.73m².

Epidemiology and risk factors for ARC will be studied in order to confirm and clarify past studies which have mostly been done in rather small and specific subsets of patients. A predictive algorithm for ARC will be trained and subsequently validated for use in clinical practice. Moreover this algorithm will be compared to existing scoring systems, which have not yet found their way into clinical practice. This algorithm will provide the ability to anticipate ARC on the intensive care unit. Also use of formulae estimating renal function will be evaluated in this population. These estimators will be compared to the CrCl24h, which is considered the golden standard in clinical practice. A cutoff for the best estimating formula in order to detect ARC will be calculated. Finally the impact of ARC on serum levels of hydrophilic molecules likes vancomycine and aminoglycosides will be studied. As this research follows a retrospective design these levels will be used a surrogate marker for clinical impact. This will potentially point out some opportunities for future research on the clinical impact of ARC.

Study Timeline

• Study Start: 2018-10-01
• Status Verified: 2019-05
• Study First Submitted: 2019-05-03
• Study First Submitted QC: 2019-05-15
• Last Update Submitted: 2019-05-15
• Study First Posted: 2019-05-17
• Last Update Posted: 2019-05-17
• Primary Completion: 2019-08
• Study Completion: 2019-09

Recruitment & Eligibility

• Status: UNKNOWN
• Sex: All
• Target Recruitment: 10000

Inclusion Criteria

• Having at least one 24h creatinine clearance measurement available

Exclusion Criteria

• Any form of renal replacement therapy

Study & Design

• Study Type: OBSERVATIONAL
• Study Design: Not specified

Arm && Interventions

Group	Intervention	Description
Critically ill patients with a CrCl24h	no intervention	Patients admitted to any intensive care unit (surgical, medical or cardiac) and having at least one 24h creatinine clearance measurement available.

Primary Outcome Measures

Name	Time	Method
Most precise formula using Bland-Altman agreement analysis	Retrospective analysis between January 2013 and December 2015	Bland-Altman agreement analysis between CrCl24h and 3 commonly used serum creatinine based formulae estimating renal function (CKD-EPI, C\&G, MDRD) will be used to identify the formula with the best precision (SD of the bias).
ARC daily prevalence	Retrospective analysis between January 2013 and December 2015	Daily prevalence of ARC (% of ARC days per ICU admission day)
Performance of the best cutoff for ARC using ROC curve analysis	Retrospective analysis between January 2013 and December 2015	Performance of the best cutoff for ARC using ROC curve analysis on the most precise formula estimating renal function.
ARC incidence per admission	Retrospective analysis between January 2013 and December 2015	Incidence of ARC in % of ICU admissions: with ARC incidence defined as at least once, min. 50% of the measurements, 100% of the measurements during ICU admission)
Logistic regression with ARC as dependent variable	Retrospective analysis between January 2013 and December 2015	Risk factors associated with ARC will be identified through logistic regression analysis on demographic and clinical data.
Predictive algorithm for ARC	Retrospective analysis between January 2013 and December 2015	An algorithm predicting ARC on the next day(s) will be created using a backward selection logistic regression model on the risk factors associated with ARC detected in this study and/or in previously published studies.
Exploration of clinical impact of ARC via surrogate markers	Retrospective analysis between January 2013 and December 2015	Vancomycin and aminoglycoside (amikacin \& gentamycin) serum concentrations will be used as surrogate markers to evaluate potential clinical impact of ARC.
ARC incidence per day	Retrospective analysis between January 2013 and December 2015	Incidence of ARC per 100 ICU days
Duration and course of ARC episodes	Retrospective analysis between January 2013 and December 2015	ARC episodes: number of episodes (count), length of the episodes (days) and both combined to obtain relative contribution to ARC as a % ((count\*length)/total ARC days)

Trial Locations

Locations (1)

UZLeuven; 🇧🇪 Leuven, Belgium