Discordance of renal drug dosing using estimated creatinine clearance and measured urine creatinine clearance in hospitalized adults: A retrospective cohort study

Abstract

Background:

Assessment of kidney function is fundamental to optimize drug dosing. The Cockcroft–Gault (CG) equation is widely used but has questionable validity for females, changing renal function, and the critical ill. Eight-hour urine collections (U_8h) offer direct measurement of creatinine clearance (CrCl) but lack the data for drug dosing. The primary objective of this study was to determine if there was a difference in renal drug dosing based on the estimation of CG CrCl (CrCl_CG) versus 8-h CrCl (CrCl_8h).

Methods:

This was an observational, retrospective cohort study of adult patients admitted between March 2018 and September 2018 with a collection U_8h during hospitalization. The primary outcome was discordance of renal drug dosing defined as the percentage of U_8h for which at least one different active medication CrCl dosing cutoff would result using the CrCl_CG versus CrCl_8h. The secondary outcomes were correlation between CrCl_CG and CrCl_8h and percentage of CrCl_CG values outside ± 20% of the CrCl_8h.

Results:

One hundred collections drawn from 85 unique patients (50.6% male, median age 55 [41–70] years, intensive care unit 88%) were included in the analysis. Median serum creatinine was 0.76 (0.52–1.06) mg/dL and blood urea nitrogen was 20 (14–28) mg/dL at time of collection_8h. Median CrCl_8h was 86.2 (43.5–140.3) mL/min versus 99.7 (56.5–166.9) mL/min CrCl_CG(P < 0.001) and discordance was 25%. The correlation between CrCl_8h and CrCl_CG was 0.76 (P < 0.001). Only 31% of CrCl_CG values were within ± 20% of the CrCl_8h value.

Conclusion:

We found 25% discordance for drug dosing between CrCl_8h and CrCl_CG. Further studies are needed to determine the impact on clinical outcomes.

INTRODUCTION

The assessment of kidney function is fundamental to determine the dose adjustments of renally eliminated medications to maximize efficacy and minimize adverse effects. The Cockcroft–Gault (CG) equation was developed in a study conducted at a Veterans Hospital in the 1970s to provide a simple calculation of creatinine clearance (CrCl) to estimate a patient's kidney function without the need for urine collection.[1] The CG study originally planned to use 505 urine collections, but only 249 were reproducible and arbitrarily determined to be good approximations of actual clearance. Furthermore, 96% of the study participants were male; patients with variable serum creatinine (SCr) values were excluded; and the patients were mainly located in medical wards. Therefore, the validity of the CG equation in patients who are female, have fluctuating SCr values, and/or are critically ill is unknown.[1]

The Food and Drug Administration (FDA) recommended the use of the CG equation for developing drug dosing guidelines and pharmacokinetic studies in 1998.[2] More recently, the FDA draft guidance in 2010 recommended using both the CG-estimated CrCl (CrCl_CG) and the Modification of Diet and Renal Disease estimation of glomerular filtration rate (eGFR) for study design in patients with impaired renal function.[3] In 2017, meropenem/vaborbactam became the first drug in the United States labeled for dosing adjustment based on eGFR.[4] Dosage adjustments based on kidney function are essential to appropriate drug therapy, but limited data assessing the various estimates of CrCl and potentially different drug dosing exist.[5,6]

As an alternative to estimating CrCl, urine collections provide a method to directly measure CrCl.[7] Twenty-four hour urine collections have been used to assess kidney function, and the CG equation was originally validated using this method.[1] More recently, the use of the 8-h urine collection (U_8h) has evolved due to faster results, decreased workload, and decreased potential for collection errors. The comparison of 8-h urine CrCl (CrCl_8h) with 24-hour urine CrCl (CrCl_24h) has shown a good correlation in the surgical intensive care unit (ICU), trauma, and hospitalized elderly patients.[8,9] However, U_8h has limited studies addressing the use for drug dosing in comparison with the CG equation.

At our institution, U_8h is a relatively new method used to assess a patient's renal function in addition to CG estimates. The primary aim of this study was to determine if there was a difference in renal drug dosing for selected renally eliminated drugs based on the estimation of CrCl_CG versus CrCl_8h in adult hospitalized patients.

METHODS

Design

This study was a single-center, observational, retrospective cohort design.

Participants and setting

The University Medical Center is a 1506 bed academic medical center consisting of seven hospitals. Patients hospitalized between March 1, 2018, and September 30, 2018, were evaluated. Patients were included if they were 18 years of age or older and had a U_8h during their hospital admission. Patients were excluded for the following: pregnancy/lactation, incarceration, renal replacement therapy within 72 h before collection, diuretic administration within 8 h before or during the collection, past medical history (PMH) of renal transplant or admission for renal transplant, presence of ileal conduit, factors that do not allow for CG calculation (SCr below lower limit of detection of 0.2 mg/dL, height < 60″, SCr not collected within 24 h of collection, height and weight not recorded), and no active orders to select renally adjusted medications [Table 1]. These medications were chosen because of common use in the settings where U_8h are utilized at our institution. This study was approved by our university institutional review board in accordance with the ethical standards set forth in the Helsinki Declaration in 1975 (IRB Number 2018E0698).

Table 1.

Select Renally Adjusted Medications

Medication	CrCL Dosing Breakpoints (mL/min)
Aminoglycosides	≥60	40-59	20-39	<20
Ampicillin/sulbactam	≥50	30-49	15-29	<15
Cefazolin	≥30	10-29	<10
Cefepime	≥60	30-59	<30
Ceftaroline	≥50	30-49	15-29	<15
Ceftolozane/tazobactam	≥50	30-49	<30
Daptomycin	≥30	<30
Enoxaparin	≥30	<30
Lacosamide	≥30	<30
Levetiracetam	>80	50-80	30-49	<30
Levofloxacin	>50	20-50	<20
Meropenem	≥50	26-49	10-25	<10
Piperacillin/tazobactam	≥20	<20
Rivaroxaban (atrial fibrillation)	>50	15-50	<15
Vancomcyin	≥60	30-59	<30

Data retrospectively collected from the electronic medical records (EMR, EPIC, Madison, Wisconsin, USA) included age, sex, race, height, actual body weight, PMH of chronic kidney disease, patient location, date/time of U_8h, SCr, and corresponding blood urea nitrogen (BUN) within 24 h before urine collection, urine volume, CrCl_8h calculated by the EMR, and active medication orders [Table 1] on the day of urine collection. Medications were placed within CrCl cutoffs originating from our institution's adult dosing guideline using the CrCl_8h compared to the CrCl_CG. If a medication was not included in our institution's guideline, the package insert or societal guideline was used.[10,11,12] The study data were collected and managed using Research Electronic Data Capture.[13]

The EMR automatically calculates the CrCl_CG utilizing the actual body weight (Wt_Actual) if Wt_Actual is less than the ideal body weight (IBW). If the Wt_Actual is ≥ 120% of the IBW, then adjusted body weight (Wt_adjusted) is used. All other situations use IBW. The SCr drawn closest to the time of urine collection within the previous 24 h was used in the CG calculation and is referred to as the reference SCr. The following equations were used for calculating CrCl_CG:

Male IBW(kg)=50+ (2.3 × inches over 5 feet)

Female IBW(kg) = 45.5 + (2.3 × inches over 5 feet)

The EMR also displays the calculated urine CrCl for the U_8h using the following equation:

Outcomes

The primary outcome was the discordance rate of renal drug dosing, defined as the number of urine collections for which at least one different medication CrCl dosing breakpoint would result using the CrCl_CG versus CrCl_8h divided by the total number of collections multiplied by 100. The secondary outcomes were the correlation of the CrCl_CG and CrCl_8h and percentage of CrCl_CG values outside ± 20% of the CrCl_8h. Augmented renal clearance (ARC) was defined as a CrCl_8h≥ 130 mL/min.[14]

Statistical analysis

The sample size was not calculated as there was a paucity of data comparing drug dosing using CrCl_8h to CrCl_CG. Data were reported using the descriptive statistics. Dichotomous variables were analyzed using Fisher's exact test or the Chi-squared test as appropriate. Normally distributed continuous variables were presented as mean (±standard deviation) and analyzed using the paired Student's t-test, whereas nonnormally distributed data were presented as a median (25%–75% interquartile range) and analyzed using the Mann-Whitney U test. Spearman rank correlation was used to determine the correlation between CrCl_8h and CrCl_CG. Statistical significance was defined as a P < 0.05. Data were analyzed using the Statistical Package for the Social Sciences software version 24.0 (SPSS^®), Inc. (Chicago, IL, USA).

RESULTS

A total of 150 U_8h were completed between March 2018 and September 2018. Fifty collections were excluded [Figure 1], leaving 100 collections drawn from 85 unique patients, with a median age of 55 (41–70) years. A little more than half of the patients were male (50.6%), and a majority (75.3%) were Caucasian. At the time of urine collection, 88% were in an ICU, but the median SCr and BUN were 0.76 (0.52–1.06) mg/dL and 20 (14–28) mg/dL, respectively. The median CrCl_8h was 86.2 (43.5–140.3) mL/min versus 99.7 (56.5–166.9) mL/min CrCl_CG, P < 0.001. Other patient characteristics at the time of urine collection are included in Table 2.

Figure 1.

Inclusion and exclusion of urine collections. †SCr undetectable, height <60 inches, undocumented height/weight. CG: Cockcroft–Gault

Table 2.

Patient characteristics at the time of urine collection

Characteristics	Values
Height (inches)	67 (64-71)
Admission weight (kg)	84.4 (71.7-102)
BMI (kg/m2)	28.5 (24.6-34)
SCr (mg/dL)	0.76 (0.52-1.06)
Increase in SCr of ≥0.3 mg/dL in previous 48 h (%)	13
BUN (mg/dL)	20 (14-28)
ICU status (%)	88
CrCl8h (mL/min)	86.2 (43.5-140.3)
CrClCG (mL/min)	99.7 (56.5-166.9)
ARC† (%)	26

All data represented as median (25%–75% IQR) unless noted otherwise. ^†Augmented renal clearance defined as a CrCl8h >130 mL/min. BMI: Body mass index, SCr: Serum creatinine, BUN: Blood urea nitrogen, ICU: Intensive care unit, CrCl8h: Eight-hour creatinine clearance, ARC: Augmented renal clearance, IQR: Interquartile range

The most common active medication orders at the time of urine collection were vancomycin (52), cefepime (37), piperacillin/tazobactam (33), levetiracetam (22), enoxaparin (22), and meropenem (15) [Table 3]. Seventeen percent of all medications resulted in a different drug dose which was similar for the top six most common drugs, with the exception of piperacillin/tazobactam and enoxaparin which had 3% and 0% discordances, respectively. The analysis of differing drug doses by individual medication is included in Table 3. The primary outcome resulted in a 25% discordance rate, meaning 25% of the urine collections would have resulted in at least one different drug dose for active medications when using CrCl_8h versus CrCl_CG. The correlation between CrCl_8h and CrCl_CG was 0.76 (P < 0.001), with an R²= 0.581. Only 31% of CrCl_CG values were within ± 20% of the CrCl_8h value. Over half (51%) of CrCl_CG values were >120% of CrCl_8h and 18% were below 80% of CrCl_8h [Figure 2].

Table 3.

Active medication orders on the day of urine collection

Medication	Number of active orders	Discordance by medication (%)†
Vancomycin	52	11 (21.2)
Cefepime	37	8 (21.6)
Piperacillin/tazobactam	33	1 (3)
Levetiracetam	22	6 (27.3)
Enoxaparin	22	0
Meropenem	15	4 (26.7)
Lacosamide	9	0
Daptomycin	8	4 (50)
Cefazolin	4	0
Levofloxacin	4	0
Ampicillin/sulbactam	1	1 (100)
Ceftaroline	1	0
Ceftolozane/tazobactam	1	1 (100)
Rivaroxaban	1	0
Aminoglycosides	2	0
Total	212	36 (17%)

^†Discordance by medication calculated as the number of the specified medication orders with a different CrCl8h cutoff than CrClCG divided by the total number of active orders for that medication. CrCl8h: Eight-hour creatinine clearance, CrClCG: Cockcroft–Gault creatinine clearance

Figure 2.

CrCl_CG and CrCl^8h Comparison. CrCl (CG): Cockcroft–Gault creatinine clearance, CrCl (8h): Eight-hour creatinine clearance

DISCUSSION

Our comparison of CrCl_CG and CrCl_8h resulted in at least one discordant renal drug dose for one in every four patients who had a U_8h. Previous studies comparing CrCl_8h and CrCl_CG focused on correlation, bias, and precision or limited CrCl_8h use to identify ARC.[14,15,16,17] Our study is unique because we evaluated the effect on drug dosing at the time of urine collection.

It is not surprising that antibiotics were the most common class of renally eliminated medications at the time of urine collection, especially in a cohort of predominantly critical care patients. There are increasing data suggesting that dosing of antibiotics, especially beta-lactams, needs to be individualized in ICU patients.[18,19] We found that the rates of discordance for the most commonly observed active medication orders [Table 3] were similar to the overall medication discordance, except for piperacillin/tazobactam (3%) and enoxaparin (0%). Lower discordance rates with these medications are likely related to only having a single CrCl cutoff versus a medication like cefepime with three dosing adjustments based on CrCl. Regardless, the use of U_8h to calculate CrCl_8h may be an appropriate surrogate for dose adjustments that would otherwise require therapeutic drug monitoring (for example, with beta-lactam antibiotics).

Differences between the CrCl_8h and the CrCl_CG were not readily apparent based on the correlation in our study. The correlation we observed (0.76) was only slightly lower than that observed in the CG study (0.83) and still represents a fair correlation.[1] However, looking at the distribution of the data revealed that 69% of the CrCl_CG values fell outside ±20% of the urine CrCl values as opposed to only 33% in the original CG study.[1] A few variances in our study may explain our different results. For one, 88% of our patients were in the ICU, whereas the CG study reported their patients were “mainly on medical wards” without further elaboration.[1] Our patients were evenly balanced between male and female patients, compared to almost an entirely male population (96%) in the CG study. Recommendations for the use of the 0.85 correction factor for females in the CG equation came from an estimate of fat and muscle differences between the sexes mentioned in the discussion of the CG study, but this proposed adjustment was not evaluated.[1] The use of this correction could lead to inaccuracies in the estimation of CrCl, especially for female patients, and may lead to inadequate medication adjustments.

A number of limitations need to be considered for our study. First, the CrCl_8h is not the established gold standard for estimating CrCl. Furthermore, our study was retrospective and conducted at a single center using a small study population. The decision to order U_8h was not standardized and left to clinician discretion. Applying our exclusion criteria also removed 50 collections of the 150 records reviewed from the analysis. We collected apriori list of medications and did not record other active orders for renally adjusted medications. Furthermore, the CrCl_8h calculation in our EMR assumes a urine collection time interval of 8 h. The nurse does not record the actual amount of time for the collections, which could lead to over or underestimation of CrCl_8h. The CG study selected patients with renal function at a steady state by excluding patients with SCr values that differed by >20%. Based on our median SCr of 0.76 (0.52–1.06) and only 13% having a ≥0.3 mg/dL SCr increase in 48 h before collection, few had an acute kidney injury.[1] Furthermore, our study did not assess the clinical outcomes for our patients.

CONCLUSION

We determined that there was 25% discordance in drug dosing between CrCl_8h and CrCl_CG in our study population. We also determined that over two-thirds of U_8h resulted in CrCl_CG values falling outside ± 20% of the CrCl_8h. Our findings support the need for subsequent studies comparing CrCl_CG to CrCl_8h for renal drug dosing and resulting clinical outcomes.

Research quality and ethics statement

The authors of this manuscript declare that this scientific work complies with reporting quality, formatting, and reproducibility guidelines set forth by the EQUATOR Network. The authors also attest that this clinical investigation was determined to require Institutional Ethics Committee, Research Cell, King George's Medical University, Lucknow and appropriate approval (84^th ECM II-B-Thesis) was granted by the Research Cell, King George's Medical University, Lucknow.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.

Ethical conduct of research

This study was approved by the Institutional Review Board / Ethics Committee. The authors followed applicable EQUATOR Network (http://www.equator-network.org/) guidelines during the conduct of this research project.