Sample Processing and Stability for Urine Biomarker Studies

Abstract

Background

Current methods of processing and storing urine samples have not been compared systematically to determine optimal conditions for advancing research on urinary biomarkers. Often, preanalytical handling is nonideal, especially considering the COVID-19 pandemic; consequently, we compared the effects of different short-term storage and processing methods on urinary biomarker measurements.

Methods

Spot urine samples were collected via a Foley catheter from 20 hospitalized patients from the Yale New Haven Hospital within 48 hours postcardiac surgery. The effects of 3 urine storage and processing methods on biomarkers were tested: (a) 48-hour temporary storage at 4 °C prior to freezing at −80 °C, (b) 48-hour temporary storage at 25 °C prior to freezing at −80 °C, and (c) no centrifugation and immediate storage at −80 °C. Established Meso-Scale Device assay methods were used to measure the urine concentrations of 18 biomarkers: interferon gamma (IFN-ɣ), interleukin (IL)-10, IL-12p70, IL-13, IL-1β, IL-2, IL-4, IL-6, IL-8, IL-18, tumor necrosis factor alpha (TNF-α), epidermal growth factor (EGF), neutrophil gelatinase-associated lipocalin (NGAL), osteopontin (OPN), uromodulin (UMOD), kidney injury molecule-1 (KIM-1), monocyte chemoattractant protein-1 (MCP-1), and chitinase-3-like protein 1 (YKL-40).

Results

Measurements of most biomarkers investigated remained stable after temporary storage at 4 °C. IL-6, IL-8, KIM-1, MCP-1, YKL-40, EGF, and NGAL were stable across all 3 processing conditions. IL-12p70 and IL-4 demonstrated significant differences in all tested conditions compared to the reference standard.

Conclusions

We identified several notable biomarkers that are robust to variations in preanalytical techniques and can be reliably investigated with nonideal handling conditions.

Impact Statement

Research on urinary biomarkers of acute kidney injury (AKI) is an expanding field to better understand diagnosis, progression, and treatment of the condition. With the current COVID-19 pandemic, studies aimed to investigate such biomarkers must frequently use nonideal specimen handling and processing procedures. Consequently, it is important to further elucidate how different storage and processing methods affect measurements of urinary biomarker levels. Our findings suggest several biomarkers that can be reliably investigated with the nonoptimal preanalytical handling of samples, supporting continued research to advance the understanding and clinical application of urinary biomarkers in the setting of AKI.

Introduction

Urinary biomarkers are important for research on early diagnosis, risk stratification, and prognosis of acute kidney injury (AKI) (1), with more than 65 000 published studies investigating various plasma and urine biomarkers in AKI (2). Most studies do not analyze biomarkers immediately after sample collection; instead, samples are stored for future batch analyses. Sample processing and storage protocols vary across research sites based on conventional practice, rather than on evidence. The current standard is immediate centrifugation and freezing at −80 °C (3). In practice, some samples are temporarily left at room temperature or refrigerated, or not centrifuged. In the context of the COVID-19 pandemic, sample handling is compromised with restrictions on research personnel in patient areas, protocols limiting the number of laboratory personnel, and lack of accessibility of adequate facilities not attached to research facilities. We previously examined the effects of centrifugation and short-term storage methods on measurements of 5 select biomarkers (4). We have expanded our studies to 18 urinary biomarkers of value in studies of AKI, including epidermal growth factor (EGF), neutrophil gelatinase-associated lipocalin (NGAL), osteopontin (OPN), uromodulin (UMOD), interleukin-18 (IL-18), kidney injury molecule-1 (KIM-1), monocyte chemoattractant protein-1 (MCP-1), chitinase-3-like protein 1 (YKL-40), and a panel of 10 proinflammatory biomarkers.

Materials and Methods

Patient Population

Written consent was obtained from 20 patients from the Yale New Haven Hospital. Eligible participants had at least 1 risk factor for AKI: emergency surgery, preoperative serum creatinine >2 mg/dL, ejection fraction <35% or grade 3 or 4 left ventricular dysfunction, age >70 years, diabetes mellitus, concomitant coronary artery bypass graft and/or valve surgery, or repeat revascularization surgery. A random 50-mL spot urine sample was collected from each participant via a Foley catheter within 48 hours postcardiac surgery between March 2012 and April 2012. This study was approved by the Yale’s research ethics committee at time of sample collection, and currently approved by the Johns Hopkins University IRB committee.

Processing

Urine samples were processed within 4 hours of collection with either the reference protocol [immediate centrifugation (5000g at 4 °C for 10 min), aliquoting into 1-mL aliquots, and storage at −80 °C] or 1 of 3 test conditions: condition A, immediate centrifugation and aliquoting followed by 48-hour temporary storage at 4 °C before transfer to −80 °C; condition B, immediate centrifugation and aliquoting followed by 48-hour temporary storage at 25 °C before transfer to −80 °C; and condition C, no centrifugation but immediate aliquoting and storage at −80 °C (Fig. 1).

Figure 1.

Schematic representation of the experimental protocol.

Biomarker Analysis

Urine dipstick tests were performed on all samples. Analysis was conducted between November 2017 and October 2020. Each panel of biomarkers (Table 1 in the online Data Supplement) was analyzed in a single batch using established multiplex protocols on Meso-Scale Device (MSD) assays by researchers blinded to the processing condition. Biomarkers tested were interferon gamma (IFN-ɣ), interleukin (IL)-10, IL-12p70, IL-13, IL-1β, IL-2, IL-4, IL-6, IL-8, IL-18, tumor necrosis factor alpha (TNF-α), KIM-1, MCP-1, YKL-40, EGF, NGAL, OPN, and UMOD. MSD assay detection limits are reported for each biomarker (Supplemental Table 1).

Statistical Methods

Statistical analyses were performed using MATLAB R2019a (MathWorks, Natick, MA). To determine the strength of the relationship between biomarker measurements after the reference protocol and conditions A–C, Pearson correlation coefficients (r) were calculated. Concordance correlation coefficients (CCC) (5), ranging from −1 (perfect negative agreement) to 1 (perfect agreement), were also calculated to quantify the agreement between the same values (6). For biomarkers with less than 25% of values below detection limits, the lower detection limit was imputed for measurements below the limit. Data was visualized using correlation and Bland–Altman plots. Significance of the difference between the reference protocol and conditions A–C was determined by visual evaluation of these figures along with the slope of the best fit line, CCC, and the mean/median bias.

Results

Twelve biomarkers had no values below detection limits, and all but IL-13 had less than 25% of values below detection limits (Supplemental Table 1). IL-13 was excluded from correlation analysis due to high levels of undetectable values; the proportion of detectability across all 4 processing conditions is presented in Supplemental Table 2. Table 1 summarizes biomarker stability metrics under the tested conditions. No significant differences were observed in levels of 7 biomarkers (IL-6, IL-8, KIM-1, MCP-1, YKL-40, EGF, and NGAL) across all 3 processing conditions compared to the reference standard. IL-4 and IL-12p70 were not correlated across all conditions. Correlation and Bland–Altman plots for each biomarker are provided (Supplemental Figs S1–S18).

Table 1.

Agreement between 3 preanalytic conditions.

	Urine biomarker	Reference	Condition A			Condition B			Condition C
			(4 °C)			(25 °C)			(No centrifugation)
		Median (IQR) (pg/mL)a	r b	CCCc	Significant differenced	r	CCC	Significant difference	r	CCC	Significant difference
Biomarkers with stability across all conditions	IL-6	1.72 (0.93, 2.39)	0.995	0.995	None	0.997	0.984	None	0.996	0.996	None
	IL-8	3.18 (0.91, 8.84)	0.998	0.998	None	0.998	0.995	None	0.979	0.989	None
	KIM-1	1.38 × 103 (0.39 × 103, 2.43 × 103)	0.998	0.998	None	0.983	0.965	None	0.936	0.962	None
	MCP-1	540.99 (150.03, 812.81)	0.999	0.999	None	0.987	0.978	None	0.997	0.997	None
	YKL-40	2.17 × 103 (0.32 × 103, 4.95 × 103)	0.995	0.997	None	0.995	0.996	None	0.945	0.968	None
	EGF	1.86 × 103 (1.20 × 103, 3.64 × 103)	0.997	0.997	None	0.996	0.988	None	0.996	0.997	None
	NGAL	1.07 × 104 (0.21 × 104, 3.73 × 104)	1.000	0.999	None	0.998	0.999	None	0.993	0.995	None
Biomarkers with Stability across one or 2 conditions	IL-2	0.32 (0.25, 0.42)	0.871	0.932	None	0.618	0.732	None	0.498	0.659	Increase
	OPN	4.29 × 105 (1.08 × 105, 9.58 × 105)	0.990	0.975	None	0.811	0.693	Decrease	0.864	0.915	None
	IFN-ɣ	0.26 (0.16, 0.43)	0.729	0.882	None	0.620	0.673	Increase	0.629	0.575	Increase
	IL-10	0.040 (0.030, 0.072)	0.832	0.912	None	0.648	0.741	Increase	0.822	0.713	Increase
	TNF-α	0.15 (0.10, 0.18)	0.812	0.900	None	0.736	0.789	Increase	0.865	0.709	Increase
	IL-1β	0.19 (0.13, 0.32)	0.972	0.985	None	0.957	0.966	Increase	0.916	0.903	Increase
	IL-18	104.29 (34.17, 161.72)	0.933	0.926	Increase	0.317	0.441	Decrease	0.993	0.995	None
	UMOD	1.30 × 106 (7.32 × 106, 2.46 × 106)	0.931	0.961	None	0.868	0.832	Decrease	0.793	0.789	Increase
Biomarkers without stability across any conditions	IL-12p70	0.064 (0.045, 0.087)	0.380	0.453	Increase	0.577	0.356	Increase	0.168	0.156	Increase
	IL-4	0.027 (0.021, 0.032)	0.427	0.596	Increase	0.538	0.441	Increase	0.267	0.202	Increase

^aThe median concentration and interquartile range (IQR) measured after processing with the reference protocol.

^bPearson correlation coefficient. r>0.9 represents excellent correlation. r between 0.7 and 0.9 denotes good correlation. r between 0.3 and 0.5 indicates low correlation. r<0.3 indicates very poor correlation.

^cLin’s concordance correlation coefficient. A CCC of 0.90 or greater is considered excellent agreement, as recommended by Lin (5). Using our prior experience and visual plots, we established a CCC of 0.80 or higher as good agreement.

^dSignificant difference between values measured after the reference protocol and each condition. If a significant difference exists, the direction of the difference is detailed.

Seven urinary biomarkers from the inflammation pathway (IFN-g, IL-10, IL-1β, IL-2, IL-6, IL-8, and TNF-α) showed excellent agreement between measurements from samples processed with the reference standard and those temporarily stored at 4 °C. Significant differences in measurements after storage at 4 °C were observed with IL-18. Inflammatory biomarkers not affected by short-term storage at 25 °C included IL-2, IL-6, IL-8, and TNF-α. No significant differences in biomarker levels were observed for IL-6, IL-8, and IL-18 measurements between samples processed by the reference standard and without centrifugation.

Of the noninflammatory biomarkers, KIM-1, MCP-1, YKL-40, EGF, NGAL, OPN, and UMOD showed high agreement in biomarker levels between the reference protocol and short-term storage at 4 °C. KIM-1, MCP-1, YKL-40, EGF, NGAL measurements in samples temporarily stored at 25 °C were also in excellent agreement with reference standard measurements. OPN and UMOD urine measurements were no longer stable in this condition. Excluding UMOD, all noninflammatory biomarkers analyzed were not affected by lack of centrifugation.

Discussion

No significant differences were observed in levels of IL-6, IL-8, KIM-1, MCP-1, YKL-40, EGF, and NGAL with all tested conditions compared to the reference standard. This suggests versatility in processing methods for studies focusing on these biomarkers, allowing continued research when rigorous preanalytic handling procedures cannot be performed. In contrast, IL-12p70 and IL-4 demonstrated instability under all conditions, suggesting the need for the reference protocol when targeting these biomarkers. For largely all remaining biomarkers analyzed, measurements after temporary storage at 4 °C did not significantly differ from measurements obtained with the reference protocol.

In half of tested biomarkers, stability at 25 °C was significantly altered compared with the reference standard, suggesting that leaving samples in nonrefrigerated conditions does not provide reliable measurements for many biomarkers. Most biomarkers showed significantly higher values after temporary storage at room temperature compared to the reference protocol, indicating that observed biomarker levels may be systemically inflated within analyzed biosamples. At room temperature, changes in the quaternary structures of these proteins may increase the degree of binding with assay antibodies, similar to changes in the R and T forms of hemoglobin that affect its affinity for oxygen (7). However, observed values of IL-18, OPN, and UMOD were significantly lower after short-term storage at 25 °C relative to handling with the reference standard, likely due to degradation or polymerization.

Biomarkers’ stability without sample centrifugation was similar to that at 25 °C, with only 9 analyzed biomarkers showing high agreement with the reference standard. Most deviations were significantly higher than the reference standard measurements.

The observed stability of NGAL after short-term storage at 4 °C (8, 9) and 25 °C (9) in this study is consistent with previous studies. The stability of KIM-1 after temporary storage at 4 °C is also consistent with past work (8). For IL-18, we also observed a decrease in concordance between short-term storage at 4 °C and 25 °C, as previously described (4). However, deviating from previous findings (4), measurements of IL-18 after temporary storage at 4 °C were significantly increased compared with the reference protocol, despite high concordance.

A 2014 study on UMOD found similar results, with significantly elevated levels measured in noncentrifuged samples (10). Notably, enzyme-linked immunosorbent assays were used (10), whereas we used MSD assays, suggesting that results related to the effects of no centrifugation on urinary UMOD measurements hold across different immunoassay platforms. This work further supports our finding that measurements of UMOD significantly decrease after storage at room temperature (20 °C) (10). However, contrary to our results, they reported a significant decrease in observed UMOD concentrations after storage at 4 °C (10). This may be explained by their prolonged storage time of 1 week at 4 °C compared with 48 hours in our protocol.

Certain limitations to the study must be addressed. First, the results can only be generalized to the studied biomarkers after measurement by the specific assays used as different assays may produce different results. A comparison between immediate storage at −80 °C and fresh, unfrozen tissue was also not included. This may be important to investigate because in clinical settings, biosamples are often analyzed immediately after collection. Short-term storage was limited to 48 hours, as this seemed feasible, even when accounting for transportation between facilities. Further investigation into prolonged temporary storage at 4 °C, 25 °C, and −80 °C may be insightful for biospecimen handling and storage protocols in clinical research. Due to the small sample size, disease status of enrolled participants may have affected results. A final limitation is that effects of protease inhibitors were not evaluated.

In conclusion, we found several important biomarkers to be robust and stable to preanalytic handling concessions during the pandemic or generally when immediate sample processing is not feasible.

Supplemental Material

Supplemental material is available at The Journal of Applied Laboratory Medicine online.

Glossary

Nonstandard Abbreviations

IFN-ɣ

interferon gamma

IL

interleukin

TNF-α

tumor necrosis factor alpha

EGF

epidermal growth factor

NGAL

neutrophil gelatinase-associated lipocalin

OPN

osteopontin

UMOD

uromodulin

KIM-1

kidney injury molecule-1

MCP-1

monocyte chemoattractant protein-1

YKL-40

chitinase-3-like protein 1

AKI

acute kidney injury.

 

Disclaimer: Opinions detailed in this study reflect those of the authors and not those of the funding agency.