Stability of volatile organic compound metabolites in urine at various storage temperatures and freeze-thaw cycles for 8 months

Abstract

The urinary concentrations of mercapturic acid metabolites of volatile organic compounds (VOCs) have been used as biomarkers of human exposure to this class of chemicals. However, long-term stability of these VOC metabolites (VOCMs) in urine at various storage conditions such as temperature, duration, and freeze-thaw cycles is not known. In this study, spot urine samples collected from three volunteers, stored at 22°C (room temperature: RT), 4°C (refrigerator) and −20°C (freezer) for up to 240 days were analyzed at weekly to monthly interval for a total of 19 time points. Samples stored at 4°C and −20°C underwent 18 freeze-thaw cycles at RT for 30 min at each of the time points. Among 38 VOCMs analyzed, up to 18 metabolites were detected at concentrations above their respective detection limits on Day 0 (baseline concentration), and the concentrations of several VOCMs declined with the storage duration. Eight to ten VOCMs were lost completely within 240 days of storage at RT, compared to between two and five at 4°C and between one and seven at −20°C. The loss rate varied depending on the sample, storage temperature, VOCM, and number of freeze-thaw cycles. Storage of urine at RT led to a rapid loss of VOCMs in comparison to that stored at 4°C or −20°C. Among VOCMs measured, CEMA, SBMA, GAMA, DHBMA, AMCC, TCVMA, and HPMMA were lost more rapidly than the other metabolites. CMEMA, a major VOCM found in all three urines at baseline, showed a rapid loss in those of two volunteers but not of the other volunteer, suggesting to sample to sample variation in lose rates. Freeze-thaw cycles considerably affected VOCMs concentrations in urine stored at 4°C or −20°C. It is recommended that urine samples are analyzed for VOCMs within a couple of months of collection and stored at temperatures below −20°C, with minimal or no freeze-thaw cycles. This study highlights the need for appropriate storage conditions to maintain the integrity of samples for biomonitoring studies.

Graphical Abstract

1. Introduction

Volatile organic compounds (VOCs) are ubiquitous environmental pollutants. The International Agency for Research on Cancer (IARC) has classified VOCs such as benzene, trichloroethylene, and 1,3-butadiene as known carcinogens; acrylamide, acrylonitrile, N,N-dimethylformamide, ethylbenzene, isoprene, and styrene as probable carcinogens; and toluene, acrolein, crotonaldehyde, and xylene as ‘unclassifiable’ due to the lack of data (Pal et al., 2022 a, b; Li et al., 2021; WHO, 2010). The United States Environmental Protection Agency (EPA) has categorized 1,3-butadiene, propylene oxide, toluene, styrene, xylene, trichloroethylene, tetrachloroethylene, acrolein, acrylonitrile, acrylamide, carbon disulfide, and vinyl chloride as hazardous air pollutants (US EPA, 2017). Human exposure to VOCs is widespread and is a significant public health concern.

Human exposure to VOCs is linked to increased risks for respiratory illnesses, birth defects, leukemia, neurocognitive deficits, reproductive impairment, and cancers (Li et al., 2021; McGraw et al., 2021). Following inhalation exposure, VOCs are absorbed by the lungs, where they diffuse through the capillaries and alveolar type I cells, and quickly equilibrate between the capillary blood and alveoli (NRC, 2009). The absorbed VOCs are then distributed via systemic circulation, reach the liver, and undergo hepatic metabolism. VOCs are biotransformed into more water-soluble metabolites via the action of hepatic cytochrome P450 enzymes (Frigerio et al., 2019). VOCs and/or their metabolites conjugate with glutathione, leading to the formation of water-soluble mercapturic acids, (referred to as mercapturic acid metabolites of VOCs or VOCMs), which are then excreted via urine (Li et al., 2021; Pal et al., 2023).

The traditional approach to assess human exposure to VOCs is through the analysis of these chemicals in inhaled air or exhaled breath, but the measurements are prone to uncertainties associated with the volatile nature of these chemicals. Nevertheless, recent developments in liquid chromatography-tandem mass spectrometry (LC-MS/MS) permit direct analysis of VOCMs in urine, which is a measure of internal dose of exposure. In comparison to parent VOCs, VOCMs are less volatile and can be measured reliably in urine. Furthermore, urinary VOCMs have longer biological half-lives than the parent compounds measured in blood (Boyle et al., 2016). The urinary biomarkers of VOCMs have been widely used as biomarkers of exposure in epidemiologic studies. Nevertheless, the stability of VOCMs in urine stored at different temperatures and duration is not well understood.

Prospective epidemiological and exposure assessment studies collect and store biospecimens such as urine for analysis at a later time, as not all target analytes are recognized at the beginning of a study. Although recommended practices include storage of specimens under frozen conditions in small aliquots, it is not always feasible. It is important to examine whether the concentrations of urinary chemicals are affected by storage temperature, duration and freeze-thaw cycles. Urinary metabolites of environmental chemicals can be altered by temperature and microorganisms, which can affect concentration profiles (Wang et al., 2019; Laparre et al., 2017). High temperatures can lead to evaporation or concentration, and microbial growth contribute to degradation of analytes (Mohr et al., 2016). Information pertaining to stability of analytes at different storage conditions is useful to develop proper storage protocols for biospecimens.

Very few previous studies have assessed stability of environmental chemicals in biological specimens. For example, stability of bisphenol A (BPA) conjugates was examined in urine for up to six months of storage at room temperature (RT), 4°C and −70°C (Ye et al., 2007). Conjugated forms of BPA stored at −70°C were stable for six months whereas storage at RT led to degradative loss within a day. Stability of BPA dimethylacrylate and triethylene glycol dimethacrylate was assessed following storage of saliva at −20°C and −70°C (Atkinson et al., 2002). The concentrations of these chemicals were unchanged in saliva stored at −70°C in comparison to that stored at −20°C. A stability study that evaluated phenols in serum stored at 4°C, 25°C, or 37°C reported that BPA and other phenols were stable at 37°C for 30 days (Ye et al., 2009). The stability of phthalate metabolites and BPA in urine stored at RT (20°C) showed that the concentrations decreased by 15% to 44% after 4 weeks (Guo et al., 2013).

This is the first study to investigate the stability of VOCMs in urine at three storage temperatures, namely, room temperature (RT; 22°C), 4°C and −20°C and up to 18 freeze-thaw cycles for 240 days. This study provides valuable information on appropriate urine storage conditions for accurate and reliable quantification of VOCMs for exposure assessment and epidemiologic studies.

2. Materials and methods

2.1. Sample collection

Urine samples were selected from a pool of freshly collected specimens available from several anonymous volunteers. Spot urine samples collected in 2022 (approximately 25 mL) from three healthy, non-smoking individuals with no history of occupational exposure (ages 22–49 yrs) residing in New York State, USA, were selected for this study. Volunteer 1 was a female and volunteers 2 and 3 were males. No other demographic information was available for these specimens. The project met the criteria for exemption with approval from the Institutional Review Board of New York State Department of Health. The urine samples were aliquoted and stored without any preservative at three storage temperatures: room temperature (22°C), 4°C (Thermo Scientific laboratory refrigerator; Waltham, MA, USA) and −20°C (Norlake Scientific laboratory freezer; Hudson, WI, USA). The urine aliquot taken on the day of collection was marked as Day 0. An aliquot of urine was drawn for up to 240 days during March to December 2022, into a glass vial, at weekly to monthly interval for a total of 19 time-points, from each of the three samples stored at three different temperatures. A total of 342 samples were analyzed for 38 VOCMs. The samples were analyzed within 24 h of aliquoting at different time points. The urine samples that were stored at 4°C and −20°C endured up to 18 freeze-thaw cycles, for aliquoting at a given time point, prior to analysis. The thawing cycle included removal of samples from the cold storage, and maintained at RT for 30 min prior to aliquoting and analysis.

2.2. Target analytes

A detailed list of 38 target VOCMs and the corresponding labeled internal standards (ISs; n = 35) used in the analysis are provided in Supplementary Text 1 (Text S1).

2.3. Determination of VOCMs in urine

The urinary concentrations of VOCMs were determined by following a method described elsewhere, with slight modifications (Pal et al., 2023). Briefly, 50 μL of a 500 ng/mL IS mixture was spiked into a mixture of 50 μL of urine and 400 μL of 15 mM ammonium acetate buffer (pH ~6.8) taken in a 2-mL snap-cap polypropylene (PP) microcentrifuge tube (Costar^®, Corning Incorporated, Corning, NY, USA). The sample mixture was vortexed and centrifuged (MiniSpin^®, Eppendorf, Enfield, CT, USA) at 10,000g for 3 min. The supernatant was transferred into a 300-μL glass insert that had been placed in a 2-mL vial, for high-performance liquid chromatography–tandem mass spectrometry (HPLC-MS/MS) analysis. Identification and quantification of target analytes were accomplished by electrospray ionization (ESI) MS/MS using a Waters ACQUITY UPLC system (Waters, Milford, MA, USA) interfaced with an AB SCIEX QTRAP 4500 mass spectrometer (Applied Biosystems, Foster City, CA, USA). The HPLC-MS/MS analysis was performed in a negative ionization multiple reaction monitoring mode (for details see Tables S1 and S2). Chromatographic separation of target analytes was accomplished using a Zorbax Eclipse Plus C18 column (100 mm × 4.6 mm × 3.5 μm; Agilent, Santa Clara, CA, USA) serially connected to a Betasil C18 guard column (20 mm × 2.1 mm, 5 μm; Thermo Scientific, West Palm Beach, FL, USA). The mobile phases were 0.1% acetic acid in methanol (A) and 0.1% acetic acid in water (B) maintained at a flow rate of 400 μL/min. The mobile-phase gradient flow was set as: 85% B from 0.0 to 2.0 min, 60% B from 2.0 to 7.0 min, 30% B from 7.0 to 9.0 min, 5% B from 9.0 to 12.5 min and 85% B from 12.5 min to 15 min. The total run time was 15 min. The HPLC column was maintained at 40°C using a heater (Cole-Palmer Instrument Company, Vernon Hills, IL, USA). The sample injection volume was 5 μL.

Analytes were quantified using an isotope dilution method, by taking the ratio of absolute response of each target analyte to that of the corresponding isotope-labeled IS. Peak integration, calibration, and quantification were performed using Analyst^® software (version 1.7.2, AB Sciex, Framingham, MA, USA). A linear regression model was used (1/x weighting, where x is the standard concentration) for fitting the calibration curve.

2.4. Quality assurance (QA) and quality control (QC)

The HPLC syringe was rinsed twice with a mixture of methanol and water (50:50, v/v) before and after each injection. Methanol was used in place of urine for procedural blanks, which were analyzed with every batch of 10 samples. None of the target analytes was found in procedural blanks at concentrations above the limit of detection (LOD). A 16-point standard calibration curve, prepared at a concentration range of 0.01–1000 ng/mL (in 15 mM ammonium acetate buffer) with a regression coefficient of >0.99 for each analyte, was used for the quantification of target analytes. Recoveries of VOCMs through the analytical procedure were determined by spiking a known amount of target analytes into a sample of synthetic urine (Surine^® Negative Urine Control; Cerilliant Corporation, TX, USA) at three different concentrations (low, 10 ng/mL; medium, 50 ng/mL; high, 100 ng/mL). The relative recoveries of analytes were in the range of 70–130% (Table S3). The relative standard deviation (RSD%) of repeated analysis (n = 9) of the QC samples was ≤ 20% for all VOCMs. The method LODs were in the range of 0.087–17.2 ng/mL, except for MU (23.5 ng/mL) (Table S3). The performance of the analytical method was determined through the analysis of National Institute of Standards and Technology (NIST) smokers’ (3672) and non-smokers’ urine (3673) Standard Reference Materials (SRMs), and the measured concentrations were in the range of 70–130% of the certified values (Table S4). The data quality was further validated through participation in German External Quality Assessment Scheme (G-EQUAS) proficiency testing (PT) by measuring 15 VOCMs, for which reference values exist at two different concentrations (environmental and occupational levels) in human urine samples and the results for all analytes were within the tolerance range (Table S5).

2.5. Data analysis

The concentrations of analytes below the LODs were substituted with a value of LOD divided by square root of 2. Statistical analyses were performed using SPSS 19.0 (SPSS Inc., Chicago, IL, USA) and GraphPad Prism v.9.1.1 (GraphPad Software, San Diego, CA, USA). The significance level was set at p < 0.05. The summed concentrations of VOCMs present in urine are presented as ∑VOCMs. Since many of the variables were not normally distributed, non-parametric tests were applied for statistical analysis. The Mann-Whitney test and Kruskal-Wallis test were used in comparing analyte concentrations between various categories. Correlations among VOCMs were determined by Spearman’s rank correlation. Analytes with detection frequencies (DFs) < 60% were excluded from further statistical analysis.

3. Results and discussion

3.1. Loss of VOCMs at different storage conditions

Thirty-eight VOCMs were analyzed in urine samples stored at three different temperatures for 8 months. Eighteen VOCMs were found in the urine of volunteers 1 and 2, whereas 16 VOCMs were found in the urine of volunteer 3, at concentrations above the LOD at baseline (day 0) (Table 1). The baseline (Day 0) and Day 240 concentrations of VOCMs found in urine stored at three different temperatures from the three volunteers are presented in Table 1. The VOCM concentrations and profiles varied among the three volunteers at baseline (Day 0) and at the completion of the study (Day 240) for all three storage temperatures (Figure 1). Bromopropane and methylating agents that were found in the urine of volunteers 1 and 2 at baseline (day 0) were not found in the urine of volunteer 3. TTCA (metabolite of carbon disulfide) was elevated in the urine of volunteer 2. The sum concentrations of 18 VOCMs (∑VOCM) in urine of three volunteers at baseline varied by an order of magnitude, and ranged from 661 to 6210 ng/mL (Table 1). The Day 0 concentration of VOCMs varied for volunteer 1 from 2.09 (SBMA) to 601 (CMEMA) ng/mL; for volunteer 2 from 3.74 (CYMA) to 2010 (CMEMA) ng/mL and for volunteer 3 from 1.91 (CYMA) to 158 (CMEMA) ng/mL. In general, CMEMA (a metabolite of crotonaldehyde) was the most abundant VOCM found in the urine of all three volunteers, accounting for 24–32% of ∑VOCM (Figure 1) and that concentration varied by an order of magnitude among the three volunteers.

Table 1.

Comparison of urinary concentrations (ng/mL) of volatile organic compound metabolites (VOCMs) at baseline (Day 0) and day 240 at three storage temperatures. RT= room temperature (22° C). < LOD: below limit of detection.

Parent VOC	VOCM	Volunteer 1				Volunteer 2				Volunteer 3
		Baseline conc. (day 0)	Conc. on day 240 at RT	Conc. on day 240 at 4° C	Conc. on day 240 at −20° C	Baseline conc(day 0)	Conc. on day 240 at RT	Conc. on day 240 at 4° C	Conc. on day 240 at −20° C	Baseline conc (day 0)	Conc. on day 240 at RT	Conc. on day 240 at 4° C	Conc. on day 240 at −20° C
Acrolein	CEMA	21.3	< LOD	< LOD	< LOD	232	294	195	143	67.8	< LOD	< LOD	< LOD
Acrolein	3HPMA	187	34.7	20.6	22.7	686	149	126	130	126	22.9	11.6	6.52
Propylene oxide	2HPMA	73.5	46.8	70.7	50.2	259	230	219	197	48.7	40.0	50.5	33.1
Toluene	SBMA	2.09	< LOD	0.902	0.656	11.4	9.56	8.66	7.05	2.06	< LOD	0.904	0.207
Acrylamide	AAMA	30.6	< LOD	12.7	8.87	145	84.3	88.4	58.4	11.3	0.433	5.85	0.370
Acrylamide	GAMA	9.33	39.6	3.85	1.13	42.3	15.7	119.0	69.3	1.92	< LOD	< LOD	< LOD
Acrylonitrile	CYMA	4.73	4.86	3.54	1.82	3.74	2.08	6.63	2.11	1.91	< LOD	< LOD	< LOD
Xylene	2MHA	45.4	61.6	63.4	31.1	30.7	30.4	63.8	42.0	4.42	14.6	16.5	11.7
Xylene	3+4 MHA	56.2	37.2	31.3	18.1	69.1	23.2	54.4	20.4	9.49	22.2	11.4	13.8
1,3-Butadiene	DHBMA	345	< LOD	101	43.8	599	343	221	277	68.85	<LOD	11.0	<LOD
1,3-Butadiene	MHBMA 1+2+3	24.0	9.50	13.4	11.5	126	141	173	99	6.40	8.46	8.98	2.81
Crotonaldehyde	CMEMA	601	1.95	301	186	2010	1325	1450	1030	158	33.7	84.7	19.7
N,N-Dimethylforma mide	AMCC	50.4	< LOD	48.8	21.0	154	< LOD	< LOD	< LOD	16.9	< LOD	< LOD	< LOD
1-Bromopropane	BPMA	8.07	< LOD	3.55	2.07	259	23.9	50.5	61.9	< LOD	< LOD	< LOD	< LOD
Carbon disulfide	TTCA	376	< LOD	209	122	1230	1410	20.8	153	79.6	18.0	35.7	6.50
Tetrachloroethylene	TCVMA	19.3	< LOD	3.65	1.34	19.7	5.18	5.34	3.3	3.68	< LOD	< LOD	< LOD
Methylating agents	MMA	20.5	< LOD	< LOD	2.5	13.0	15.0	1.72	2.79	< LOD	< LOD	< LOD	< LOD
Crotonaldehyde	HPMMA	98.5	< LOD	59.6	28.1	322	280	334	232	53.8	< LOD	15.5	< LOD
∑VOCM		1970	234	947	552	6210	4380	3140	2530	660	126	253	94.7

38 VOCM metabolites were analyzed. The table presents only those that were found in samples. AAMA-Sul, CMEMA, HEMA, 1,2- & 2,2-DCVMA, 2,4-, 2,5- & 3,4-DPMA, ATCA, PGA, 1,2- & 2,2-PHEMA, MA, SPMA, MU, EMA, NANPC, and IPM 1 & IPM 3 were not detected in any of the urine samples. Samples stored at 4°C and −20°C endred 18 freeze-thaw cycles.

Figure 1.

Composition of VOCMs in urine strored at room temperature at baseline (Day 0) and on day 240 from three volunteers.

The ∑VOCM concentrations in urine stored at RT on Day 240 ranged from 234 to 4380 ng/mL (Table 1). There was a considerable inter-individual variability in the % loss in ∑VOCM among the three volunteers. Urine from volunteers 1 and 3 exhibited 80–88% loss in ∑VOCM whereas that of volunteers 2 showed 30% loss (Table 1). The differences in loss rates of ∑VOCM between individuals may be related to microbial composition of urine.

In general, as the storage duration and the number of freeze-thaw cycle increased, the concentrations of several VOCMs declined gradually and the concentrations of some VOCMs were below the LOD at certain time points (Figures 2, S4 and S5). The loss rates of ∑VOCM in urine stored at 4°C and −20°C (after 18 freeze-thaw cycles) for 240 days were in the range of 50–62% and 60–86% of the baseline concentrations (Table 1). It is worth to note that significant loss of ∑VOCM was found even for urine stored at −20°C, but after 18 freeze-thaw cycles. Freezing of samples is expected to increase the stability of environmental chemicals by reducing the likelihood of microbial growth and slowing degradation processes, particularly those mediated by the presence of water – such as hydrolysis._Freezing alone was not likely the cause of reduction in VOCMs, but it is that 18 freeze-thaw cylcles of storing samples at RT for 30 min that affected VOCM concentrations. It is likely that thawing at RT favored microbial growth that resulted in VOCM degradation.

Figure 2.

Temporal changes in the concentration of volatile organic compound metabolites (VOCMs) in urine collected from volunteer 1 stored at three storage temperatures (room temperature, 4° C and −20° C) for 240 days. Note: The samples stored at 4 °C and −20 °C were subjected to multiple freeze-thaw cycles. X-axis represent days of storage.

The rate of loss of VOCMs (i.e., < LOD in comparison to Day 0 concentrations) was more pronounced following storage at RT than those that were stored at 4°C and −20°C (Figures 2, S4, S5; Tables 2, S25, S26 and S27). For instance, urine from volunteer 1 lost 10 VOCMs, namely, CEMA, SBMA, AAMA, DHBMA, AMCC, BPMA, TTCA, TCVMA, MMA and HPMMA, within 240 days, when stored at RT, in comparison to the loss of 2 VOCMs (CEMA, MMA) when stored at 4°C while none was lost when stored at −20°C (Table 2). Similarly, urine from volunteer 3 stored at RT lost 8 VOCMs in 240 days in comparison to the loss of 5 metabolites at 4°C and 7 metabolites at −20°C (Table 2). Urine from volunteer 2 exhibited lower rates of loss and only AMCC was lost completely at the three storage temperatures (Table 2; Figure S4). These results imply that loss VOCMs in urine is related not only to the storage temperature and duration but also from sample to sample and number of freeze-thaw cycles. However, storage of urine at −20°C slowed the loss rate in comparison to storage at RT, for majority of the VOCMs analyzed (Figures 3 and 4). It should be noted that baseline VOCM concentrations were notably higher (2–16 fold) in the urine of volunteer 2 than those of other volunteers (Table 1). Urine specimen that contained elevated concentrations of VOCMs exhibited lower rates of loss at all three temperatures (Figure S4). Higher concentrations could plausibly provide a protective effect, slowing down the degradation process or extending the time required to reach instrumental detection limits. However, the loss of metabolites can also be affected by factors such as composition and profiles of microbiota (Wang et al., 2019; Laparre et al., 2017; Mohr et al., 2016). Overall, our results point to sample-specific loss of VOCMs in urine during storage. Further research is needed to explore the mechanisms underlying degradation rates as related to initial VOCM concentrations.

Table 2.

Number of days to reach VOCM concentrations below the limit of detection (LOD) in urine stored at three temperatures collected from three volunteers in over 240 days. NA: Not applicable, since no complete loss was found at that time point. Only VOCMs that were found in urine above the LOD are included. RT= room temperature (22 °C).

VOCM	Volunteer 1			Volunteer 2			Volunteer 3
	RT	4° C	−20° C	RT	4° C	−20° C	RT	4° C	−20° C
CEMA	60	120	180	NA	NA	NA	90	84	60
SBMA	150	NA	NA	NA	NA	NA	240	NA	NA
AAMA	240	NA	NA	NA	NA	NA	NA	NA	NA
GAMA	NA	NA	NA	NA	NA	NA	105	180	180
CYMA	NA	NA	NA	NA	NA	NA	180	180	180
DHBMA	210	NA	NA	NA	NA	NA	150	NA	105
AMCC	54	NA	NA	54	54	105	12	15	90
BPMA	69	NA	NA	NA	NA	NA	NA	NA	NA
TTCA	150	NA	NA	NA	NA	NA	NA	NA	NA
TCVMA	180	NA	NA	NA	NA	NA	105	210	105
MMA	60	180	NA	NA	NA	NA	NA	NA	NA
HPMMA	180	NA	NA	NA	NA	NA	240	NA	210

38 VOCM metabolites were analyzed. The table presents only those that were found in samples. AAMA-Sul, CMEMA, HEMA, 1,2- & 2,2-DCVMA, 2,4-, 2,5- & 3,4-DPMA, ATCA, PGA, 1,2- & 2,2-PHEMA, MA, SPMA, MU, EMA, NANPC, and IPM 1 & IPM 3 were not detected in any of the urine samples. No 100% loss of 3HPMA, 2HPMA, 2MHA, 3+4MHA, MHBMA and CMEMA was found in any of the samples and are not included on this table.

Samples stored at 4°C and −20°C endred 18 freeze-thaw cycles in 240 days.

Figure 3.

Rate of loss (slope of the concentration curve) of CMEMA and CYMA in urine from volunteer 1 stored at room temperature (RT), 4°C and −20°C for 240 days. Samples stored at 4°C and −20°C endred 18 freeze-thaw cycles and each data point refers to analysis after freeze-thaw cycles.

Figure 4.

Rate of loss (slope of the concentration curve) of CMEMA and CYMA in urine from volunteer 3 stored at room temperature (RT), 4°C and −20°C for 240 days. Samples stored at 4°C and −20°C endred 18 freeze-thaw cycles and each data point refers to analysis after freeze-thaw cycles.

The storage of urine at −20°C slowed the loss of majority of VOCM concentrations (Figures 2–4; based on the slopes). Significant loss of a few VOCMs, especially the two major ones namely, CMEMA and DHBMA, was observed in the urine of volunteer 3, following 8 freeze-thaw cycles in 60 days (Tables S25, S26 and S27). The sum concentrations of CMEMA and DHBMA in the urine of volunteer 3 at baseline were 227 ng/mL whereas that on Day 60 was 104 ng/mL (54% decline). In the urine of other two volunteers, no net reduction in the concentrations of CMEMA, DHBMA, and ∑VOCMs was found for up to 60 days at −20 °C following 8 freeze-thaw cycles (Tables S25, S26 and S27). These results suggest that VOCM concentrations in some urine samples can be significantly affected by freeze-thaw cycles. Studies have shown that freeze-thaw cycles of urine stored at −20°C can affect the stability of some metabolites (Chen et al., 2022; Stevens et al., 2019; Rotter et al., 2017; Klimowska et al., 2023). Freeze-thaw cycles can disrupt the cell membrane of various biomolecules in urine by forming intracellular ice crystals during the freezing process (Bojic et al., 2021; Stevens et al., 2019; Zhang et al., 2015). Therefore, it is recommended that freeze-thaw cycles should be minimized when urine samples are stored at freezing temeratures for trace organic chemical analysis (Klimowska et al., 2023; Chen et al., 2022; Rotter et al., 2017). It is also interesting to note that the concentrations of some VOCMs increased following freeze-thaw cycles for some samples stored at 4°C and −20°C, and this could be related to the transformation of parent VOCs into metabolites. However, this increase in VOCM concentrations was not consistent across all samples and storage temperatures. In general, two of the three samples had no loss of ∑VOCM for up to 8 freeze-thaw cycles when stored at −20 °C.

3.2. Profiles of loss in VOCM concentrations in urine from three volunteers

Among VOCMs, CEMA (metabolite of acrolein) was lost completely within 60 days of storage at RT, whereas this compound was found at concentrations above the LOD for up to 120 days when stored at 4°C and up to 180 days at −20°C (Table 2). SBMA (metabolite of toluene), AAMA (metabolite of acrylamide), DHBMA (metabolite of 1,3-butadiene), AMCC (metabolite of N, N-dimethylformamide), BPMA (metabolite of 1-bromopropane), TTCA (metabolite of carbon disulfide), TCVMA (metabolite of tetrachloroethylene) and HPMMA (metabolite of crotonaldehyde) were all lost following storage at RT but were all found at concentrations above the LOD at 4°C and −20°C for up to 8 months. In general, CEMA, SBMA, GAMA, DHBMA, AMCC, TCVMA, and HPMMA were prone to rapid loss from the urine of volunteers 1 and 3 when stored at RT. AMCC was lost within 54 days of storage of urine at RT for all three volunteers.

The loss of urinary VOCM concentrations was rapid at RT storage followed by 4°C and −20° C (Figures 3 and 4). We calculated the rate of decline in VOCMs concentrations (% loss) at three storage temperatures using the following equation: (Tables S6–S14):

$$\text{\% loss }=\text{ VOCM concentration at }(\text{Day 0 }-\text{ Day 240) }\ast 100/[\text{Day 0]}$$ (1)

The loss rate of VOCMs varied from sample to sample. Apart from those analytes that were lost 100% within 240 days, the loss rate (a positive value) varied from 16.9 to 99.7% at RT, from 16.9 to 98.5% at 4°C and from 20.7 to 97.4% at −20°C for other VOCMs. These results suggest that several VOCMs are subject to degradation at different rates following storage. The rate of loss was more pronounced at RT followed by 4°C and −20°C (Tables S25–S27). Our results are in line with a previous study that found considerable variations (increase/decrease) in concentrations of 55 urinary metabolites including hippurates, benzoates and creatinine with storage duration (Saude and Sykes, 2007).

Although the urine from healthy individuals is presumed to be sterile, studies have reported the occurrence of microflora including bacteria in urine (Kogan et al., 205; Wang et al., 2019; Thongboonkerd and Saetun, 2007; Slupsky et al., 2007; Saude and Sykes, 2007). Bacterial species such as Acinetobacter, Citrobacter, Enterobacter, Escherichia coli, Klebsiella, Proteus, Pseudomonas, Salmonella, Staphylococcus, and Streptococcus have been reported to be present in human urine, and these microbes are capable of transforming aromatic compounds (Wang et al., 2019; Thongboonkerd and Saetun, 2007; Slupsky et al., 2007).

To examine the relationships among VOCM concentrations at different storage conditions, we calculated Spearman’s rank correlation coefficients (Tables S15–S23) and generated heat maps (Figures S1–S13). Urinary concentrations of several VOCMs showed significant positive correlations at RT, 4°C and −20°C (range: ρ = 0.01–0.98), which suggested loss at different storage temperatures is also related to chemical species. Some VOCMs are possibly susceptible to degradation over others (Figures 3 and 4); i.e., compound-specific degradation. Furthermore, a VOCM that would degrade at RT may also degrade at 4°C and −20°C, but at slower rates at lower temperatures. Further research is needed to validate these findings.

Storage of urine at higher temperatures can result in evaporative water loss, bacterial growth and/or contamination (Mohr et al., 2016; Antonini et al., 2012). Our results (days to reach VOCM concentrations below the limit of detection (LOD) for majority of VOCMs, Table 2) are in line with previous studies suggesting that urinary concentrations are more stable at freezing temperatures (−20° C or below) than at RT (Table S24) (Wang et al., 2019; Laparre et al., 2017; Adams et al., 2017; Remer et al., 2014; Alwis et al., 2012; Ye et al., 2009, 2007; Saude and Sykes, 2007). One study examined the stability of VOCMs (n = 28) in urine stored at 4°C and −20°C, but the study duration was only for a week (Alwis et al., 2012). The analytes were stable in urine samples for a week that were stored at −20°C; that study showed that the concentrations of TTCA and PGA were not stable during storage at 4°C for a week. However, the samples were not subject to freeze-thaw cycles in that study. Our results suggest that temperature, duration, sample-specific and chemical-secific factors and freeze-thaw cycles affect the stability of VOCMs in urine. Therefore, we recommend that urine samples should be stored frozen (−20°C or below) immediately after collection and should be analyzed within a few months of storage with minimal or no freeze-thaw cycles.

4. Conclusions

To the best our knowledge, this is the first study to determine the stability of VOCM concentrations in urine samples stored at three different temperatures (22°C, 4°C and −20°C) for 240 days. Several VOCMs declined at all three storage conditions and compound-specific and sample-specific factors appear to influence the loss. Eight to ten VOCMs were lost at room temperature compared to 2 to 5 VOCMs at 4°C and 1 to 7 VOCMs at −20°C in 240 days. CEMA, SBMA, AAMA, DHBMA, GAMA, CYMA, BPMA, TTCA, TCVMA, MMA and HPMMA were prone to loss during long-term storage. Storage of urine at −20°C reduced the loss rate of VOCMs, but freeze-thaw cycles contributed to loss of VOCMs. Our study has some limitations, and therefore caution should be exercised in generalizing the findings. First, spot urine samples were analyzed for only 3 individuals and notable inter-individual variability in loss rates of VOCMs were found. Second, we did not test storage at – 80°C and this temperature may improve the stability of VOCMs. Third, creatinine concentrations were not measured in spot urine samples. It is not known if creatinine normalization may correct for the loss of VOCMs. Fourth, the study design lacks the measurement of VOCMs in urines that were not subjected to any freeze-thaw cycles. Further studies are needed to better understand how freeze-thaw cycles alone can affect the concentrations of urinary VOCMs. Nevertheless, this study provides baseline information on the stability of VOCMs and offers information for future studies that use archived urine samples for exposure assessment and epidemiologic investigations.

Highlights.

• Stability of VOC metabolites was determined in urine stored at 22°C, 4°C and −20°C.

• Loss of VOC metabolites occurred at all three temperatures.

• Loss rates of VOCMs were faster in urines stored at room temperature.

• Freeze-thaw cycles significantly affected urinary VOCM concentrations.

• Inter-individual variability in loss of urinary VOCMs was found.