BIOPRESERVATION AND BIOBANKING Volume 13, Number 2, 2015 ª Mary Ann Liebert, Inc. DOI: 10.1089/bio.2014.0072
Evaluation of the Short-Term Stability of Specimens for Clinical Laboratory Testing Katsuyoshi Ikeda,1 Kiyoshi Ichihara,2 Teruto Hashiguchi,3 Yoh Hidaka,4 Dongchon Kang,5 Masato Maekawa,6 Hiroyuki Matsumoto,7 Kazuyuki Matsushita,8 Shigeo Okubo,9 Tatsuyuki Tsuchiya,10 Koh Furuta,11 on behalf of The Committee for Standardization, The Japanese Society of Laboratory Medicine ( JSLM)
Background: A major concern in both the laboratory-medicine and research communities is the quality of human specimens for analysis. However, there is insufficient scientific evidence regarding optimal conditions for handling and storing routine specimens, especially those in liquid form. Thus, we investigated the stability of clinically relevant samples stored under various conditions. Materials and Methods: Ten clinical laboratories in Japan conducted analyses of the stability of post-clinical (left over after analysis) test samples in relation to temperature and storage duration. We examined serum, whole blood, and urine samples submitted to each laboratory for routine testing. In this study, at least 5 samples for each of 35 tests were analyzed at each laboratory. After completion of routine testing, specimens with sufficient residual volume and values between LL - R/2 (lower limit of reference interval) and UL + R/2 (upper limit) were divided into 300 mL aliquots, where R = UL - LL. Aliquots of serum specimens were stored at either room temperature (23C), 4C, - 20C, or - 80C without light exposure. Aliquots of whole blood and urine specimens were stored at either 23C or 4C. The storage time was either 1, 3, or 7 days. Average differences between pre- and post-storage test results were evaluated for each laboratory test by two-way ANOVA. Fvalues for between-day variations were used for judging the statistical significance of storage-related changes in test values, whereas the ratio of between-day SD to between-individual SD (one-fourth of reference interval) was used to indicate the practical significance of the change. Results and Conclusion: Sample denaturation is clearly temperature- and storage-duration dependent for almost all analytes. In general, specimens were most susceptible to denaturation at 23C, then 4C, - 20C, and - 80C. This study confirmed the accumulated routine, practice-based, detailed knowledge regarding specimen stability and will help to ensure the reliability of laboratory test results.
Introduction
A
major concern for both the laboratory-medicine and research communities is the quality of specimens, especially those obtained from human sources. Thanks to continuous efforts to improve quality control in clinical laboratories, the precision and accuracy of laboratory test
results have improved greatly in recent years in Japan. Although experience and anecdotal evidence help to maintain high analytical quality in clinical laboratories, there is insufficient scientific evidence regarding the optimal conditions for handling and storing routine samples, especially those in liquid form. Several investigations have focused on denaturation of human specimens, particularly liquid
1
Department of Clinical Laboratory, Kumamoto University Hospital Faculty of Health Sciences, Kumamoto, Japan. Department of Clinical Laboratory Sciences, Graduate School of Medicine, Yamaguchi University, Yamagushi, Japan. Department of Laboratory and Vascular Medicine, Graduate School of Medical and Dental Sciences, Kagoshima University, Kagoshima, Japan. 4 Department of Laboratory Medicine, Osaka University Graduate School of Medicine, Osaka, Japan. 5 Department of Clinical Chemistry and Laboratory Medicine, Kyushu University Graduate School of Medical Sciences, Fukuoka, Japan. 6 Department of Laboratory Medicine, Hamamatsu University School of Medicine, Hamamatsu, Japan. 7 Department of Medical Technique, Nagoya University Hospital, Nagoya, Japan. 8 Department of Molecular Diagnosis and Division of Clinical Genetics and Proteomics, Graduate School of Medicine, Chiba University, Chiba, Japan. 9 Department of Clinical Laboratory, The University of Tokyo Hospital, Tokyo, Japan. 10 Department of Laboratory Medicine, Keiyu Hospital, Yokohama, Japan. 11 Division of Clinical Laboratories, National Cancer Center Hospital, Tokyo, Japan. 2 3
135
136
samples.1–9 These studies inspired us to investigate the stability of human-derived liquid clinical samples following storage at various temperatures. All analytical measurements are method dependent. However, concerns remain about the relationship between the original quality of the sample and the data that results from analysis.10 Samples that are of good quality when they are obtained from patients could be denatured rapidly by inappropriate handling. Data based on poor quality samples affect decisions regarding diagnostic and therapeutic strategies. Furthermore, poor quality data also affect the outcome of research and clinical trials. Patients’ lives, as well as significant financial investments in biomarker and drug development, can be seriously jeopardized by poor quality data derived from poor quality samples. The only way to avoid this potentially disastrous situation is to ensure that samples are of adequate quality. Thus, those conditions leading to sample denaturation must be identified. Based on this knowledge, we can then analyze samples using procedures that will prevent denaturation as much as possible. Therefore, the aim of this multi-center study was to determine appropriate storage conditions for human-derived liquid samples with respect to temperature and duration. Of 35 routine clinical test analytes, we found that some are denatured relatively rapidly, whereas others showed negligible rates of denaturation. By defining how test results might change over time, we can standardize sample handling procedures in order to minimize the denaturation of clinically important analytes.
Materials and Methods Clinical laboratories associated with eight institutions (Kumamoto University, Kagoshima University, Osaka University, Kyushu University, Hamamatsu University, Nagoya University, Chiba University, and The University of Tokyo) collaborated with our laboratory (National Cancer Center Hospital) in this study. Specimens of serum, whole blood, and urine remaining after routine clinical testing were evaluated. Hb (hemoglobin), PLT (platelet count), RBC (red blood cell count), and WBC (white blood cell count) were determined using whole blood. Urinary sugar and urinary protein data were determined using urine samples. All other data were determined using serum. No plasma samples were examined. At least five samples per analyte were analyzed at each laboratory. However, for ease of statistical evaluation (see below), specimens were carefully chosen for each test based on the following criteria: (1) test value was within or close to the reference interval (RI), not exceeding one-half of the RI below or above the RI limits; and (2) the value was within the measurement range providing optimal analytical precision. Samples selected based on these criteria were subjected to storage at various temperatures for varying durations, and analyte stability was evaluated using 35 routine tests. Patient identity information was removed from all samples used in this study. The study was approved by the Institutional Review Board (IRB) of each facility. Because each laboratory needed to acquire individual IRB approval for this study, 6 months elapsed between the time of the first laboratory’s approval and that of the last. The samples utilized for this study were collected and processed in-hospital in a strict temperaturecontrolled environment by each of the participating laboratories. Each analytical device and reagent used in this study at each facility was recorded. Uncertainty was confirmed and
IKEDA ET AL.
shared among participating facilities. ‘‘Uncertainty’’ was defined as a parameter associated with the result of a measurement that characterizes the range of values within which the value of the quantity being measured is expected to lie.11 The types of sample collection and storage tubes used at each facility were also recorded. To harmonize the experimental conditions, sources of variation related to sample collection and measurement were listed and shared among the collaborating facilities. Immediately after completion of routine assays, specimens with sufficient remaining volume were divided into 300-mL aliquots. All of the participating laboratories processed the samples within 4 hours of collection; samples were stored at room temperature until processing, in all cases for less than 4 h. Aliquots of serum were placed in freezing containers and stored at room temperature - RT (23C), 4C, - 20C, or - 80C, without light exposure. Whole blood and urine samples were stored only at 23C and 4C. Typically, urine samples are not stored below 0C. In addition to routine initial tests (day 0 tests), other analytical tests were performed after various time periods, as specified below. Specifications regarding blood-collection tubes, urine-collection cups, sample storage devices, refrigerators, and freezers were recorded. Temperatures during storage were recorded. Samples were stored under the various temperature conditions for 1 day (24 h), 3 days (72 h), or 7 days (168 h). After storage, frozen samples were thawed under running water for 10 min. The thawed samples were then kept at room temperature for 10 min. All sample containers were gently inverted 10 times for mixing just before testing. Individual analyses were performed within 1 h of the scheduled time including the time required for thawing and the time samples were kept at room temperature. The temperature of running water used to thaw frozen samples was recorded in the experimental notes. The stability of samples at each time point (days 1, 3, and 7) was expressed as the mean ( – SD) difference in test results from those acquired on day 0 for each of the four temperature conditions. The statistical significance of any observed changes in test results were evaluated using two-way ANOVA. The F-value (representing variance due to storage time divided by the residual variance) and corresponding P value were computed. Because the large number of degrees of freedom associated with our dataset for computing the Fvalues led to too much power for detecting changes, we considered p = 0.01 as indicating statistical significance. In order to quantify the practical magnitude of timedependent changes in test results, we first computed the net SD due to storage time (nSDtime) by subtracting the residual variance (Vres = SD2res) from the crude variance due to the storage time (Vtime = SD2time) as follows: rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Vtime Vres nSDtime ns 1 where ns denotes the number of specimens used for the analysis. The SD ratio (SDR) was calculated as the ratio of the nSDtime to the SD of the RI of the test item for computation, SDRI, which is approximately one-fourth of the RI: SDRtime ¼
nSDtime SDRI
SPECIMEN STORAGE CONDITIONS FOR LAB TESTS
The SDRtime parameter provides an estimate of the practical significance of time-dependent changes in test results. The theoretical basis for the use of SDRtime is as follows. SDRI is equivalent to the SD representing biological variability, which comprises the between-individual SD (SDG) and within-individual SD (SDI), SDRI ¼
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi SD2G þ SD2I
In evaluating the magnitude of analytical bias in the field of laboratory medicine, it is customary to regard a bias < 0.125 · SDRI as optimal, a bias < 0.25 · SDRI as allowable, and a bias < 0.375 · SDRI as the minimum allowable. An SDRtime < 0.125 is regarded as optimal, indicating an essentially negligible level of storage time-dependent changes, whereas 0.125 £ SDRtime < 0.250 is regarded as a noticeable level of variation affecting interpretation of test results, and 0.250 £ SDRtime < 0.375 or 0.375 £ SDRtime are regarded as indicating a moderate or prominent effect on test results. In interpreting two-way ANOVA F-values, we regarded p < 0.01, p < 0.0001, and p < 0.000001 as indicative of slight, moderate, and prominent time-dependent changes, respectively.
Results Figure 1A shows the statistical results for each analyte stored under various conditions. Analytes showing ‘‘prominent’’ changes following storage at RT include Tbil (total bilirubin), AST (aspartate aminotransferase), ALT (alanine aminotransferase), LD (lactate dehydrogenase), CK (creatine kinase), TG (triglyceride), Elastase I, Hb (hemoglobin), PLT (platelet count), and WBC (white blood cell count). Analytes showing prominent changes at 4C include ALT, LD, HDL-C (high-density-lipoprotein cholesterol), Hb, PLT, and WBC. Analytes showing prominent changes at -20C include Tbil and ALP (alkaline phosphatase). No analytes showed prominent changes at -80C. Analytes showing ‘‘moderate’’ changes following storage at RT include Glu (glucose), insulin, urinary sugar, and RBC (red blood cell count). Analytes showing moderate changes at 4C include Tbil, K (potassium), Cl (chloride), TG, Tcho (total cholesterol), insulin, and RBCs. Analytes showing moderate changes at - 20C include BUN (blood urea nitrogen), Cl, AST, ALT, and CK. Only BUN showed moderate changes following storage at - 80C. Analytes showing ‘‘slight’’ changes following storage at RT include BUN, IgG, amylase, HDL-C, FT4 (free thyroxine), and urinary protein. Analytes showing slight changes at 4C include BUN, Na (sodium), AST, amylase, ALP, and FT4. Analytes showing slight changes at - 20C include CRE (creatinine), Na, K, LD, amylase, HDL-C, insulin, FT4, and urinary sugar. Analytes showing slight changes at - 80C include Tbil, IgG (immunoglobulin G), TG, Elastase I, FT4, and urinary sugar. At all temperatures, Ca (calcium), Alb (albumin), Ferr (ferritin), LDL-C (low-density-lipoprotein cholesterol), TPLA (Treponema pallidum latex agglutination), AFP (alpha-fetoprotein), CEA (carcinoembryonic antigen), and PSA (prostate-specific antigen) showed no statistically significant changes as determined by ANOVA. Test results
137
were stable at - 80C for all analytes except BUN, Tbil, IgG, TG, Elastase I, FT4, and urinary sugar. The SDR, which is indicative of the practical magnitude of time-dependent changes, agreed with the results of twoway ANOVA (Fig. 1B). The analytes showing ‘‘minimal’’ changes following storage at RT based on SDR were AST and insulin. Analytes showing minimal changes at 4C were LD and insulin. Only ALT showed minimal changes at -20C, and no analytes showed minimal changes at - 80C. Analytes showing ‘‘allowable’’ changes following storage at RT include Tbil, ALT, PLT, and WBC. Analytes showing allowable changes at 4C include FT4 and WBC. Analytes showing allowable changes at - 20C include LD and insulin. No analytes showed allowable changes at - 80C. Analytes showing ‘‘optimal’’ changes following storage at RT include PSA, Elastase I, FT4, and Hb. Analytes showing optimal changes at 4C include Na, Cl, ALT, and PLT. Analytes showing optimal changes at - 20C include Na, Cl, AST, and FT4. No analytes showed optimal changes at - 80C. With storage at - 80C, no remarkable denaturation was observed. At all temperatures, Glu, BUN, CRE, K, Ca, Alb, IgG, Ferr, amylase, CK, ALP, TG, Tcho, HDL-C, LDL-C, AFP, CEA, and RBC show no significant changes with respect to SDRtime values. The stability of samples under different storage conditions was analyzed by averaging the differences between pre- and post-storage test results, as shown in Figures 1C1, 1C2, and 1C3.
Discussion The objective of this multi-center study was to improve quality control for clinical samples by elucidating the storage conditions that lead to instability of analytes. The multiinstitution nature of the study made it possible to gather results from different analyzers and reagents, and thus obtain an unbiased picture of specimen instability. In this study we evaluated the stability of commonly tested analytes in liquid samples stored under different conditions. Changes in stability were assessed by analyzing samples stored under the various conditions and determining the average differences in test values pre- and post-storage (Fig. 1C1, 1C2, and 1C3). The statistical significance of the differences was assessed using two-way ANOVA (Fig. 1A), and practical significance was estimated by determining the ratio of SD due to time in storage (SDtime) to the SD of the RI (SDRRI) (Fig. 1B). We did not assess whether there were significant differences among participating institutions with respect to individual tests per se. Instead, each institution was asked to submit documents regarding the performance, accuracy, and uncertainty of their assays. Additionally, it is of note that all the participating institutions are ISO15189–certified for the quality of their analytical services. Besides ISO15189 certification, two other factors are relevant. One is participation in the external quality assurance program and the other is the movement toward standardization by utilizing certified reference materials. With respect to the former, all of the laboratories involved in this study underwent monitoring by several domestic external quality assurance programs. All of the laboratories also worked to standardize their routine practices. Although protocols to minimize inter- and intra-
FIG. 1. (A) F-values determined using two-way ANOVA for each analyte in samples stored under various conditions. The F-value, representing time-dependent variance in analyte values divided by the residual variance, is listed for each storage temperature: RT (room temperature: 23C), 4C, - 20C, and - 80C. Statistical significance ( p-value) is indicated by the color of the box. (B) SD ratio (SDR) of time-dependent changes in test results for each analyte in samples stored under various conditions. The SDR for time-dependent changes is listed for each storage condition. The value represents the ratio of the SD due to time-dependent changes to the SD of the reference interval (RI) of the test item, which is approximately one-fourth of the RI. The SDR is indicative of the practical significance of time-dependent changes in comparison with between-individual variability. (C1, 2, 3) Stability of samples stored under various conditions, expressed as average ( – SD) differences between analyte values on days 1, 3, and 7 and values determined before storage. TPLA (infection) is omitted because of insufficient data. A color version of this figure is available in the online article at www.liebertpub.com/bio 138
SPECIMEN STORAGE CONDITIONS FOR LAB TESTS
FIG. 1.
laboratory variation in clinical testing in Japan are not as yet uniform in a strict sense, because of the efforts described above, progress is being made in this regard. Thus, practical inter- and intra-laboratory variations are small enough to ignore among leading hospitals in Japan, such as those participating in this study. Similarly, although each laboratory has its own individualized reference range for routine practice, essentially all laboratories participate in standardization efforts, and some, but not all, of the reference ranges are now uniform. For this study, 35 tests for which the reference ranges were made uniform were selected. Figure 1C1, 1C2, and 1C3 reflect the rough trends observed in this study. Based on our evaluations, the following conclusions were made:
139
(Continued).
1. Amino acids are mostly stable; however, Tbil stored at RT is not. 2. Electrolytes are mostly stable, even when stored at RT. Na and Cl show some decrease when stored at 4C and - 20C. 3. Proteins are mostly stable, although ferritin shows some decrease after 3 days of storage in various temperatures except - 80C. 4. Enzymes such as AST and ALT are readily denatured if stored at RT. ALT and LD show marked inactivation when stored at 4C and - 20C. Of note, ALT shows marked inactivation when stored at any temperature above - 80C. Other enzymes, such as amylase, CK, and ALP are mostly stable when stored at all temperatures.
140
IKEDA ET AL.
FIG. 1.
5. With storage at RT and 4C, lipids show no remarkable alteration. 6. Tumor markers, except for PSA and Elastase I, are relatively stable even for 7 days. 7. The hormones insulin and FT4 are unstable at all temperatures above - 80C. Insulin shows a marked decrease when samples are stored at RT. 8. Urinary sugar is unstable at RT but relatively stable at lower temperatures. Urinary protein is relatively stable at all temperatures. 9. CBCs, especially PLT and WBC, are readily lysed when samples are stored at RT and 4C. RBC and Hb are mostly stable.
(Continued).
This study shows that sample degradation is clearly temperature dependent. This does not mean that storage duration is not critical in sample degradation. Most of the samples begin to degrade as the storage duration increases at all the temperatures examined in this study. Denaturation is more pronounced at RT, followed by 4C, - 20C, and -80C. This is consistent with accumulated, general, experience-based empirical knowledge regarding specimen stability in clinical laboratories.12,13 In the previous study, ALT and AST showed denaturation following storage at - 30C. 13 It is interesting that even in our study, large differences in stability were observed for samples stored frozen between - 20C and - 80C.
SPECIMEN STORAGE CONDITIONS FOR LAB TESTS
FIG. 1.
Although we do not know the precise mechanism underlying this phenomenon, it is possible that it could involve interaction between heat, water, and protein. Water usually freezes at 0C; however, in the presence of NaCl, the freezing point of water is lower, around - 21C, what cryobiologists call the eutectic temperature.14,15 This means that at - 20C, water is not completely frozen if it contains NaCl. Thus, samples stored at - 20C might contain liquid water, although the amount would be very small. The presence of a small amount of water could be sufficient to promote hydrolysis, even at - 20C. In contrast, samples stored at - 80C would have almost no liquid water, therefore no hydrolysis would occur. In addition, water that
141
(Continued).
has changed into ice forms crystals that enlarge and shrink with changes in temperature. Ice crystals can affect the three-dimensional structure of proteins, such as transforming an alpha helix to a beta sheet. Such changes could have caused the denaturation observed with the samples in our study.16 Kang et al. examined the effect of various pre-analytical stresses on 10 biochemical analytes (ALT, AST, Tchol, GGT, TG, LD, CRP, Cre, Glu, and BUN) in human blood, although they did not investigate the effects of temperature and storage time. 17 They reported that the serum levels of GGT (gamma-glutamyl transferase) and LD were significantly changed, depending on both the time interval
142
between blood collection and fractionation, and the time interval between fractionation and freezing of serum and plasma samples. Thus, they concluded that GGT and LD could be index markers for various quality control studies. Unfortunately, we did not analyze GGT in the present study; however, our data regarding LD support their conclusions. To date, many reports examining the effects of stress on cells have been published;18 however, not many reports of a study examining liquid samples have been published.19 We understand that the present study has a number of limitations; however, we believe our results will stimulate further research in this field and that many interesting reports will follow.
Conclusions In this study, prompted by shared concerns regarding sample quality, 10 medical facilities in Japan conducted analyses of the stability of clinical test samples in relation to storage temperature and duration. The results of 35 routine clinical tests using samples stored under various conditions show that some analytes denature relatively rapidly, whereas others denature slowly. Serum samples stored at - 80C were stable for up to 7 days. Some serum samples stored at temperatures other than - 80C showed a decrease or increase in test values. This is an important study for clinical laboratories that conduct routine clinical tests, confirming that the accumulated experience-based knowledge regarding sample stability is correct. Further studies of other quality-related and transportation-related parameters will provide for more reliable laboratory testing, ensure the quality of clinical trials, and promote translational research, such as biomarker exploration. These studies will also provide information concerning index molecules in human-derived liquid samples as related to various aspects of the life sciences.
Additional note This study was an official project of the Japanese Society of Laboratory Medicine ( JSLM). The study began in May 2012. Because of the utilization of patient samples, each facility was required to secure approval of its IRB; this took approximately 8 months. The general scheme of this project was previously reported at the JSLM conferences in November 2012 and 2013.
Acknowledgments We thank the following medical technologists for their contribution to this study: Kagoshima University: Kanakko Sato, Tomohisa Takenoshita, Chikako Ishi, Masakaze Matsushita, Osaka University: Masahiko Matsui, Osamu Kabutomori, Matsuo Deguchi, Keiko Takeoka, Masatomo Kamata, Nobuaki Hata, Wataru Kobayashi, National Cancer Center Hospital: Takao Miura, Naoki Maesawa, Kouji Ono, Hiroshi Yamakawa, Akashi Koseki, Satoe Miyaki, Chiba University: Yuji Sawabe, Toshihiko Yoshida, Hamamatsu University: Sachinori Uchiyama, Etsuko Hamada, Kunihiro Iwahara, Yasuyuki Suga, Yukiko Itoh, Aya Shimoda, Ayano Fujiwara. In addition to these medical technologists, those at Kumamoto University, Kyusyu
IKEDA ET AL.
University, Nagoya University, and Tokyo University also contributed to this study. This study was supported in part by the National Cancer Center Research and Development Fund (26-A-1), Japan.
Author Disclosure Statement No conflicting financial interests exist.
References 1. Clark S, Youngman LD, Chukwurah B, et al. Effect of temperature and light on the stability of fat-soluble vitamins in whole blood over several days: Implications for epidemiological studies. Intl J Epidemiol 2004;33: 518–525. 2. Sapre N, Hong MKH, Huang JG, et al. Bladder cancer biorepositories in the ‘‘-Omics’’ era: Integrating quality tissue specimens with comprehensive clinical annotation. Biopreserv Biobank 2013;11:166–172. 3. Kushnir MM. Are samples in your freezer still good for biomarker discovery? Am J Clin Pathol 2013;140: 287–288. 4. Dunn WB, Broadhurst D, Ellis DI, et al. A GC-TOF-MS study of the stability of serum and urine metabolomes during the UK Biobank sample collection and preparation protocols. Intl J Epidemiol 2008;37:i23–i30. 5. Gaye A, Peakman T, Tobin MD, Burton PR. Understanding the impact of pre-analytic variation in haematological and clinical chemistry analytes on the power of association studies. Intl J Epidemiol 2014;43:1633–1644. 6. Breier M, Wahl S, Prehn C, et al. Targeted metabolomics identifies reliable and stable metabolites in human serum and plasma samples. PLoS ONE 2014;9:e89728. 7. Hubel A, Spindler R, Skubitz APN. Storage of human biospecimens: Selection of the optimal storage temperature. Biopreserv Biobank 2014;12:165–175. 8. Specimen stability in storage. BioCor:http://biocor.umn .edu/assets/files/StorageTemp.pdf. Last accessed March 25, 2015. 9. Scott KL, Lecak J, Acker JP. Biopreservation of red blood cells: Past, present, and future. Transfus Med Rev 2005; 19:127–142. 10. Lim MD, Dickherber A, Compton CC. Before you analyze a human specimen, think quality, variability, and bias. Anal Chem 2010;83:8–13. 11. Burtis CA, Ashwood ER, Bruns DE. Fundamentals of Clinical Chemistry: Sixth edition, Saunders Elsevier; St. Lous, Missouri, 2008; p. 202. 12. Unhale SA, Skubitz APN, Solomon R, Hubel A. Stabilization of tissue specimens for pathological examination and biomedical research. Biopreserv Biobank 2012;10:493–500. 13. Ichihara K, Kouguchi K. Practical Manual for Evidencebased Laboratory Diagnosis. [Japanese]. Kyoto: Japan Education Research Center; 2011. 14. Rasmussen DH, MacKenzie AP. Phase diagram for the system water-dimethylsulphoxide. Nature 1968;220:1315– 1317. 15. Cocks FH, Brower WE. Phase diagram relationships in cryobiology. Cryobiology 1974;11:340–358. 16. Yamauchi A, Aoki T. Study on the stock of serum samples for ALT activity determination. Jap J Med Technol 1995;44:907–910. 17. Kang HJ, Jeon SY, Park J-S, et al. Identification of clinical biomarkers for pre-analytical quality control of blood samples. Biopreserv Biobank 2013;11:94–100.
SPECIMEN STORAGE CONDITIONS FOR LAB TESTS
18. Ragoonanan V, Less R, Aksan A. Response of the cell membrane–cytoskeleton complex to osmotic and freeze/ thaw stresses. Part 2: The link between the state of the membrane–cytoskeleton complex and the cellular damage. Cryobiology 2013;66:96–104. 19. Furuta K, Less R, Aksan A, Hubel A. Response of albumin solutions to stresses of freezing and thawing and processing (abstract). Biopreserv Biobank 2011;9:99. 20. [Internet]: Available from: http://en.wikipedia.org/wiki/ Cold_chain. Accessed on October 6, 2014.
143
Address correspondence to: Koh Furuta, MD, PhD Division of Clinical Laboratories National Cancer Center Hospital 5-1-1, Tsukiji, Chuo-ku Tokyo 1040045 Japan E-mail: [email protected]
Evaluation of the Short-Term Stability of Specimens for Clinical Laboratory Testing Katsuyoshi Ikeda,1 Kiyoshi Ichihara,2 Teruto Hashiguchi,3 Yoh Hidaka,4 Dongchon Kang,5 Masato Maekawa,6 Hiroyuki Matsumoto,7 Kazuyuki Matsushita,8 Shigeo Okubo,9 Tatsuyuki Tsuchiya,10 Koh Furuta,11 on behalf of The Committee for Standardization, The Japanese Society of Laboratory Medicine ( JSLM)
Background: A major concern in both the laboratory-medicine and research communities is the quality of human specimens for analysis. However, there is insufficient scientific evidence regarding optimal conditions for handling and storing routine specimens, especially those in liquid form. Thus, we investigated the stability of clinically relevant samples stored under various conditions. Materials and Methods: Ten clinical laboratories in Japan conducted analyses of the stability of post-clinical (left over after analysis) test samples in relation to temperature and storage duration. We examined serum, whole blood, and urine samples submitted to each laboratory for routine testing. In this study, at least 5 samples for each of 35 tests were analyzed at each laboratory. After completion of routine testing, specimens with sufficient residual volume and values between LL - R/2 (lower limit of reference interval) and UL + R/2 (upper limit) were divided into 300 mL aliquots, where R = UL - LL. Aliquots of serum specimens were stored at either room temperature (23C), 4C, - 20C, or - 80C without light exposure. Aliquots of whole blood and urine specimens were stored at either 23C or 4C. The storage time was either 1, 3, or 7 days. Average differences between pre- and post-storage test results were evaluated for each laboratory test by two-way ANOVA. Fvalues for between-day variations were used for judging the statistical significance of storage-related changes in test values, whereas the ratio of between-day SD to between-individual SD (one-fourth of reference interval) was used to indicate the practical significance of the change. Results and Conclusion: Sample denaturation is clearly temperature- and storage-duration dependent for almost all analytes. In general, specimens were most susceptible to denaturation at 23C, then 4C, - 20C, and - 80C. This study confirmed the accumulated routine, practice-based, detailed knowledge regarding specimen stability and will help to ensure the reliability of laboratory test results.
Introduction
A
major concern for both the laboratory-medicine and research communities is the quality of specimens, especially those obtained from human sources. Thanks to continuous efforts to improve quality control in clinical laboratories, the precision and accuracy of laboratory test
results have improved greatly in recent years in Japan. Although experience and anecdotal evidence help to maintain high analytical quality in clinical laboratories, there is insufficient scientific evidence regarding the optimal conditions for handling and storing routine samples, especially those in liquid form. Several investigations have focused on denaturation of human specimens, particularly liquid
1
Department of Clinical Laboratory, Kumamoto University Hospital Faculty of Health Sciences, Kumamoto, Japan. Department of Clinical Laboratory Sciences, Graduate School of Medicine, Yamaguchi University, Yamagushi, Japan. Department of Laboratory and Vascular Medicine, Graduate School of Medical and Dental Sciences, Kagoshima University, Kagoshima, Japan. 4 Department of Laboratory Medicine, Osaka University Graduate School of Medicine, Osaka, Japan. 5 Department of Clinical Chemistry and Laboratory Medicine, Kyushu University Graduate School of Medical Sciences, Fukuoka, Japan. 6 Department of Laboratory Medicine, Hamamatsu University School of Medicine, Hamamatsu, Japan. 7 Department of Medical Technique, Nagoya University Hospital, Nagoya, Japan. 8 Department of Molecular Diagnosis and Division of Clinical Genetics and Proteomics, Graduate School of Medicine, Chiba University, Chiba, Japan. 9 Department of Clinical Laboratory, The University of Tokyo Hospital, Tokyo, Japan. 10 Department of Laboratory Medicine, Keiyu Hospital, Yokohama, Japan. 11 Division of Clinical Laboratories, National Cancer Center Hospital, Tokyo, Japan. 2 3
135
136
samples.1–9 These studies inspired us to investigate the stability of human-derived liquid clinical samples following storage at various temperatures. All analytical measurements are method dependent. However, concerns remain about the relationship between the original quality of the sample and the data that results from analysis.10 Samples that are of good quality when they are obtained from patients could be denatured rapidly by inappropriate handling. Data based on poor quality samples affect decisions regarding diagnostic and therapeutic strategies. Furthermore, poor quality data also affect the outcome of research and clinical trials. Patients’ lives, as well as significant financial investments in biomarker and drug development, can be seriously jeopardized by poor quality data derived from poor quality samples. The only way to avoid this potentially disastrous situation is to ensure that samples are of adequate quality. Thus, those conditions leading to sample denaturation must be identified. Based on this knowledge, we can then analyze samples using procedures that will prevent denaturation as much as possible. Therefore, the aim of this multi-center study was to determine appropriate storage conditions for human-derived liquid samples with respect to temperature and duration. Of 35 routine clinical test analytes, we found that some are denatured relatively rapidly, whereas others showed negligible rates of denaturation. By defining how test results might change over time, we can standardize sample handling procedures in order to minimize the denaturation of clinically important analytes.
Materials and Methods Clinical laboratories associated with eight institutions (Kumamoto University, Kagoshima University, Osaka University, Kyushu University, Hamamatsu University, Nagoya University, Chiba University, and The University of Tokyo) collaborated with our laboratory (National Cancer Center Hospital) in this study. Specimens of serum, whole blood, and urine remaining after routine clinical testing were evaluated. Hb (hemoglobin), PLT (platelet count), RBC (red blood cell count), and WBC (white blood cell count) were determined using whole blood. Urinary sugar and urinary protein data were determined using urine samples. All other data were determined using serum. No plasma samples were examined. At least five samples per analyte were analyzed at each laboratory. However, for ease of statistical evaluation (see below), specimens were carefully chosen for each test based on the following criteria: (1) test value was within or close to the reference interval (RI), not exceeding one-half of the RI below or above the RI limits; and (2) the value was within the measurement range providing optimal analytical precision. Samples selected based on these criteria were subjected to storage at various temperatures for varying durations, and analyte stability was evaluated using 35 routine tests. Patient identity information was removed from all samples used in this study. The study was approved by the Institutional Review Board (IRB) of each facility. Because each laboratory needed to acquire individual IRB approval for this study, 6 months elapsed between the time of the first laboratory’s approval and that of the last. The samples utilized for this study were collected and processed in-hospital in a strict temperaturecontrolled environment by each of the participating laboratories. Each analytical device and reagent used in this study at each facility was recorded. Uncertainty was confirmed and
IKEDA ET AL.
shared among participating facilities. ‘‘Uncertainty’’ was defined as a parameter associated with the result of a measurement that characterizes the range of values within which the value of the quantity being measured is expected to lie.11 The types of sample collection and storage tubes used at each facility were also recorded. To harmonize the experimental conditions, sources of variation related to sample collection and measurement were listed and shared among the collaborating facilities. Immediately after completion of routine assays, specimens with sufficient remaining volume were divided into 300-mL aliquots. All of the participating laboratories processed the samples within 4 hours of collection; samples were stored at room temperature until processing, in all cases for less than 4 h. Aliquots of serum were placed in freezing containers and stored at room temperature - RT (23C), 4C, - 20C, or - 80C, without light exposure. Whole blood and urine samples were stored only at 23C and 4C. Typically, urine samples are not stored below 0C. In addition to routine initial tests (day 0 tests), other analytical tests were performed after various time periods, as specified below. Specifications regarding blood-collection tubes, urine-collection cups, sample storage devices, refrigerators, and freezers were recorded. Temperatures during storage were recorded. Samples were stored under the various temperature conditions for 1 day (24 h), 3 days (72 h), or 7 days (168 h). After storage, frozen samples were thawed under running water for 10 min. The thawed samples were then kept at room temperature for 10 min. All sample containers were gently inverted 10 times for mixing just before testing. Individual analyses were performed within 1 h of the scheduled time including the time required for thawing and the time samples were kept at room temperature. The temperature of running water used to thaw frozen samples was recorded in the experimental notes. The stability of samples at each time point (days 1, 3, and 7) was expressed as the mean ( – SD) difference in test results from those acquired on day 0 for each of the four temperature conditions. The statistical significance of any observed changes in test results were evaluated using two-way ANOVA. The F-value (representing variance due to storage time divided by the residual variance) and corresponding P value were computed. Because the large number of degrees of freedom associated with our dataset for computing the Fvalues led to too much power for detecting changes, we considered p = 0.01 as indicating statistical significance. In order to quantify the practical magnitude of timedependent changes in test results, we first computed the net SD due to storage time (nSDtime) by subtracting the residual variance (Vres = SD2res) from the crude variance due to the storage time (Vtime = SD2time) as follows: rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Vtime Vres nSDtime ns 1 where ns denotes the number of specimens used for the analysis. The SD ratio (SDR) was calculated as the ratio of the nSDtime to the SD of the RI of the test item for computation, SDRI, which is approximately one-fourth of the RI: SDRtime ¼
nSDtime SDRI
SPECIMEN STORAGE CONDITIONS FOR LAB TESTS
The SDRtime parameter provides an estimate of the practical significance of time-dependent changes in test results. The theoretical basis for the use of SDRtime is as follows. SDRI is equivalent to the SD representing biological variability, which comprises the between-individual SD (SDG) and within-individual SD (SDI), SDRI ¼
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi SD2G þ SD2I
In evaluating the magnitude of analytical bias in the field of laboratory medicine, it is customary to regard a bias < 0.125 · SDRI as optimal, a bias < 0.25 · SDRI as allowable, and a bias < 0.375 · SDRI as the minimum allowable. An SDRtime < 0.125 is regarded as optimal, indicating an essentially negligible level of storage time-dependent changes, whereas 0.125 £ SDRtime < 0.250 is regarded as a noticeable level of variation affecting interpretation of test results, and 0.250 £ SDRtime < 0.375 or 0.375 £ SDRtime are regarded as indicating a moderate or prominent effect on test results. In interpreting two-way ANOVA F-values, we regarded p < 0.01, p < 0.0001, and p < 0.000001 as indicative of slight, moderate, and prominent time-dependent changes, respectively.
Results Figure 1A shows the statistical results for each analyte stored under various conditions. Analytes showing ‘‘prominent’’ changes following storage at RT include Tbil (total bilirubin), AST (aspartate aminotransferase), ALT (alanine aminotransferase), LD (lactate dehydrogenase), CK (creatine kinase), TG (triglyceride), Elastase I, Hb (hemoglobin), PLT (platelet count), and WBC (white blood cell count). Analytes showing prominent changes at 4C include ALT, LD, HDL-C (high-density-lipoprotein cholesterol), Hb, PLT, and WBC. Analytes showing prominent changes at -20C include Tbil and ALP (alkaline phosphatase). No analytes showed prominent changes at -80C. Analytes showing ‘‘moderate’’ changes following storage at RT include Glu (glucose), insulin, urinary sugar, and RBC (red blood cell count). Analytes showing moderate changes at 4C include Tbil, K (potassium), Cl (chloride), TG, Tcho (total cholesterol), insulin, and RBCs. Analytes showing moderate changes at - 20C include BUN (blood urea nitrogen), Cl, AST, ALT, and CK. Only BUN showed moderate changes following storage at - 80C. Analytes showing ‘‘slight’’ changes following storage at RT include BUN, IgG, amylase, HDL-C, FT4 (free thyroxine), and urinary protein. Analytes showing slight changes at 4C include BUN, Na (sodium), AST, amylase, ALP, and FT4. Analytes showing slight changes at - 20C include CRE (creatinine), Na, K, LD, amylase, HDL-C, insulin, FT4, and urinary sugar. Analytes showing slight changes at - 80C include Tbil, IgG (immunoglobulin G), TG, Elastase I, FT4, and urinary sugar. At all temperatures, Ca (calcium), Alb (albumin), Ferr (ferritin), LDL-C (low-density-lipoprotein cholesterol), TPLA (Treponema pallidum latex agglutination), AFP (alpha-fetoprotein), CEA (carcinoembryonic antigen), and PSA (prostate-specific antigen) showed no statistically significant changes as determined by ANOVA. Test results
137
were stable at - 80C for all analytes except BUN, Tbil, IgG, TG, Elastase I, FT4, and urinary sugar. The SDR, which is indicative of the practical magnitude of time-dependent changes, agreed with the results of twoway ANOVA (Fig. 1B). The analytes showing ‘‘minimal’’ changes following storage at RT based on SDR were AST and insulin. Analytes showing minimal changes at 4C were LD and insulin. Only ALT showed minimal changes at -20C, and no analytes showed minimal changes at - 80C. Analytes showing ‘‘allowable’’ changes following storage at RT include Tbil, ALT, PLT, and WBC. Analytes showing allowable changes at 4C include FT4 and WBC. Analytes showing allowable changes at - 20C include LD and insulin. No analytes showed allowable changes at - 80C. Analytes showing ‘‘optimal’’ changes following storage at RT include PSA, Elastase I, FT4, and Hb. Analytes showing optimal changes at 4C include Na, Cl, ALT, and PLT. Analytes showing optimal changes at - 20C include Na, Cl, AST, and FT4. No analytes showed optimal changes at - 80C. With storage at - 80C, no remarkable denaturation was observed. At all temperatures, Glu, BUN, CRE, K, Ca, Alb, IgG, Ferr, amylase, CK, ALP, TG, Tcho, HDL-C, LDL-C, AFP, CEA, and RBC show no significant changes with respect to SDRtime values. The stability of samples under different storage conditions was analyzed by averaging the differences between pre- and post-storage test results, as shown in Figures 1C1, 1C2, and 1C3.
Discussion The objective of this multi-center study was to improve quality control for clinical samples by elucidating the storage conditions that lead to instability of analytes. The multiinstitution nature of the study made it possible to gather results from different analyzers and reagents, and thus obtain an unbiased picture of specimen instability. In this study we evaluated the stability of commonly tested analytes in liquid samples stored under different conditions. Changes in stability were assessed by analyzing samples stored under the various conditions and determining the average differences in test values pre- and post-storage (Fig. 1C1, 1C2, and 1C3). The statistical significance of the differences was assessed using two-way ANOVA (Fig. 1A), and practical significance was estimated by determining the ratio of SD due to time in storage (SDtime) to the SD of the RI (SDRRI) (Fig. 1B). We did not assess whether there were significant differences among participating institutions with respect to individual tests per se. Instead, each institution was asked to submit documents regarding the performance, accuracy, and uncertainty of their assays. Additionally, it is of note that all the participating institutions are ISO15189–certified for the quality of their analytical services. Besides ISO15189 certification, two other factors are relevant. One is participation in the external quality assurance program and the other is the movement toward standardization by utilizing certified reference materials. With respect to the former, all of the laboratories involved in this study underwent monitoring by several domestic external quality assurance programs. All of the laboratories also worked to standardize their routine practices. Although protocols to minimize inter- and intra-
FIG. 1. (A) F-values determined using two-way ANOVA for each analyte in samples stored under various conditions. The F-value, representing time-dependent variance in analyte values divided by the residual variance, is listed for each storage temperature: RT (room temperature: 23C), 4C, - 20C, and - 80C. Statistical significance ( p-value) is indicated by the color of the box. (B) SD ratio (SDR) of time-dependent changes in test results for each analyte in samples stored under various conditions. The SDR for time-dependent changes is listed for each storage condition. The value represents the ratio of the SD due to time-dependent changes to the SD of the reference interval (RI) of the test item, which is approximately one-fourth of the RI. The SDR is indicative of the practical significance of time-dependent changes in comparison with between-individual variability. (C1, 2, 3) Stability of samples stored under various conditions, expressed as average ( – SD) differences between analyte values on days 1, 3, and 7 and values determined before storage. TPLA (infection) is omitted because of insufficient data. A color version of this figure is available in the online article at www.liebertpub.com/bio 138
SPECIMEN STORAGE CONDITIONS FOR LAB TESTS
FIG. 1.
laboratory variation in clinical testing in Japan are not as yet uniform in a strict sense, because of the efforts described above, progress is being made in this regard. Thus, practical inter- and intra-laboratory variations are small enough to ignore among leading hospitals in Japan, such as those participating in this study. Similarly, although each laboratory has its own individualized reference range for routine practice, essentially all laboratories participate in standardization efforts, and some, but not all, of the reference ranges are now uniform. For this study, 35 tests for which the reference ranges were made uniform were selected. Figure 1C1, 1C2, and 1C3 reflect the rough trends observed in this study. Based on our evaluations, the following conclusions were made:
139
(Continued).
1. Amino acids are mostly stable; however, Tbil stored at RT is not. 2. Electrolytes are mostly stable, even when stored at RT. Na and Cl show some decrease when stored at 4C and - 20C. 3. Proteins are mostly stable, although ferritin shows some decrease after 3 days of storage in various temperatures except - 80C. 4. Enzymes such as AST and ALT are readily denatured if stored at RT. ALT and LD show marked inactivation when stored at 4C and - 20C. Of note, ALT shows marked inactivation when stored at any temperature above - 80C. Other enzymes, such as amylase, CK, and ALP are mostly stable when stored at all temperatures.
140
IKEDA ET AL.
FIG. 1.
5. With storage at RT and 4C, lipids show no remarkable alteration. 6. Tumor markers, except for PSA and Elastase I, are relatively stable even for 7 days. 7. The hormones insulin and FT4 are unstable at all temperatures above - 80C. Insulin shows a marked decrease when samples are stored at RT. 8. Urinary sugar is unstable at RT but relatively stable at lower temperatures. Urinary protein is relatively stable at all temperatures. 9. CBCs, especially PLT and WBC, are readily lysed when samples are stored at RT and 4C. RBC and Hb are mostly stable.
(Continued).
This study shows that sample degradation is clearly temperature dependent. This does not mean that storage duration is not critical in sample degradation. Most of the samples begin to degrade as the storage duration increases at all the temperatures examined in this study. Denaturation is more pronounced at RT, followed by 4C, - 20C, and -80C. This is consistent with accumulated, general, experience-based empirical knowledge regarding specimen stability in clinical laboratories.12,13 In the previous study, ALT and AST showed denaturation following storage at - 30C. 13 It is interesting that even in our study, large differences in stability were observed for samples stored frozen between - 20C and - 80C.
SPECIMEN STORAGE CONDITIONS FOR LAB TESTS
FIG. 1.
Although we do not know the precise mechanism underlying this phenomenon, it is possible that it could involve interaction between heat, water, and protein. Water usually freezes at 0C; however, in the presence of NaCl, the freezing point of water is lower, around - 21C, what cryobiologists call the eutectic temperature.14,15 This means that at - 20C, water is not completely frozen if it contains NaCl. Thus, samples stored at - 20C might contain liquid water, although the amount would be very small. The presence of a small amount of water could be sufficient to promote hydrolysis, even at - 20C. In contrast, samples stored at - 80C would have almost no liquid water, therefore no hydrolysis would occur. In addition, water that
141
(Continued).
has changed into ice forms crystals that enlarge and shrink with changes in temperature. Ice crystals can affect the three-dimensional structure of proteins, such as transforming an alpha helix to a beta sheet. Such changes could have caused the denaturation observed with the samples in our study.16 Kang et al. examined the effect of various pre-analytical stresses on 10 biochemical analytes (ALT, AST, Tchol, GGT, TG, LD, CRP, Cre, Glu, and BUN) in human blood, although they did not investigate the effects of temperature and storage time. 17 They reported that the serum levels of GGT (gamma-glutamyl transferase) and LD were significantly changed, depending on both the time interval
142
between blood collection and fractionation, and the time interval between fractionation and freezing of serum and plasma samples. Thus, they concluded that GGT and LD could be index markers for various quality control studies. Unfortunately, we did not analyze GGT in the present study; however, our data regarding LD support their conclusions. To date, many reports examining the effects of stress on cells have been published;18 however, not many reports of a study examining liquid samples have been published.19 We understand that the present study has a number of limitations; however, we believe our results will stimulate further research in this field and that many interesting reports will follow.
Conclusions In this study, prompted by shared concerns regarding sample quality, 10 medical facilities in Japan conducted analyses of the stability of clinical test samples in relation to storage temperature and duration. The results of 35 routine clinical tests using samples stored under various conditions show that some analytes denature relatively rapidly, whereas others denature slowly. Serum samples stored at - 80C were stable for up to 7 days. Some serum samples stored at temperatures other than - 80C showed a decrease or increase in test values. This is an important study for clinical laboratories that conduct routine clinical tests, confirming that the accumulated experience-based knowledge regarding sample stability is correct. Further studies of other quality-related and transportation-related parameters will provide for more reliable laboratory testing, ensure the quality of clinical trials, and promote translational research, such as biomarker exploration. These studies will also provide information concerning index molecules in human-derived liquid samples as related to various aspects of the life sciences.
Additional note This study was an official project of the Japanese Society of Laboratory Medicine ( JSLM). The study began in May 2012. Because of the utilization of patient samples, each facility was required to secure approval of its IRB; this took approximately 8 months. The general scheme of this project was previously reported at the JSLM conferences in November 2012 and 2013.
Acknowledgments We thank the following medical technologists for their contribution to this study: Kagoshima University: Kanakko Sato, Tomohisa Takenoshita, Chikako Ishi, Masakaze Matsushita, Osaka University: Masahiko Matsui, Osamu Kabutomori, Matsuo Deguchi, Keiko Takeoka, Masatomo Kamata, Nobuaki Hata, Wataru Kobayashi, National Cancer Center Hospital: Takao Miura, Naoki Maesawa, Kouji Ono, Hiroshi Yamakawa, Akashi Koseki, Satoe Miyaki, Chiba University: Yuji Sawabe, Toshihiko Yoshida, Hamamatsu University: Sachinori Uchiyama, Etsuko Hamada, Kunihiro Iwahara, Yasuyuki Suga, Yukiko Itoh, Aya Shimoda, Ayano Fujiwara. In addition to these medical technologists, those at Kumamoto University, Kyusyu
IKEDA ET AL.
University, Nagoya University, and Tokyo University also contributed to this study. This study was supported in part by the National Cancer Center Research and Development Fund (26-A-1), Japan.
Author Disclosure Statement No conflicting financial interests exist.
References 1. Clark S, Youngman LD, Chukwurah B, et al. Effect of temperature and light on the stability of fat-soluble vitamins in whole blood over several days: Implications for epidemiological studies. Intl J Epidemiol 2004;33: 518–525. 2. Sapre N, Hong MKH, Huang JG, et al. Bladder cancer biorepositories in the ‘‘-Omics’’ era: Integrating quality tissue specimens with comprehensive clinical annotation. Biopreserv Biobank 2013;11:166–172. 3. Kushnir MM. Are samples in your freezer still good for biomarker discovery? Am J Clin Pathol 2013;140: 287–288. 4. Dunn WB, Broadhurst D, Ellis DI, et al. A GC-TOF-MS study of the stability of serum and urine metabolomes during the UK Biobank sample collection and preparation protocols. Intl J Epidemiol 2008;37:i23–i30. 5. Gaye A, Peakman T, Tobin MD, Burton PR. Understanding the impact of pre-analytic variation in haematological and clinical chemistry analytes on the power of association studies. Intl J Epidemiol 2014;43:1633–1644. 6. Breier M, Wahl S, Prehn C, et al. Targeted metabolomics identifies reliable and stable metabolites in human serum and plasma samples. PLoS ONE 2014;9:e89728. 7. Hubel A, Spindler R, Skubitz APN. Storage of human biospecimens: Selection of the optimal storage temperature. Biopreserv Biobank 2014;12:165–175. 8. Specimen stability in storage. BioCor:http://biocor.umn .edu/assets/files/StorageTemp.pdf. Last accessed March 25, 2015. 9. Scott KL, Lecak J, Acker JP. Biopreservation of red blood cells: Past, present, and future. Transfus Med Rev 2005; 19:127–142. 10. Lim MD, Dickherber A, Compton CC. Before you analyze a human specimen, think quality, variability, and bias. Anal Chem 2010;83:8–13. 11. Burtis CA, Ashwood ER, Bruns DE. Fundamentals of Clinical Chemistry: Sixth edition, Saunders Elsevier; St. Lous, Missouri, 2008; p. 202. 12. Unhale SA, Skubitz APN, Solomon R, Hubel A. Stabilization of tissue specimens for pathological examination and biomedical research. Biopreserv Biobank 2012;10:493–500. 13. Ichihara K, Kouguchi K. Practical Manual for Evidencebased Laboratory Diagnosis. [Japanese]. Kyoto: Japan Education Research Center; 2011. 14. Rasmussen DH, MacKenzie AP. Phase diagram for the system water-dimethylsulphoxide. Nature 1968;220:1315– 1317. 15. Cocks FH, Brower WE. Phase diagram relationships in cryobiology. Cryobiology 1974;11:340–358. 16. Yamauchi A, Aoki T. Study on the stock of serum samples for ALT activity determination. Jap J Med Technol 1995;44:907–910. 17. Kang HJ, Jeon SY, Park J-S, et al. Identification of clinical biomarkers for pre-analytical quality control of blood samples. Biopreserv Biobank 2013;11:94–100.
SPECIMEN STORAGE CONDITIONS FOR LAB TESTS
18. Ragoonanan V, Less R, Aksan A. Response of the cell membrane–cytoskeleton complex to osmotic and freeze/ thaw stresses. Part 2: The link between the state of the membrane–cytoskeleton complex and the cellular damage. Cryobiology 2013;66:96–104. 19. Furuta K, Less R, Aksan A, Hubel A. Response of albumin solutions to stresses of freezing and thawing and processing (abstract). Biopreserv Biobank 2011;9:99. 20. [Internet]: Available from: http://en.wikipedia.org/wiki/ Cold_chain. Accessed on October 6, 2014.
143
Address correspondence to: Koh Furuta, MD, PhD Division of Clinical Laboratories National Cancer Center Hospital 5-1-1, Tsukiji, Chuo-ku Tokyo 1040045 Japan E-mail: [email protected]
