The impact of pre‐analytical variations on biochemical analytes stability: A systematic review

Abstract Objective A common problem in clinical laboratories is maintaining the stability of analytes during pre‐analytical processes. The aim of this study was to systematically summarize the results of a set of studies about the biochemical analytes stability. Methods A literature search was performed on the Advanced search field of PubMed using the keywords: “(stability) AND (analytes OR laboratory analytes OR laboratory tests OR biochemical analytes OR biochemical tests OR biochemical laboratory tests).” A total of 56 entries were obtained. After applying the selection criteria, 20 articles were included in the study. Results In the 20 included references, up to 123 different analytes were assessed. The 34 analytes in order of the most frequently studied analytes were evaluated: Alanine aminotransferase, aspartate aminotransferase, potassium, triglyceride, alkaline phosphatase, creatinine, total cholesterol, albumin, lactate dehydrogenase, sodium, calcium, γ‐glutamyltransferase, total bilirubin, urea, creatine kinase, inorganic phosphate, total protein, uric acid, amylase, chloride, high‐density lipoprotein, magnesium, glucose, C‐reactive protein, bicarbonate, ferritin, iron, lipase, transferrin, cobalamin, cortisol, folate, free thyroxine, and thyroid‐stimulating hormone. Stable test results could be varied between 2 hours and 1 week according to the type of samples and/or type of blood collection tubes on a basic classification set as refrigerated or room temperature. Conclusions Biochemical analytes stability could be improved if the best pre‐analytical approaches are used.


| INTRODUC TI ON
Clinical chemistry and laboratory medicine are fundamental parts of the diagnosis of human diseases. It is clear that the minimized interval between sample collection and processing is the best strategy to prevent changes in analytes activities and concentrations. However, some of the recommendations from guidelines 1,2 are difficult to apply in routine practice. Therefore, a common problem in clinical laboratories is maintaining the stability of serum/plasma analytes during pre-analytical sample handling (collection and preparations) and next, during post-analytical sample handling (storage time and temperature).
Due to increased separation process and decreased hemolysis, many clinical laboratories for routine analytes have been using plasma or serum separator tubes. The gel used in these tubes is relatively inert; however, it may affect analyte concentrations or stability. More importantly, commonly ordered analytes (eg, potassium and phosphate) are known to be sensitive to delayed centrifugation and temperature.
Thus, type of blood collection tube, the time interval between sample collection and analysis, and finally the time and temperature at which the samples are stored constitutes important variables that may affect analysis results and may lead to wrong clinical decisions. 3 Stability limits should be defined for each analyte and sample and be used as a cause of rejection before processing the sample. Alongside these challenges, the evolving scenario of consolidation of smaller laboratories into larger ones has become more popular nearly all over the world. One of the leading consequences of the core labs creation is that patients' samples affect the extra-preanalytical factors such as different transportation conditions, prolonged storage at high or low temperature, and improper handling.
On the other hand, research biobanks usually aimed to perform observational epidemiological studies, and/or interventional projects, collect biological samples from a large population and freeze the samples for long-term storage for future analyses. Therefore, to mimic the pre-analytical sample handling takes longer than usual in core laboratories or in research biobanks, we performed a systematic review study and evaluated the effect of storage time from 45 min to 12 months at different storage temperatures on the stability of biochemical analytes. To our knowledge, this study is the first of its type which systematically evaluated the effect of storage conditions on more than 30 biochemical analytes testing.

| Search strategy
A PubMed Advanced Search was performed without any time limit using the following keywords in the Title Field: "(stability) AND (analytes OR laboratory analytes OR laboratory tests OR biochemical analytes OR biochemical tests OR biochemical laboratory tests)." A total of 56 entries were obtained. The acceptable articles were selected by the two authors by checking the titles and abstracts.

| Study selection
The process of study selection was as follows: (1)  After applying the selection criteria, 20 articles 5-24 were included in the study.

| Stability conditions
In order to classify reviews of the papers, a variables checklist including the following items was defined: (1) sample size of the study; (2) sample type of the study; (3) sample container; (4) instruments used in the study; (5) metabolites; (6) baseline value used in the study; (7) tested values investigated in the study; and (8) statistical analysis. These items were precisely reviewed and recorded for 20 included articles (Table 1).

| Data analysis
To quantitatively summarize the reviewed data, we conducted a descriptive statistical analysis. Then, we evaluated the most studied analytes assessed in the least five included papers (Table 2). According to the related articles, we listed the different stability status of each included analyte. Then, we presented the maximum acceptable delays obtained from the relevant articles.

| RE SULTS
In the 20 included references, up to 123 different analytes were assessed. For the accurate evaluations, we focused on analytes  For the 34 biochemistry analytes, the maximum acceptable delays obtained from the relevant articles (Table 2) are as follows. The analytes are presented in order of the analyte frequency evaluations in studies. It should be noted that the presented maximum acceptable delays might not be universally reproducible because they are according to the defined approaches of the included studies 25 (Table 1). Most analytes were stable at −20 or −80°C such that some studies had used the −80°C storage condition as a baseline value. 11,20,22 Hereupon, we focused on a basic classification set as refrigerated or room temperature (RT).

| Alanine aminotransferase (ALT)
Nineteen of the 20 articles studied the ALT stability in different storage conditions. Accordingly, ALT activity changed substantially at room temperature. However, it could be stable up to 7 days in chilled blood. ).
TCL (total change limit) CVa, analytical imprecision; CVg, between-subject variation; CI, confidence interval; ANOVA, one-way analysis of variance; SEM, standard error of mean; SD, standard deviation; TE, total allowable error; CVb, within-subject variation. *The metabolites abbreviations are given in the Analytes Abbreviations section

| Aspartate aminotransferase (AST)
AST, the second most frequently studied analyte, was stable at least up to 56 hours in whole blood samples and up to 2 weeks in fridge serum.

| Potassium (K + )
The maximum stability for K + was 24 hours in whole blood at RT. The increase in K + after 24 hours was attributable to Na + /K + -ATPase pump failure, with diffusion of K + from the erythrocytes.

TA B L E 2
The most thirty-four studied analytes and related included articles

| Triglyceride (TG)
Seventeen of the 20 articles studied the TG stability in different storage conditions. Accordingly, it could be measured reliably in whole blood samples kept at room temperature for at least a week before plasma separation. TG in fridge serum samples was stable for two weeks.

| Alkaline phosphatase (ALP)
ALP activity was stable either in serum separator gel tube or in plain tube and could be measured reliably in serum samples kept at RT or at 4°C for at least 3 days.

| Creatinine (CREA)
One-week delay for chilled whole blood samples and two-week delay for fridge serum samples were the reported maximum acceptable delays for CREA.

| Total cholesterol (TC)
Total cholesterol, such as TG, is a stable analyte and could be measured reliably in whole blood samples kept at room temperature for at least a week. TC in fridge serum samples was stable for two weeks.

| Albumin (AL)
One-week delay for whole blood samples at RT and two- week delay for fridge serum samples were the maximum acceptable delays for AL that has been reported by the included studies.

| Lactate dehydrogenase (LDH)
LDH activity was stable as long as one week in the serum gel tube at both RT and 4°C.

| Sodium (Na + )
Fifty-six-h delay for whole blood samples at RT and two-week delay for fridge serum samples were the maximum acceptable delays for Na + that have been reported by the included studies.

| Calcium (Ca 2+ )
The maximum acceptable delay for Ca 2+ was 24 hours storage at 4°C and RT in whole blood and in plasma or serum samples.

| γ-glutamyltransferase (GGT)
For GGT, there are 14 reports of either slight or prominent changes in its activity. The maximum stability reported for GGT was 7 days at 4°C.

| Total bilirubin (TB)
The percentage changes in total bilirubin were not clinically meaningful when stored in serum gel tubes either at RT or chilled up to one week.

| Urea (BUN)
Urea in plasma and serum that were exposed to prolonged contact with cells and in double-spun specimens remained stable over the 56-h period. It was stable up to 30 days at −20°C.

| Creatine kinase (CK)
For CK, there are 12 reports of either slight or prominent changes in its activity. The maximum stability reported for CK was 7 days at 4°C.

| Inorganic phosphate (Pi)
A number of studies have reported a significant decrease in phosphate concentration in whole blood samples in the early hours after phlebotomy due to glycolysis. 1,13 However, most studies have reported an increase in Pi concentration due to hydrolysis of intraerythrocytic phosphate esters and leakage of the inorganic part. 1,15 Therefore, Pi is one of the most unstable analytes.

| Total protein (TP)
Concentrations of total protein did not differ either in serum separator gel tube or in plain tube and could be measured reliably in samples kept at RT or at 4°C for at least three days after centrifugation. It was also stable in whole blood samples up to one week at RT.

| Uric acid (UA)
For uric acid, gel tubes showed enhanced stability compared to plain tubes and the stability can be maximized by refrigeration at 4°C.

| Amylase (AM)
The results of amylase were somewhat contradictory. Most studies have found that AM is generally very stable when it was stored at 4°C or below. However, one study showed decreased levels of serum AM in prolonged storage at −20°C.

| Chloride (Cl)
Chloride could be measured reliably in whole blood and in plasma/serum samples kept at room temperature for at least 24 hours.

| High-density lipoprotein (HDL)
HDL was very stable when stored below 4°C, but some fluctuations were noted at 25°C. Nevertheless, it has been demonstrated that HDL could be measured reliably in whole blood samples kept at room temperature for at least a week before plasma separation.

| Magnesium (Mg 2+ )
Magnesium stability in whole blood was dependent on storage temperature. It is less stable at RT (a few hours) and more stable at 4°C (up to 48 hours). In double-spun plasma and serum centrifuged immediately, it was reported that Mg 2+ could be stable up to 56 hours.

| Glucose (GLU)
Glucose was an unstable analyte that could only be kept stable at RT and 4°C up to 24 hours using tubes containing glycolysis inhibitors.

| C-reactive protein (CRP)
CRP was unaffected when stored at 4°C for up to 72 hours after centrifugation. It was stable up to 24 hours when stored at RT in whole blood samples.

| Bicarbonate (BIC)
Bicarbonate was one of the analytes that was not stable after centrifugation and its amount decreased in plasma samples.

| Ferritin (FERR)
Ferritin could be measured reliably in whole blood samples kept at RT for at least 24 hours before cell separation. But it did not remain stable at higher temperatures.

| Iron (Ir)
Iron was stable after 24 hours storage at 4°C and RT in whole blood and in serum/plasma samples.

| Lipase (LIP)
Except for decreased activity at temperature fluctuation events, lipase was stable in all analytical circumstances studied.

| Transferrin (TF)
Transferrin was stable in all analytical circumstances studied.

| Cobalamin (B12)
Cobalamin was not significantly affected up to 72 hours in whole blood and serum or plasma at RT and refrigerated temperatures in plain glass and any kind of tubes.

| Cortisol (COR)
Cortisol also did not significantly change up to 72 hours in whole blood and serum or plasma at RT and refrigerated temperatures in plain glass and any kind of tubes.

| Folate (FOL)
In decapped plasma samples stored at RT, folate was stable for 2 hours. In whole blood, folate was not significantly affected up to 72 hours at 4°C and up to 48 hours at 25°C. It is stable in serum or plasma at room or refrigerated temperatures up to 72 hours.

| Free thyroxine (FT4) and Thyroid-stimulating hormone (TSH)
The same five articles studied FT4 and TSH stabilities. FT4 and TSH hormones were not significantly affected up to 72 hours in whole blood and serum or plasma at room or refrigerated temperatures.
The two hormones were also stable at 35°C until 24 hours. In decapped plasma samples stored at RT, FT4 was stable for 4 hours and TSH was stable for 6-8 hours.

| D ISCUSS I ON
A comprehensive discussion about studies reached to the different results is complicated. Different panels of analytes, measured in different matrices, collected from different populations, variety in sample sizes, different storage times, different temperature conditions, different technologies and analyzers, different acceptance limits, or statistical models are the papers variables that make it difficult to obtain an intact output and conclusion. However, dividing the problem into its components would probably lead to solve the puzzle.

| Type of blood collection tube
Blood collection tubes consist of tube walls, rubber stoppers, lubricants, anticoagulants, separator gels, clot activators, and surfactant, all of which can affect the quality of the specimens, accuracy, and precision of laboratory tests. 27

| Delays before centrifugation
Plasma and serum should ideally be separated from cells as quickly as possible to prevent cellular metabolism and analytes movement between cellular parts and the plasma or serum. Prolonged contact of plasma or serum with blood cells occurring by delays before centrifugation is a common cause of false test results. 5 Before the centrifugation, the storage time and temperature the whole blood samples were stored in are very important items for reduction of the false results. According to the reviewed papers, K + , Pi, Mg 2+ , Ca 2+ , Ir, LDH, GLU, CREA, BUN, and FERR are the analytes most influenced by delayed centrifugation, because they are present in cells and they leak out of the cell over the time. 19,21 Temperature and time storage of whole blood samples are important for improving the analyte stability. Although some analytes such as Pi, Mg 2+ , and GLU are not stable in whole blood at RT for more than 24 hours, they become more stable when samples stored at 4°C compared to RT. 19,28

| Delays after centrifugation
At clinical laboratories especially hospital laboratories, serum/ plasma samples are often stored for a given time for potential later analysis. Knowing about the reliability and stability of analytes in samples stored at RT or lower temperatures for reanalysis is very important.
According to the reviewed studies, LDH and bicarbonate were the analytes with the lowest stability after centrifugation; therefore, any reanalysis of these analytes in centrifuged tubes cannot be allowed. 9 However, many analytes in the double-spun specimens could be stable for up to 56 hours at ambient temperature. 5

| CON CLUS ION
Given the results of this study, analytes including K + , Pi, Mg 2+ , GLU, AST, BUN, HDL, and UA have different stability depending on the tube type selected. K + , Pi, Mg 2+ , Ca 2+ , Ir, LDH, GLU, CREA, BUN, and FERR have variable stability when centrifugation is delayed. Any reanalysis of analytes including LDH and bicarbonate in centrifuged tubes are not recommended.
It is necessary to emphasize that this study only combines the results of 20 stability studies and summarizes their outcomes systematically to generate maximum analyte stability information.
This study may be useful for definition of acceptable delay times and temperatures when pre-analytical sample handling takes longer than usual. However, there are some discrepancies between the results of the included studies may reflect differences in the analytical methodology. A meta-analysis study is recommended for future studies to estimate the effect of storage conditions on biochemical testing.