Human blood gas stability data is limited to small sample sizes and questionable statistical techniques. We sought to determine the stability of blood gases under room temperature and slushed iced conditions in patients using survival analyses.
Whole blood samples from ∼200 patients were stored in plastic syringes and kept at room temperature (22–24 °C) or in slushed ice (0.1–0.2 °C) before analysis. Arterial and venous pO2 (15–150 mmHg), pCO2 (16–72 mmHg), pH (6.73–7.52), and the CO-oximetry panel [total hemoglobin (5.4–19.3 g/dL), percentages of oxyhemoglobin (O2Hb%, 20–99%), carboxyhemoglobin (COHb, 0.1–5.4%) and methemoglobin (MetHb, 0.2–4.6%)], were measured over 5-time points. The Royal College of Pathologists of Australasia’s (RCPA’s) criteria determined analyte instability. Survival analyses identified storage times at which 5% of the samples for various analytes became unstable.
COHb and MetHb were stable up to 3 h in slushed ice and at room temperature; pCO2, pH was stable at room temperature for about 60 min and 3 h in slushed ice. Slushed ice shortened the storage time before pO2 became unstable (from 40 to 20 min), and the instability increased when baseline pO2 was ≥60 mmHg. The storage time for pO2, pCO2, pH, and CO-oximetry, when measured together, were limited by the pO2.
When assessing pO2 in plastic syringes, samples kept in slushed ice harm their stability. For simplicity’s sake, the data support storage times for blood gas and CO-oximetry panels of up to 40 min at room temperature if following RCPA guidelines.
The results of arterial blood gas analyses are used for many patient management decisions. As such, the samples analyzed must be free of preanalytical changes.
Several preanalytical changes can affect blood gas values. Blood gas values are affected by factors such as storage time , , , , , ,  and temperature , , [3, 5, 8], , , , presence of air bubbles in the syringe [8, 9, 12, 13], sample volume [14, 15], syringe composition , , , , leukocytosis or thrombocytosis , , , , , , , , , , , and the initial oxygen pressure (pO2) of the sample [1, 4, 13, 27].
Plastic syringes are used for virtually all blood gas sampling. The primary issue with plastic syringes used in blood gas analysis is that plastic syringes are gas permeable, unlike glass syringes. Nevertheless, plastic syringes are used everywhere today as they are inexpensive, less breakable, and disposable compared to glass syringes. Therefore, labs should define the maximal allowable storage times for blood gases/CO-oximetry at their institution based on up-to-date, sound research studies using plastic syringes.
Despite the relatively large number of studies that have determined the allowable storage time before blood gases become unstable (i.e., see online Supplementary Tables S1 and S2), there is no consensus. Indeed, there is a discrepancy among several studies. Some studies say that when blood is in plastic syringes, blood gases should be stored at room temperature. The analysis should be performed immediately [3, 28] or within 10 min , 15 min , 20 min [11, 12], 30 min [6, 13, 30, 31], or 60 min [8, 32] of the draw time. Furthermore, other studies recommend that samples be stored on ice if there is a delay in sampling time. Even in these studies, the storage time recommendations vary between “right away” and 95 min [1, 2, 5, 9, 10, 27, 31, 33, 34].
Some international guidelines , , ,  and manufacturer bulletins suggest analysing samples in a plastic syringe within 30 min. If a delay in analysing samples is more than 30 min, storage in ice water should be considered . Nevertheless, stability is worsened when the sample is placed on ice for some analytes (i.e., potassium  and pO2 ), so if a combination of analytes are to be assessed together, then the storage time needs to be determined by the analyte with the shortest stability time of the group.
The discrepancies in the storage time recommendations for blood gases depend on several factors. Some of these factors include the different plastic syringes used (with their different permeabilities [5, 34]), the sample volume in each syringe , and the presence of bubbles in each syringe [10, 12]. Also, the chosen allowable error value permitted before samples are deemed unstable varies between studies. For example, one study deemed that pO2 instability was defined as a change that is >9.3% from the initial value ; another study defined instability as a change in pO2 that was ≥7.5 mmHg from the initial value . These criteria differ from the analytical performance specifications from The College of American Pathologists (CAP) or the Royal College of Pathologists of Australasia (RCPA) [40, 41]. The initial pO2 of the sample also affects storage time recommendations. The sample should be analysed immediately in studies where the initial pO2 is >400 mmHg [2, 5]. However, if the sample pO2 is about 100 mmHg, the recommendation for some studies has been 60 min [8, 32]. Furthermore, the storage time recommendations can change when different statistical analyses are used. If the recommendations are based on mean changes, then the storage time is roughly based on the instability of the average sample . So about 50% of the samples would show a pO2 change that is less than the mean change, and about 50% would show a change that is more than the mean change. Stability can also be defined as the storage time for which 5% of the samples exceed the stability threshold [29, 39]. Finally, the blood gas analyser’s performance specifications also affect recommendations. Some of today’s blood gas analyzers have a total analytical error of ≤5 mmHg for pO2 in whole blood (when pO2 is ≤145 mmHg) [42, 43], which is much lower than even 20 years ago.
Survival analyses have rarely been used for stability determination. This statistical technique has only been used twice before in blood/electrolyte stability studies [29, 39]. This study’s premise is that this analysis will determine more accurate storage times for blood gases under room temperature and slushed ice conditions.
As such, the purpose of this study was to determine the storage time at which 5% of blood gas samples exceeded the performance specifications (total allowable error) as deemed by the Royal College of Pathologists of Australasia (RCPA) (Table 1) . We specifically determined the storage time for individual analytes, as well as in combination with other analytes under slushed ice and room temperature conditions.
|Blood gas||Lower limit||Upper limit|
|pO2||±5 mmHg for samples that have a pO2 of ≤83 mmHg||±6% change for samples that have a pO2 of >83 mmHg|
|pCO2||±2.0 mmHg for samples that have pCO2 of ≤34.0 mmHg||±6% change for samples that have pCO2 of >34.0 mmHg|
|Metabolites||Lower limit||Upper limit|
|Lactate||±0.5 mmol/L for samples that have a concentration of ≤5 mmol/L||±8.0% for samples that have a concentration of >5 mmol/L|
|Co-oximetry||Lower limit||Upper limit|
|Hemoglobin concentration||±5 g/L for samples that have a concentration of ≤100 g/L (±0.5 g/dL for samples that have a concentration of <10 g/dL)||±5% change for samples that have a concentration >100 g/L (±5% for samples that have a concentration of <10 g/dL)|
|Oxyhemoglobin percentage (O2Hb%)||±3 units for samples ≤75.0%||±4 units for samples >75.0%|
|Carboxyhemoglobin percentage (COHb)||±1.0 units for samples ≤5.0%||±20% change for samples >5.0%|
|Methemoglobin percentage (MetHb)||±2.0 units for samples ≤20%||±10% change for samples >20%|
August 2021. Royal College of Pathologists of Australasia .
Materials and methods
This study was completed along with two other companion studies examining whole blood lactate stability and electrolytes using identical specimens obtained for patient care [39, 44]. The UC Davis Institutional Review Board (IRB) administration reviewed the project (IRB ID 1469859-1). It determined that this research was IRB-exempt as it did not involve human subjects, and no patient-identifying information was obtained.
Whole blood venous and arterial patient specimens were obtained from the blood gas lab at the University of California, Davis Medical Center, in Sacramento, CA, between October 2019 and February 2020. The blood gas lab receives samples from all over the hospital, including the emergency department and all the intensive care units, and neurosurgical units. The blood gas lab is accredited by the State of California Department of Public Health (Lab ID CDF0002547; CLIA Number: 05D0615654) and the College of American Pathologists (CAP Number 2422006). Both venous and arterial samples were included in the study if the pO2 ranged from 10 to 150 mmHg. The high limit of 150 mmHg was selected to represent the maximum possible blood pO2 at sea level when the sample is exposed to 21% oxygen and stored at room temperature. When blood samples are exposed to atmospheric pO2, the movement of oxygen tends to go from high to low partial pressures. Thus, blood samples with an initial pO2 higher than 150 mmHg would decrease when exposed to ambient air, while samples with a pO2 lower than 150 mmHg would tend to increase when exposed to ambient air. Thus, we did not want conflicting results if the baseline pO2 of the blood samples began at >150 mmHg. The samples obtained were sent to the lab via a pneumatic tubing system. All specimens were analyzed using the Radiometer ABL 90 Flex blood gas analyzer (Radiometer Medical, Brønshøj, Denmark). The within-analyzer precision of the analyzer is presented in Supplementary Table S3.
Samples were obtained from two different syringes: (A) 3 mL Portex Line Draw Arterial Blood Sample Syringes that contained 23.5 IU of dry lithium heparin neutralized for ionized calcium per mL (Ref: 4042-2, Smiths Medical, ASD, Inc.); (B) vented 3 mL Portex Pro-Vent® Arterial Sampling Kits that contained 23.5 IU of dry lithium heparin neutralized for ionized calcium per mL (Ref: 4598P-2, Smiths Medical, ASD, Inc.).
Samples were stored at room temperature (22–24 °C) or on slushed ice (0.1–0.2 °C) over an average of 80–90 min (and up to 3 h). The storage time was limited to two to 3 h as the lab rarely processed samples above 3 h. However, when the whole blood specimens were inserted into the analyzer, the analyzer measured all specimens at 37 °C. Measurements were obtained at five different time points: baseline (minute 0), then approximately 20–30, 40–60, 60–80, and 90–180 min after receiving the sample. Each blood sample at each timepoint was mixed thoroughly for 5 s in upright and inverted positions before inserting the sample into the analyzer. All bubbles were removed before analysis. If the specimen was in slushed ice, the syringe was placed vertically in a container containing slushed ice. The temperature of the slushed ice bath was measured via two thermometers of the same brand (Fisherbrand™ Traceable™ Refrigerator/Freezer Plus Thermometer, Thermo Fisher Scientific, Pittsburgh, PA), and the temperature of the two thermometers was averaged. The reported accuracy of the thermometers was ±0.5 °C.
A repeated-measures analysis of variance (rmANOVA) was used to identify if the average values for pO2, pCO2, and pH (measured at 37 °C) and CO-oximetry (total Hb, %O2Hb, COHb, MetHb) differed over five different timepoints and two conditions (room temperature and slushed ice). A rmANOVA was also used to compare the changes in each analyte at approximately 20, 40, 60, and 80 min post-draw time. For pO2, we further compared stability in specimens with a pO2<60 mmHg to those with a pO2≥60 mmHg. The low and high pO2 groups were split this way because the oxyhemoglobin dissociation curve begins to flatten out at a pO2 of ∼60 mmHg and is steeper when pO2 is <60 mmHg , which in turn affects the oxygen buffering capacity of hemoglobin and the rate of change in pO2 when samples are stored in ice-water . Bonferroni correction was used to adjust for multiple comparisons and to determine post-hoc differences. If Mauchly’s Test Sphericity was statistically significant, then a Greenhouse–Geisser adjustment was used.
Data screening was utilized to identify outliers from each rmANOVA. Any data point with a Cook’s Distance dissimilar to the other data points via visualization was eliminated. Additionally, any studentized residual that was ≥±2.5 SD units during the screening was eliminated.
A Kaplan–Meier estimator (survival analysis) was also performed to estimate the “time to event” of only the analytes that show meaningful changes over time [46, 47]. In our case, the time to event involved computing the time at which 5% of the samples exceeded the performance specifications per the RCPA  (Table 1). We defined the storage time at which 5% of the samples exceeded performance specifications as the time of instability since there is general acceptance that 95% of samples must fall within the total allowable error [48, 49]. In addition, it is a conservative estimate used elsewhere in studies where survival analysis was performed [39, 50].
After the elimination of samples with a baseline pO2 of >150 mmHg, 99 samples (54 venous, 45 arterial) patient whole blood (heparinized) samples were used to examine the stability of pO2, pCO2, pH, and CO-oximetry over time under room temperature conditions (22–24 °C); another 93 (49 venous, 44 arterial) whole blood samples were used to examine the stability of those same analytes in slushed ice conditions (0.1–0.2 °C). From there, one outlier was removed for each of pCO2, %O2Hb, COHb, and MetHb; six outliers were removed for pH, and 19 outliers were removed for pO2 (primarily due to oxygen contamination). Baseline values are presented in Supplementary Table S4. About 17% of the arterial samples were considered “critical values” for pO2 (Supplementary Table S5). About 7 and 34% of the samples were critical for pH and pCO2 (Supplementary Table S5). None of the CO-oximetry values was considered critical at baseline in any specimen. When the groups were further subdivided based on pO2, the pO2 for the <60 mmHg group displayed a mean (SD) of 40 (10) mmHg, with a range of 15–59 mmHg. The pO2≥60 mmHg group exhibited a mean (SD) of 85 (25) mmHg, with a range of 60–150 mmHg.
According to the RCPA (Table 1), the acceptable threshold for pO2 is 5.0 mmHg when the baseline pO2 is ≤83.0 mmHg and 6.0% when the baseline pO2 is >83.0 mmHg. The survival analysis demonstrated that 5% of the samples stored at room temperature became unstable after 40 min, while samples stored in slushed ice became unstable after 20 min of storage; but this is when one considers the full range of oxygen pressures together (15–150 mmHg) (Figure 1).
However, when comparing the mean overall pO2 between conditions, there was no statistical significance between samples stored in slushed ice vs. storing samples at room temperature. The mean difference was −3 mmHg (95% CI –10 to 3 mmHg) between the conditions overall (p=0.27).
Nevertheless, when examining the mean change in pO2 over ∼85 min, there were significant differences between room temperature and slushed ice conditions, exacerbated when the initial pO2 was ≥60 mmHg (Figure 2). In Figure 2, it can be seen that storing samples in slushed ice worsens stability compared to room temperature conditions [mean difference in the increase in pO2 compared to room temperature when averaging all time points=+10 mmHg (95% CI=8–12 mmHg)], which is more exacerbated when the initial PO2 is ≥60 mmHg (+18 mmHg) compared to when the initial pO2 is <60 mmHg (+2 mmHg) and the sample is stored for extended periods (Figure 2). Main effects present: time, condition, and group (p<0.001) (Figure 2). There were also interaction effects (condition × group; time × condition; and time × group, all p<0.001) (Figure 2). There were 55 samples stored at room temperature with the initial pO2<60.0 mmHg, there were 32 samples stored at room temperature with the initial pO2≥60.0 mmHg, there were 42 samples stored in slushed ice with the initial pO2<60.0 mmHg, and there were 44 samples stored in slushed ice with the initial pO2≥60.0 mmHg (Figure 2). As well, 15 out of 32 samples that had an initial mean pO2 of 109 (SD 18 mmHg) increased to 166 (SD 14) mmHg when stored in slushed ice for 86 (SD 6) min.
According to the RCPA (Table 1), the acceptable threshold for %O2Hb is 3.0 units for samples ≤75.0% and 4.0 units for samples >75.0%. The survival analysis revealed that stability for %O2Hb was not significantly different between conditions, at about 40–45 min, when considering the full range of %O2Hb values together (20–98.8%) (Figure 3). The mean difference in %O2Hb was −1.3% (95% CI= –4.8 to 2.1%) between the conditions overall (p=0.46).
However, when separating the %O2Hb into high and low initial baseline values (<90% or ≥90%), ice shortened stability compared to room temperature conditions [mean difference in the increase in %O2Hb compared to room temperature when averaging all time points=+1.3% (95% CI=0.9–1.7%), which is more exacerbated when the initial O2Hb is <90% (+1.8%) compared to when the initial %O2Hb is <90% (+0.8%) and the sample is stored for extended periods (Figure 4). Main effects present: time, condition, and group (p<0.001) (Figure 4). There were also interaction effects (condition × group; time × condition; and time × group, all p<0.001) (Figure 4). There were 51 samples stored at room temperature with the initial O2Hb%<90%, there were 42 samples stored at room temperature with the initial O2Hb%≥90%, there were 41 samples stored in slushed ice with the initial O2Hb%<90%, and there were 53 samples stored in slushed ice with the initial O2Hb%≥90% (Figure 4). When O2Hb% was less than 90%, stability was reduced regardless of whether the sample was stored on ice (Figure 4, Supplementary Figure S2).
The mean differences in Hb concentration did not clinically change over ∼85 min. There was a mean increase of +0.1 (95% bootstrapped CI=0.0 to 0.2) g/dL after ∼85 min of storage under room temperature conditions and +0.1 (95% bootstrapped CI=0.1 to 0.2) g/dL after ∼85 min stored in slushed ice. The survival analysis demonstrated stability was ∼40 min under slushed ice conditions and 46 min when stored at room temperature. There was no statistical difference between conditions (Supplementary Figure S3).
One hundred percent of the samples remained stable for pH, pCO2, COHb, and MetHb, in slushed ice conditions, even up to ∼3 h (Supplementary Figures S4–S8). As with pO2, the average pCO2 did not differ between conditions, but the change in pCO2 did (Supplementary Figure S6). Storing samples in slushed ice is similar to storing samples at room temperature: the mean difference was −1 mmHg (95% CI –4 to 3 mmHg) between the conditions overall (p=0.76). There was a main effect of time, with time 0 significantly different from 23-, 44-, 65-, and 85-min post-storage (p<0.001). There was also a time × condition interaction effect, p<0.001). Furthermore, under room temperature conditions, 5% of the samples became unstable by ∼60 min (Supplementary Figure S4). Regarding pH, 5% of the samples became unstable at room temperature after 65 min of storage (Supplementary Figure S5). Specifically, the mean pH decreased by 0.02 pH units (95% CI=−0.03 to −0.02 pH unit change, p<0.01) after a mean storage time of 82 min at room temperature and by −0.01 pH units (95% CI=−0.01 to – 0.01 pH unit change, p<0.01) after a mean storage time of 90 min in slushed ice. No specimen stored in slushed ice ever exceeded a change in pH of more than 0.04 pH units.
Based on these findings, summary tables are presented. Table 2 displays the recommended storage time when a combination of analytes is measured, and Table S6 displays the recommended storage time when only one analyte is measured.
|Analytes analyzed and reported in combinations||Room temperature||Slushed ice|
pO2 (15–150 mmHg)
pCO2 (16–72 mmHg)
|40 min [storage time limited by the pO2]||20 min [storage time limited by the pO2]|
pO2 (<60 mmHg)
pCO2 (16–72 mmHg)
|60 min [all analytes had similar storages times, separately]||40 min [storage time limited by the pO2]|
pO2 (≥60 mmHg)
pCO2 (16–72 mmHg)
|25 min [storage time limited by the pO2]||20 min [storage time limited by the pO2]|
|CO-oximetry panel only
Total Hb (5.4–19.3 g/dL)
|45 min [storage time limited by the O2Hb%]||40 min|
|Metabolite panel only
Lactate (0.6–13.5 mmol/L)
Glucose (2–17 mmol/L)
|40 min [storage time limited by lactate]||∼180 min|
|Electrolyte panel only
K+ (2.2–11.7 mmol/L)
Ca2+ (0.73–1.50 mmol/L)
Cl− (77–130 mmol/L)
Na+ (97–167 mmol/L)
|120 min||∼70 min [storage time limited by potassium concentration]|
|Blood gas panel and CO-oximetry panel together, with or without the metabolite panel (pO2<150 mmHg)||40 min [storage time limited by the pO2 and lactate]||20 min [storage time limited by the pO2]|
|Blood gas panel and CO-oximetry panel together (pO2<60 mmHg)||45 min [storage time limited by the O2Hb%]||40 min [storage time limited by the pO2]|
|Blood gas panel and CO-oximetry panel together, including the metabolite panel (pO2<60 mmHg)||40 min [storage time limited by lactate]||40 min [storage time limited by the pO2]|
|Blood gas panel and CO-oximetry panel, together with or without the metabolite panel (pO2≥60 mmHg)||25 min [storage time limited by the pO2]||20 min [storage time limited by the pO2]|
|pH values in pleural fluid (pH 7.00–7.61)||60 min||135 min|
The recommended allowable storage time (min) is rounded to the nearest 5-min value before 5% of the specimens become “unstable.” Each specimen was kept inside a plastic syringe with no air bubbles. Unstable means that >5% of samples for the analyte have exceeded the performance specifications per the Royal College of Pathologists of Australasia . The numbers within parentheses are the ranges under which each analyte was assessed. Most labs report either the blood gas panel or the combination of blood gases and CO-oximetry. As such, storing samples in ice shortens the stability considering that most analytes are measured together. In other words, avoid ice when analyzing the blood gas panel or the combination of blood gases and CO-oximetry. This table was created from the current study and two other companion studies that used identical specimens [39, 44]. In addition, another study published by the same group developed storage times for pleural fluid pH .
The purpose of this study was to determine the storage time at which 5% of blood gas samples exceeded the performance specifications (total allowable error) as deemed by the RCPA (Table 1) . We determined the storage time for individual analytes and a combination of analytes under slushed ice and room temperature conditions using Kaplan–Meier survival analysis, a technique rarely used in stability studies. This study demonstrated that when the complete blood gas and CO-oximetry panels were analyzed, the acceptable storage time was 25 min at room temperature, but only when the pO2 was >60 mmHg. If the pO2 was ≤60 mmHg, the storage time for blood gases and CO-oximetry panels was 45 min. If the complete range of pO2’s is considered together (15–150 mmHg) with no consideration for initial pO2, then a storage time of up to 40 min can be acceptable at room temperature, depending on the laboratory’s total allowable error and RCPA criteria (Table 1). Storing specimens on ice worsened the stability of pO2 compared to samples stored at room temperature.
When blood gas and CO-oximetry analytes were measured individually, pCO2, pH, COHb, and MetHb remained stable in plastic syringes stored in slushed ice for 3 h. The high stability of pH and pCO2 stored under cold temperatures is confirmed in other studies [1, 3, 11, 12, 27].
The confusion in interpreting stability from various studies (i.e., studies presented in Tables S1 and S2) can be partly due to the statistical analysis performed and the pO2 ranges evaluated together or separately. We believe that presenting data in terms of mean values and standard deviation of the mean values at various storage time points, as so many studies have done (i.e. [3, 6, 7, 15, 32, 51]), does not tell the whole story. For example, we have shown no statistically significant difference between conditions for pO2 if presenting the mean value (and 95% CI of the mean value) over five different time points). However, if the mean change in pO2 is compared to the baseline, there is a statistical difference between conditions (Figure 2). Nevertheless, even when presenting the figure this way, there are some misleading numbers. For example, one could think that a mean increase in pO2 of 4 mmHg over 85 min of storage time at room temperature would suggest that the pO2 is stable over 85 min since the analyte performance specifications for pO2 are within ±5 mmHg or 6% . However, this would be incorrect as the survival analysis shows that 5% of the samples exceed this threshold by ∼40 min post-draw time when looking at specimens with oxygen pressures varying from 15 to 150 mmHg (Figure 1). Nonetheless, in Figure 1, the initial pO2 was not considered. If the initial pO2 was considered by separating the groups into low (≤60 mmHg) and high baseline pO2 values (>60 mmHg), the pO2 stability increases to ∼60 min when considering samples with low oxygen pressures and decreases to ∼25 min when considering specimens with high oxygen pressures (Supplementary Figure S1). Thus, the various findings between studies presented in Supplementary Tables S1 and S2 can be due to the type of statistical analysis and the range of oxygen pressures lumped together in the analyses.
Furthermore, the use of ice in stability studies is also confusing to interpret. When one uses plastic syringes, the oxygen can move from the slushed ice through the plastic syringe and into the blood due to the syringe’s permeability. As blood is cooled, the oxyhemoglobin dissociation curve shifts left, resulting in hemoglobin having an increased affinity for oxygen . At the same time, the solubility of oxygen increases when blood is cooled . The oxygen solubility in the blood and plasma approximately doubles when blood is cooled from 37 to about 0.2 °C [52, 53]. In combination, these two phenomena lead to an influx of exogenous oxygen in the sample when stored in a semipermeable container such as a plastic syringe . This explains why in the current study, nearly 50% of the samples with an initial pO2>60 mmHg ended up with a final pO2 of >150 mmHg after nearly 90 min of slushed ice storage. Moreover, when the sample is reheated to 37 °C in a blood gas analyzer, the exogenous oxygen added to the blood and plasma while cooled results in a falsely increased pO2 . We have shown this to be true (Figure 2, bottom panel), confirming other reports [4, 27]. Interestingly, we observed that the magnitude of this change is dependent on the initial pO2 of the sample, as was also shown by Mahoney et al. . They attributed this to hemoglobin’s loss of oxygen buffering capacity at a higher initial pO2.
The recommendations for storage times in plastic syringes are presented in Table 2 (combined analytes) and Supplementary Table S5 (individual analytes). Those tables are based on the decision limits for stability by determining when 5% of the samples exceed the thresholds specified by the RCPA . It is also understood that laboratories may wish to have a straightforward recommendation that works for all analytes without being perfectly precise. There would be no need to partition pO2 into low and high categories. When specimens are stored without air inside plastic syringes, a storage time of up to 40 min at room temperature would seem acceptable “enough” for all analytes, depending on the total allowable error and RCPA’s performance criteria (Table 1).
In two companion studies, whole blood lactate and electrolyte stability were determined from identical specimens [39, 44], but the lactate study did not use a survival analysis. We have now updated the stability of lactate based on survival analysis (Supplementary Figure S9). Nevertheless, Supplementary Figure S9 shows similar results to the original paper , i.e., whole blood lactate stored at room temperature in a blood gas syringe for less than 45 min remains stable. Indeed, blood-gas analyzers can assess blood gases, CO-oximetry, metabolites, and electrolytes. Therefore, Table 2 provides the precise storage times when metabolites and electrolytes are included in the analyses, along with blood gases, CO-oximetry, and pleural fluid. When the combination of analytes is measured, the stability depends on the most unstable analyte. For example, suppose the complete electrolyte panel is assessed, and the sample is stored on ice. In that case, potassium’s storage time is limited as it is the first electrolyte to exceed the stability criteria (Supplementary Table S5 and Zavorsky et al. ).
There may be some concern that white blood cell or platelet counts were not simultaneously measured along with the blood gases in this study. A high white blood cell (WBC) or platelet count in syringes could consume enough oxygen in vitro to artificially lower the pO2 in the syringe, resulting in “spurious hypoxemia” . In patients with leukocytosis or thrombocytosis, the decline in pO2 is greater when stored at room temperature compared to on ice [21, 23, 24, 26], but regardless of the storage medium, the pO2 decline occurs almost immediately when blood is placed in a syringe , , [23, 26, 55]. At room temperature, pO2 declines by ∼10–50 mmHg within ∼0–15 min of the draw-time in patients with leukocytosis , , [23, 25, 55] vs. 0–11 mmHg when stored on ice [16, 21, 23]. However, the WBC or platelet counts must be sufficiently high for a rapid decline in pO2 to occur. There was extreme leukocytosis in nine case reports totaling 12 patients (median leukocyte count was 276,000 cells per mm3) when falsely low pO2 was manifested , , , , , , , [23, 26]. Extreme thrombocytosis (>500,000 cells per mm3) also must be observed before spurious hypoxemia appears . Specifically, the WBC count needs to be ≥∼50,000 cells per mm3 or the platelet count needs to be >500,000 cells per mm3 for any chance of spurious hypoxemia in blood gas samples [18, 23], , , [26, 56]. Spurious hypoxemia and pseudo hyperkalemia are rare since only ∼1% of patients admitted to a major hospital has either >500,000 platelets per mm3 or >50,000 leukocytes per mm3 or both . Finally, only one of our samples had a drop in pO2 of 10 mmHg within the first 20 min of storage at room temperature, and another sample stored on ice decreased by one mmHg over the same period. Thus, we are confident that spurious hypoxemia was unlikely to occur in more than 1% of our samples.
In conclusion, this study provides a novel method to determine appropriate storage times of blood in plastic syringes using survival analysis. Another strength of this study was the large sample size (∼200 human blood specimens measured over five different time points under room temperature and slushed ice conditions). We show that 95% of all whole blood specimens obtained from humans remain stable at room temperature when the complete CO-oximetry panel or the combination of the CO-oximetry panel and pO2, pCO2, and pH are assessed together within 45 min of the draw time, but only when the baseline pO2<60 mmHg. If the baseline pO2 is ≥60 mmHg, 95% of the samples remain stable within 25 min of the draw time. When baseline pO2≥60 mmHg, the reduction in stability time can be due to the decreased buffering capacity of hemoglobin since hemoglobin is more saturated at higher oxygen pressures, reducing its buffering capacity. When assessing multiple analytes together, ice usually worsens stability and shortens the storage time before 5% of the samples become unstable. Nevertheless, if a laboratory wishes to have one practical, simple recommendation on blood gas stability, which includes most analytes obtained from a blood-gas analyzer, 30 min of storage time at room temperature can be recommended. However, a storage time of up to 40 min is acceptable under room temperature conditions, depending on the lab’s total allowable error and stability criteria used.
Radiometer America provided one high-volume solution pack and one high-volume sensor cassette to conduct this study. The authors also thank the staff of the UC Davis blood gas lab for their help in data collection.
Research funding: None declared.
Author contributions: A literature search was performed by XVW. GZ performed the study design, data collection, and data analysis. Manuscript preparation was performed by XVW, and GZ. Those responsible for drafting the work and revising it critically were GZ and XVW. All authors have accepted responsibility for the entire content of this manuscript and approved its submission.
Competing interests: GSZ is the Vice-Chair of the Clinical & Laboratory Standards Institute (CLSI) Document Development Committee (DDC) for Blood Gas and pH Analysis and Related Measurements (C46). XVW is a Member of CLSI DDC for Blood Gas and pH Analysis and Related Measurements (C46) and is a Board member of the American Association of Clinical Chemistry Industry Division. XVW is also an employee of Beckman Coulter (part of Danaher).
Informed consent: Not applicable.
Ethical approval: Not applicable. The local Institutional Review Board deemed the study exempt from review.
1. Arbiol-Roca, A, Imperiali, CE, Dot-Bach, D, Valero-Politi, J, Dastis-Arias, M. Stability of pH, blood gas partial pressure, hemoglobin oxygen saturation fraction, and lactate concentration. Ann Lab Med 2020;40:448–56. https://doi.org/10.3343/alm.2020.40.6.448.Search in Google Scholar PubMed PubMed Central
2. Smeenk, FW, Janssen, JD, Arends, BJ, Harff, GA, van den Bosch, JA, Schonberger, JP, et al.. Effects of four different methods of sampling arterial blood and storage time on gas tensions and shunt calculation in the 100% oxygen test. Eur Respir J 1997;10:910–3. https://doi.org/10.1183/09031936.97.10040910.Search in Google Scholar
3. Knowles, TP, Mullin, RA, Hunter, JA, Douce, FH. Effects of syringe material, sample storage time, and temperature on blood gases and oxygen saturation in arterialized human blood samples. Respir Care 2006;51:732–6.Search in Google Scholar
4. Mahoney, JJ, Harvey, JA, Wong, RJ, Van Kessel, AL. Changes in oxygen measurements when whole blood is stored in iced plastic or glass syringes. Clin Chem 1991;37:1244–8. https://doi.org/10.1093/clinchem/37.7.1244.Search in Google Scholar
5. Pretto, JJ, Rochford, PD. Effects of sample storage time, temperature and syringe type on blood gas tensions in samples with high oxygen partial pressures. Thorax 1994;49:610–2. https://doi.org/10.1136/thx.49.6.610.Search in Google Scholar PubMed PubMed Central
7. Smajic, J, Kadic, D, Hasic, S, Serdarevic, N. Effects of post-sampling analysis time, type of blood samples and collection tubes on values of blood gas testing. Med Glas (Zenica) 2015;12:108–12. https://doi.org/10.17392/823-15.Search in Google Scholar PubMed
8. Cuhadar, S, Ozkanay-Yoruk, H, Koseoglu, M, Katircioglu, K. Detection of preanalytical errors in arterial blood gas analysis. Biochem Med (Zagreb) 2022;32:020708. https://doi.org/10.11613/bm.2022.020708.Search in Google Scholar
9. Harsten, A, Berg, B, Inerot, S, Muth, L. Importance of correct handling of samples for the results of blood gas analysis. Acta Anaesthesiol Scand 1988;32:365–8. https://doi.org/10.1111/j.1399-6576.1988.tb02746.x.Search in Google Scholar PubMed
10. Biswas, CK, Ramos, JM, Agroyannis, B, Kerr, DN. Blood gas analysis: effect of air bubbles in syringe and delay in estimation. Br Med J (Clin Res Ed) 1982;284:923–7. https://doi.org/10.1136/bmj.284.6320.923.Search in Google Scholar PubMed PubMed Central
11. Nanji, AA, Whitlow, KJ. Is it necessary to transport arterial blood samples on ice for pH and gas analysis? Can Anaesth Soc J 1984;31:568–71. https://doi.org/10.1007/bf03009545.Search in Google Scholar
12. Madiedo, G, Sciacca, R, Hause, L. Air bubbles and temperature effect on blood gas analysis. J Clin Pathol 1980;33:864–7. https://doi.org/10.1136/jcp.33.9.864.Search in Google Scholar PubMed PubMed Central
13. O’Connor, TM, Barry, PJ, Jahangir, A, Finn, C, Buckley, BM, El-Gammal, A. Comparison of arterial and venous blood gases and the effects of analysis delay and air contamination on arterial samples in patients with chronic obstructive pulmonary disease and healthy controls. Respiration 2011;81:18–25. https://doi.org/10.1159/000281879.Search in Google Scholar PubMed
14. Kume, T, Sisman, AR, Solak, A, Tuglu, B, Cinkooglu, B, Coker, C. The effects of different syringe volume, needle size and sample volume on blood gas analysis in syringes washed with heparin. Biochem Med (Zagreb) 2012;22:189–201. https://doi.org/10.11613/bm.2012.022.Search in Google Scholar
15. Hedberg, P, Majava, A, Kiviluoma, K, Ohtonen, P. Potential preanalytical errors in whole-blood analysis: effect of syringe sample volume on blood gas, electrolyte and lactate values. Scand J Clin Lab Invest 2009;69:585–91. https://doi.org/10.1080/00365510902878716.Search in Google Scholar PubMed
16. Loke, J, Duffy, TP. Normal arterial oxygen saturation with the ear oximeter in patients with leukemia and leukocytosis. Cancer 1984;53:1767–9. https://doi.org/10.1002/1097-0142(19840415)53:8<1767::aid-cncr2820530826>3.0.co;2-e.10.1002/1097-0142(19840415)53:8<1767::AID-CNCR2820530826>3.0.CO;2-ESearch in Google Scholar
17. Mizock, BA, Franklin, C, Lindesmith, P, Shah, PC. Confirmation of spurious hypoxemia using continuous blood gas analysis in a patient with chronic myelogenous leukemia. Leuk Res 1995;19:1001–4. https://doi.org/10.1016/0145-2126(95)00117-4.Search in Google Scholar
18. Gorski, TF, Ajemian, M, Hussain, E, Talhouk, A, Ruskin, G, Hanna, A, et al.. Correlation of pseudohypoxemia and leukocytosis in chronic lymphocytic leukemia. South Med J 1999;92:817–9. https://doi.org/10.1097/00007611-199908000-00016.Search in Google Scholar
21. Wong, KF, Leung, VK, Ma, SK, Ma, YH. Spurious anoxaemia in a patient with chronic myeloid leukaemia. Clin Lab Haematol 1992;14:263–4. https://doi.org/10.1111/j.1365-2257.1992.tb00374.x.Search in Google Scholar
22. Prasad, KN, Manjunath, P, Priya, L, Sasikumar, S. Overcoming the problem of pseudohypoxemia in myeloproliferative disorders: another trick in the bag. Indian J Crit Care Med 2012;16:210–2. https://doi.org/10.4103/0972-5229.106504.Search in Google Scholar PubMed PubMed Central
23. Gartrell, K, Rosenstrauch, W. Hypoxaemia in patients with hyperleukocytosis: true or spurious, and clinical implications. Leuk Res 1993;17:915–9. https://doi.org/10.1016/0145-2126(93)90037-l.Search in Google Scholar PubMed
24. Hess, CE, Nichols, AB, Hunt, WB, Suratt, PM. Pseudohypoxemia secondary to leukemia and thrombocytosis. N Engl J Med 1979;301:361–3. https://doi.org/10.1056/nejm197908163010706.Search in Google Scholar
25. Chillar, RK, Belman, MJ, Farbstein, M. Explanation for apparent hypoxemia associated with extreme leukocytosis: leukocytic oxygen consumption. Blood 1980;55:922–4. https://doi.org/10.1182/blood.v55.6.922.bloodjournal556922.Search in Google Scholar
27. Beaulieu, M, Lapointe, Y, Vinet, B. Stability of PO2, PCO2, and pH in fresh blood samples stored in a plastic syringe with low heparin in relation to various blood-gas and hematological parameters. Clin Biochem 1999;32:101–7. https://doi.org/10.1016/s0009-9120(98)00098-8.Search in Google Scholar PubMed
28. Scott, PV, Horton, JN, Mapleson, WW. Leakage of oxygen from blood and water samples stored in plastic and glass syringes. Br Med J 1971;3:512–6. https://doi.org/10.1136/bmj.3.5773.512.Search in Google Scholar PubMed PubMed Central
29. Picandet, V, Jeanneret, S, Lavoie, JP. Effects of syringe type and storage temperature on results of blood gas analysis in arterial blood of horses. J Vet Intern Med 2007;21:476–81. https://doi.org/10.1111/j.1939-1676.2007.tb02993.x.Search in Google Scholar
30. Higgins, V, Nichols, M, Gao, H, Maravilla, N, Liang, E, Su, J, et al.. Defining blood gas analysis stability limits across five sample types. Clin Biochem 2022. https://doi.org/10.1016/j.clinbiochem.2022.09.006.Search in Google Scholar PubMed
31. Woolley, A, Hickling, K. Errors in measuring blood gases in the intensive care unit: effect of delay in estimation. J Crit Care 2003;18:31–7. https://doi.org/10.1053/jcrc.2003.yjcrc7.Search in Google Scholar
32. Mohammadhoseini, E, Safavi, E, Seifi, S, Seifirad, S, Firoozbakhsh, S, Peiman, S. Effect of sample storage temperature and time delay on blood gases, bicarbonate and pH in human arterial blood samples. Iran Red Crescent Med J 2015;17:e13577. https://doi.org/10.5812/ircmj.13577.Search in Google Scholar PubMed PubMed Central
33. Brito, MV, Cunha, IC, Aragon, MG, Braga, TG, Lima, FD. Effects of blood storage on ice in biochemical and arterial blood gas analysis of rats. Acta Cir Bras 2008;23:462–8. https://doi.org/10.1590/s0102-86502008000500013.Search in Google Scholar PubMed
34. Muller-Plathe, O, Heyduck, S. Stability of blood gases, electrolytes and haemoglobin in heparinized whole blood samples: influence of the type of syringe. Eur J Clin Chem Clin Biochem 1992;30:349–55. https://doi.org/10.1515/cclm.19220.127.116.119.Search in Google Scholar PubMed
35. Burnett, RW, Covington, AK, Fogh-Andersen, N, Kulpmann, WR, Maas, AH, Muller-Plathe, O, et al.. International Federation of Clinical Chemistry (IFCC), Committee on pH, Blood Gases and Electrolytes: approved IFCC recommendation on definitions of quantities and conventions related to blood gases and pH. Eur J Clin Chem Clin Biochem 1995;33:399–404.Search in Google Scholar
36. Davis, MD, Walsh, BK, Sittig, SE, Restrepo, RD. AARC clinical practice guideline: blood gas analysis and hemoximetry: 2013. Respir Care 2013;58:1694–703. https://doi.org/10.4187/respcare.02786.Search in Google Scholar PubMed
37. Dukic, L, Kopcinovic, LM, Dorotic, A, Barsic, I. Blood gas testing and related measurements: national recommendations on behalf of the Croatian Society of Medical Biochemistry and Laboratory Medicine. Biochem Med (Zagreb) 2016;26:318–36. https://doi.org/10.11613/bm.2016.036.Search in Google Scholar PubMed PubMed Central
38. Clinical and Laboratory Sciences Institute (CLSI). Blood gas and pH analysis and related measurements. Number 8. C46-A2, Approved Guideline-Second Edition, D’Orazio, P, et al.., editors. Wayne Pennsylvannia; 2009, vol 29.Search in Google Scholar
39. Zavorsky, GS, van Wijk, XMR, Gasparyan, S, Stollenwerk, NS, Brooks, RA. Stability of whole blood electrolyte specimens at room temperature vs. slushed ice conditions. J Appl Lab Med 2022;7:541–54. https://doi.org/10.1093/jalm/jfab089.Search in Google Scholar PubMed
40. Royal College of Pathologists of Australasia. Quality assurance programs. In: Chemical pathology analytical performance specifications: blood gases 2021. Sydney, Australia; 2021:1–16 pp. Available from: https://rcpaqap.com.au/resources/chemical-pathology-analytical-performance-specifications/.Search in Google Scholar
41. College of American Pathologists. Surveys and Anatomic Pathology Education Programs: critical care blood gas with chemistry. Northfield, IL: AQ-C2020, College of American Pathologists; 1–20 pp.Search in Google Scholar
42. Radiometer Medical. ABL 90 Flex Plus. Instructions for use (from software version 3.4). Brønshøj, Denmark: Radiometer; 2018:406 pp.Search in Google Scholar
43. RAPIDPoint 500 System Operator’s Guide. REF. 11537381, 2022, Siemen’s Healthcare Diagnostics: Tarrytown, NY.Search in Google Scholar
44. Zavorsky, GS, Gasparyan, S, Stollenwerk, NS, Brooks, RA. Stability of whole blood lactate specimens at room temperature versus slushed ice conditions. Respir Care 2021;66:494–500. https://doi.org/10.4187/respcare.08023.Search in Google Scholar PubMed
45. Astrup, P, Engel, K, Severinghaus, JW, Munson, E. The influence of temperature and pH on the dissociation curve of oxyhemoglobin of human blood. Scand J Clin Lab Invest 1965;17:515–23. https://doi.org/10.3109/00365516509083359.Search in Google Scholar
47. Rich, JT, Neely, JG, Paniello, RC, Voelker, CC, Nussenbaum, B, Wang, EW. A practical guide to understanding Kaplan-Meier curves. Otolaryngol Head Neck Surg 2010;143:331–6. https://doi.org/10.1016/j.otohns.2010.05.007.Search in Google Scholar PubMed PubMed Central
48. Westgard, JO, Carey, RN, Wold, S. Criteria for judging precision and accuracy in method development and evaluation. Clin Chem 1974;20:825–33. https://doi.org/10.1093/clinchem/20.7.825.Search in Google Scholar
49. Roys, EA, Husoy, AM, Brun, A, Aakre, KM. Impact of different sampling and storage procedures on stability of acid/base parameters in venous blood samples. Clin Chem Lab Med 2021;59:e370–3. https://doi.org/10.1515/cclm-2021-0202.Search in Google Scholar PubMed
50. Zavorsky, GS. The stability of pleural fluid pH under slushed ice and room temperature conditions. Clin Chem Lab Med 2022;61:e22–4. https://doi.org/10.1515/cclm-2022-0669.Search in Google Scholar PubMed
51. Wan, XY, Wei, LL, Jiang, Y, Li, P, Yao, B. Effects of time delay and body temperature on measurements of central venous oxygen saturation, venous-arterial blood carbon dioxide partial pressures difference, venous-arterial blood carbon dioxide partial pressures difference/arterial-venous oxygen difference ratio and lactate. BMC Anesthesiol 2018;18:187. https://doi.org/10.1186/s12871-018-0655-9.Search in Google Scholar PubMed PubMed Central
52. Christoforides, C, Hedley-Whyte, J. Effect of temperature and hemoglobin concentration on solubility of O2 in blood. J Appl Physiol 1969;27:592–6. https://doi.org/10.1152/jappl.1918.104.22.1682.Search in Google Scholar PubMed
53. Christoforides, C, Laasberg, LH, Hedley-Whyte, J. Effect of temperature on solubility of O2 in human plasma. J Appl Physiol 1969;26:56–60. https://doi.org/10.1152/jappl.1922.214.171.124.Search in Google Scholar PubMed
56. Grzych, G, Roland, E, Beauvais, D, Maboudou, P, Lippi, G. Leukocytosis interference in clinical chemistry: shall we still interpret test results without hematological data? J Med Biochem 2020;39:66–71. https://doi.org/10.2478/jomb-2019-0005.Search in Google Scholar PubMed PubMed Central
57. Ranjitkar, P, Greene, DN, Baird, GS, Hoofnagle, AN, Mathias, PC. Establishing evidence-based thresholds and laboratory practices to reduce inappropriate treatment of pseudohyperkalemia. Clin Biochem 2017;50:663–9. https://doi.org/10.1016/j.clinbiochem.2017.03.007.Search in Google Scholar PubMed
This article contains supplementary material (https://doi.org/10.1515/cclm-2022-1085).
© 2023 the author(s), published by De Gruyter, Berlin/Boston
This work is licensed under the Creative Commons Attribution 4.0 International License.