Many preanalytical variables of specimen collection, transport and storage may affect coagulation test results. The Clinical and Laboratory Standards Institute (CLSI) recommends analyzing fresh prothrombin time international normalized ratio (PT-INR) specimens within 24 h and activated partial thromboplastin time (APTT) and other coagulation analyses within 4 h, or a reduction in factor activity may occur. If analysis is postponed, the CLSI recommends that platelet poor plasma samples should be quickly frozen and kept at −20°C for up to 2 weeks or −70°C for up to 6 months . Shipping on dry ice (solid carbon dioxide, CO2) appears to be a common way to transport frozen samples, but guidelines concerning such transport are lacking, and little attention has been paid to dry ice as a preanalytical factor. In 2013, Murphy et al.  reported that exposure to dry ice had an acidifying effect on pH in antibody solutions. They proposed that this effect was caused by diffusion of CO2 through sample tubes or non-integral seals into sample headspace and that subsequent formation of carbonic acid and decreased sample pH occurred during thawing. If temperatures remained below −40°C, CO2 did not interact with the sample. Placing samples into a −70°C freezer for 96 h after dry ice exposure allowed the CO2 to dissipate, and prohibited acidification.
The CLSI guidelines for lupus anticoagulant (LA) analysis, updated April 2014 , now recommends that plasma samples for coagulation analyses shipped on dry ice should be thawed uncapped at 37°C for at least 15 min to avoid CO2 retention and decreased sample pH. However, this is impractical in a routine laboratory, and it increases the risk of sample contamination. We recently found that plasma samples exposed to dry ice displayed a drop in pH and significantly prolonged clotting times in LA analysis if they were placed in a −20°C freezer after being exposed to dry ice. This was avoided if samples were stored at −80°C after dry ice exposure . The aim of this study was to investigate whether dry ice exposure had similar effects on other commonly used coagulation analyses, and to evaluate whether placing samples in −80°C freezer after dry ice exposure could counteract such effects.
Materials and methods
Sample preparation and storage
Venous blood was collected as part of routine blood donation from 30 healthy blood donors, none on anticoagulant therapy, into four Vacuette sample collection tubes containing 3.2% trisodium citrate (BD, Franklin Lakes, NJ, USA), with minimal stasis. Samples were collected over 2 days, with 15 samples each day. Samples were centrifuged at room temperature (RT) at 2500 g for 15 min to produce platelet poor plasma (platelet count <10×109/L), as recommended . For each individual, supernatant plasma from three of the four Vacuette tubes was pipetted, leaving 1 cm plasma above the blood cells and dispensed into 3×5 mL polypropylene tubes with skirted push cap (Sarstedt Inc., Newton, NC, USA). pH measurements and coagulation analyses in fresh citrate plasma were performed directly in the fourth Vacuette tube from each individual, within 2 h after sample collection. In total, four aliquots per individual were allocated to four different preanalytical regimes:
R1: Fresh citrate plasma analyzed immediately.
R2: Stored at −20°C for 48 h until analysis.
R3: Stored at −20°C for 24 h, then placed in a rack in a styrofoam box with ample dry ice at ambient RT for 24 h until analysis.
R4: Stored at −20°C for 24 h, then placed in a rack in a styrofoam box with ample dry ice at ambient RT for 24 h, followed by storage at −80°C for 24 h until analysis, simulating transport of frozen samples.
In our previous work , samples that had been exposed to dry ice were placed at −80°C for 96 h to prohibit CO2 retention and acidification, as recommended by Murphy et al. . However, we later performed a pilot study to evaluate how long citrated plasma exposed to dry ice had to be kept at −80°C to avoid acidification. Based on the results from this study, we decided to store samples at −80°C for 24 h after dry ice exposure, as longer storage times were unlikely to significantly reverse pH any further.
pH analysis was performed on an automated blood gas analyzer (ABL 725; Diamond Diagnostics, Holliston, MA, USA). pH was measured in one Vacuette tube from each individual immediately after centrifugation, while the polypropylene tubes were handled according to regime 2, 3 and 4. Before analysis, frozen plasma samples were thawed in a water bath with cap on at 37°C for 10–15 min along with other samples in the same regime, in batches of five samples. According to CLSI-recommendations, samples were analyzed immediately after thawing and gentle mixing . The intermediate intralaboratory analytical coefficients of variation (CV) are listed in Table 1.
After pH measurements, coagulation analyses were performed on an automated coagulation analyzer (ACL top 750; Instrumentation Laboratory, Milan, Italy). APTT, fibrinogen, antithrombin, protein C and protein S was analyzed using reagents from Instrumentation Laboratory while PT-INR was analyzed using STA-SPA+ reagent (Diagnostica STAGO, Paris, France). APTT was measured using HemosIL® SynthASiL, antithrombin using HemosIL Liquid Antithrombin, and fibrinogen was measured using HemosIL QFA Thrombin (bovine). Protein C and protein S were measured using HemosIL protein C – and HemosIL Free Protein S reagent. The analytic principle for PT-INR, APTT and fibrinogen is based on optical clot detection, while analysis of antithrombin and protein C are chromogenic methods. Protein S was measured by an immunological turbidimetric method. The intermediate intralaboratory analytical CV for the coagulation analyses are listed in Table 1.
Data are presented as median test results, and Wilcoxon signed-rank test was used to test if the median difference≠zero. Statistical analyses were performed using STATA (version 13.1 for Windows, StataCorp LP, TX, USA). A value of p<0.001 was considered statistically significant, to account for the effects of multiple testing (Bonferroni correction).
The study was carried out in full accordance with the Ethical Principles of the Declaration of Helsinki. As it was part of a method validation study, no ethical committee approval was required, according to national Norwegian regulations.
Samples exposed to dry ice for 24 h (R3) had a statistically significantly lower median pH (6.13) compared to fresh plasma samples (R1, pH 7.16), samples that had been frozen at −20°C (R2, pH 7.42) and samples exposed to dry ice but stored at −80°C for 24 h afterwards (R4, pH 7.36) (see Table 2 and Figure 1). pH was significantly lower in fresh citrate plasma (R1) than in samples that had been frozen at −20°C (R2) and then thawed before analysis. This is probably due to CO2’s temperature-dependent solubility; as the temperature falls during freezing, CO2 diffuses out of the sample and pH increases .
Clotting times for PT-INR and APTT analyses were statistically significantly prolonged in samples exposed to dry ice (R3) compared to fresh samples (R1), samples that had been frozen at −20°C (R2) and samples exposed to dry ice but placed in a −80°C freezer afterwards (R4) (median prolongation of 0.07, 0.06 and 0.05 [corresponding to 1.65, 1.6, and 1.5 s] for PT-INR and 13.2, 13.7 and 13.3 s for APTT, respectively). The same was observed for fibrinogen, where clotting time is inversely proportional to fibrinogen concentration. Accordingly, we observed 0.3–0.36 g/L lower median fibrinogen concentration in samples exposed to dry ice (R3), compared to the other regimes. Levels of protein C was significantly lower (8.5%–11%) in samples exposed to dry ice (R3) than in samples handled according to the other regimes. Levels of antithrombin and protein S were not significantly altered in samples exposed to dry ice (R3) compared to the other regimes, but protein S levels were statistically significantly higher in samples that were frozen in −20°C (R2) compared to samples from the other three regimes. For APTT, 24 of 30 samples exposed to dry ice (R3) had clotting times above the normal reference interval. All results in fresh citrated plasma (R1) were within the reference interval. The greatest observed change in a single patient was an increase of 27.7 s in clotting time for APTT in regime R3 compared to R1. Despite the increased clotting times in regime R3, no samples had PT-INR and fibrinogen levels outside the upper reference interval.
As can be seen in Figure 1, there were three outliers with respect to pH in regime R3. These samples had not been acidified as much as the others when exposed to dry ice. To evaluate the effect of the outliers, we analyzed the dataset with and without them. No statistically significant differences were found. A possible pH-dependent increase in clotting times for APTT could be indicated by our findings, as depicted in Figure 2. However, more data are needed to confirm this.
In this work, we have shown that citrate plasma samples temporarily exposed to dry ice (R3) display a relatively large drop in pH, and have prolonged clotting times in analysis of PT-INR, APTT and fibrinogen as well as lower levels of protein C. However, these effects are avoided if samples are stored for 24 h at −80°C after dry ice exposure (R4). Twenty-four (80%) of 30 samples had APTT above the upper reference interval, demonstrating that exposure to dry ice can have consequences of clinical relevance. We previously reported that samples could be kept at −80°C for 96 h to avoid CO2 retention and acidification , but in this work, we show that 24 h is sufficient, providing a feasible alternative to thawing samples uncapped after dry ice exposure which is currently recommended by CLSI .
The effects of dry ice exposure has previously been noted in two short reports, where the authors found that prothrombin time (PT) [6, 7], but not APTT  was prolonged in samples stored on dry ice, coinciding with an acidification of the samples. More recently, Gosselin et al.  evaluated whether storing citrate plasma in polypropylene tubes in either −70°C or dry ice for 16 h had an effect on coagulation test results. Similar to our findings, they found that samples stored on dry ice had prolonged clotting times in PT analyses, but only in samples that were thawed capped after dry ice storage. They observed no change in antithrombin concentrations, coinciding with our findings. Surprisingly, they reported shorter clotting times for APTT in samples exposed to dry ice compared to samples stored at −70°C, but these changes were minimal (<1 s). In addition, they saw no effect of dry ice exposure for fibrinogen or protein C analyses. It is possible that the discrepancies between ours and their results may be due to the time the samples were exposed to dry ice. We kept samples on dry ice for 24 h, whereas Gosselin et al.  only stored them for 16 h. Moreover, they did not confirm that exposure to dry ice had an effect on sample pH, and it is conceivable that 16 h may not be sufficient to cause sample acidification, depending on the amount of dry ice used, as well as the ambient temperature dry ice and samples were stored in. We stored samples in ample dry ice in RT for exactly 24 h, and 27 of 30 samples had a pH drop >1.2 units. We observed a possible pH-dependent prolongation of APTT clotting time, as indicated by Figure 2. However, more data are needed to confirm this.
How acidification leads to prolonged clotting times in PT-INR, APTT, fibrinogen as well as LA analyses based on the SCT-and DRVVT tests  is not clear, but it is conceivable that acidification could lead to altered function or damage to clotting factors, causing a delayed clot formation. In 1978, Chaimoff et al.  demonstrated that citrate plasma adjusted with hydrochloric acid from pH 7.6 to 5, had prolonged clotting times in PT and APTT analyses, and that the increase in clotting times correlated with the degree of drop in pH. They found that the concentration of fibrinogen was relatively stable in the pH range of 7.6 to 6, but fell when pH <6. Okude et al.  found that fibrin polymerization, measured by turbidity and tromboelastography, was affected by pH <6.5 in a dose-dependent manner. Maximum fibrin polymerization was decreased, time until onset of clotting was delayed and clot elasticity was reduced at pH <6.5 in fibrinogen concentrates diluted in buffer solution. Based on these and our findings, one could speculate that the final step in the coagulation cascade could be the common denominator prolonging clotting time in coagulation analyses of samples exposed to dry ice.
Whether dry ice affects samples during transport probably depends on the type of sample tube material, type of seal/cap, the length of time exposed to dry ice, ambient temperature, thawing process, etc. For example, in another pilot study, we found that only one of 10 Vacuette tubes (polyethylene tubes with rubberized cap) became acidified after exposure to dry ice for 24 h. Plumhoff et al.  found that plasma samples kept in both polypropylene and polystyrene tubes, but not siliconized glass tubes, were affected if they were exposed to dry ice for 24 h. This indicates that both sample tube material and cap type determines whether CO2 diffuses into the sample tube and affects pH.
In conclusion, we have demonstrated that exposure to dry ice (R3) has a significant effect on sample pH, and leads to prolonged clotting time in analyses of PT-INR, APTT and fibrinogen, as well as decreased protein C levels. For APTT, the changes were of such a magnitude that 80% of samples fell outside the normal reference interval if exposed to dry ice. Dry ice is an important preanalytical factor that laboratories receiving frozen samples should take into account. It has been recommended that samples exposed to dry ice should be thawed uncapped after dry ice exposure , but we propose that samples should instead be placed at −80°C for 24 h afterwards to prevent acidification and alteration in subsequent analyses.
Clinical and Laboratory Standards Institute (CLSI). Collection, Transport, and Processing of Blood Specimens for Testing Plasma-Based Coagulation Assays and Molecular Hemostasis Assays; Approved Guideline – Fifth Edition. CLSI document H21-A5 (ISBN 1-56238-657-3). Wayne, PA: CLSI, 2008. Google Scholar
Murphy BM, Swarts S, Mueller BM, van der Geer P, Manning MC, Fitchmun MI. Protein instability following transport or storage on dry ice. Nat Methods 2013;10:278–9. Web of SciencePubMedCrossrefGoogle Scholar
Clinical and Laboratory Standards Institute (CLSI). Laboratory testing for the lupus anticoagulant; Approved guideline. CLSI document H60-1 (ISBN 1-56238-960-2). Wayne, PA: CLSI, 2014. Google Scholar
Vis R, van den Besselaar AM, Witteveen E, Meeuwisse-Braun J, Schaefer-van Mansfeld H. Artificial prolongation of the prothrombin time of lyophilized plasma induced by transportation with solid carbon dioxide. Thromb Haemost 1992;67:725. PubMedGoogle Scholar
Okude M, Yamanaka A, Akihama S. The effects of pH on the generation of turbidity and elasticity associated with fibrinogen-fibrin conversion by thrombin are remarkably influenced by sialic acid in fibrinogen. Biol Pharm Bull 1995;18:203–7. CrossrefPubMedGoogle Scholar
About the article
Published Online: 2017-08-28
Published in Print: 2017-11-27
Author contributions: All the authors have accepted responsibility for the entire content of this submitted manuscript and approved submission.
Research funding: None declared.
Employment or leadership: None declared.
Honorarium: None declared.
Competing interests: The funding organization(s) played no role in the study design; in the collection, analysis, and interpretation of data; in the writing of the report; or in the decision to submit the report for publication.