Effects of the start date of refrigeration on functions of delayed cold‐stored platelets

Abstract

Background and Objectives

The functions of cold‐stored platelets (PLTs) have recently been re‐evaluated. In Japan, cold storage may be started immediately after blood collection (Day 0, control group), after the period for improving bacterial screening sensitivity has ended (Day 2, DCS2) and after the expiration date of room‐temperature storage has ended (Day 5, DCS5). The present study investigated the effects of different cold storage start dates on PLT function.

Materials and Methods

Concentrated PLTs, collected using an apheresis device, were placed in an additive solution (Terumo‐platelet additive solution+, Terumo BCT) and then divided into three groups with different cold storage start dates (days 0, 2 or 5). The three groups were stored until Day 21, and various PLT function parameters—including aggregation (adenosine diphosphate [ADP] + collagen‐induced and ristocetin‐induced aggregation), viscoelasticity (maximum clot firmness [MCF], clotting time and α‐angle) and PLT activation (CD62P, PAC‐1, CD63 and Annexin V)—were measured and compared.

Results

Comparisons of results until Day 14 revealed that among the PLT function parameters examined, only ADP + collagen‐induced aggregation and MCF showed a deterioration in DCS2 and DCS5. However, aggregation and MCF values in DCS2 and DCS5 on Day 10 were not markedly different from those in the control group on Day 14 (US expiration date).

Conclusion

PLTs stored at room temperature for 2 or 5 days after collection and then refrigerated may be preserved until Day 10.

Highlights.

• Delayed cold‐stored platelets (PLTs) after 2 or 5 days of storage at 20–24°C maintained desirable levels of PLTs, pH, PLT activation markers and ristocetin‐stimulated aggregation until at least Day 14.

• ADP + collagen aggregation and maximum clot firmness in delayed cold‐stored PLTs on Day 10 were not significantly different from PLTs stored in the cold at Day 0 and studied on Day 14.

• PLTs stored at room temperature for 2 or 5 days after blood collection and then refrigerated may be stored until Day 10.

INTRODUCTION

The standard practice for transfusing platelets (PLTs) to maintain PLT counts in patients with thrombocytopenia involves the use of room temperature (RT)‐stored platelet concentrates (PCs), some of which show long in vivo survival after transfusion. Stolla et al. reported that the post‐transfusion recovery rate of fresh PLTs was 60% on Day 1 after transfusion and decreased to 20% by Day 7 [1]. The effects of refrigerated storage have also been examined [2]. Although refrigerated storage has the disadvantage of a short survival time after transfusion [2, 3, 4], it reduces the risk of bacterial infection by inhibiting bacterial proliferation and improving haemostatic function [5]; its effects have recently been re‐evaluated [6, 7, 8, 9, 10, 11, 12]. The Food and Drug Administration (FDA) has authorized the use of refrigerated PLTs derived from apheresis, which may be stored for 14 days after collection, for the treatment of bleeding when RT‐stored PLTs are unavailable or impractical [13]. Clinical trials on refrigerated PCs have been conducted in Norway and the United States, and their effectiveness has been reported [11, 12, 14, 15, 16]. The findings of these clinical trials have indicated the clinical utility of refrigerated PCs stored for up to 14 days. The recently completed Chilled Platelet Study (CHIPS; NCT04834414) [17, 18]—a Phase III, multicentre, international clinical trial—compared the efficacy and safety of RT‐stored PCs with those of refrigerated PCs stored for up to 21 days. Although the findings of this trial have yet to be published, they are expected to support the extension of the storage duration for refrigerated PCs. In addition, the limited use of PCs refrigerated after the completion of RT storage is being considered to promote the efficient utilization of PLT products [19, 20, 21, 22]. In response to PC shortages during the COVID‐19 pandemic, the FDA in 2020 authorized the use of pathogen‐inactivated platelet concentrates suspended in platelet additive solution (PAS‐PCs) stored at RT for 5 days followed by refrigeration for up to 9 days at the Mayo Clinic in the United States [21]. Furthermore, when refrigerated PLTs were administered to patients, sufficient haemostasis with no adverse events was achieved [21]. In Canada, a study comparing pathogen‐inactivated PAS‐PCs stored at RT with those stored at RT for 4 days followed by refrigeration for up to 10 days is currently under way [23, 24].

In Japan, PCs are 100% plasma PCs, which are universally irradiated. Following the introduction of bacterial screening in 2025, the shelf life of PCs was extended from 3 days after collection to 5 days. This change was made to improve the sensitivity of screening by implementing a 2‐day RT holding period at the blood centre after collection. If refrigerated PLTs are put into practical use, it is desirable to start refrigeration immediately after blood collection (Day 0) for the inhibition of bacterial growth. However, there are also options to start refrigeration at the end of the holding period (Day 2) or at the expiration date (Day 5) for the effective use of PCs, and it is unclear how much effect these options have on PLT function.

PAS‐PCs are prepared by replacing 60%–70% of the plasma in PCs with an additive solution. PAS‐PCs have the advantage of causing fewer transfusion‐related adverse events than PCs made from 100% plasma [25, 26, 27], and the replaced plasma may be used as a raw material for plasma‐fractionated products. Therefore, PAS‐PCs may be introduced in Japan in the near future. In addition, cold stored PAS‐PCs have been shown to suppress the formation of PLT aggregates compared to refrigerated storage in 100% plasma [9]. However, the quality of apheresis‐derived, irradiated PAS‐PCs stored at RT for 2 or 5 days and then refrigerated has yet to be evaluated.

In the present study, three groups of PAS‐PCs (start of refrigeration: Day 0, 2 or 5) were cold‐stored until Day 21, and the effects of the timing of the start of cold storage on PLT storage performance were examined.

MATERIALS AND METHODS

PAS‐PC preparation and storage

Hyperconcentrated PLTs were collected from 10 healthy volunteers using an apheresis collection system (Trima Accel, Terumo BCT, Tokyo, Japan). PAS‐PCs were prepared by diluting hyperconcentrated PLTs to a final ratio of 65% Terumo‐platelet additive solution+ (T‐PAS+) (Terumo BCT, Tokyo, Japan)/35% plasma through the addition of T‐PAS+ to the PLT bag after collection (Figure 1). The components were rested for more than 1 h, and were then stored for more than 1 h at RT (20–24°C) with constant agitation (60 cycles/min) on a shaker (EBAC, Ebara, Tokyo, Japan).

FIGURE 1.

Preparation and storage of blood components.

Two PLT products of the same ABO group were pooled and divided into three equal aliquots using storage bags integrated into the apheresis collection kit (Figure 1). During preparation of PAS‐PC, if multiple PLT aggregates were observed after a 1‐h resting period followed by 1 h of agitation, the aggregates were removed prior to dividing the samples into three aliquots. Residual white blood cells in each aliquot were measured using a Sysmex XN haematology analyser with the Blood Bank mode (XN‐1000, Sysmex Corporation, Kobe, Japan).

One of the aliquots was stored agitated at RT for 2 days before being transferred to cold storage (4°C) without agitation (DCS2) (Figure 1). Another aliquot was stored agitated at RT for 5 days before being transferred to cold storage (DCS5) (Figure 1). The remaining aliquot was stored cold without agitation from Day 0 (control group) (Figure 1). The times from PAS‐PC preparation to refrigeration for the five control units were 6 h 19 min, 4 h 47 min, 4 h 49 min, 4 h 17 min and 6 h 28 min, respectively.

On Day 2 post‐collection, PAS‐PCs were X‐irradiated at 15 Gy using an X‐irradiator (MBR‐1530A‐TW, Hitachi Healthcare Systems, Inc., Tokyo, Japan) (Figure 1). This study was approved by the Institutional Research Board of the Japanese Red Cross Society (approval number: 2021‐025).

Sampling bags (KBP‐50C‐2, SB‐KAWASUMI Laboratories, Inc., Kawasaki, Japan) were connected to storage bags under sterile conditions and 4 mL (Day 0) or 11 mL (days 2, 5, 10, 14 and 21) was transferred for in vitro testing. Day 2 sampling and testing were performed after the irradiation of PCs in the control group and DCS2. Prior to sampling, cold units were agitated at RT for 10 min. Samples for analyses of DCS2 and DCS5 on Day 2 by PLT aggregation, rotational thromboelastometry (ROTEM), flow cytometry (FCM) and measurements of glucose levels were prepared by mixing equal parts of DCS2 and DCS5. On Day 2, DCS2 and DCS5 were both maintained under identical RT storage conditions and originated from the same manufacturing process prior to the divergence of storage protocols. Therefore, we assumed that their in vitro characteristics at this time point were similar and equal volumes of DCS2 and DCS5 were pooled for the in vitro evaluation.

To assess PLT aggregation and conduct ROTEM and FCM, platelet‐rich plasma (PRP) was prepared by adjusting the PRP concentration to 300,000/μL with autologous plasma that was simultaneously collected with the hyperconcentrated PLTs described above.

PLT counting and pH measurement

The PLT count and mean PLT volume were measured with a haematology analyser (XS‐1000i, Sysmex Corporation, Kobe, Japan). pH was analysed at 22°C on a blood gas analyser (GASTAT 700 series, Techno Medica, Yokohama, Japan).

Volume of PAS‐PCs and residual plasma concentration

The volume of PAS‐PCs was calculated by subtracting the tare weight from the weight of PAS‐PCs and dividing by the specific gravity. The residual plasma concentration was calculated as follows:

$$\text{Plasma concentration}\left(\%\right)=\begin{array}{l}100\times (1-[\text{the volume of added}\mathrm{T}-\mathrm{PAS}+]\\ /[\text{the volume of}\mathrm{PAS}-\mathrm{PC}]).\end{array}$$

Measurement of glucose

The supernatant of PAS‐PCs was collected following two rounds of centrifugation (1900×g, RT, 10 min) and was frozen at −80°C for later analyses. Glucose concentrations were measured using a commercial kit for glucose (Glucose CII Test Wako, FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan) with a detection limit of 3.8 mg/dL.

FCM

To measure the expression of CD62P and CD63 on PLTs and the binding of PAC‐1 to PLTs, we diluted PRP with HEPES Tyrode's buffer (KCl, 2.7 mM; NaCl, 127 mM; NaH₂PO₄, 0.5 mM; glucose, 5.6 mM; NaHCO₃, 12 mM; HEPES, 5 mM; 0.35% bovine serum albumin; MgSO₄, 0.5 mM; pH 7.3) to a concentration of 25,000/μL. Twenty microlitres of the diluted PRP suspension was mixed with a peridinin chlorophyll protein (PerCP)‐conjugated anti‐CD61 monoclonal antibody (MoAb), phycoerythrin (PE)‐conjugated anti‐CD62P MoAb, fluorescein isothiocyanate (FITC)‐conjugated PAC‐1 (all obtained from BD Biosciences, San Jose, CA, USA) and Pacific Blue‐conjugated anti‐CD63 MoAb (Biolegend, San Diego, CA, USA). PerCP‐conjugated anti‐mouse immunoglobulin G (IgG), PE‐conjugated anti‐mouse IgG, FITC‐conjugated anti‐mouse IgM (all obtained from BD Biosciences) and Pacific Blue‐conjugated anti‐mouse IgG (Biolegend) were used as negative controls. The reaction mixture was then incubated at RT for 20 min in the dark.

After the incubation, PLTs were fixed with cold 1% paraformaldehyde and stored at 4°C for at least 1 h, and then washed with phosphate‐buffered saline (PBS). PLTs were then resuspended in 500 μL of PBS.

To measure phosphatidylserine (PS) exposure, we diluted PRP (300,000/μL) with PBS to a concentration of 25,000/μL. Four microlitres of the diluted PRP suspension was incubated with FITC‐conjugated Annexin V (Biolegend, CA, USA) and PE‐conjugated anti‐CD42a (BD Biosciences, NJ, USA) in Annexin V binding buffer (Biolegend, CA, USA) at RT for 15 min in the dark. As a negative control, the diluted PRP suspension was stained in PBS. The mixture was incubated at RT for 15 min in the dark. After the incubation, PLTs were resuspended in 500 μL of Annexin V binding buffer or PBS, respectively.

Samples were analysed by FCM (CytoFLEX, Beckman Coulter, CA, USA). The PLT population was identified by forward and side scatter characteristics and CD61‐ or CD42a‐positive events. Fluorescence data from 10,000 PLT events were collected to assess the ratio of CD62P‐, CD63‐, PS‐ or PAC‐1‐positive PLTs.

Measurement of PLT aggregation

In vitro PLT aggregation was evaluated using light transmission aggregometry (PRP313M, Taiyo Instruments, Osaka, Japan). A total of 180 μL of PRP samples was pre‐incubated with CaCl₂ (4 mmol/L) at 37°C for 1 min on the aggregometer. The following agonists were added to PRP samples and incubated at 37°C for 7 min to measure PLT aggregation: 1.5 mg/mL ristocetin (American Biochemical and Pharmaceutical Ltd., Marlton, NJ, USA), or a combination of both 2.5 mg/L collagen (Collagen reagent ‘Horm’, Nycomed Pharma GmbH, Munich, Germany) and 5 μmol/L adenosine diphosphate (ADP) (‘MCM’ ADP; Sakae Corporation, Tokyo, Japan).

Thromboelastometry

Clot quality and strength were assessed using ROTEM (ROTEM delta, Tem International GmbH, Munich, Germany). PRP samples were mixed with EXTEM reagents according to the manufacturer's instructions. The following parameters were reported: coagulation time (s), α‐angle (°) and maximum clot firmness (MCF) (mm).

Statistical analysis

Data were analysed using statistical analysis software (Kaleida Graph, HULINKS, Tokyo, Japan). A one‐way repeated measures analysis of variance (ANOVA) with Dunnett's post hoc test was used to compare the control group, DCS2 and DCS5 at each time point from Day 10 onwards. In addition, for aggregation and MCF, the same statistical approach was applied to compare the control group on Day 14 with DCS2 and DCS5 on Day 10. In all statistical analysis, p < 0.05 was considered to be significant.

RESULTS

Preparation of PAS‐PCs

Table 1 shows the volume, PLT concentration and total PLT count of the three PAS‐PC groups on Day 0. No significant differences were observed in any of the parameters examined between the three groups.

TABLE 1.

Volume, platelet concentration and platelet count of each prepared product.

Parameter	Control group	DCS2	DCS5
Volume (mL)	203.4 ± 3.1	202.9 ± 2.3	203.0 ± 2.8
PLT concentration (×109/L)	1217 ± 56	1224 ± 61	1220 ± 54
PLT count (×1011/bag)	2.475 ± 0.100	2.484 ± 0.117	2.477 ± 0.100

Note: n = 5, mean ± SD.

Abbreviations: ANOVA, analysis of variance; PLT, platelet.

In all prepared PAS‐PCs before pooling (Figure 1), the residual plasma concentration ranged from 34.9% to 35.1%, while the number of contaminating leukocytes ranged from 0 to 0.12 × 10⁶/bag.

PLT concentration, glucose and pH

No significant changes were noted in the PLT concentration until Day 10 in the three groups, and this was followed by a gradual decrease (Table 2). Regarding aggregate formation in PCs during storage, one aggregate (approximately 2–3 mm in diameter) was observed in 1 of the 5 units in the control group on Day 21. In addition, in 1 of the 5 units in DCS5, approximately a dozen aggregates (≤1 mm in diameter) were detected from days 7 to 21. Between days 10 and 21, PLT concentrations were significantly higher in DCS2 and DCS5 than in the control group.

TABLE 2.

Haematological and metabolic analyses of three groups in storage.

Parameter	Day	Control group	DCS2	DCS5
PLT concentration (×109/L)	0	1217 ± 57	1225 ± 61	1221 ± 55
	2	1261 ± 54	1214 ± 58	1214 ± 59
	5	1251 ± 61	1259 ± 53	1212 ± 60
	10	1204 ± 62	1239 ± 63*	1234 ± 62*
	14	1139 ± 58	1205 ± 54*	1219 ± 63*
	21	999 ± 71	1068 ± 45*	1129 ± 68*
pH (22°C)	0	7.30 ± 0.02	7.31 ± 0.01	7.30 ± 0.03
	2	7.30 ± 0.02	7.42 ± 0.03	7.42 ± 0.04
	5	7.27 ± 0.04	7.31 ± 0.05	7.41 ± 0.06
	10	7.15 ± 0.05	7.13 ± 0.08	7.18 ± 0.09
	14	7.02 ± 0.07	6.97 ± 0.10	7.04 ± 0.08
	21	6.76 ± 0.09	6.84 ± 0.07	7.05 ± 0.08*
Glucose (mM)	0	7.2 ± 0.4	7.2 ± 0.4	7.2 ± 0.4
	2	6.4 ± 0.3	6.4 ± 0.3	6.4 ± 0.3
	5	5.5 ± 0.3	5.0 ± 0.4	4.4 ± 0.6
	10	3.8 ± 0.2	2.8 ± 0.6*	1.8 ± 0.8*
	14	2.5 ± 0.3	1.1 ± 0.6*	0.2 a
	21	0.3 ± 0.3 b	ND	ND

Note: n = 5, mean ± SD.

Abbreviations: ANOVA, analysis of variance; ND, not detected; PLT, platelet.

^a n = 2, mean.

^b n = 3, mean ± SD.

^* p < 0.05 (vs. control group) (one‐way repeated measures ANOVA with Dunnett's post hoc test).

pH decreased in a time‐dependent manner in all three groups after Day 5, but remained above 6.2, the regulatory requirement in Japan, during the storage period (Table 2).

In the control group and DCS2, glucose concentrations were detectable until Day 14, while three of the five experiments in DCS5 had no measurable glucose on Day 14 (Table 2). On Day 21, measurable glucose concentrations remained detectable in 3 of the 5 units in the control group, whereas no glucose was detected in DCS2 or DCS5.

PLT activation markers

The results obtained on PLT activation markers are shown in Figure 2. No significant differences were observed among the three groups on Day 10 or Day 14 (Figure 2a). PAC‐1 was lower after Day 10, while CD63 was lower after Day 14 in DCS2 and DCS5 than in the control group (Figure 2b,c). No significant differences were noted in Annexin V among the three groups during the storage period (Figure 2d).

FIGURE 2.

Platelet (PLT) activation marker expression in PAS‐PCs during 21 days of storage. PAS‐PCs were stored cold for 21 days (control group, blue column) or at room temperature for 2 days (DCS2, green column) or 5 days (DCS5, yellow column) before cold storage. The striped column on Day 2 shows the measured value of a sample that was a mixture of equal amounts of DCS2 and DCS5. Positive expression (%) was reported for (a) CD62P, (b) PAC‐1, (c) CD63 and (d) Annexin V. Differences among the control group, DCS2 and DCS5 at each time point from Day 10 were analysed using a one‐way repeated measures analysis of variance (ANOVA) with Dunnett's post hoc test. p‐values <0.05 were considered to be significant (n = 5). *p < 0.05 versus the control group at the same time point.

PLT aggregation

The results obtained on PLT aggregation are shown in Figure 3. Following a co‐stimulation with ADP and collagen, aggregation decreased in a time‐dependent manner in the three groups after Day 10 (Figure 3a). Values were significantly lower in DCS2 on Day 21 and in DCS5 after Day 10 than in the control group. No significant differences were observed in the values between DCS2 and DCS5 on Day 10 and the control group on Day 14 (a paired t‐test) (Figure 3a). The decrease in ristocetin‐induced aggregation was very small until Day 14, and there was no significant difference among the three groups up to Day 14 (Figure 3b).

FIGURE 3.

Platelet aggregation during 21 days of storage. PAS‐PCs were stored cold for 21 days (control group, blue column) or at room temperature for 2 days (DCS2, green column) or 5 days (DCS5, yellow column) before cold storage. The striped column on Day 2 shows the measured value of a sample that was a mixture of equal amounts of DCS2 and DCS5. Platelet aggregation was induced by ADP and collagen (a) or ristocetin (b). Differences among the control, DCS2 and DCS5 groups at each time point from Day 10 onwards, and comparisons between the control group on Day 14 and the DCS2 and DCS5 groups on Day 10, were analysed using a one‐way repeated measures analysis of variance (ANOVA) with Dunnett's post hoc test. p‐values <0.05 were considered to be significant (n = 5). *p < 0.05 versus the control group at the same time point. ^† p < 0.05 versus the control group on Day 14.

ROTEM

The results of the analysis using ROTEM are shown in Figure 4. The clotting time and α‐angle showed no significant change until Day 21, and there was also no significant difference between the three groups at least until Day 14 (Figure 4a,b). On the other hand, MCF decreased in a time‐dependent manner (Figure 4c). Values were lower in DCS2 and DCS5 than in the control group after Day 10. Values in DCS2 on Day 10 were similar to those in the control group on Day 14. While values were significantly lower in DCS5 on Day 10 than in the control group on Day 14, the difference was modest in magnitude (Figure 4c).

FIGURE 4.

The hemostatic capacity of PAS‐PCs during 21 days of storage. PAS‐PCs were either stored cold for 21 days (control group, blue column) or at room temperature for 2 days (DCS2, green column) or 5 days (DCS5, yellow column) before cold storage. The striped column on Day 2 shows the measured value of a sample that was a mixture of equal amounts of DCS2 and DCS5. ROTEM was used to assess (a) the clotting time, (b) α‐angle and (c) maximum clot firmness. Differences among the control, DCS2 and DCS5 groups at each time point from Day 10 onwards, and comparisons between the control group on Day 14 and the DCS2 and DCS5 groups on Day 10, were analysed using a one‐way repeated measures analysis of variance (ANOVA) with Dunnett's post hoc test. p‐values <0.05 were considered to be significant (n = 5). *p < 0.05 versus the control group at the same time point. ^† p < 0.05 versus the control group on Day 14.

DISCUSSION

We here investigated the effects of delayed refrigeration on PLT function in irradiated, apheresis‐derived PAS‐PCs. DCS2 and DCS5 were X‐irradiated prior to refrigeration, whereas the control group was irradiated after refrigeration. Previous studies reported that irradiation did not significantly affect the in vitro PLT quality of PAS‐PCs stored in PAS‐E during RT or cold storage [28, 29]. Therefore, temperature differences at the time of irradiation are unlikely to have had a marked impact on PLT quality.

PLT concentrations did not markedly change in any of the three groups until Day 10, but showed higher values from Day 10 in the groups with later refrigeration (Table 2). The reason for the higher PLT concentration in the delayed cold storage group than in the control group remains unclear. However, the results showing that the PLT concentration in cold‐stored PCs decreased after approximately Day 10 [30] and delayed cold‐stored PCs demonstrated higher values, or a tendency towards higher values, compared with cold‐stored PCs on days 14 and 21, consistent with our results [20, 22]. In contrast, Wood et al. reported no significant difference in PLT counts between cold‐stored and delayed cold‐stored PCs [19]. Although the reason for the discrepancy between the results of Wood et al. and the aforementioned studies, including our own, remains unclear, differences in PLT count measurement methods, including variations in anticoagulant use and the analytical principles of haematology analysers among institutions, may have been contributing factors. Glucose concentrations were detectable in all cases (n = 5) examined up to Day 14 in the control group and DCS2 and up to Day 10 in DCS5. The fastest depletion of glucose concentrations was observed in DCS5; however, pH in DCS5 tended to remain the highest through to Day 21. Increased glucose consumption and the subsequent accumulation of lactate are generally considered to contribute to a pH decline during PC storage. However, in a previous study that compared the in vitro quality of cold‐stored and delayed cold‐stored PAS‐PCs, glucose concentrations declined faster in the latter during storage; nevertheless, pH generally remained higher in delayed cold‐stored PCs through to Day 21 than in cold‐stored PCs, with values remaining above 7.0 [20, 22]. Notably, in PCs subjected to delayed cold storage after 7 days at RT, glucose was no longer detectable by Day 21 [20]. The present results are consistent with and support these findings. On Day 21, glucose was depleted in 2 of the 5 units in the control group and in all units in DCS2 and DCS5; however, certain degrees of PLT aggregation and clot formation were still observed. PLT aggregation and a hypotonic shock response have been shown to occur even after glucose depletion [31]. In that study, as well as in the present study, PRP for in vitro quality testing was prepared by resuspending PLTs in autologous plasma that had been collected and frozen on the day of donation. Therefore, it is possible that glucose present in PRP contributed to the partial restoration of aggregation and the capacity for thrombus formation. Accordingly, although the clinical efficacy of Day 21 PLTs remains unclear, it is conceivable that transfused PLTs may regain some degree of aggregation and thrombus‐forming ability in vivo in the presence of plasma glucose in the recipient.

PLT activation markers did not significantly differ between the control and delayed cold‐stored groups from days 10 to 21, or were maintained at low levels in the delayed cold‐stored groups. The results, showing that CD62P expression and PS exposure in the delayed cold‐stored groups were not significantly different from those in the control group and that PAC‐1 binding was lower in the delayed cold‐stored groups than in the control group, were in accordance with previous findings [19]. To the best of our knowledge, CD63 expression has yet to be compared between delayed cold‐stored groups and a control group. The increased expression of CD62P and CD63 reflect the secretion of α‐granules and dense granules, respectively, in PLTs. PLTs have been shown to release factors that are involved in transfusion reactions upon activation [32, 33, 34]. Therefore, the risk of transfusion‐related adverse events associated with PLT‐derived factors in delayed cold‐stored PCs may be similar or lower compared with cold‐stored PC.

In the United States, if RT‐stored PLTs cannot be used for some reason, refrigerated PLTs with a 14‐day expiration date may be used to treat bleeding [13]. In the present study, desirable levels of PLTs, pH, PLT activation markers and ristocetin‐stimulated aggregation were maintained in DCS2 and DCS5 until at least Day 14. Glucose depletion was not observed in any of the cases examined until Day 14 in DCS2 and Day 10 in DCS5. ADP + collagen aggregation and MCF in DCS2 and DCS5 on Day 10 were not markedly different from those in the control group on Day 14. Therefore, the results of this study indicate that PLTs stored at RT for 2 or 5 days after blood collection and then refrigerated may be stored until Day 10.

In a study that examined the in vitro properties of delayed cold‐stored PCs, Wood et al. [19] demonstrated that PCs refrigerated from the day after blood collection and PCs stored at RT for 4 days after blood collection and then refrigerated maintained similar activation marker levels and aggregation activity (ADP or collagen) up to 21 days after blood collection. Braathen et al. [20] compared PCs refrigerated from the day of blood collection with PCs stored at RT for 7 days after blood collection and then refrigerated until Day 21. They reported that delayed cold storage initiated on Day 7 resulted in poorer outcomes than continuously cold‐stored PCs. Braathen et al. [22] also compared PCs refrigerated from the day after blood collection and PCs stored at RT for 5 days after blood collection and then refrigerated up to 21 days after blood collection. Although delayed cold‐stored PCs exhibited decreased aggregation activity (arachidonic acid stimulation and TRAP stimulation), glucose was not depleted up to 14 days after blood collection, and increased CD62P expression upon a stimulation was maintained, indicating that delayed cold‐stored PCs may be stored for up to 14 days after blood collection. In addition to the present study, the aforementioned studies differed in several methodological aspects. The source of PAS‐PCs was apheresis‐derived PCs in the present study and Braathen et al. [20], whereas others used whole blood–derived PCs [19, 22]. The timing of refrigeration in the control group was on the day of collection in the present study and Braathen et al. [20], whereas others initiated refrigeration on the day after collection [19, 22]. The ratio of plasma to PAS in PAS‐PCs was 35:65 in the present study and 30:70 [19], 37:63 [20] and 42:58 [22] in previous studies. The PLT concentration in the present study was approximately 1220 × 10⁹/L, whereas it was approximately 900 × 10⁹/L [19], 1330 × 10⁹/L [20] and 980 × 10⁹/L [22] in previous studies. Among the three studies, Braathen et al. [20] assessed the viscoelastic properties using intrinsic pathway activation, whereas in the present study we evaluated the coagulation capacity using an extrinsic pathway stimulation. Because of these differences, it is difficult to compare the present results with previous findings. One possible explanation for the differences of the present results, which suggest storage feasibility up to 10 days, and those of Wood et al., which showed no significant differences in quality between cold‐stored and delayed cold‐stored PCs up to 21 days [19], with the findings of Braathen et al. proposing a shelf life of 14 days for delayed cold‐stored PCs [22], may be differences in PLT concentrations. Since high PLT concentrations lead to increased metabolism and a decline in PLT function, reducing the PLT concentration to a value similar to that in previous studies may extend the optimal storage period to 14 or 21 days in DCS2 and DCS5 in this study.

Klompas et al. [35] used 300 units of PLTs stored at RT for 5 days followed by refrigeration until Day 14 for transfusion. They concluded that although the PLT count did not sufficiently increase, emergency use was still possible [35]. Therefore, the conclusion of the present study that PLTs stored at RT for 2 or 5 days and then refrigerated may be stored until Day 10 cannot be denied. Although substantial evidence for the utility of cold‐stored PLTs has been accumulated and the associated challenges have been increasingly clarified, information on the clinical utility of delayed cold‐stored PLTs remains limited [36]. To establish the optimal PLT component source (apheresis‐derived or whole blood–derived), type of additive solution, plasma content and PLT concentration for delayed cold‐stored PLTs, further accumulation of data, including the findings of ongoing clinical trials, is needed.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.

Fuse H, Wakamoto S, Arisawa F, Kanashiki T, Fujihara M, Akino M, et al. Effects of the start date of refrigeration on functions of delayed cold‐stored platelets. Vox Sang. 2026;121:989–998.

DATA AVAILABILITY STATEMENT

The authors confirm that the data supporting the findings of this study are available within the article.