Manufacturing stresses do not differentially impact red blood cells from donors with diabetes

Abstract

Background and Objectives

With the rising prevalence of diabetes and expanded blood donor criteria in Canada, individuals with diabetes are increasingly contributing to the blood supply. However, little is known about how routine manufacturing affects red blood cells (RBCs) from this group. This study examined RBC differences in donors with type 1 (T1D) or type 2 diabetes (T2D) following processing to generate red cell concentrates (RCCs).

Materials and Methods

Whole blood (WB) donations were collected from voluntary T1D (n = 12), T2D (n = 11) and non‐diabetic age/sex‐matched (n = 23) donors. Donations were processed via red cell filtration to generate RCCs. At donation, 2.7‐mL of WB was collected into EDTA tubes, and 70 mL of processed RCCs was aliquoted into satellite bags. WB‐EDTA tubes and RCC satellite bags were characterized on Day 2 post collection.

Results

Donors with T1D and T2D had similar, but higher glycated haemoglobin (HbA1c) levels than matched controls (p < 0.001). Processing increased RBC count, haemoglobin and haematocrit in all groups (p < 0.0001). Donors with T2D had decreased mean corpuscular haemoglobin (MCH) and mean corpuscular haemoglobin concentration (MCHC) compared to controls, both pre and post processing (p < 0.05), with a similar trend in p50 (pre: p < 0.01; post: p < 0.05).

Conclusion

Blood component manufacturing did not exacerbate stress on RBCs from donors with diabetes. Donors with T2D had altered MCH, MCHC and p50 compared to matched controls, which persisted after processing. These findings emphasize the importance of donor health on blood product quality.

Highlights.

• Blood manufacturing does not differentially impact red blood cells (RBCs) from donors with type 1 diabetes or type 2 diabetes (T2D).

• RBCs from donors with T2D show persistent haemoglobin content and oxygen affinity alterations, independent of processing.

• The changes to RBC characteristics from donors with T2D likely stem from metabolic alterations specific to the condition and are independent of glycaemic control.

INTRODUCTION

Blood products are in constant demand worldwide for therapeutic use in both chronic and acute contexts. With over 85 million red blood cell (RBC) transfusions administered annually, ensuring product quality and safety is vital [1]. As such, stringent policies and regulations govern blood donor screening and eligibility to help ensure both the quality and safety of an RBC product. Despite these regulations, RBC units exhibit variable characteristics, including extracellular vesicle formation, altered haematological indices and oxygen affinity [2, 3, 4, 5]. These differences can be attributed to factors such as manufacturing techniques, storage conditions and donor factors [2, 3, 4, 5].

Diabetes, characterized by chronically elevated blood sugar, is a growing health concern, which was estimated to affect over 589 million adults, aged 20–79, in 2024 and is projected to increase by 45% by 2050 [6, 7]. Recently, blood donor eligibility criteria in Canada have expanded to allow most individuals living with diabetes, either type 1 (T1D) or type 2 (T2D), to donate whole blood (WB) [8]. As a result, understanding how these health conditions affect blood product quality and variability has become increasingly important. This work is also increasingly relevant, as the European Directorate for the Quality of Medicines and Health Care just published the 22nd edition of the Blood Guide, which revises the stringent 2004 European Directive and recommends that insulin use should no longer be a criterion for deferral [9]. Under these updated guidelines, it is recommended that individuals with diabetes are eligible to donate as long as their condition is well controlled and they have no complications [9]. However, specific policies vary across countries, and many countries outside of Europe defer donors taking insulin [10]. The heterogeneity in donor eligibility criteria demonstrates a need for evidence‐based guidelines [10].

T1D and T2D have distinct pathophysiological origins, which may further impact their RBC properties. T1D is an autoimmune disorder resulting in insulin deficiency and immune‐mediated β‐cell destruction [6, 7, 11, 12]. The β‐cell destruction is driven by autoreactive T cells, which can lead to localized inflammation (insulitis) [12]. In contrast, T2D develops later in life as a result of insulin resistance and can often go undiagnosed for an extended period [6, 7, 13]. T2D is commonly associated with chronic low‐grade inflammation, which is primarily driven by tissue‐derived cytokines (e.g., IL‐6, TNF‐α) [12]. This prolonged exposure to systemic inflammation may exacerbate RBC dysfunction, possibly contributing to more significant alterations to RBC characteristics [12].

The extent to which these changes persist or evolve once RBCs are processed and stored as red cell concentrates (RCCs) for transfusion medicine remains poorly understood. Reports from Turpin et al. [14] and Ebenuwa et al. [15] have suggested that RBCs from individuals with diabetes can exhibit increased oxidative stress, increased glycation and poor deformability, all of which may compromise their resilience to processing and storage. However, in the Turpin et al. [14] paper, the presence of comorbidities in donors with T2D was not reported. In practice, individuals with diabetes must meet additional eligibility criteria to donate blood, which may result in a healthier subset of donors in whom these cellular abnormalities may be less pronounced. Increased oxidative stress and inflammation are known to contribute to microvascular damage, endothelial dysfunction and impaired oxygen delivery in diabetes [16]; however, it remains unclear whether these effects are transferred to transfusion recipients via blood products from donors with diabetes. Furthermore, increased glycated haemoglobin (HbA1c), a widely used marker of average glycaemic control over a 2–3 month period, has been associated with increased membrane peroxidation, altered oxygen affinity and reduced deformability in RBCs [17, 18, 19, 20].

In this study, we investigated the impact of diabetes (T1D and T2D) on RBC characteristics shortly after manufacturing, as an early‐phase investigation to determine whether RBCs from donors with diabetes are differentially affected by manufacturing stressors. Based on the current understanding of RBCs from donors with diabetes, we hypothesized that RBCs from donors with diabetes, namely T2D, will exhibit altered haematological indices, oxygen affinity and deformability compared to non‐diabetic donors. This study will provide insight into how donor health influences blood product quality by analysing parameters such as RBC indices, oxygen affinity and rheological properties before and after processing. With the increasing prevalence of diabetes worldwide, characterizing these effects are essential to ensuring blood product quality and safety and to help inform other blood services worldwide about the implications of expanding donor eligibility criteria.

MATERIALS AND METHODS

Blood collection and processing

This study was approved by the Canadian Blood Services Research Ethics Board (Protocol 2022.016). WB units were collected from voluntary donors without self‐reported diabetes (sex/age‐matched controls; n = 23) and with self‐reported T1D (n = 12) or T2D (n = 11) at the Canadian Blood Services Blood4Research facility (Vancouver, BC, Canada) into a quadruple blood bag system (MacoPharma, LQT710X) containing citrate phosphate dextrose (CPD).

To assess the effects of processing, 2.7 mL of WB was collected into EDTA tubes (BD Biosciences, 367856) for characterization. The WB units were then processed using a semi‐automated, modified buffy coat method to yield single‐donor platelet concentrate, RCC and fresh frozen plasma. Briefly, following centrifugation, red cells were collected into PVC‐DEHP (MacoPharma, LQT710X) bags containing the additive solution saline‐adenine‐glucose‐mannitol (SAGM), and then leukoreduced. After processing, 70 mL of RCCs was aliquoted into PVC‐DEHP Compoflex satellite bags (Fresenius Kabi, R6R2001). WB‐EDTA tubes and RCC satellite bags were transported to an external Canadian Blood Services Lab (Edmonton, AB, Canada) at 2–6°C and characterized on Day 2 post collection and processing.

RBC characterization

HbA1c and lipid profiles (total cholesterol, high‐density lipoprotein [HDL] and triglycerides) were measured in EDTA WB using the cobas b101 analyser (Roche Diagnostics, Mannheim, Germany) at the time of donation. A haematology analyser (DxH 520, Beckman Coulter Ireland, Co. Clare, IE) was used to assess the RBC count, haemoglobin (HGB) levels, haematocrit (HCT), mean corpuscular volume (MCV), mean corpuscular haemoglobin (MCH), mean corpuscular haemoglobin concentration (MCHC) and red cell distribution width (RDW_SD) in WB and RCCs. WB and RCCs were also assessed for oxygen affinity (Hemox‐Analyser Model B, TCS Scientific, New Hope, USA), haemolysis, and deformability and osmoscan by an ektacytometer (LORRCA, RR Mechatronics Manufacturing B.V., The Netherlands) using previously described methods [4, 21].

For deformability, RBCs were diluted 9:1000 in an isotonic polyvinylpyrrolidone (PVP) solution, mixed and subjected to shear stress ranging from 0.95 to 30.0 Pa at 37°C. EI_max and K _EI measurements were obtained by transforming the deformability curves using the Eadie–Hofstee linearization technique [22]. For osmoscan, RBCs were diluted 1:20 in an isotonic PVP solution and subjected to an osmotic gradient ranging from 100 to 600 mOsm/kg at a constant shear stress of 30 Pa. Additionally, all WB samples were centrifuged (2200g, 10 min, 4°C) to increase the HCT of the sample to >50% prior to running on the osmoscan. The following osmoscan indices were extrapolated by plotting the elongation index (EI) of RBCs against an osmolality gradient: EI_max (maximum deformability), O _min (osmolality at which cells have minimum deformability), O _EImax (osmolality at which cells have maximum deformability) and O _hyper (cellular response in hyperosmotic conditions, reflecting hydration status).

Statistical analysis

All statistical analyses were performed using IBM SPSS Statistics 29.0 (IBM Corp, USA). A two‐way ANOVA with post hoc Bonferroni‐adjusted multiple comparisons was used to compare differences between WB and RCCs and between donor groups (T1D, T2D and their respective matched controls).

Exploratory subgroup analyses were performed using the Mann–Whitney U test because of limited sample sizes and non‐normal distributions. These included comparisons between donor groups (T1D, T2D and their respective matched controls) stratified by donor sex, as well as sex‐based comparisons within each group and product type (e.g., female vs. male within RCCs). These subgroup comparisons were conducted regardless of whether the overall group comparisons (ANOVA) were statistically significant. Statistical significance was defined as p < 0.05.

RESULTS

Blood donors with diabetes exhibit higher levels of HbA1c at the time of donation

Blood donors with diabetes (T1D or T2D) had higher HbA1c levels at donation than controls, as expected (Table 1). Donors with T2D also had significantly higher body mass index (BMI) than matched control donors at the time of donation (Table 1; p < 0.01). Interestingly, total cholesterol, HDL and triglycerides were not significantly different between donors with T1D or T2D and their matched controls (Table 1). When examined by sex, male donors with T2D showed particularly elevated HbA1c values (7.00% ± 1.09%) compared to their non‐diabetic counterparts (5.28% ± 0.22%), although not statistically significant (Table S1). Additionally, female donors with T2D exhibited significantly higher BMI (p < 0.05; Table S1). Male T2D donors displayed higher triglyceride levels (4.33 ± 3.34 mmol/L) compared to control males (1.11 ± 0.40 mmol/L), although this difference was not statistically significant (Table S1). Together, these data showed that the subjects in our study generally had the expected characteristics of individuals with T1D and T2D.

TABLE 1.

Demographics and clinical characteristics of matched samples at the time of donation.

	Matched controls	T1D	Matched controls	T2D
Samples collected	12	12	11	11
Age, mean (range)	31 (22, 49)	29 (21, 43)	63 (50, 71)	62 (44, 70)
Females, N (%)	8 (67%)	8 (67%)	7 (64%)	7 (64%)
HbA1c (%)	5.09 (0.25)	6.92 (0.74)**	5.39 (0.25)	6.49 (0.81)*
BMI (kg/m2)	25.5 (3.2)	27.7 (4.3)	24.5 (3.5)	29.1 (3.6)*
Total cholesterol (mmol/L)	4.63 (0.87)	4.06 (0.55)	4.77 (1.07)	4.28 (1.05)
HDL cholesterol (mmol/L)	1.50 (0.45)	1.42 (0.31)	1.67 (0.35)	1.32 (0.41)
Triglycerides (mmol/L)	1.63 (0.94)	1.13 (0.59)	1.37 (0.63)	2.51 (2.39)

Note: Data presented as mean (SD), unless stated otherwise.

Abbreviations: BMI, body mass index; HbA1c, glycated haemoglobin; HDL, high‐density lipoprotein; SD, standard deviation; T1D, type 1 diabetes; T2D, type 2 diabetes.

^* p < 0.01: significantly different compared to matched control donors.

^** p < 0.0001: significantly different compared to matched control donors.

RBC processing increases RBC count, HGB concentration and HCT in all donor types

As expected, RBC count, HGB and HCT were increased after processing of RCCs in all donors (p < 0.0001; Figure 1a–c). No differences in indices were observed between subjects before or after processing. Upon further analysis, the changes in RBC count and HGB due to processing were significantly greater in male donors with T1D than in female donors with T1D (Table S2; p < 0.01). These differences were not observed in the T1D‐matched controls (Table S2). For donors with T2D, the changes in RBC count, HGB and HCT due to processing between males and females were not significant (Table S3).

FIGURE 1.

Red blood cell (RBC) indices before (fresh whole blood; WB) and after (red cell concentrate; RCC, tested at Day 2 of hypothermic storage) processing in control donors and donors with type 1 diabetes (T1D) or type 2 diabetes (T2D). Violin plots represent distributions for (a) RBC count, (b) haemoglobin (HGB) and (c) haematocrit (HCT). Solid lines indicate median and dotted lines indicate interquartile range. Statistical comparisons were made between T1D and matched controls and between T2D and matched controls only. Significance is indicated as *p <0.05; **p <0.01; ***p <0.001; ****p <0.0001.

RBCs from donors with T2D show decreased HGB concentration before and after processing

RBCs from donors with T2D showed lower MCH compared to their matched controls before (WB stage) and after (RCC stage) processing (p < 0.05; Figure 2a). No significant differences in MCHC were observed between T1D donors and their respective controls in either WB or RCC (Figure 2b). For donors with T2D, MCHC significantly decreased (p < 0.05) after processing, and these donors also had significantly lower MCHC than their controls in both WB and RCC (p < 0.05; Figure 2b). Interestingly, MCV did not differ significantly between control and the diabetes groups in either WB or RCC (Figure 2c). However, red cell size variability (RDW_SD) increased across all groups, including T1D and matched controls (p < 0.01) and T2D and matched controls (p < 0.05; Figure 2d).

FIGURE 2.

Red blood cell (RBC) indices before (fresh whole blood; WB) and after (red cell concentrate; RCC, tested at Day 2 of hypothermic storage) processing in control donors and donors with type 1 diabetes (T1D) or type 2 diabetes (T2D). Violin plots represent distributions for (a) mean corpuscular haemoglobin (MCH), (b) mean corpuscular haemoglobin concentration (MCHC), (c) mean corpuscular volume (MCV) and (d) red cell distribution width (RDW_SD). Solid lines indicate median, and dotted lines indicate interquartile range. Statistical comparisons were made between T1D and matched controls and between T2D and matched controls only. Significance is indicated as *p <0.05; **p <0.01; ***p <0.001; ****p <0.0001.

Although there were no differences in MCHC in donors with T1D, further analysis revealed that the post‐processing decrease in MCHC in female donors with T1D was significantly greater than that of males (p < 0.01; Table S2). For donors with T2D and their matched controls, only female donors showed decreased MCHC after processing (p < 0.05; Table S3). Additionally, RDW_SD was significantly increased after processing in only female donors with T2D (p < 0.05; Table S3).

T2D RBCs exhibit decreased oxygen affinity following processing

Oxygen affinity (p50) significantly decreased in all donor groups after processing (p < 0.0001; Figure 3a). Donors with T2D showed significantly increased p50 values (decreased oxygen affinity) in both WB and RCC than controls (WB: p < 0.01; RCC: p < 0.05; Figure 3a). Furthermore, no significant differences in deformability (EI_max) and rigidity were observed between controls, T1D and T2D at either the WB or RCC stage (Figure 3b,c).

FIGURE 3.

Assessment of red blood cell (RBC) properties before (fresh whole blood; WB) and after processing (red cell concentrate; RCC, tested at Day 2 of hypothermic storage) in control donors and donors with type 1 diabetes (T1D) or type 2 diabetes (T2D). Violin plots show distributions for (a) oxygen affinity, (b) deformability and (c) rigidity. Solid lines indicate median and dotted lines indicate interquartile range. Statistical comparisons were made between T1D and matched controls and between T2D and matched controls only. Significance is indicated as *p <0.05; **p <0.01; ***p <0.001; ****p <0.0001.

RBC processing increased O _hyper in all donor groups

Osmoscan parameters, including EI_max, O _min and O _EImax, showed no significant differences between the control, T1D and T2D groups in WB or after RBC processing (Figure 4a–c). All donor groups showed a significant increase in O _hyper after RBC processing (p < 0.05); however, no significant differences were observed between the controls and the diabetic groups (Figure 4d). When broken down by sex, only female donors with T1D showed significant increases in O _hyper after processing (p < 0.05; Table S2). Additionally, female donors with T2D and matched control donors also had increased O _hyper after processing (p < 0.05; Table S3). Lastly, to further assess membrane integrity, we also measured haemolysis in all donor groups at both the WB and RCC stages. Haemolysis levels remained low and did not differ significantly between donors with diabetes and their matched controls after RCC manufacturing (Table S4).

FIGURE 4.

Assessment of red blood cell (RBC) properties before (fresh whole blood; WB) and after processing (red cell concentrate; RCC, tested at Day 2 of hypothermic storage) in control donors and donors with type 1 diabetes (T1D) or type 2 diabetes (T2D). Violin plots show distributions for osmoscan parameters: (a) EI_max, (b) O _min, (c) O _EImax and (d) O _hyper. Solid lines indicate median, and dotted lines indicate interquartile range. Statistical comparisons were made between T1D and matched controls and between T2D and matched controls only. Significance is indicated as *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001.

DISCUSSION

In this study, we investigated the response of RBCs from donors with T1D or T2D or without diabetes to RCC blood component manufacturing. We confirmed that processing universally alters RBC indices, oxygen affinity and membrane properties, but that this is not different for RBCs from donors with diabetes. While all groups showed expected manufacturing‐induced changes to RBC indices, oxygen affinity and membrane properties, only donors with T2D showed modest but persistent differences in RBC characteristics, suggesting intrinsic alterations beyond glycaemic control alone.

One major finding was that donors with T2D showed consistently lower MCH and MCHC compared to matched control donors before and after processing. This may reflect altered iron metabolism due to chronic inflammation in T2D [23, 24]. In T2D, inflammatory cytokines such as IL‐6 act to stimulate hepcidin production and, in turn, limit iron mobilization and suppress HGB synthesis [24, 25]. Long‐term exposure to elevated glucose levels may also further compromise HGB stability, making it more susceptible to degradation [26]. However, given that donors with T1D and T2D had comparable HbA1c levels, the observed changes in RBC characteristics likely reflect metabolic alterations specific to T2D, such as increased BMI and variability in triglyceride levels. Additionally, metformin, a common medication used to treat T2D, was also found to reduce MCHC significantly and has been independently associated with anaemia [27]. Our donor cohort showed similar results, with significant differences observed between blood donors with T2D taking metformin (n = 8) and control donors, but not donors with T2D not taking metformin (n = 3) (Figure S5). Together, these findings suggest that both metabolic changes and the effects of common medications such as metformin contribute to the altered RBC indices observed in this donor population. However, further work to examine this is required.

Another key finding was that donors with T2D showed elevated p50 values, indicating decreased RBC oxygen affinity. A reduction in oxygen affinity has previously been thought to result from a compensatory response to tissue hypoxia or anaemia [28, 29]. While tissue hypoxia can occur in both T1D and T2D, the observed increase in p50 was limited to donors with T2D. In T1D, tissue hypoxia is more tightly linked to chronic hyperglycaemia [30]. Thus, the absence of increased p50 in T1D, despite similar levels of HbA1c, further supports the possibility that non‐glycaemic factors specific to T2D may be underlying these changes to p50. In T2D, hypoxia may be linked to glycaemia and metabolic factors such as dyslipidaemia [30, 31]. Although triglyceride variability was noted in T2D donors, no definitive associations could be established in this study. Additionally, an early study has suggested that 2,3‐diphosphoglycerate levels in diabetic individuals may rise in association with vascular complications, potentially reflecting altered RBC metabolism in T2D [29]. These observations reinforce that T2D introduces additional metabolic stresses that are sufficient to alter RBC properties beyond the effects of chronic hyperglycaemia.

Interestingly, RBC deformability did not differ between donors with and without diabetes. Previous studies have frequently reported reduced RBC deformability in individuals with diabetes, often attributed to chronic hyperglycaemia, oxidative stress and glycation‐induced alterations in membrane structure and function [32, 33, 34]. However, a key difference in our study may be the health status of the blood donor population. As blood donors are subject to additional eligibility criteria, it likely excludes most individuals with very poorly controlled diabetes or significant co‐morbidities such as cardiovascular disease, both of which have been linked to impaired RBC rheology [14, 15, 16]. The absence of differences in RBC deformability may reflect the selective health criteria applied to the blood donor population.

While RBC deformability remained unchanged between the groups, processing steps appeared to influence other aspects of RBC properties, such as O _hyper, a marker of cellular hydration [35, 36]. The decreased O _hyper in WB relative to RCC may reflect RBC swelling in response to prolonged EDTA exposure resulting from Na+ influx and K+ efflux [37]. In contrast, CPD anti‐coagulant and SAGM additive solution in an RCC unit may provide a more osmotically stable environment for RBCs [38]. Although a previous study found no significant changes in O _hyper in EDTA WB stored at 4°C for up to 80 h (conditions matching those in our study) [37], a direct comparison of the CPD WB to the manufactured RCC could have eliminated the potential impact of EDTA exposure. Such a comparison may have offered additional insights into the impacts of blood processing, but was not feasible within the current study design. Alternatively, increases in O _hyper may be due to leukoreduction, a standard step in blood processing. Leukoreduction has been shown to help preserve RBC membrane integrity, reduce oxidative damage and slow the development of storage lesions, all of which may help maintain or even improve cellular hydration post processing [39, 40]. In addition, haemolysis levels were comparable across all donor groups, further supporting that RBC membrane integrity was maintained during processing regardless of diabetic status (Table S4). Thus, changes to RBC deformability and integrity after processing are likely influenced by manufacturing procedures such as leukoreduction, rather than underlying donor characteristics.

Building on these observations, an exploratory analysis was conducted to examine whether donor sex further modified RBC responses to processing [5, 41]. Female donors with T1D exhibited exaggerated processing‐induced changes to RBC count, HCT and MCHC compared to males, suggesting possible sex‐based differences. On the other hand, only female donors with T2D showed significant changes to MCHC, RDW and p50 after processing. However, the sample size in all donor groups strongly limits the sex‐based analysis. Despite this limitation, these findings highlight the compounding effects of donor sex and health on RBC characteristics.

Several other limitations should be considered for this study. First, the sample size may have had limited power to identify more subtle interactions between diabetes subtype and sex. Second, this study assessed only WB and RCC products on Day 2 of hypothermic storage. Monitoring these parameters over a 42‐day storage period is of interest, as a longitudinal study could allow for the detection of post‐processing effects that become more pronounced later during storage. A parallel study is currently under way to examine how donor diabetes affects RBC function and RCC quality over the routine 42‐day storage period. This data will complement the findings from this early‐phase investigation by providing further insight into storage‐related changes. Additionally, RCC products were transported in satellite bags, introducing additional variables related to storage conditions, which may have influenced the measured RBC characteristics. Thus, our approach does not fully replicate standard blood bank storage practices. Together, these considerations limit the generalizability of our findings and demonstrate the need for more follow‐up studies on RBCs from donors with T1D or T2D.

In summary, our study suggests that RCC manufacturing does not differentially affect blood products from donors with T1D and T2D. Additionally, donors with T2D show intrinsic differences in MCH, MCHC and p50, which persisted through manufacturing. While these changes likely reflect T2D‐specific metabolic changes and immunological factors, their clinical significance for transfusion medicine remains uncertain. Furthermore, the observed sex‐specific responses in donors with T1D and T2D warrant further investigation, because of the limited sample size. As the prevalence of diabetes continues to rise worldwide [6] and eligibility criteria are evolving across countries [9], future studies should assess storage‐related changes and the possible clinical significance of these findings.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.

Supporting information

Table S1. Demographics and clinical characteristics of matched samples at the time of donation between males and females.

Table S2. Red blood cell characteristics between males (n = 4) and females (n = 8) in donors with type 1 diabetes.

Table S3. Red blood cell characteristics between males (n = 4) and females (n = 7) in donors with type 2 diabetes.

Table S4. Haemolysis levels between donors with diabetes and their matched controls.

Figure S1. Red blood cell (RBC) indices before (fresh whole blood; WB) and after (red cell concentrate; RCC, tested at Day 2 of hypothermic storage) processing in control donors and donors with type 1 diabetes (T1D) or type 2 diabetes (T2D), stratified by sex. Violin plots represent distributions for (A) RBC count, (B) haemoglobin (HGB) and (C) haematocrit (HCT). Solid lines indicate median and interquartile range. Circles represent male donors, and diamonds represent female donors. Statistical comparisons were made between T1D and matched controls and between T2D and matched controls only. Significance is indicated as *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001.

Figure S2. Red blood cell (RBC) indices before (fresh whole blood; WB) and after (red cell concentrate; RCC, tested at Day 2 of hypothermic storage) processing in control donors and donors with type 1 diabetes (T1D) or type 2 diabetes (T2D), stratified by sex. Violin plots represent distributions for (A) mean corpuscular haemoglobin (MCH), (B) mean corpuscular haemoglobin concentration (MCHC), (C) mean corpuscular volume (MCV) and (D) red cell distribution width (RDW_SD). Solid lines indicate median and interquartile range. Circles represent male donors, and diamonds represent female donors. Statistical comparisons were made between T1D and matched controls and between T2D and matched controls only. Significance is indicated as *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001.

Figure S3. Assessment of red blood cell (RBC) properties before (fresh whole blood; WB) and after processing (red cell concentrate; RCC, tested at Day 2 of hypothermic storage) in control donors and donors with type 1 diabetes (T1D) or type 2 diabetes (T2D), stratified by sex. Violin plots show distributions for (A) oxygen affinity, (B) deformability and (C) rigidity. Solid lines indicate median and interquartile range. Circles represent male donors, and diamonds represent female donors. Statistical comparisons were made between T1D and matched controls and between T2D and matched controls only. Significance is indicated as *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001.

Figure S4. Assessment of red blood cell (RBC) properties before (fresh whole blood; WB) and after processing (red cell concentrate; RCC, tested at Day 2 of hypothermic storage) in control donors and donors with type 1 diabetes (T1D) or type 2 diabetes (T2D), stratified by sex. Violin plots show distributions for osmoscan parameters: (A) EI_max, (B) O _min, (C) O _EImax and (D) O _hyper. Solid lines indicate median and interquartile range. Circles represent male donors, and diamonds represent female donors. Statistical comparisons were made between T1D and matched controls and between T2D and matched controls only. Significance is indicated as *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001.

Figure S5. Mean corpuscular haemoglobin concentration (MCHC) in red blood cells from control donors and donors with type 2 diabetes (T2D), stratified by metformin use. Measurements were taken before (fresh whole blood; WB) and after processing (red cell concentrate; RCC, tested at Day 2 of hypothermic storage). Circles represent T2D donors using metformin (n = 8), diamonds represent T2D donors not on metformin (n = 3) and triangles represent matched controls (n = 11). Solid lines indicate median and interquartile range. Significance is indicated as *p < 0.05.

Phan C, Rowe EM, Yazdanbakhsh M, Kurach J, Brandon‐Coatham M, Devine DV, et al. Manufacturing stresses do not differentially impact red blood cells from donors with diabetes. Vox Sang. 2025;120:1093–1102.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.