Patients receiving blood aged 35 to 48 days did not have a significantly lower ∆PPV after transfusion of pRBC. However, in severely critically ill patients (i.e., SOFA score ≥ 10) PPV increased less when older versus fresh pRBC was transfused. In this group of patients, age of pRBC may be of importance, whereby severely ill ICU patients with poor MCF may profit from shorter stored blood. Two possible explanations could be 1) the pre-existing impairment of the patient’s own erythrocytes due to critical illness, and 2) impaired erythrocyte-to-capillary interaction . Therefore, administered pRBC containing storage lesions might not enter the already altered microcirculation and might not bring this group of patients the same positive effects as fresh pRBC. These findings jeopardize today’s arbitrary haemoglobin-derived transfusion thresholds where neither MCF impairment or severity of illness nor the age of the transfused blood is taken into account.
Microcirculatory blood vessels (i.e., arterioles, capillaries, and venules) are those vessels with a diameter be < 100 µm. Homogenous MCF is crucial for tissue oxygenation . Thus, understanding the role of MCF and implementation of corresponding measurements and study results in the treatment of patients is of great importance but confirmatory data from prospective trials are lacking. In septic patients, for example, persistent MCF alteration correlates with adverse outcome [11, 20–23], thus, warranting treatment .
For tissue and microcirculatory perfusion, blood flow is of greater importance than blood pressure . In a prospective study of 20 septic shock patients, MCF alterations could not be improved after elevation of mean arterial pressure (MAP) with norepinephrine . In the second consensus on the assessment of sublingual microcirculation , the authors listed heterogeneous blood flow, haemodilution, and stagnant microcirculatory flow due to arterial vasoconstriction or oedema with prolonged oxygen diffusion distances as reasons for discrepant macro- and microcirculatory flow patterns.
Two previous studies examining the relationship between the age of pRBC and the MCF produced divergent findings. Weinberg and colleagues reported decreased perfused capillary density (PCD) after transfusion of older pRBC with consecutive changes in regional microvascular perfusion in a cohort of 93 patients . In contrast, Yürük and colleagues detected no such effects in their 20 patient cohort , thereby, concluding that although the impact of storage lesions on haemorheology is well-known, but its clinical relevance remains unclear.
While the average age of pRBC at the time of transfusion is between 16–21 days [1, 29], many countries allow a storage time for up to 42 days . Of note, a decrease in oxygen-delivering capacity was seen after storage of five to six weeks (35–42 days) .
With a median storage age of 28.5 days, age of pRBC used in our study was even higher than both our hospital average (mean = 22.9 days) and the international average described above. In order not to waste any blood products, it is a common practice to transfuse the oldest available pRBC first with a minimum time to process a unit of pRBC being two days .
Overall, there is a slight tendency to a less pronounced increase of blood flow in the microcirculation after transfusion of older blood. However, our data show no statistical significance in the correlation between the age of pRBC up to a maximum storage of 49 days and ∆MCF values. Unequal distribution of data may explain the difference to the previously mentioned inverse correlation with the SOFA subgroups.
Several authors have reported no difference in MCF in studies comparing transfusion of younger (7–20 days old) to older (21 to 42 days old) pRBC [4, 13–15, 28, 31]. Our own comparisons of fresh, medium, and old blood also did not detect a significant difference between MCF and pRBC age. However, these numbers do not respect severity of illness.
A more clinical approach was used in several large randomized controlled trials. Patient outcome, such as mortality, was investigated in the ABLE , RECESS  or TRANSFUSE  trials. Results of these studies as well as of other trials were summarized in a Cochrane analysis in 2018 , and revealed no clear difference in the risk of death after transfusion of blood closer to the expiration date in adults. Based on these findings, transfusion of relatively old blood was considered safe . Nevertheless, these studies also failed to consider the severity of illness. Moreover, mass transfusions were not examined. Studies on the infusion of large volumes of pRBC are thus necessary to shed light on the relevance of storage lesions .
Mismatch in sex of donor and recipient of pRBC is described in the literature as a possible factor influencing the outcome of transfusion . We could not detect a benefit in matching pRBC to the recipient’s sex in terms of MCF and also did not find a negative impact of mismatch on flow in capillaries. In addition, no significant differences of donor’s sex and change in MCF could be detected, possibly also due to the small sample size of our study.
Finally, so-called storage lesions due to prolonged storage are well described and understood in in-vitro as well as in in-vivo animal models, but the clinical significance remains unclear [7, 28, 38, 39]. Therefore, the findings of this investigation can only be seen as a jigsaw piece in the field of microcirculation studies with considerable uncertainty. The growing importance of individualized patient treatment in medicine supports bedside measurement of perfusion within the smallest vessels in real-time (MCF), thus serving as an indirect predictor of erythrocyte flow and oxygen delivery to the organs [26, 40].
Our study has several limitations. First, as a retrospective analysis, collection of data about transfused pRBCs was difficult due to missing values in patient charts. Second, the sample size was not powered for the presented research question, and data were not distributed equally due to the standard practice of blood banks to transfuse older pRBCs first. This explains the underrepresentation of fresh blood in this trial. Third, variability of measurements was high. Here, an automated software analysis could help in the future. In addition, distinctions between different groups may be larger with transfusion of fresh blood being even younger than eight days.