Elderly blood donors
Zeiler T, Lander-Kox J, Eichler H et al.
The safety of blood donation by elderly blood donors
Vox Sang 2011; ahead of print (doi: 10.1111/j.1423-0410.2011.01492.x).
Current age limits for blood donation are more based on tradition than solid evidence. A minimum age around 18 years can be justified, perhaps, by the need to obtain valid consent but a maximum age around 65 years is based only on a presupposition of increased frailty or on the fear that the consequences of eventual side effects could be worse in elderly people. However, some studies showed that moderate and severe reactions are actually less frequent in elderly donors. The present article provides further interesting data on this subject.
In Germany, the maximum age for blood donors was raised to 68 years in 1996 and, by 2005, even older donors were allowed to donate, if the physician responsible approved. The Authors of this study from the German Red Cross Blood Service West prospectively collected information on donor deferrals and reactions from September 2006 to December 2009. Reactions occurring on-site were documented on a standard form. Those occurring off-site (within 48 hours of donation) were collected by means of a questionnaire, which was given to the donor together with a pre-paid envelope. Reactions and injuries were classified as mild, moderate or severe according to standard criteria. Reactions requiring medical treatment were considered severe. Only repeat donors of whole blood were included. The volume collected was 500 mL plus 40 mL for laboratory tests. The study compared two elderly groups (66 to 68 years and 69 to 71 years) with a control group of donors aged 50 to 52 years. The age of the subjects in the control group was chosen because donors in that age group had a similar donation frequency and were assumed to be equally experienced as the those in the older groups.
A total of 64,260 donations were studied, more than 20,000 in each of the three groups. The male/female (M/F) ratio increased with age, from 2.1 to 3.6. Reactions were more frequent in females in all groups: 50–52 years, F: 1.90%, M: 0.38%; 66–68 years, F: 1.71%, M: 0.33%; 69–71 years, F: 0.83%, M: 0.23%. If only severe reactions were considered, the rates were as follows: 50–52 years, F: 0.09%, M: 0.02%; 66–68 years, F: 0.06%, M: 0.02%; 69–71 years, F: 0.10%, M: 0.03%. Mild reactions decreased with age, while severe reactions occurred at approximately the same frequency. Most (93%) of the reactions occurred off-site and were mainly mild haematomas and fatigue. On-site reaction rates (%) in the three groups were 0.10, 0.04, and 0.14, respectively, in females, and 0.03, 0.02, and 0.02, respectively, in males.
In contrast, deferrals increased greatly with age in both male and female donors: 50–52 years, F: 1.81%, M: 1.12%; 66–68 years, F: 3.17%, M: 2.18%; 69–71 years, F: 8.74%, M: 6.56% (p<0.001). In the control group, deferrals were mainly (52%) due to a low haemoglobin concentration. This was also the reason for 52% of the deferrals in the group 66–58 years old, but for only 35% in the group 69–71 years old. Blood pressure problems and ignoring extant deferrals increased in the oldest group. Systolic and diastolic blood pressure increased with age, while venous haemoglobin concentration did not change appreciably. Body weight decreased significantly in male donors only (from 89.9 kg to 84.9 kg).
All three age groups had a very high annual donation frequency (from 2.9 to 3.5). Accordingly, the number of previous donations was also high: 50–52 years: 44±30; 69–71 years: 74±36. The oldest group contributed 3.6% of the annual blood donations.
The Authors commented that the higher deferral rates may have influenced the donor reaction rate significantly, as donors at risk may not have been allowed to donate. However, this selection was performed entirely by the physicians of their Institution, without recourse to the family physician, whose involvement had been advocated by others, but not without problems. The current policy at the German Red Cross Blood Service West is to actively invite repeat donors up to their 72nd birthday. Older repeat donors are also accepted, if approved by the attending physician. The Authors also commented that this change of policy was very favourably received, as expected by such a committed population of donors.
Elderly repeat donors constitute an experienced group, whose members have been through a sort of "natural selection" which conceivably selected for those healthier and less reaction-prone. The Authors also provided, in the Discussion, some data about first time donors, who were outside the scope of the study: the reaction rates were 3.64% in donors aged 18–20 years, 1.69% in those between 41 and 50 years, 1.78% in donors between 51 and 60 years, and 2.25% in the donors over 60 years old.
The results of this study concur with those previously reported on the safety of accepting elderly repeat donors, who may also be safer donors as regards transmissible diseases. In our opinion, maintaining active the elderly repeat donors represents the most practical and immediate response to the feared contraction of the donor pool, caused by the ageing of the population. This study demonstrates that the donor selection process, as currently applied, may reliably detect and defer those subjects who might develop significant age-related problems.
Rh incompatible platelet concentrates
Cid J, Carbassé G, Pereira A et al.
Platelet transfusions from D+ donors to D− patients: a 10-year follow-up study of 1014 patients
Transfusion 2011;51:1163–9.
Rhesus (Rh) antigens are not expressed on platelets but D+ platelet concentrates may immunise D− recipients because of the presence of contaminating red cells. The actual immunisation rate varies in the literature from 0 to 19%. The present study includes the largest retrospective case series published so far on this topic and examines several factors potentially affecting the immunisation rate.
The Authors reviewed clinical and transfusion records of all D− patients transfused at their institution with D+ platelet concentrates in the period 1999–2009. Patients were included if they had not previously been exposed to D+ blood components in the same institution and anti-D had not been detected before the first transfusion taken into account.
Patients were broadly divided into immunosuppressed and immunocompetent, according to diagnosis. The first group was further subdivided into those with a haematological disease or a solid neoplasm. The clinical diagnosis of "immunocompetent" patients was not specified. Rh immunoglobulin was not transfused to avoid immunisation.
A total of 1,014 D− patients were initially included (615 males, 399 females; 342 with haematological diseases, 147 with solid neoplasms, and 525 immunocompetent subjects). The prevalence of males was plausibly explained by the attempt to reserve D− platelets for women with childbearing potential. The patients received 6,043 platelet concentrates (5,128 D+). Anti-D was detected in 49 (5.7%) of the 867 patients from whom at least one sample was studied after the first D+ transfusion. A further 29 patients (3.3%) developed other specificities. The Authors pointed out that 37 of the 49 patients who developed anti-D, did so within 3 weeks of the first D+ transfusion. They commented that, according to previous experimental data, primary anti-D immunisation should become detectable no earlier than 4 weeks after the D+ transfusion. They, therefore, recalculated the immunisation rate, only taking into account the cases studied at least 4 weeks after the D+ transfusion. There were 12 positive cases (3.8%) out of a total of 315 patients.
I am not altogether convinced that the Authors’ reasoning is sound. Firstly, they referred to literature published 30 to 40 years ago, when less sensitive immunohaematological techniques were probably being used. Secondly, the Authors did not clarify whether the 315 patients were studied both before and after 4 weeks, and whether the 12 positive cases were negative before 4 weeks. Presumably, they were only studied after 4 weeks, not excluding in any case a secondary immunisation. In my opinion, in a retrospective setting it is impossible to distinguish primary and secondary immunisation. In this context, it would have been useful to know how many already immunised patients were excluded from the study: taking into account that anti-D is the specificity less prone to disappearance over time (Blood Transfusion 2008; 6: 225–34), that percentage would represent the upper limit of the secondary immunisation rate. However, the follow-up varied widely from 0 to 718 weeks (median, 29 weeks) and it is also possible that a few cases of primary immunisation were missed.
The Authors examined three factors possibly affecting the immunisation rate: ABO compatibility, the amount of red cells contaminating the platelet concentrates, and immunosuppression. Pooled platelet concentrates were ABO incompatible (major or bidirectional incompatibility) in 12% of cases, apheresis platelets in 20%. It is a well established fact that ABO incompatibility exerts a protective effect on D immunisation in pregnancy, but the Authors did not detect any such effect in the present study. In part, this could be explained by the relative rarity of incompatible transfusions. However, I suspect that another factor may have played a role: red cells are not optimally stored in the storage conditions of platelet concentrates. Although there are no experimental data, as far as I know, it is probable that contaminating red cells would not survive in the recipient’s circulation if transfused after the first days of in vitro storage. This could significantly influence the rate of immunisation, by a similar mechanism to that assumed for ABO incompatibility, i.e. the rapid disappearance of red cells from the circulation. Unfortunately, this aspect has been ignored in the pertinent literature and has not been mentioned in the present study, either.
Citing quality control data from their institution, the Authors calculated that pooled platelet concentrates contained 0.036 mL of red cells, on average, while apheresis platelets contained as little as 0.00043 mL. As the minimum immunising amount has been estimated to be around 0.03 mL, apheresis platelets should not be able to stimulate a primary response. In fact, no patient who exclusively received apheresis platelets developed red cell antibodies. However, there were very few such patients (N=32). The patients who developed anti-D had received seven D+ platelet concentrates (six of which were pooled), on average, while the other patients had received only four (three of which were pooled), but the difference was not statistically significant (p=0.1).
Patients with haematological disease were younger and received more immunosuppressive therapy. The three groups of patients (haematological disease, solid neoplasm, and immunocompetent) developed anti-D in 3.4%, 12.9%, and 1.9% of cases, respectively. The difference is not statistically significant and in any case it does not suggest a lower rate of immunisation in the immunosuppressed patients. However, patients with haematological diseases received significantly more platelet concentrates than the others (p<0.001).
Rh(D) is the most immunogenic red cell antigen (besides A and B) but the very high immunisation rates (around 80%) obtained in deliberate experiments in volunteers are not observed in clinical experience. The Authors conclude that the practice of giving D+ platelet concentrates to D− recipients is safe and that the immunoprophylaxis with Rh immune globulin should be reserved for women of childbearing potential.
Platelet inventory management
de Kort W, Janssen M, Kortbeek N et al.
Platelet pool inventory management: theory meets practice
Transfusion 2011; ahead of print (doi: 10.1111/j.1537-2995.2011.03190.x).
Platelet concentrates have a high outdate rate, ranging from 15% to 20% in large organisations and probably higher in smaller blood banks. This notwithstanding, it is difficult to avoid occasional shortages. Two factors complicate the inventory management of platelet concentrates: the short storage life (most commonly 5 days) and the fact that a single therapeutic unit, unless obtained by apheresis, is derived from five or six whole blood donations. A high outdate rate is, therefore, generally believed an unavoidable consequence of guaranteeing ready availability of concentrates. The economic consequences, in terms of labour and disposables wasted, are obvious.
However, one should not only expect economic improvements from attempts to decrease outdate rates without increasing shortage rates. Platelets deteriorate during storage: function and in vivo recovery decrease and older platelet concentrates also cause more non-haemolytic transfusion reactions (although leucoreduction may have affected the latter aspect). A better alignment between production and consumption could conceivably decrease the average storage age of the platelets at the moment of transfusion, directly improving the quality of the transfusion therapy for the patient.
The present study illustrates the application of operational research to the inventory management of platelet concentrates. Operational research is a mathematical discipline which studies the optimal solutions to complex problems. Typically, the solutions are arrived at using dynamic programming (subdividing the problem into simpler subproblems, under certain conditions) or, when the problem is computationally too heavy, approximated through simulation. Inventory management is a classical field of application of operational research. In the case of platelet concentrates, the problem can be stated in this way: finding the optimal production policy so as to minimise both outdate and shortage rates. As shortages may have not only economic but also clinical consequences, they are assigned a higher "cost" than outdates.
Unfortunately, the inventory management of platelet concentrates suffers from the so-called "curse of dimensionality", i.e. it is too complex, computationally, to be solved directly. Early applications of operational research resorted to computer simulation to overcome this difficulty (Katz AJ et al. Vox Sang 1983; 44: 31–6). In a previous paper (Haijema R et al. Int J Production Economics 2009; 121: 464–73), the Authors of the present article criticised the exclusive use of simulation, as it offered no guarantee of optimality (nor any measure of the difference between the approximation and the optimal solution), and its results were highly dependent on the particular case studied and its assumptions. The present Authors’ approach was to reduce the computational burden by "downsizing" it, i.e. considering multiples of therapeutic units, calculate the optimal policy, and verify through simulation how much a practical, simplified policy approximated the optimal one.
The basic assumptions were as follows:
- the setting is a regional blood bank (Sanquin Blood Bank Southeast Region, the Netherlands), which collects 150,000 whole blood donations and, each year, supplies 11,000 platelet concentrates (98% from pools of five buffy coats) to the hospitals in the area. Local inventories in the hospitals are ignored.
- Production takes place from Monday to Saturday. There are no limits to production. The daily production is decided every morning. Production losses (positive screening tests, breakages, etc.) are ignored.
- Platelet concentrates are available for transfusion on the day after production. Their shelf-life is 7 days.
- Platelet concentrates are issued following the first-in first-out rule. Blood group is ignored.
- Shortage "costs" are set at five times the outdating costs.
The distribution of the daily demand had been calculated from actual data (year 2007), separately for each day of the week. Previous studies had shown that neither the month-to-month nor the week-to-week variability was statistically significant, but that the differences between the days of the week were substantial (Katz AJ et al, cited above).
The Authors concluded that a simple "order-up-to" rule could be devised, which yielded results within 2% of optimality. This rule is a simple prescription that production should be equal to the difference between the current inventory level and that predefined for the day. The predefined level is different for each day of the week. The residual shelf-life of the extant platelet concentrates can be safely ignored.
The Authors compared the theoretical results of this simple policy with the actual 2007 data. Shortages would have amounted to four platelet concentrates in a year (0.04%; there were actually 150 [1.4%]), while 25 pools would have been outdated (0.25%; in reality, there were 157 [1.4%]; another 266 units outdated in the hospitals). It is interesting that the mean storage age of the platelet concentrates would have been 3.75 days, 0.65 days less than that actually recorded.
From August 2007 to August 2008, the Authors developed and introduced into practice a software tool (TIMO: thrombocyte inventory management optimizer) designed to help the daily decision process, performing the necessary calculations and obviating the need, on the part of the blood bank personnel, to understand the underlying theory. Comparing the years from 2005 to 2010 (real data), the outdate rate showed a significant decline in correspondence to the introduction of TIMO (−1.7%, p=0.005). Shortages also decreased in the study period from five occasions in 2007 to two per year in 2008 and 2009 and none in the first half of 2010.
The attentive reader will certainly note that the results were actually very good, compared with data in the literature, even before the introduction of TIMO. This aspect was not discussed by the Authors, even though their previous publications reported baseline data of 15–20% outdate rate and 1% shortage rate (regarding a different Dutch blood bank, however). They also made a passing reference to the implementation of empirical order-up-to levels in 2006 but, according to their analysis, this mainly affected the mean storage age of the transfused platelet concentrates. One is left with the impression that the already excellent baseline rates deserved an explanation.
In conclusion, however, the results of the practical application confirmed the theoretical expectations. These good results are not obvious, as the Authors had introduced in the theoretical model a number of simplifications. For example, the blood group of the platelet concentrates had been ignored. However, the same Authors had previously argued (Haijema R et al, cited above) that their approach would have stood the test in practice, provided a perfect match was not considered necessary. The requirements were that production should be preferentially directed towards the most compatible concentrates and the least compatible ones should be released earlier.
The Authors also examined the possible effect of a 4-day shelf life, as opposed to the 7-day one currently used in their blood bank. They predicted a 0.8% shortage rate together with an outdate rate of more than 5%. They also noted that "economy of scale" (in this setting, a large reference area of the blood bank) greatly helps in reducing both rates. In fact, it is known that the coefficient of variation of the demand decreases as the mean increases. In other words, the greater the demand, the smaller the daily variations (in percentage terms).
An interesting and potentially dangerous effect of the order-up-to policy is that shortages may be caused by panic-driven excess production. The Authors explained that when in a single day the production greatly exceeds the real demand, in the following days no concentrates are produced, in order to comply with the replenishment rule. This causes the progressive ageing of the extant concentrates, which may finally exceed the expiry date at the weekend, when no production is possible. It is evident that, in such a case, the assumption that the storage age of the concentrates in the inventory can be safely ignored is not true.
In another paper (Haijema R et al, cited above), the Authors showed that their approach could also be adapted to cope with special (short) periods of suspended production, such as Christmas and Easter. However, it is not clear what would be the impact on their model of negating the assumption that the daily production level is virtually unlimited. In my blood service, for example, 56% of whole blood donations are collected on Sundays and Fridays and on those days production is limited by the availability of personnel and logistic constraints. On the other days, production is limited by the availability of donations. Moreover, donations follow a seasonal pattern, with the lowest point in summer. I suspect that in those circumstances, a fixed order-up-to rule would be inadequate and the rule should also consider the age of the concentrates in the inventory.
In conclusion, the Authors claim that the common belief that shortages can be avoided only by accepting a substantial rate of outdating is wrong. Furthermore, surprisingly low shortage and outdate rates can be obtained by such a simple and straightforward policy as an order-up-to rule. This is in agreement with results from previously reported, similar, if less sophisticated, approaches.
Fuller AK, Uglik KM, Braine HG, King KE
A comprehensive program to minimize platelet outdating
Transfusion 2011; ahead of print (doi: 10.1111/j.1537-2995.2010.03039.x).
The previous article in this Press Review reported that the majority (63%) of the outdated platelet concentrates expired in the hospitals, not in the blood bank. The approach described in that article was, however, focused exclusively on the blood bank, thereby leaving aside the main source of the problem. The present article usefully complements the previous one, as it addresses the same topic, the inventory management of platelet concentrates, from the standpoint of the hospital.
In the Authors’ institution (the Johns Hopkins Medical Institutions, Baltimore, MD, USA), the number of transfused platelet concentrates increased recently from 12,000 in the year 2000 to 19,000 in the year 2009, but the outdate rate remained uniformly low, varying in the same decade from a minimum of 0.8% to a maximum of 1.6% per year. Such a remarkable result was obtained by means of a particular organisation.
Most (95%) of the platelet concentrates are acquired from the regional Red Cross. They are leucoreduced, apheresis products and the platelet content of each of them is known. Both parts of a split apheresis donation are usually delivered together. At least 80% of the products are HLA-A and -B typed. However, the most particular aspect is the presence of so-called platelet transfusion coordinators. These are a team of physician assistants and medical technologists who play a pivotal role in the management of platelet transfusion. As regards the patient, they meet every day with the patient’s physician to determine the appropriate transfusion threshold and plan the next transfusions. They choose the unit(s) to assign to the patient, on the basis of the platelet content of the unit, the patient’s needs and previous transfusion increments. To this end, they maintain a computerised record for each patient, called the "platelet transfusion history", which contains information on diagnosis, treatment, ABO and HLA type, HLA antibody screen, height, weight, body surface area, cytomegalovirus status, special transfusion requirements, and platelet units transfused (including ABO and HLA type and platelet content). In addition, the coordinators enter the patient’s bleeding status daily.
The software retrieves the daily maximum body temperature from the hospital information system and the blood counts from the laboratory information system and calculates the corrected count increment for each transfusion, using platelet counts before and after transfusion, body surface area, and the platelet content of the transfused unit.
The coordinators evaluate the increments and decide whether the patient requires HLA matched platelets.
In a typical day, an inventory of 100–140 apheresis platelets is available for about 150 thrombocytopaenic oncology or haematology patient. In addition, a separate inventory of 10–12 units serves other clinical indications, from operating rooms, intensive care units and emergency departments. These requests, often being urgent, are not directly managed by the coordinators, but they adjust the number of units available in the subinventory, after reviewing the operating room schedule of the next day.
ABO compatibility is respected. Rh compatibility is respected for women ≤50 years and children ≤18 years, but it is ignored in oncology patients of all ages.
As regards the provision of the platelet concentrates, the coordinators inform the supplier on Wednesday of the number of platelet concentrates to be used in each day of the following week, detailing ABO and HLA type. An 85% compliance for the ABO type is expected.
The Authors reported in detail the reasons for non-utilisation of platelet concentrates in 2009. A total of 158 (0.8%) units were not transfused, but only 32 of them expired because of an excessive inventory. In other cases there were technical reasons (broken bag, etc.) or transfusion was delayed by error or cancelled. When too many platelet units approached the expiry date, the coordinators used a few strategies to avoid wasting them: split apheresis products were used as if they were a single unit. In this way, the patient benefited from a prolonged transfusion interval without increasing donor exposure. Moreover, if the coordinators predicted, on the basis of the clinical situation and the blood counts, that a patient would meet the requirements to be transfused the following day, they would ask permission from the patient’s physician to perform the transfusion in advance, in order not to waste a platelet unit soon to expire. Finally, short-dated products could be sent to other hospitals. No data were reported on how frequently these strategies were used.
As regards the cost, the Authors calculated that, each year, the whole system required 6.5 full-time equivalents (FTE) for the transfusion coordinators, 1.5 FTE for the data managers and 0.5 FTE for the maintenance and occasional programming of the software. However, the Authors noted that if their institution had the average outdate rate of apheresis platelets in the USA (10.9%), the loss would be around $1,000,000 per year.
This article left me with a mixed feeling. Such a complex and expensive organisation would hardly be justified if it had the exclusive aim of avoiding the waste of resources, however precious. Unfortunately the Authors did not report on any quality assessment, not even the average storage age of the transfused units or the corrected count increments which they routinely recorded. Did their system avoid unnecessary transfusions? Were the clinical objectives reached? How do their clinical results compare with those in the literature? After all, a much greater saving would have been obtained using non-HLA-matched platelet pools instead of HLA-typed apheresis platelets for the patients who were not immunised (the vast majority, presumably). Moreover, if patients’ interests are given the maximum importance, the policy of advance transfusion not to waste a short-dated unit seems at least dubious.
The main message of both the previous articles is that the outdate rate of platelet concentrates can be reduced to a minimum without incurring shortages. The approach studied in the first article is essentially based on the fine tuning of production. It does not require expensive investments and it is definitely worth trying. However, its performance is probably strictly dependent on how much two requisites are satisfied in the real situation: flexibility in the daily volume of production and limited variability of demand. The latter requisite is generally dependent on the dimension of the blood bank (the larger, the better).
Solving the problem in the blood bank is, however, insufficient. The second article strongly suggests that, in order to solve the problem in the hospital, it is necessary to develop a better link between the hospital transfusion service and the clinical departments. I doubt that many institutions would consider implementing a sort of intermediate structure such as platelet transfusion coordinators. I also think it quite improbable that clinical departments have or will develop a keen interest in transfusion medicine. In my opinion, it is the transfusion service that should try to fill the gap. But we will not be successful unless we are able to demonstrate that our efforts are directed towards, and result in, better care for patients.