My Opinion: Issues in Transfusion Medicine Software and Component Production

I anticipate that there are several innovations coming or in the process of coming to mainstream blood component production and software.  Some of these I have already addressed in some of my previous posts:

Pathogen inactivation:  We have had this for over a decade.  However, with new emerging pathogens, this will become more important so I expect it will be adopted in many centers where it is not currently being used.  I expect we will close the loop and pathogen-inactivated RBCs will be available so all components will be treated.  Still, the first-generation pathogen-inactivated RBCs may have reduced shelf life compared to regular, untreated units.

Automated component production:  Although this is expensive, it does provide excellent GMP production.  It is fast and may provide higher yields, especially for platelets.  I expect more centers will adopt this technology, especially in combination with pathogen inactivation.

Blood bank computer software:  This software must be considered as dynamically changing, and considerable resources are needed to keep in compliance with ever-changing international regulations and the latest epidemiologic data.  Production rules can be strictly and mercilessly enforced by a dedicated blood bank computer software.  It can also ensure that the final ISBT label is not applied unless all the production rules (registration, collection, processing, and testing) are met.  Manual processing is extremely risky nowadays with all the parameters to be monitored.

Patient Blood Management:  Current blood bank software does not adequately address the need for prospective review of component orders.  I expect that collaboration will occur between laboratory and blood bank software vendors to fill this gap.

Refrigerated platelets:  The pendulum swings back to this component which was used over 40 years ago.  Refrigerated platelets suspended in additive solution may be effective up to 14 days for hemostasis in the trauma setting.  These platelets are activated so standard 20-24C stored platelets may be preferred for prophylactic transfusions.

Low-titer group A universal plasma:  This is already available, but its use will increase because of the low numbers of group AB units available and increased demand.  This includes its production for COVID convalescent plasma.  Your transfusion medical director must decide what “low titer” means.  Also you need a robust way of performing anti-B titers, this may require use of an immunohematology analyzer with titration built-in.

Low titer group O whole blood:  Use of this product may reduce the need for components in massive transfusion settings but it requires performing anti-A and anti-B titers on large numbers of units.  Your transfusion medical director must decide what “low titer” means.  Also you need a robust way of performing anti-A and anti-B titers, this may require use of an immunohematology analyzer with titration built-in.  Also, you must decide whether to leukodeplete the whole blood units:  few whole blood filters are platelet-sparing.

Reveos and Atreus Automated Blood Component Processing: My Experience

This post is about my over 10 years of experience with automated component processing using Terumo equipment, first Atreus and then Reveos at HMC Qatar.  The Reveos system is still in use at that institution.  There is also a previous post about Mirasol riboflavin-based pathogen inactivation.

We were the first place in the world to combine the automated component production Atreus with the Mirasol pathogen inactivation.  Their synergism was very important in the rapid throughput of component production for Qatar.

Terumo has two programs, 2C (C for components) to yield plasma and RBCs and 3C for yielding RBCs, plasma, and platelets.  The 2C program is faster but no platelets are separated.

We used the Atreus since 2010 and later replaced it with the Reveos in 2016.  Both systems use a special blood bag set that collects the whole blood in European CPD.  The kit is carefully placed in the machine.  Atreus machines accepted one blood bag set, the Reveos can accept up to 4 sets.  In both cases, the whole blood is processed to yield packed RBCs, leukoreduced plasma (<1E6 residual WBCs), platelets, and a special WBC bag (i.e. the residual buffy coat, which is not for clinical use.

The Atreus took about 10 minutes to process the one bag set whereas the Reveos processes 4 bag sets in slightly more than 20 minutes.  Thus, the throughput from the Reveos is twice that of Atreus.

We had 4 Atreus and later 4 Reveos machines and these were handled by up to 4 technologists, depending on the number of the units.  While the machines were running, the staff were filtering the RBCs and platelet pools, pooling the platelets, and performing the PAS-Mirasol pathogen inactivation.  The workflow was not hectic and staff were not stressed out by the multiple tasks.  Normally 1 staff member ran the Reveos or Atreus machines at any one time.

When the processing was complete, the RBC bags were filtered with an integral leukodepletion filter designed to leave a residual of <1E6 WBCs in accordance with the CE Standard.  The platelets were combined to give a target yield of >= 2.4E11 absolute number of platelets.  Then the pool was leukodepleted by filtration to a residual of <1E6 WBCs.

Both Reveos and Atreus measured the RBC, platelet, and plasma volume yields.  Additionally, for platelets a Platelet Yield Index PYI was calculated as a relative measure of the platelet yield.  To reach a goal of 2.4E11 platelets, the PYI indices for the individual platelet bags were added so that the total exceeded 240.

When combined with the Mirasol system, the component volumes for the plasma and platelets needed to be within specified ranges.  Both systems could easily meet these requirements.

When we switched from Atreus to Reveos, our platelet yields increased.  The transition period was only two weeks.   When we adopted platelet additive solution PAS at the same time, the Reveos had a special program to make “dry” platelets with less volume so that the PAS could be added and still stay within the acceptable range for pathogen-inactivation.

Throughout these years, Terumo sent us special engineers to handle the Atreus then Reveos, Mirasol, and PAS processing.  All staff were trained by Terumo initially before we finalized their competency assessments. 

We had excellent local service:  we never had downtimes due to equipment failures.  During the COVID pandemic, all materials (kits, filters, Mirasol solution, and PAS) have been provided without interruption.

We went live with Medinfo Hematos IIG software for the entire blood donor center and hospital blood banks in 2013.  From the first day 30/6/13 we had bidirectional interfaces first to the Atreus and later to the Reveos—the world’s first.  Likewise, the Mirasol and PAS processing were fully integrated with Medinfo when they were activated.

The residual buffy coat was not used for patient care.  However, it has proven invaluable as a quality control material for the stem cell laboratory.  In addition, many researchers have used it to establish cell lines for investigational use.

Proper handling the collected whole blood units is critical to success:

  1. Maintain the temperature below 25C.
  2. Carefully stack the whole blood units in the blood containers—do not play “ring toss” and just throw them into the container.

In summary, I am very pleased with using this system for over 10 years.  In a few weeks, the production laboratory was fully GMP compliant using a diverse group of staff with varying technical backgrounds.

The following are some pictures of the Reveos and its prepared blood components.

Reveos Machine has 4 chambers to process the 4 whole blood units.
Reveos takes up little space: this crowded corner processed all whole blood for Qatar.
Buffy-coat platelets processed by the Reveos
Close-up of Reveos buffy coat platelets: notice there are NO streaks of RBCs.

My Ten Years of Experience with Mirasol

I originally adopted this technology because of the plethora of new emerging pathogens.  In addition, I have been concerned about unknown pathogens that have not yet been discovered.  It is not what we know, it is what we don’t know that bothered me.

It is now 10 years since I started using riboflavin-based pathogen inactivation.  Our adoption of the technology was as follows:

  • 2010 pooled buffy coat and apheresis platelets, both suspended in plasma
  • 2012 whole-blood derived plasma and apheresis plasma
  • 2015 pooled buffy coat and apheresis platelets, both suspended in platelet additive solution PAS

Buffy coat platelet pools and whole-blood-derived plasma were both prepared with automated blood component technology, originally with the Terumo Atreus and later with Terumo Reveos system.  We were the first site worldwide to use automated production with the Mirasol system.

After 10 years and over 300,000 donor collections, no documented infectious agent transmission has been noted.  Our average platelet loss has been 4%.  There has been no increase in adverse reactions to plasma or platelets compared the time before we adopted these technologies.  Physicians accepted the products readily.

Mirasol adoption allowed us to discontinue irradiation of platelets and extend our outdate to 7 days.  We did not need a specific bacterial detection system.  Pending regulations in the USA will require stringent bacterial detection processes that are not necessary if a pathogen-inactivation system is being used.

Terumo sent its own engineers to set up and validate the system.  They also trained all the staff in the actual pathogen-inactivation processes and helped us to perform the validations.

When adopting pathogen inactivation, we compared Mirasol with its competitor and selected it for the following reasons:

  • Loss of platelets is low (about 4%)—lower than its competitor product.
  • There is no need to remove the riboflavin from the final product.

In our system, our goal was rapid processing of units.   With Mirasol, we did not have remove the riboflavin from the final product.  The competitor product requires at least 6 hours post-treatment to remove the psoralen agent.  We could immediately use the Mirasol product after treatment!

In our Reveos-Mirasol system, we can process whole blood into packed red cells in SAGM, buffy coat platelets, and plasma in a total of 5 hours including all testing with Mirasol treatment and platelet additive solution PAS.

We originally used the system manually, but in 2013, Terumo in conjunction with Medinfo Hematos IIG developed an interface to the Mirasol illuminator.  The latter device would transmit the successful completion of the illumination to the software.  Any errors in the illumination would block release of the blood component from Medinfo.  Medinfo also monitored the component volumes to prevent treatment of units outside Terumo’s recommendation ranges.

Adoption of platelet additive solution PAS gives us a final product with minimal residual plasma which potentially can reduce plasma reactions and TRALI/TACO.  It also minimizes our need to reduce the volume of platelet components for pediatric patients, especially in cases with ABO-incompatible plasma

Adopting any system of pathogen inactivation requires meticulous monitoring of component volumes to ensure they are within the range for the treatment.  The use of a blood bank software greatly facilitates this.

We make both buffy coat and apheresis platelets. The change from plasma-suspended to PAS-suspended platelets went smoothly.  Special training for Trima apheresis staff to use the new processes was provided by Terumo.

Throughout this time interval, Terumo has provided excellent technical support and educational activities for all staff.  Despite the COVID pandemic, Terumo has been able to deliver supplies to meet our needs so there was no interruption in production.

We started COVID convalescent plasma CCP production at the end of winter 2020.  We set up a parallel but separate quarantine system of collection and processing, originally manual but later controlled by the dedicated blood bank software Medinfo Hematos IIG.  All CCP units have been treated with Mirasol.

In the future, I hope Mirasol will close the loop by providing a pathogen-inactivation process for red blood cells so all components can be treated.  The CE mark for pathogen inactivation of whole blood is exciting and I hope that component preparation from this product will be offered.

In summary, our blood bank system had an excellent, synergistic relationship with Terumo and Medinfo to provide the highest quality product that is currently licensed.  I hope we will all continue to work together to improve the patient care.

International Perspective

When I first moved overseas from the United States, I brought the perspective of my American training and experience.  I saw everything in my new blood bank through those eyes.

Yet, most of my staff were not American or even North American.  Few were even native in English, and most of those  were not American.  They had different qualifications, many of which would not have been accepted by the American schemes.  Still, they functioned well.

I also worked with the US military technologist staff during Gulf War One.  Some did not even have a Bachelor’s degree;  yet, they performed the work well.

I used many technologies that were not yet (or never) US FDA approved such as gel or glass bead typings and pooled buffy coat platelet production.  There were rare reagents I could buy off the shelf (e.g. anti-Tja/PP1Pk).

Later, I adopted pathogen-reduction technology (Mirasol), automated component production (Atreus then Reveos), and platelet additive solution.  I achieve a level of good manufacturing practice that would have been difficult to achieve by the FDA-approved methods.

My perspective had changed.  In the Middle East, I studied many frameworks and came to the conclusion that the best approach was to customize them to our local needs.  My particular experience was to start with one framework, i.e. Council of Europe CE, and then localize it.

To do this, I could not use an American turnkey blood bank software for either the donor or patient operations.  I needed a flexible system that could be customized to my needs.  Again, I chose a CE-marked system, Medinfo Hematos IIG that had already been adapted to many frameworks.

It is much easier to work solely within one system such as FDA.  However, if I had done that, I would have lost so much flexibility and not had a system optimized for local conditions.  I would not have used Mirasol, Reveos, Diamed, and many other reagents.

One big disappointment at such international meetings is the perspective by one country’s regulatory agency that they feel its regulations and framework will work well overseas.  I would wager that those people were not well acquainted with international conditions.

Another frustration was attending another international meeting in which the presenters apologized for the source of information since it came from a foreign country (France) and not their own (United States).

No country has a monopoly on what is best for everyone.  To share our experiences and compare is so valuable.  No one assume his way is the best.  In my career, I have had the richest experiences studying other perspectives and my organizations have benefited greatly from the exchange.  We can all learn from each other.  We are citizens of the world.

Therapeutic Apheresis Volume Calculations

You can get the values off the therapeutic apheresis machine, but in the middle of the night when you have to write orders, it is convenient to estimate the volumes (whole blood, plasma, RBCs).  These are the values from my lectures to hematology fellows while I was at HMC Doha:

Whole Blood:

Weight in kg X 70 ml/kg = whole (whole) blood volume adult

Weight in kg X 85 ml/kg =whole blood volume for child (prepubertal)

Weight in kg X 100 ml/kg = whole blood volume for neonates/premature

Example:  70 kg adult has 4900 ml blood volume (I round up to 5 liters)

Plasmacrit + hematocrit = 1.00 in fractions (100%), ignore buffy coat volume

Plasmacrit = 1- hematocrit

Plasma volume:

Plasma volume = whole blood volume x plasmacrit = whole blood volume X (1-hematocrit)

RBC volume:

RBC volume = whole blood volume x hematocrit

Estimates for blood components:

The volumes will depend on the original amount collected (e.g. 450 vs 500 ml),  original preservative solution used (e.g. CPD), use of automated component production such as Terumo Atreus or Reveos, use of RBC additive solution (e.g. SAGM), leukodepletion, platelet additive solution, pathogen inactivation.

At HMC Doha, the average values were:

300 ml for leukodepleted RBCs in SAGM prepared by Reveos

300 ml for platelet pools in Mirasol and platelet additive solution (residual WBC < 1E6)

300 ml for plateletpheresis concentrate (2.4E11) in Mirasol and platelet additive solution

250 ml for leukodepleted, pathogen inactivated plasma

Summary of Accomplishments at HMC 2011-2020

I resigned from HMC on 16/4/20.  Here are a set of my major accomplishments during that period. None of my work after this date has any relationship to HMC.


Established automated component production using Atreus technology, plasma and platelet pathogen inactivation (Mirasol)—made HMC component production Good Manufacturing System GMP compliant

Adopted non-PCR-based NAT technology (Grifols/Novartis Tigress) and Qatar becomes world reference site for this

Based on the above, Qatar can now completely process all whole blood into blood components (red cells, platelets, and plasma) in as little as 5 hours from collection!


Prepared policies and procedures for the hospital blood banks/transfusion services, blood donor center, therapeutic apheresis, and laboratory information systems to bring HMC in compliance with the Council of Europe, international AABB, and other standards.  I customized our own standards for our local needs based on them.


Implemented custom build of the multilingual blood bank computer system (Medinfo) for both patient and donor services, including development of interfaces to all production equipment including Atreus and Mirasol (world’s first) and a direct link to Ministry of the Interior to obtain patient demographics in English and Arabic—Qatar became the world’s first site to combine fully-interfaced, automated component production with pathogen inactivation:  Qatar becomes world reference site for this.


Built, validated, and implemented laboratory build of hospital information system, Cerner Millennium


Replaced and updated Atreus with Reveos automated component production to allow faster throughput and capacity with a full bidirectional interface (world’s first), introduced platelet-additive solution PAS with pathogen inactivation (Mirasol)—Medinfo interfaces updated to Reveos for all equipment:  this doubles the capacity to process whole blood into components using the same physical space


Updated dedicated blood bank software Medinfo Hematos IIG by several versions using Division Head, LIS, and internally trained Super Users—at great cost savings to HMC by not using outside consultants (e.g. Dell Consulting)


Established column absorption technology using Terumo Optia therapeutic apheresis machine for treatment of ABO-incompatible renal transplants:  I validated using the Ortho Vision MAX to perform ABO antibody titers for this system and correlated it with the reference method at Karolinska Institutet in Stockholm (manual gel) to bring rapid throughput and labor savings—Qatar being the first-site in the world to do this.  We saved money by using the same apheresis machine to use this column absorption technology (no need for second machine to use the columns)


Expedited setup (two weeks total) of COVID-19 convalescent plasma production, initially manual and then fully integrated into the Medinfo computer system as a customized module with separate quarantine collection, production, and transfusion service functions


I was awarded two HMC Star of Excellence Awards:

2013—Liver Transplantation Transfusion Support

2019—ABO-Incompatible Renal Transplantation Support

TRALI/TACO Policy and Process

The following was my process at HMC Doha for TRALI/TACO.  It includes proactive measures to minimize the risk of TACO and the procedure for surveillance and workup of such cases.

In the Medinfo blood bank computer system, we did not prepare plasma or platelets from female donors.  If approved by a transfusion medicine physician, a manual override was made in exceptional cases (e.g. mother donating platelets for her child in neonatal alloimmune thrombocytopenia cases.).  In some other countries, they do HLA antibody testing to allow females to donate platelets.

I emphasize that the diagnosis of TRALI and/or TACO is clinical, but the transfusion medicine physicians must always consider the possibility whenever there is an adverse effect associated with progressive respiratory distress.


Since TACO and TRALI are major causes of serious adverse effects from transfusions, this policy outlines actions being pro-actively taken to mitigate the risks in Transfusion Medicine.  TACO and TRALI may be difficult to distinguish so this policy addresses both.


  1. Implement measures to minimize TRALI and TACO
  2. Track cases of transfusion-associated acute lung injury TRALI and TACO
  3. Develop algorithms for suspected cases of TRALI and TACO


  1. All transfusion reactions are reviewed by the Division Head, Transfusion Medicine or his designee on a STAT basis, 24 hours a day, 7 days a week
  2. Any reactions with respiratory distress are reported as “rule-out TRALI/TAC” to the clinician
  3. All transfusion reactions are recorded in Medinfo HIIG for tracking and reporting.

Risk Management:

  1. Female blood donors routinely are only used for making RBC components (i.e. not for FFP, cryoprecipitate, or platelets).
  2. RBCs are in additive solution SAGM so only 35 ml residual plasma is present per unit.
  3. All platelet components are in platelet additive solution with only 35 ml residual plasma per component.
  4. All platelets and plasma are pathogen-inactivated which may reduce the risk of TRALI.
  5. If the female has a rare phenotype (e.g. IgA deficient, rare platelet antigen typing) she will only be considered for a directed donation of platelets or FFP/cryoprecipitate for that special-needs patient if she does not have HLA antibodies (anti-human-neutrophil antibody testing to be implemented in such cases when it is available on-site).
  6. Solvent-detergent treated plasma SDP is available for patients with a confirmed or suspected history of TRALI.
  7. All cellular components are leukodepleted (< 1E6 residual WBCs/component as per CE Standards) in the blood bank at the time of production.
  8. Blood bank computer system Medinfo Hematos IIG limits the number of components released at any one time (excluding emergencies).


  1. Transfusion Medicine TM will notify the outside blood supplier of any units implicated or associated with TRALI.
  2. A transfusion medicine physician will notify the most responsible physician of any workup results suggesting the possibility of TRALI/TACO.
  3. TM will notify all donors of their disqualification from blood donation based on the following algorithm.

Algorithm for Diagnosis and Management of Donors:

  1. Evaluation of the Donor and Recipient in Suspected TRALI:
    1. The medical technologist will process all transfusion reactions as STAT and immediately contact the TMS Director or physician designate with the results.
    2. The medical technologist will convey information to the TMS Director or designate about ALL blood components issued recently, especially in the last 6 hours prior to the event.
    3. The TMS Director or designate will specifically check if there is evidence of respiratory distress listed on the transfusion reaction investigation form.  If so, he will contact the responsible clinician immediately for further assessment.
    4. If the signs and symptoms suggest ALI (see Table 1 above), the TMS Director or designate will inquiry about the left atrial pressure to rule out left-sided heart failure as a cause for the pulmonary edema.
    5. Based on the clinical information, the TMS Director or designate may elect to order any or all of the following tests if available:
      1. Quarantine all remaining components from possibly implicated/associated donor(s) while the workup is in progress.
      2. Recipient HLA, platelet, and/or granulocyte antibody screen, or a crossmatch between recipient plasma/serum and donor leukocytes
      3. Donor HLA and/or platelet antibody screen, granulocyte antibody screen, crossmatch donor plasma/serum and recipient leukocytes, inter-donor crossmatch between plasma/serum of one and leukocytes of another donor.
      4. The usual algorithm to be followed is as follows:
        1. Consult patient medical record and clinical care physician to determine if the diagnosis of TRALI is likely.
        2. Check all components transfused within 6 hours prior to the onset of symptoms.’
        3. Immediately quarantine other components from the same donations and contact outside blood suppliers if indicated.
        4. Obtain donor antibody testing of only highly suspect cases, based on the clinical manifestation and initial diagnostic tests:
          1. If multiple units transfused within hours, only investigate components donated by multiparous females and/or last two units transfused.
          2. First test for presence of HLA class I and class II antibodies in donor components.
          3. If antibody positive, HLA type recipient’s lymphocytes to detect corresponding antigen or perform crossmatch with donor plasma and recipient lymphocytes.
          4. If HLA antibody negative, proceed with neutrophil-specific antibody testing of donor plasma.
        5. If matching antigen-antibody identified or if positive crossmatch, defer implicated donor immediately.
        6. If no such concordance found or if crossmatch is negative, donor eligible to continue donating.
        7. If no antibodies found in donor plasma, test recipient plasma for antibodies to HLA class I and II antigens:
          1. If recipient antibody positive, HLA type donor’s lymphocytes to detect corresponding antigen or perform crossmatch with recipient plasma and donor lymphocytes.
          2. If recipient HLA antibody negative, proceed with neutrophil-specific antibody testing of recipient plasma
  2. Donor Disposition:
    1. For donors implicated in TRALI or associated with multiple events of TRALI, one or more of the following options may be selected at the discretion of the Head, Transfusion Medicine or designate:
      1. Defer donor from donation
      2. Divert plasma for fractionation or discard plasma from future whole blood donations from that Blood and Apheresis Donor Main Questionnaire
      3. Manufacture no platelet or plasma components from that donor
      4. Wash or freeze/deglycerolize RBCs from that donor
      5. Permanently defer the donor from future plasmapheresis or plateletpheresis donations
      6. Evaluate the previous donations from that Blood and Apheresis Donor Main Questionnaire  Avoid giving the same recipient future transfusions from the same donor implicated in TRALI
      7. If the implicated unit(s) are from another facility, that blood center should be notified to initiate a workup for possible TRALI in the donor.
  3. Interpretation:
    1. The diagnosis of TRALI is not clear-cut:
      1. The AABB interim standard does not apply.  It is at the discretion of the TMS Director or designate whether to conduct donor assessments.
    2. The donor is associated with a single event of TRALI:
      1. This applies where the diagnosis of TRALI has been established based on clinical and radiographic findings:
      2. Each donor from each and every component associated with TRALI must be identified and traced.
      3. Co-components from the current donation and components from previous donations should be evaluated for recipient complications.
      4. The donors medical history should be evaluated for previous pregnancies, transfusions or other events that may have resulted in antibody development.
      5. Based on the results of this investigation, the Head, Transfusion Medicine or designate should decide:
        1. Whether to perform laboratory testing
        2. Whether to discard the remaining blood components from the donor
        3. Whether to allow or indefinitely defer the donor
    3. The donor is associated with multiple events of TRALI:
      1. This applies where the diagnosis of TRALI has been established based on clinical and radiographic findings:
      2. Each donor from each and every component associated with TRALI must be identified and traced.
      3. Co-components from the current donation and components from previous donations should be evaluated for recipient complications.
      4. The donors medical history should be evaluated for previous pregnancies, transfusions or other events that may have resulted in antibody development.
      5. Based on the results of this investigation, the TMS Director or designate should decide:
        1. Whether to perform laboratory testing
        2. Whether to discard the remaining blood components from the donor
        3. Whether to allow or indefinitely defer the donor
    4. Triage based on laboratory testing for TRALI:
      1. The donor associated with TRALI is antibody-negative:
        1. The donor may continue to donate.
      2. The donor associated with TRALI is antibody-positive but the specificity is NOT directed against a recipient antigen by either antigen typing or crossmatching (i.e. the donor is NOT implicated in TRALI—see definition above):
        1. Indefinitely defer the donor from all donations OR
        2. Allow donation of washed/frozen-deglycerolized RBCs only
      3. The donor is implicated in TRALI (see definition above):
        1. Indefinitely defer the donor from all donations OR
        2. Allow donation of washed/frozen-deglycerolized RBCs only
      4. The recipient has antibodies implicated in TRALI (determined by crossmatch or antibodies directed against specific HLA class I, HLA class 2, and/or human neutrophil antigens):
        1. The recipient must receive leukodepleted blood components
    5. TACO
      1. TACO is due to cardiac overload.  Our mitigations are to restrict release of the number of components outside emergency events.


  1. AABB Association Bulletin 14-02, TRALI, Bethesda, MD, USA
  2. Han Y. and Goldfinger D., Transfusion Medicine TM 07-5 (TM-297) Checksample, American Society for Clinical Pathology, Chicago, IL, USA. July 2007
  3. Goldman M, Webert, KE, Arnold DM, et al., Transfusion Med Rev  2005; 19:2-31.
  4. Fung YL, Goodison KA, Wong JK, Minchinton RM., Investigating Transfusion-Related Acute Lung Injury (TRALI), Intern Med J. 2003 Jul;33(7):286-90.
  5. Standards for Blood Banks and Transfusion Services, Current Edition, AABB, Bethesda, MD, USA
  6. AABB Association Bulletin #05-09, Transfusion-Associated Acute Lung Injury, 11/8/05
  7. AABB Association Bulletin #05-04, Proposed Interim Standard for Deferral of Donors Implicated in TRALI, 9/3/05.
  8. TRM.42110, CAP Transfusion Medicine Checklist, 15/6/09