Recently one of our members shared with the e-network that her institution was "still reeling from a 1998 FDA inspection where the institution was cited for not having reported a possible transfusion-related fatality. As a result of this citation the facility instituted a policy that requires their nurses to report to the laboratory each time a patient expires, if that patient received blood in the final 24 hours of life. Upon being informed of a patient's death, the laboratory performs a standard transfusion reaction investigation to rule out the possibility that the death was due to the transfusion therapy. The inquiring member asked the e-network if any other institutions have such a policy, or if other institutions merely rely on the clinician to determine that a patient's death might have been transfusion-related? The inquiring member also wanted to know what other institutions did if a patient expired directly as a result of trauma (but got blood before dying) or if a nurse wrote in the "vital sign" section of Transfusion Record "patient coded". Do any institutions automatically work up such deaths as possible transfusion reactions? Finally, the inquiring member wanted to know if other Blood Banks obtained a copy of the completed Transfusion Record (from the patient's medical chart, including the vital signs), so Blood Bank staff can independently review for increases in temperature. etc.?

To which the following replies have been received (NOTE: some of the replies have been slightly edited for the sake of clarity):

1. We are a large three-hospital system. We have never had such a proactive approach to what is admittedly a potentially problematic issue. While such an approach would be expected to have some level of case yield, I believe that it would be so small as have little merit - at least to undertake it voluntarily. We do receive copies of the transfusion requisitions (not invariably), but we often find that vital signs have been entered in the patient's chart instead of on the requisitions. In our facilities, reviewing these would not be a good means of uncovering unreported reactions.

2. No, at our institution we don't have such a policy and I think implementation of such a policy is not warranted either. In the situation under discussion, temporary followup on fatalities that coincide with blood transfusions may be helpful as an audit tool to ensure that staff is following appropriate procedures. Beyond that, I would concentrate on a thorough review of all transfusion-related procedures and policies and formulation of a corrective actions plan, if warranted. I would channel energy and resources into prevention of transfusion accidents and errors -that should also take care of any worries that something bad is happening and you don't hear about it.

3. Our Blood Bank does not attempt to establish cause of death by an independent chart review. Also, we do not initiate investigations of possible transfusion-related fatalities on all deaths because the patient was coincidentally receiving a transfusion. We feel that the cause of death is most accurately established by the patient's physician. Having said that, I do not exclude the theoretical possibility that someday there could be a specific adverse laboratory finding, such as red plasma in a tube sent to the chemistry laboratory for analysis following our issuing Red Blood Cells to that patient's roommate, that would raise concern and such an independent investigation. In that theoretical situation, I would take each step with awareness that I'll probably be explaining these actions someday in a courtroom.

4. Working up all deaths as possible transfusion reactions because transfusion has been received in the 24 hours prior to death is "over the top".

5. We do not perform routine surveillance as described by the inquiring member for the possibility of fatal reactions. The clinical staff is responsible for "calling" a transfusion reaction, although questions regarding this can forwarded to the blood bank staff and pathologist. It would be a Herculean task to do a transfusion reaction workup on all deceased patients who had blood 24 hours prior to death, at least at our institution. It seems that this as much a clinical problem as a laboratory one, and that an effort to educate the clinical staff about transfusion reactions might be more rewarding. If one assumes a death rate from hemolytic reactions as 1:100,000 to 1:600,000 RBC's transfused, and that most of these will be suspected clinically anyway, one might go through a lifetime(s) of screening to discover the very rare "silent" fatal transfusion reaction.

6. As the transfusion service director in a county hospital with a Level I trauma center (not the Editor's place!), I attempted to monitor whether transfusions given within 24 hours of death were in any way related to death. The conclusion was that in most cases, death soon after admission was due to trauma and not to the multiple transfusions received by in extremis patients; however, one possible case of TRALI was picked up in a non-trauma patient and reported. A study of patients with ARDS to rule out TRALI would be more productive.

7. The tip of another iceberg and one reason why FDA gets so very few reports compared to the reality. Follow-up is performed from a serological perspective and intervention of the Transfusion Service. We sometimes hear rumors or will call to find out what happened when we hear a patient we transfused has died. Transfusion medicine physicians make the call about what to report and when; we've routinely had to "strongly encourage" nurses to report their findings to lab even in routine reaction reporting. Without a knowledgeable staff who go that extra mile, know where to probe for the information, and are willing to stand firmly for doing the right thing, many transfusion-related concerns do not get followed up. This is the very issue which most concerns me when using a "generalist" staffing method. Generalists can be trained but the depth of their experience is often the issue.

8. We rely on the clinician to determine the cause of death. We do, however, review transfusion slips which have the pre and post transfusion vital signs.

9. We routinely screen the transfusion tag for abnormal vitals. We also have access to the daily admission, death and discharge log, so that we can respond with a workup to any nursing or physician suspicion of a fatality. In the past 15 years, we have never had a fatal transfusion reaction reported to us that had not been reported contemporaneously with a transfusion. I think perhaps the facility in question had no policy, procedure and no education in place for medical staff and perhaps that was why their FDA inspector cited them.

10. We do not audit all deaths in our hospital to see if patients were getting transfused at the time they died and then working up those cases as a suspected transfusion reaction. We rely on those caring for the patient to inform us of transfusion reactions and of suspected deaths due to transfusion. I have always found the patient care areas to not hesitate to inform the laboratory if they think a patient's death was attributable to the transfusion. I think general education of the nursing and physician staffs concerning the recognition and treatment of transfusion reactions is more efficacious, in the long run, than searching for zebras when there aren't any. Having said that, I think that the short term monitoring and gathering of data can be useful as a quality assurance/compliance audit. For example, one can look at a number of issues related to proper completion of paperwork, proper transfusion techniques, proper monitoring of the patient's vital signs, proper recognition of transfusion reactions by patient care personnel, etc. If such an audit would demonstrate, for example, that transfusion reactions are not being properly recognized and worked up, then an educational effort might be needed. This would seem to be a better approach in the long run rather than the transfusion service assuming the role of transfusion reaction "police".

11. We absolutely do not pro-actively look for deaths in patients who have received transfusions. Our hospital supports large hematology/oncology and BMT programs, as well as complex cardiovascular surgery and a trauma center. When we try to notify recipients in lookback programs, we find that almost 1/4 of our transfusion recipients died during the same hospitalization in which they were transfused, related to their underlying diseases. With regard to transfusion reaction reports, part of our protocol is that when a report of a "possible transfusion reaction" is received by the transfusion service, a transfusion service technologist calls the reporting nursing unit and inquires as to the current status of the patient. There are criteria for contacting the transfusion service physician, which include dyspnea, cyanosis, hypotension, shock, death (and of course, testing that indicates the presence of a hemolytic reaction). The transfusion service physician will follow-up on any of these "worrisome" symptoms.

12. And finally, a Note from the Editor: At USC we do not routinely perform a transfusion reaction workup on a patient just because he/she died within 24 hours of being transfused. Such a policy would result in an enormous amount of unnecessary work, because all too often a trauma patient expires in close relationship to a transfusion. However, if a clinical suspicion of a cause and effect that a transfusion might have contributed to a patient's death is raised by a clinician, nurse, family member or anyone who might have a legitimate opinion, an investigation is done to determine the cause of death and the role (if any) that thetransfusion might have played as a contributing cause. In the absence of a clerical error leading to administration of ABO-incompatible blood or of a clinically overt reaction attributed to a transfusion (i.e., acute hemolysis, anaphylaxis, acute respiratory distress or sepsis), it is unlikely that a transfusion would be responsible for death in the acute setting. Furthermore, if a forensic investigation to determine the cause of death including a review of the medical records and the performance of an autopsy reveal no link to a transfusion, then a transfusion-related death is most likely excluded. On the other hand, if a clinical or laboratory investigation reveals evidence of one of the following: [1] post-transfusion hemolysis (such a finding in a massive trauma victim must be interpreted with caution, as hemoglobinemia and hemoglobinuria may result directly or indirectly from a traumatic injury or may be confused with myoglobinuria), [2] an incompatible transfusion, [3] clinical evidence of an anaphylactic or acute pulmonary reaction to transfusion, or [4] a bacterially contaminated blood component, then there may have been a causal link between the patient's death and a transfusion. In the event a blood transfusion is proven or highly likely to be related to a patient's death, we will report the fatality to the FDA, as required.

ADDENDUM July 25, 2000

13. The following comment was submitted by a tertiary care facility that specializes in children. This facility does not do a transfusion reaction workup for a patient who died unless a record review suggests the need. Here is what the Transfusion Service Director said they do:

"I, or my designee, review the medical records of all patients who expire at our hospital, either on the date of expiration or on the following day, to determine if there are issues requiring a Coroner referral, a referral to Hospital Risk Management, or if there is a potential that the death represented a fatal transfusion reaction, either immediate or delayed. A transfusion reaction workup is done whenever a reaction is suspected.

ADDENDUM July 26, 2000

14. For 30 years we have been performing transfusion reaction workups when patients have died during transfusion. This is partly to avoid cavalier notes in charts by house staff or clinicians that the death was due to the transfusion. This policy did deflect the FDA inspector.

Please submit comments to the e-Network Forum.

Ira A. Shulman, MD
CBBS e-Network Forum Editor & Moderator