I have been trying to find good guidelines for transfusing patients with ESRD. When When
patients are on Epogen and they are not responding to it or because of acute blood loss,
at what level of hemoglobin one should give blood transfusion?
Abdus Malik, M.D., Zanesville, Ohio
Answers by Steve Fishbane, M.D., (SUNY Stonybrook School of Medicine, Mineola,
NY)
(responding to the implied question of Epogen failure) and by Robert N. Foley, M.D.
(University of Newfoundland, St. Johns, New Brunswick) (responding to the
actual question of to what Hgb to transfuse).
Dr. Fishbane
Resistance to recombinant human erythropoietin (EPO) is a very common
clinical situation which may be precipitated by several clinical conditions.
Research by the National Anemia Cooperative Project confirms the notion that
iron deficiency causes approximately 60% of EPO resistance. We have recently
found that iron deficiency contributes to EPO resistance in approximately 80%
of such patients in our population (Seminars Dialysis in press 1/98). Iron
deficiency is so common because of the incessant blood loss associated with
hemodialysis, blood sampling for laboratory testing, access bleeding, and
surgery. In addition, demand for iron is increased by the hyperstimulus to
red cell production induced by intravenous EPO therapy.
The second most important cause of EPO resistance is probably
hyperparathyroidism. Fibrosis of the marrow leads to inefficient
erythropoiesis, which can be corrected by intensive medical therapy, or by
parathyroidectomy. Inefficient dialysis has been demonstrated to block
responsiveness to EPO, as has aluminum toxicity. Infection and inflammation
are common causes of EPO resistance, probably due to a block in the
utilization of storage iron. Inflammation in dialysis patients may often be
occult, and related to the inherent bioincompatibility of the procedure.
Other issues to consider in patients who do not respond well to EPO include
occult bleeding, and secondary hematologic diseases such as hemolysis.
Thus, EPO resistance itself is not synonymous with iron deficiency. Patients
must be evaluated for all the clinical states discussed above. Emphasis
should be placed on appropriate testing for iron deficiency.
Dr. Foley:
These are very good questions. When someone has an acute bleed, what
should the target hemoglobin be? For the average hemodialysis patient,
it is prettly clear from epidemiological studies, RCTs on quality of
life and echocardiographic studies that a target Hgb of > 100 to 110 g/L
(10 - 11 g/dL) is appropriate, although the million dollar question "How much > 100 to
110?" has yet to be answered. For a single acute bleed it is not clear
what the long-term impact of say 6 weeks of an average of say 85 g/L Hgb
("starting" at 70 g/L and "finishing at 100" after six weeks) will be.
Personally, if the hemoglobin bottoms out above 70, and the bleeding is
stopped I would neither transfuse nor increase erythropoietin dose. Below that
transfusion, cautiously, is probably indicated.
The question about anemia resistant to rHuEPO/iv iron is difficult. In
the absence of data on this group the same target is probably
appropriate, all things being equal. I see this question as breaking
into 2 clear groups, those with and without transplant potential. In
those transfusion-dependent patients in whom transplant is not
indicated, 2 questions naturally arise: what is the lower target to
trigger transfusion and how high should one go beyond the target Hgb (say
100 g/L) given that the hemoglobin is likely to drift back down over time.
In our data-set, variability of hematocrit and mean level over time were
independently associated with mortality. This suggests that one sould
minimise departures from whatever target is chosen. That is, a threshold
for transfusion close to the target, with small overshoots. I am
guessing that giving a unit of packed cells when hemoglobin drops below
85 or 90 g/L would be the way to go. In the rare tranplant candidate who is
transfusion dependent, a key factor is time. There are so many
trade-offs (likely waiting time, co-morbid conditions, lifestyle) in
this group that I find it difficult to suggest an actual target.