NKF KDOQI GUIDELINES 2000
GUIDELINES FOR ANEMIA OF CHRONIC KIDNEY DISEASE
SECTION I: ANEMIA EVALUATION
Thirty-three patients with severe chronic kidney disease (creatinine clearance of 10 to 14 mL/min) were moderately anemic with mean Hgb 10g/dL.52 These patients were treated with oral iron initially, with minimal response, and then received 50 mg of IV ferric saccharate every week for a total of 1,000 mg. Twenty-two, or two thirds of the patients, were classified as responders in that their Hgb increased from 9.9 to 11.1 g/dL (Hct from 29.4% to 32.4%). The responding patients had baseline mean transferrin saturation values of 21.9% and baseline mean serum ferritin values of 94 ng/mL. The nonresponders had similar values. The authors recommended that enough IV iron should be given to increase the serum ferritin levels to 200 to 400 ng/mL and/or transferrin saturation values to between 25% and 35% before initiating Epoetin therapy.52
SECTION II: TARGET HEMOGLOBIN/HEMATOCRIT
In the Canadian study,73 three groups of patients were studied: one placebo group and two Epoetin-treated groups with target Hgb values of 10.2 ± 1.0 g/dL and 11.7 ± 1.7 g/dL, respectively. The mean values for the physiologic tests of the higher target group at baseline (Hgb 7.1 ± 1.2 g/dL) were much higher than the mean values obtained in the other treatment group which reached the target Hgb of 10.2 ± 1.0 g/dL, thereby making the authors conclusions difficult to interpret. In an Australian study,74 cardiac index and heart rate were clinically and statistically significantly better in the group of patients reaching higher target Hgb. Quality of life was also better at the higher Hgb, but the difference failed to reach statistical significance,75 possibly because of the relatively small number of patients.
Hgb levels among 107 patients without LVH were 11.9 ± 2.0 g/dL, whereas in 68 patients with LVH, Hgb levels were 11.0 ± 2.1 g/dL (P = 0.0049).86 Sixty percent of the patients with an Hgb ñ11 g/dL had LVH.
It is now clear that the increased cerebral blood flow is a physiological response to a decrease in the total oxygen content of arterial blood.381 In an in vivo study, cerebral blood flow and oxygen delivery were examined in eight normal volunteers whose Hct levels were decreased from a mean of 42.5% to 37.7% via phlebotomy and hemodilution.382 Using PET scans and oxygen (O2)15 inhalation, a decrease in total oxygen content of arterial blood from 19.1 ± 1.0 to 16.9 ± 1.0 mL/dL (P < 0.005) was demonstrated. Despite an increase in cerebral blood flow, there was still a decrease in oxygen delivery to cerebral tissue. The authors concluded that optimal Hct for the human brain is that normally found at sea level. This finding was confirmed by another study in which 27 patients with ischemic cerebrovascular disease were studied with intracarotid isotope injection to determine cerebral blood flow and oxygen delivery between a Hct range of 31% to 53%. As shown previously, there was an inverse relationship between Hct and cerebral blood flow, but a linear relationship between Hct and oxygen delivery to brain tissue, with the maximal level of oxygen delivery occurring within a Hct range of 40% to 45%.
Of 1,410 hemodialysis patients observed before Epoetin was used therapeutically in Okinawa, Japan (April 1, 1988 to March 31, 1990), 25 had strokes and 3 had acute myocardial infarctions. Following the use of Epoetin (April 1, 1990 to March 31, 1993), 61 and 12 patients out of 1,916 hemodialysis patients had strokes or acute myocardial infarctions, respectively. These were highly statistically significant differences. Unfortunately, the Hct levels were not reported in the Epoetin-treated patients, although the target level was 30%. The conclusions of this study are considered flawed as noted in a letter to the editor of Nephron.383 In response to this letter, the lead author on the study seems uncertain about his findings by stating "we do not think that EPO is the culprit of the observed increase in incidence of stroke/AMI," and yet also adds "we believe that the introduction of EPO and the changes in patient demographics caused the difference in the incidence of stroke/AMI."384 These authors had previously reported that the risk of cerebral hemorrhage in their hemodialysis patients was 10.7 times higher than the general Okinawa population,385 an incidence not observed in the United States.
SECTION III: IRON SUPPORT
This conclusion is supported by data from the USRDS Dialysis Morbidity and Mortality Study (wave 1) for 1996, in which more than 50% of 2,613 dialysis patients in 1993 had TSAT levels of <20%, 36% had serum ferritin levels of <100 ng/mL, and 56% had serum ferritin levels of <200 ng/mL. Furthermore, only 11.2% of these patients had received IV iron, and 25% had received neither IV nor oral iron.66
Hemodialysis patients typically lose up to 15 to 25 mL of whole blood at each dialysis treatment (approximately 60 mL per week) as a result of retention of blood in the dialyzer and tubing, and phlebotomy for blood testing.386,387 If such patients have a Hct of 36%, this constitutes an average loss of 3 mL of RBC per day. Average (in a normal subject) GI iron losses of 1 mg/day heighten depletion of iron stores, resulting in a total loss of at least 400 mg of iron over a 100-day period (an arbitrary duration used for calculation purposes only). Some patients, such as menstruating females, have even greater blood losses. Patients who undergo vascular access surgery or who have additional gastrointestinal bleeding often have cumulative iron losses exceeding 2 g/year. Yearly losses up to 6 to 8 grams have been documented.133
Since blood volume is approximately 8% of body weight, a 70 kg patient has a blood volume of approximately 5 to 6 L. At an Hct of 25%, this represents an RBC volume of 1,400 mL (5,600 mL _ 25%). At an Hct of 35%, a blood volume of 5 to 6 L represents an RBC volume of 2,000 mL (5,600 mL _ 35%). Thus, to increase the Hct of a 70 kg patient from 25% to 35%, a patient needs to produce 600 mL of red blood cells (2,000 mL minus 1,400 mL). Since each milliliter of newly synthesized RBCs contains 1 mg of iron, 600 mg of iron are needed to produce an increase in RBC volume of 600 mL. An alternative method for determining the iron requirement associated with Epoetin therapy has utilized a nomogram approach.388
Bone marrow biopsy specimens stained to assess iron are traditionally considered the "gold standard" measure of iron stores. It is not practical, however, to perform serial bone marrow biopsies to monitor the iron status of dialysis patients.
For example, a recent small study137 found that 12% of hemodialysis patients who were functionally iron deficient had a TSAT ñ20%, and 4% of patients who were functionally iron deficient had a TSAT Ù30%. The same investigators concluded that 12% of patients who had a TSAT <15% were not iron deficient. However the protocol these investigators used for administering Epoetin and iron (the dose of Epoetin was decreased by 25% if the Hct was >34%, and the total amount of IV iron given was limited to one gram) could have resulted in an underestimate of patients who were functionally iron deficient. This same study137 noted that 50% of patients who were functionally iron deficient had serum ferritin levels <100 ng/mL, while 10% of patients who were functionally iron deficient had serum ferritin levels >300 ng/mL. Although none of the patients who were functionally iron deficient in this small study had serum ferritin values >500 ng/mL, such patients have been reported.35,142,145
With the exception of two studies involving CAPD patients134,161 and one study involving CKD subjects,52 the studies summarized in Table IV-3 show the overall effects of regular IV iron administration in hemodialysis patients. Only mean values are depicted. The amount of IV iron is reported in milligrams per week. The frequency of iron administration varied from 3 times per week to once every other week. The duration of the studies is indicated by the number of months depicted. The first three studies showed the effects of IV iron in iron-deficient subjects. Eight studies by seven authors depicted the effects of IV iron on individuals considered to have adequate iron stores. In conjunction with the use of IV iron, the Hgb/Hct rose 19% ± 20% and the dose of Epoetin decreased by 34% ± 27% in these 8 studies. Five of these studies had baseline TSAT values between 23% and 31%. The baseline serum ferritin levels of patients who responded to IV iron varied from 99 to 403 ng/mL. The serum ferritin values did not exceed 800 ng/mL.
Acquired iron overload developed in 15 adults with normal kidney function from 70 to 210 red blood cell transfusions and was associated with serum ferritin levels of 950 to 5,000 ng/mL.389 Iron overload in dialysis patients (11 adults and 5 children) was associated with serum ferritin levels of 11,110 ± 1,679 ng/mL and 7,550 ± 3,088 ng/mL, respectively,189 and in another series of 12 adult dialysis patients with iron overload, the serum ferritin and TSAT were 8,168 ± 1,265 ng/mL and 88.3% ± 4.9%, respectively.190
One report of 670 hemodialysis patients indicated that patients (n = 118) with serum ferritin levels 521 ± 775 ng/mL had more bacterial infections than did patients (n = 489) whose serum ferritin levels were 376 ± 529 ng/mL.194 These patients were "iron-overloaded" presumably from previous transfusions, although no details were given concerning previous transfusions, parenteral iron use, or Hgb/Hct values. Only 15% of the patients were receiving Epoetin since this study was conducted over a 6-month period between September 1989 and February 1990, when Epoetin had just become available for clinical use in France. While the difference between the serum ferritin levels was significant (P = 0.028) between the two groups, the more important risk factors for infection were found to be a previous history of bacterial infection and the presence of a central venous dialysis catheter. It is difficult to exclude the role of transfusion-induced immunosuppression as the cause of the increased incidence of bacterial infections in the group with the higher serum ferritin levels. A later study by the same authors failed to confirm this relationship.198
SECTION V: INADEQUATE EPOETIN RESPONSE
Two hemodialysis patients with multiple myeloma relapsed in the setting of aggressive Epoetin therapy290; 1 patient relapsed, after a 6-month remission, with a new lytic lesion less than 1 month after initiating treatment at 100 units/kg. Urinary light chain excretion increased 10-fold (1.56 to 10.75 g/L) in a nondialysis patient following treatment with 4,000 units twice weekly; excretion reverted to pretreatment levels with discontinuation of Epoetin.293 Two observations provide physiologic rationales for these clinical observations: erythropoietin receptors are expressed on human myeloma cells,390 and nonerythroid marrow elements at the progenitor cell level respond to therapeutic doses of Epoetin in dialysis patients.391,392
SECTION VII: POSSIBLE ADVERSE EFFECTS RELATED TO EPOETIN THERAPY
The Canadian multicenter trial393 suggests that Epoetin increases the risk of thrombosis in PTFE grafts. This study matched patients on Epoetin with similar patients not on Epoetin. It reports an incidence of graft thrombosis 2.2 times higher in Epoetin-treated patients with grafts versus patients with grafts who were not treated with Epoetin. The Epoetin-treated patients had lower serum albumin levels and a higher prevalence of cardiovascular disease. The authors controlled for length of time the grafts were used prior to initiation of Epoetin therapy, but did not mention the actual age of the grafts. There were 38 pairs of patients; the authors did not state the fate of the nine patients who had grafts and received Epoetin, but for whom they could not find a paired control. In addition, there was no correlation between risk of thrombosis and the baseline Hgb, the rise of Hgb, or the dose of Epoetin. The authors suggested that their findings should be "interpreted cautiously."
© 2001 National Kidney Foundation, Inc