348
Amphotericin B Blunts Erythropoietin Response to Anemia Andrew C. Lin, Eugene Goldwasser, Edward M. Bernard, and Stanley W. Chapman
From the University of Mississippi Medical Centerand the l'eterans Administration Medical Center, Jackson,' the University of Chicago, Illinois; and Memorial Sloan Kettering Cancer Center, New lbrk, New lbrk
Normochromic normocytic anemia during amphotericin B (AmB) treatment may be due to effects on erythropoietin (EPO) production or to direct marrow suppression. In cell cultures, erythroid colony growth is inhibited by AmB concentrations of ~1.0 p.g!J;p.I, probably by its direct effecton progenitorcells [1], with enhancedgrowthat lowerconcentrations [2]. In 1978, using a bioassay method to detect EPO in the urine and serum of patients on amphotericin B (AmB), MacGregor et al. [3] found no elevationin EPO concentration at the nadir of the anemia. Consequently, they proposed that the anemia was due to AmB's effecton EPO production. To further our understanding of the interactions betweenAmB, EPO, and anemia, we evaluatedthe effectof AmB on EPO production in patients receiving prolonged AmB therapy.
Patients and Methods Patients receiving AmB for ~1 month at the University of Mississippi MedicalCenter or at the JacksonVeterans Administration Medical Center were included in the study.Those with preexisting renal disease (serum creatinine >130 JLmol/I), AIDS, or active hematologic malignancy were excluded. To control for the effects of systemicfungal disease on EPO response, patients treated with an investigational imidazole, itraconazole, were also studied serially.
Received 8 June 1989; revised 31 August 1989. Presented in part at the 1987 American Federation for Clinical Research Southern Section meeting, New Orleans, 28 January 1987 (abstract 78). Partly supported by Veterans Administration research funds. Reprints and correspondence: Dr. Stanley W. Chapman, Division of Infectious Diseases, Department of Medicine, University of Mississippi Medical Center, Jackson, MS 39216-4505. The Journal of Infectious Diseases 1990;161:348-351
© 1990 by The University of Chicago. All rightsreserved. 0022-1899/90/6102-0032$01.00
Of the three study subjects, one had cryptococcal meningitis and diffusehistiocytic lymphomain remission by standardstagingcriteria, one had cryptococcalmeningitis, and one had disseminated blastomycosis. One control had chronic cavitary histoplasmosis, Wolff-Parkinson-Whitesyndrome, and Mycobacterium scrofulaceum infection, one had pulmonary blastomycosis, and one had chronic cavitary histoplasmosis, chronic bronchitis, and Meniere'sdisease. No differences in severity of illness between AmB- and itraconazole-treated patients could be identifiedclinically. All were ambulatory and displayed microbiologic and clinical resolution of infectionon therapy. The daily and total dose of AmB, time of the dose before phlebotomy, and time of phlebotomy were recorded. Blood was drawn for baseline AmB and EPO levels at the start of treatment and at frequent intervals (weekly if possible) during therapy. Routine serum chemistry and hematologic studies were also done. The corrected reticulocyte count wascomputedusing the formula: reticulocyte count X (hematocrit/0.45) + correction factor [4]. The sera were divided into 1- to 2-m1 aliquots, labeled with a number code, and immediatelyfrozenat -30°C pendingdeterminationofEPO and AmB levels. Blindedsera wereanalyzedfor AmB levelsusinghigh-performance liquid chromatography as described [5]. Sera for erythropoietinlevels were determined using radioimmunoassay as previously described [6]. In brief, purifiederythropoietinstandardor coded unknownsamples were mixed with 12SI-labeled EPO and incubated (4 days at 4°C) with rabbit antisera to EPO in a phosphate-bufferdiluent. The free and bound 12SI-EPO were separated by a second antibody using goat anti-rabbit serum and normal rabbit serum as a carrier. Total counts per minute per tube was determined by counting randomly sampledtubes. Sampleswerecentrifugedand the pelletswashedtwice with diluent before counting in an automated gamma counter. The percentage of bound 12SI_EPO was then calculated. This percentage was corrected for nonspecific precipitation and all samples were tested in quadruplicate and at different volumes. Regression curves relating log[EPO] standard to percentage bound were determined and EPO values falling in the linear part of the curve were used to calculate the EPO content of the samples. The interassay coefficient of variation was 6%.
Downloaded from http://jid.oxfordjournals.org/ at Cambridge University on September 4, 2015
Amphotericin B causes a normochromic, normocytic anemia thought to be mediated by direct marrow toxicity or suppression of erythropoietin production. Serial hemoglobin, hematocrit, amphotericin B, and erythropoietin levels were determined before, during, and after completion of amphotericin B therapy for three patients without significant renal disease or active hematologic malignancy. Patients with systemic fungal diseases treated with itraconazole served as controls. Serum erythropoietin levels were determined by radioimmunoassay and amphotericin B by high-performance liquid chromatography. Despite anemia in all amphotericin B-treated patients, erythropoietin levels declined or remained relatively constant during therapy while erythropoietin levels in controls were appropriate for the degree of anemia. Within 2 weeks of completion of amphotericin B treatment, two patients had increasing erythropoietin levels in response to anemia. Amphotericin B appears to suppress but not abolish the erythropoietin response to anemia; this effect disappears quickly after discontinuation of the drug.
Concise Communications
JIO 1990;161 (February)
To verify that AmB did not interfere with the EPO assay, AmB in concentrations of 0.01-5 /Lg/ml was added to duplicate serum samples and assayed simultaneously with untreated serum. No differences in EPO measurements were produced by AmB in the serum (data not shown).
Results
noted at 3 weeks of therapy despite relatively unchanged serum creatinine levels. Although the creatinine increased during therapy, the blunting of EPO response continued despite stabilization of renal function. Patient 2 had a transient increase in serum creatinine in the first week of therapy, but this normalized after hydration. Erythropoietin levels, however, did not increase in the face of anemia even with a normal serum creatinine. Likewise, patient 3 also had a fall in EPO although serum creatinine remained relatively constant throughout. Erythropoietin levels done 4 and 5 days after discontinuing AmB in patients Z and 3, respectively, were unchanged from the last treatment sample. Patient 2 had serum assayed 11days after therapy, and a small increment in EPO production was noted. Finally, the EPO level for patient 1 determined 13 days after treatment showed a definite increase. Commensurate with the increasing EPO level, hemoglobin increased from 98 to 113 gil.
Table 1. Serial hemoglobin, hematocrit, amphotericin B, and erythropoietin levels in patients before, during, and after completion of treatment with amphotericin B. Subject, testing interval (w unless indicated)
Amphotericin B Hemoglobin (gIl)
Total dose (mg)
69 87 99 94 97 103 98
140 890 1340 1490 1640 1790 1890
Levels (ug/ml)
Erythropoietin Creatinine (umol/l) (munits/ml)
1.26 0.26 0.14 0.14
43 25 21 7
0.17
13
0.19 0.11
11 11
Downloaded from http://jid.oxfordjournals.org/ at Cambridge University on September 4, 2015
Table 1 shows the serial results for the three AmB-treated patients. Patients 1 and 3 had greater-than-normal EPO titers at the beginning of the study in response to their anemia. In both, the EPO fell during AmB therapy despite continued anemia. Patient 2 had low-normal EPO at the start of therapy, reflecting her normal hemoglobin. Despite a decline in hemoglobin, there was no compensatory rise in EPO. No consistent relationship between renal function, as measured by serum creatinine, and EPO production could be identified. In patient 1, a definite decrease in EPO titer was
349
350
Concise Communications
!able. 2. Serial he~oglobin, hematocrit, and erythropoietinlevels In patients treated With itraconazole (controls). Subject, testing interval (w) Control 1 0 4 13 23 43 66 Contro12 4 11 22 32 Control 3 0 3 5 9 18 22 26 31
Hemoglobin (gil)
Erythropoietin (munits/ml)
120 67 46 103 31 112 55 111 58 118 (Itraconazole ended after 52 w) 50 118
Creatinine
(umol/l) 90 70 90 80 100
NO
13 144 177 5 184 4 (Itraconazole ended after 32 w) 170 3
NO
37 33 30 19 11 23 17 23 35 w)
70 60 60 60 70
102 113 122 124 140 138 132 147 (Itraconazole ended after
90
NO 90
NO 70 60
Discussion In vivo and in vitro evidence points to erythropoietin as the ~ri~cip~ hu~oral regulator of erythropoiesis [7]. Though limited in subjectstested, our study is the first serial observa-
tion to our knowledge of EPO and AmB in patients during treatment. Our data indicate that blunting of normal erythropoietin production contributes to the anemia accompanying AmB therapy in humans. In our patients, the decreased EPO levelsoccurred after 3 weeksof therapyand were independent of the total dose of AmB given, the AmB level at the time of EPO determination, and renal function as reflected by serum creatinine. The effect also seemed to be independentof the underlying fungaldisease, as a physiologic EPO response was observed in our controls. AmB did not interfere with the EPO assay as evidenced by our results with AmB-spiked and unaltered sera. MacGregor et al. [3], using only single samples collected immediately and up to 30 daysafter completionof AmB therapy, were unable to detect EPO levels in five of six serum and three of four urine samples. Three patientshad increments in serum creatinine after therapy of 40-100 ~mol/l over pretreatment values. In contrast, all our patients had serial samplesassayed by the more sensitive radioimmunoassay technique [8,9], and elevations in serum creatininecomparedwith baseline were usually ~40 ~moll1. In addition, patients with fungal infections treated with itraconazole were included as controlsforthe effects of chronicinfection on EPOproduction. MacGregor et al. [3] found detectable EPO levels only in samples from a patient 30 days after completion of therapy. Though we did not acquire serial samples after discontinuation of AmB, our available results parallel theirs. Specifically, detectable elevations in EPO were not seen until 11 and 13 days after therapy in two of our patients. Thus, the EPO detected by MacGregor et al. [3] probably reflects a level increasing to values detectable by bioassay. The EPO resurgence noted at days 11 and 13 after treatment in our patientscorresponds to the proposed I5-day halflife of intravenous AmBin humans [10]. Serum levelsof AmB in these two patients, which do not necessarily reflect the amountof drug presentin peripheralcompartments, decreased rapidly from a value 1 day after treatment of 0.82 ~g/m1 to ?.11 ~g/m1 in. 13days,in patient 1,and from 0.82to 0.08 ~g/m1 m 11 days, m patient 2. The decision whether to transfuse the patient was made by the patient'spersonalphysician. Because some of our patients were transfused, the possibility that hypertransfusion polycythemia eliminated endogenous erythropoietin production in response to hypoxiamust be considered [8]. Our patients, nevertheless, were not transfused to polycythemic levels, and the effects on EPO levels were seen before transfusion and continued weeks after. The means by which AmB inhibits EPO production is unknown, but it may be due to effectson the kidney, the organ
Downloaded from http://jid.oxfordjournals.org/ at Cambridge University on September 4, 2015
In contrast, the EPO measured in the itraconazole-treated controls (table 2) was physiologic. Control 1 had persistently elevated EPO levelsin response to continued anemia throughout therapy. Controls 2 and 3 had a progressive decrease in EPO levels as expected with rising hemoglobin levels. Neither the schedule of AmB administration nor the levels of AmB attained during therapy correlated with the degree of anemiaor suppressionof EPO response. Patient 1 received daily AmB throughout his treatment: 30 mg/day the second week and after. Except for the peak level drawn shortly after an infusion at week 3, serum levels were determined just be- . fore the next dose. Patient 2 was given increasing daily doses of AmB for the first week and 50 mg every other day thereafter. The levels reported were trough levels obtainedjust beforethe nextdose of AmB. After a weekof daily AmB,patient 3 was maintained on 50 mg of AmB thrice weekly. All levels were drawn on Monday morning, f\J72 h after the last dose. Of the levels, 16 of 21 were
Amphotericin B Blunts Erythropoietin Response to Anemia Andrew C. Lin, Eugene Goldwasser, Edward M. Bernard, and Stanley W. Chapman
From the University of Mississippi Medical Centerand the l'eterans Administration Medical Center, Jackson,' the University of Chicago, Illinois; and Memorial Sloan Kettering Cancer Center, New lbrk, New lbrk
Normochromic normocytic anemia during amphotericin B (AmB) treatment may be due to effects on erythropoietin (EPO) production or to direct marrow suppression. In cell cultures, erythroid colony growth is inhibited by AmB concentrations of ~1.0 p.g!J;p.I, probably by its direct effecton progenitorcells [1], with enhancedgrowthat lowerconcentrations [2]. In 1978, using a bioassay method to detect EPO in the urine and serum of patients on amphotericin B (AmB), MacGregor et al. [3] found no elevationin EPO concentration at the nadir of the anemia. Consequently, they proposed that the anemia was due to AmB's effecton EPO production. To further our understanding of the interactions betweenAmB, EPO, and anemia, we evaluatedthe effectof AmB on EPO production in patients receiving prolonged AmB therapy.
Patients and Methods Patients receiving AmB for ~1 month at the University of Mississippi MedicalCenter or at the JacksonVeterans Administration Medical Center were included in the study.Those with preexisting renal disease (serum creatinine >130 JLmol/I), AIDS, or active hematologic malignancy were excluded. To control for the effects of systemicfungal disease on EPO response, patients treated with an investigational imidazole, itraconazole, were also studied serially.
Received 8 June 1989; revised 31 August 1989. Presented in part at the 1987 American Federation for Clinical Research Southern Section meeting, New Orleans, 28 January 1987 (abstract 78). Partly supported by Veterans Administration research funds. Reprints and correspondence: Dr. Stanley W. Chapman, Division of Infectious Diseases, Department of Medicine, University of Mississippi Medical Center, Jackson, MS 39216-4505. The Journal of Infectious Diseases 1990;161:348-351
© 1990 by The University of Chicago. All rightsreserved. 0022-1899/90/6102-0032$01.00
Of the three study subjects, one had cryptococcal meningitis and diffusehistiocytic lymphomain remission by standardstagingcriteria, one had cryptococcalmeningitis, and one had disseminated blastomycosis. One control had chronic cavitary histoplasmosis, Wolff-Parkinson-Whitesyndrome, and Mycobacterium scrofulaceum infection, one had pulmonary blastomycosis, and one had chronic cavitary histoplasmosis, chronic bronchitis, and Meniere'sdisease. No differences in severity of illness between AmB- and itraconazole-treated patients could be identifiedclinically. All were ambulatory and displayed microbiologic and clinical resolution of infectionon therapy. The daily and total dose of AmB, time of the dose before phlebotomy, and time of phlebotomy were recorded. Blood was drawn for baseline AmB and EPO levels at the start of treatment and at frequent intervals (weekly if possible) during therapy. Routine serum chemistry and hematologic studies were also done. The corrected reticulocyte count wascomputedusing the formula: reticulocyte count X (hematocrit/0.45) + correction factor [4]. The sera were divided into 1- to 2-m1 aliquots, labeled with a number code, and immediatelyfrozenat -30°C pendingdeterminationofEPO and AmB levels. Blindedsera wereanalyzedfor AmB levelsusinghigh-performance liquid chromatography as described [5]. Sera for erythropoietinlevels were determined using radioimmunoassay as previously described [6]. In brief, purifiederythropoietinstandardor coded unknownsamples were mixed with 12SI-labeled EPO and incubated (4 days at 4°C) with rabbit antisera to EPO in a phosphate-bufferdiluent. The free and bound 12SI-EPO were separated by a second antibody using goat anti-rabbit serum and normal rabbit serum as a carrier. Total counts per minute per tube was determined by counting randomly sampledtubes. Sampleswerecentrifugedand the pelletswashedtwice with diluent before counting in an automated gamma counter. The percentage of bound 12SI_EPO was then calculated. This percentage was corrected for nonspecific precipitation and all samples were tested in quadruplicate and at different volumes. Regression curves relating log[EPO] standard to percentage bound were determined and EPO values falling in the linear part of the curve were used to calculate the EPO content of the samples. The interassay coefficient of variation was 6%.
Downloaded from http://jid.oxfordjournals.org/ at Cambridge University on September 4, 2015
Amphotericin B causes a normochromic, normocytic anemia thought to be mediated by direct marrow toxicity or suppression of erythropoietin production. Serial hemoglobin, hematocrit, amphotericin B, and erythropoietin levels were determined before, during, and after completion of amphotericin B therapy for three patients without significant renal disease or active hematologic malignancy. Patients with systemic fungal diseases treated with itraconazole served as controls. Serum erythropoietin levels were determined by radioimmunoassay and amphotericin B by high-performance liquid chromatography. Despite anemia in all amphotericin B-treated patients, erythropoietin levels declined or remained relatively constant during therapy while erythropoietin levels in controls were appropriate for the degree of anemia. Within 2 weeks of completion of amphotericin B treatment, two patients had increasing erythropoietin levels in response to anemia. Amphotericin B appears to suppress but not abolish the erythropoietin response to anemia; this effect disappears quickly after discontinuation of the drug.
Concise Communications
JIO 1990;161 (February)
To verify that AmB did not interfere with the EPO assay, AmB in concentrations of 0.01-5 /Lg/ml was added to duplicate serum samples and assayed simultaneously with untreated serum. No differences in EPO measurements were produced by AmB in the serum (data not shown).
Results
noted at 3 weeks of therapy despite relatively unchanged serum creatinine levels. Although the creatinine increased during therapy, the blunting of EPO response continued despite stabilization of renal function. Patient 2 had a transient increase in serum creatinine in the first week of therapy, but this normalized after hydration. Erythropoietin levels, however, did not increase in the face of anemia even with a normal serum creatinine. Likewise, patient 3 also had a fall in EPO although serum creatinine remained relatively constant throughout. Erythropoietin levels done 4 and 5 days after discontinuing AmB in patients Z and 3, respectively, were unchanged from the last treatment sample. Patient 2 had serum assayed 11days after therapy, and a small increment in EPO production was noted. Finally, the EPO level for patient 1 determined 13 days after treatment showed a definite increase. Commensurate with the increasing EPO level, hemoglobin increased from 98 to 113 gil.
Table 1. Serial hemoglobin, hematocrit, amphotericin B, and erythropoietin levels in patients before, during, and after completion of treatment with amphotericin B. Subject, testing interval (w unless indicated)
Amphotericin B Hemoglobin (gIl)
Total dose (mg)
69 87 99 94 97 103 98
140 890 1340 1490 1640 1790 1890
Levels (ug/ml)
Erythropoietin Creatinine (umol/l) (munits/ml)
1.26 0.26 0.14 0.14
43 25 21 7
0.17
13
0.19 0.11
11 11
Downloaded from http://jid.oxfordjournals.org/ at Cambridge University on September 4, 2015
Table 1 shows the serial results for the three AmB-treated patients. Patients 1 and 3 had greater-than-normal EPO titers at the beginning of the study in response to their anemia. In both, the EPO fell during AmB therapy despite continued anemia. Patient 2 had low-normal EPO at the start of therapy, reflecting her normal hemoglobin. Despite a decline in hemoglobin, there was no compensatory rise in EPO. No consistent relationship between renal function, as measured by serum creatinine, and EPO production could be identified. In patient 1, a definite decrease in EPO titer was
349
350
Concise Communications
!able. 2. Serial he~oglobin, hematocrit, and erythropoietinlevels In patients treated With itraconazole (controls). Subject, testing interval (w) Control 1 0 4 13 23 43 66 Contro12 4 11 22 32 Control 3 0 3 5 9 18 22 26 31
Hemoglobin (gil)
Erythropoietin (munits/ml)
120 67 46 103 31 112 55 111 58 118 (Itraconazole ended after 52 w) 50 118
Creatinine
(umol/l) 90 70 90 80 100
NO
13 144 177 5 184 4 (Itraconazole ended after 32 w) 170 3
NO
37 33 30 19 11 23 17 23 35 w)
70 60 60 60 70
102 113 122 124 140 138 132 147 (Itraconazole ended after
90
NO 90
NO 70 60
Discussion In vivo and in vitro evidence points to erythropoietin as the ~ri~cip~ hu~oral regulator of erythropoiesis [7]. Though limited in subjectstested, our study is the first serial observa-
tion to our knowledge of EPO and AmB in patients during treatment. Our data indicate that blunting of normal erythropoietin production contributes to the anemia accompanying AmB therapy in humans. In our patients, the decreased EPO levelsoccurred after 3 weeksof therapyand were independent of the total dose of AmB given, the AmB level at the time of EPO determination, and renal function as reflected by serum creatinine. The effect also seemed to be independentof the underlying fungaldisease, as a physiologic EPO response was observed in our controls. AmB did not interfere with the EPO assay as evidenced by our results with AmB-spiked and unaltered sera. MacGregor et al. [3], using only single samples collected immediately and up to 30 daysafter completionof AmB therapy, were unable to detect EPO levels in five of six serum and three of four urine samples. Three patientshad increments in serum creatinine after therapy of 40-100 ~mol/l over pretreatment values. In contrast, all our patients had serial samplesassayed by the more sensitive radioimmunoassay technique [8,9], and elevations in serum creatininecomparedwith baseline were usually ~40 ~moll1. In addition, patients with fungal infections treated with itraconazole were included as controlsforthe effects of chronicinfection on EPOproduction. MacGregor et al. [3] found detectable EPO levels only in samples from a patient 30 days after completion of therapy. Though we did not acquire serial samples after discontinuation of AmB, our available results parallel theirs. Specifically, detectable elevations in EPO were not seen until 11 and 13 days after therapy in two of our patients. Thus, the EPO detected by MacGregor et al. [3] probably reflects a level increasing to values detectable by bioassay. The EPO resurgence noted at days 11 and 13 after treatment in our patientscorresponds to the proposed I5-day halflife of intravenous AmBin humans [10]. Serum levelsof AmB in these two patients, which do not necessarily reflect the amountof drug presentin peripheralcompartments, decreased rapidly from a value 1 day after treatment of 0.82 ~g/m1 to ?.11 ~g/m1 in. 13days,in patient 1,and from 0.82to 0.08 ~g/m1 m 11 days, m patient 2. The decision whether to transfuse the patient was made by the patient'spersonalphysician. Because some of our patients were transfused, the possibility that hypertransfusion polycythemia eliminated endogenous erythropoietin production in response to hypoxiamust be considered [8]. Our patients, nevertheless, were not transfused to polycythemic levels, and the effects on EPO levels were seen before transfusion and continued weeks after. The means by which AmB inhibits EPO production is unknown, but it may be due to effectson the kidney, the organ
Downloaded from http://jid.oxfordjournals.org/ at Cambridge University on September 4, 2015
In contrast, the EPO measured in the itraconazole-treated controls (table 2) was physiologic. Control 1 had persistently elevated EPO levelsin response to continued anemia throughout therapy. Controls 2 and 3 had a progressive decrease in EPO levels as expected with rising hemoglobin levels. Neither the schedule of AmB administration nor the levels of AmB attained during therapy correlated with the degree of anemiaor suppressionof EPO response. Patient 1 received daily AmB throughout his treatment: 30 mg/day the second week and after. Except for the peak level drawn shortly after an infusion at week 3, serum levels were determined just be- . fore the next dose. Patient 2 was given increasing daily doses of AmB for the first week and 50 mg every other day thereafter. The levels reported were trough levels obtainedjust beforethe nextdose of AmB. After a weekof daily AmB,patient 3 was maintained on 50 mg of AmB thrice weekly. All levels were drawn on Monday morning, f\J72 h after the last dose. Of the levels, 16 of 21 were
