Mineralocorticoid Insufficiency Due
`to Suramin Therapy
`
`Ken Kobayashi, M.o."’
`Roy E. Weiss, M.D., Ph.D.3
`Nicholas J. Vogelzang, Mo.“
`Everett E. Vokes, M.n.""
`Linda Janisch, n.n., M.s.ii.‘
`Mark J. Ratain, M.o."*-4
`
`‘ Committee on Clinical Pharmacology, Depart-
`ment of Medicine, University of Chicago Pritzker
`School of Medicine, Chicago, Illinois.
`
`2Section of Hematology/Oncology, University
`of Chicago Pritzker School of Medicine, Chi-
`cago, Illinois.
`
`3 Section of Endocrinology, University of Chicago
`Priizker School of Medicine, Chicago, Illinois.
`
`“ University of Chicago Cancer Research Center,
`University oi Chicago Pritzker School of Medi-
`cine, Chicago, Illinois.
`
`BACKGROUND. During a Phase I trial of suramin, a novel antineoplastic agent with
`activity against hormone-refractory prostate carcinoma, the authors observed two
`patients with clinical mineralocorticoid insufficiency in spite of hydrocortisone
`replacement therapy.
`METHODS. The authors retrospectively assessed adrenal cortical function in 20 such
`patients via adreriocorticotropic stimulation testing, measuring both cortisol and
`aldosterone responses, either at the time of treatment or immediately after discon-
`tinuation of treatment.
`
`RESULTS. Two of 9 patients (22%) treated at relatively low dose levels (S1200 mg/
`m2 on Day 1) demonstrated adrenal cortical insufficiency, as compared with 9 of
`11 patients (82%) treated with relatively high doses (>1200 mg/m2 on Day 1) (P
`= 0.03 by 1-tailed Fisher's exact test). There appeared to be a cumulative dose-
`response relationship to the development of glucocorticoid insufficiency, with no
`instances being observed at doses < 4.8 g/m2 and uniform toxicity occurring at
`doses > 7.6 gm”. Long term glucocorticoid insufficiency was present in 1 of 5
`patients (20%) tested at an interval of >90 days after discontinuation of suramin
`treatment. All instances of glucocorticoid insufficiency were associated with miner-
`alocorticoid insufficiency. Suramin did not affect the absorption or excretion of
`exogenously administered glucocorticoid in one patient.
`CONCLUSIONS. Suramin Causes both primary mineralcorticoid and primary glucocor-
`ticoid insufficiency. This may occur in a dose-dependent manner. Long term glucocor-
`ticoid insufficiency appears to occur in a minority of patients treated with low doses
`of suramin. Patients receiving high doses of suramin for treatment of advanced carci-
`noma should receive at least physiologic replacement doses of both mineralocorticoid
`and glucocorticoid. Higher doses of glucocorticoid may be required in selected pa-
`tients. Cancer 1996; 78:2411—20. © 1996 American Cancer Society.
`
`KEYWORDS: suramin, fludrocortisone, adrenal cortical insufficiency, glucocorticoid,
`mineralocorticoid, dose response, dose toxicity, hydrocortisone.
`
`Primary adrenal insufficiency occurs as a result of treatment with
`several agents,
`including aminoglutethimide,‘ mitotane (o,p’—
`
`Presented in part at the 76th Annual Meeting oi
`The Endocrine Society, Anaheim, Caliiornia, June
`15-18, 1994.
`
`the Klein Family Fund.
`
`Supported by Grant #2—T32-CA—09566—06 and
`Contract #N01-CM-07301 from the National Cancer
`institute, Grant #5-T32-GM070i 918 from the Na-
`tional lnstitute oi General Medical Sciences.
`
`Dr. Vogelzang is supported in part by the Foster
`Schlutz Memorial Fund for Cancer Research and
`
`© 1996 American Cancer Society
`
`is the recipient of an American
`Dr. Kobayashi
`Cancer Society Physicians’ Research Training
`Award (#PRTA-15). Dr. Kobayashi‘s current ad-
`dress: Division oi Oncology Drug Products,
`Food and Drug Administration, HFD—150,5600
`Fishers Lane, Rockville, MD 20857.
`
`Dr. Weiss is supported by Grant #DK02081
`
`from the National Institutes of Health and by
`Morris and Cynthia Belzberg.
`
`reprints: Nicholas J. Vogelzang,
`Address for
`M.D., Section
`of Hematology/Oncology.
`MC2115, University of Chicago Medical Center,
`5841 S. Maryland Ave., Chicago,
`IL 60637.
`
`Received March 15, 1996; revision received July
`1, 1996; accepted July 5, 1996.
`
`Amerigen Exhibit 1098
`Amerigen Exhibit 109
`Amerigen v. Janssen IPR2016-00286
`Amerien v. Janssen |PR2016—0028
`
`

`
`2412
`
`CANCER December 1, 1996/ Volume 78 I Number 11
`
`DDD),2 itraconazole? ketoconazolef metyrapone?
`etomidate,5 rifampin,7"° treosulfan,” heparin,‘2‘“‘ tri-
`lostane,” and suramin.” These drugs in general exert
`their adrenocortical effects by inhibiting the activity
`of various enzymes involved in glucocorticoid and
`mineralocorticoid biosynthesis, such as 11,6- and 18/3-
`hydroxylase, cholesterol side-chain cleavage enzyme,
`and 17,20-lyase. Suramin, a new antineoplastic agent,
`appears to interfere with the binding and effects of
`autocrine growth factors, and has attracted much at-
`tention by virtue of its activity in treating hormone-
`refractory metastatic prostate carcinoma. Suramin
`causes adrenal cortical necrosis and adrenal insuffi—
`
`ciency in some patients and as a result, all trials to
`date have routinely employed replacement doses of
`hydrocortisone. However, mineralocorticoid supple-
`mentation is not routinely prescribed. While conduct-
`ing a Phase I study of suramin, the authors observed
`two instances of addisonian crisis occurring despite
`presumably adequate replacement doses of hydrocor-
`tisone. Both patients underwent adrenocorticotropic
`hormone (ACTH) stimulation testing and were found
`to have severe mineralocorticoid insufficiency. There-
`fore a retrospective analysis was undertaken of adrenal
`function in surviving patients from whom verbal in-
`formed consent was obtained. In this report, those
`findings are described and guidelines for glucocorti—
`cold and mineralocorticoid replacement in patients
`receiving suramin are suggested.
`
`PATIENTS AND METHODS
`
`The protocol for suramin (described elsewhere”) re-
`quired that all patients receive hydrocortisone, 25 mg
`orally every morning and 10 mg orally every evening
`while receiving suramin by intermittent infusion. Pa-
`tients remained on this dose of hydrocortisone for the
`duration of suramin treatment and, except for the two
`cases noted, continued on this dose until the time
`
`of endocrine testing or death. Informed consent was
`obtained from all patients prior to entry into the proto-
`col and prior to adrenal function testing. In view of
`suramin’s extremely long half-life of 30-50 days, the
`dosing scheme, unlike previous studies, gradually used
`decreasing doses to maintain constancy of the peak
`plasma levels. These decreases were made in accor-
`dance with a prespecified scheme, rather than mea-
`sured plasma levels. A standard Phase I close escala-
`tion design was employed, in which patients were en-
`rolled successively in cohorts receiving gradually
`increasing doses of suramin until a maximally toler-
`ated dose was attained. Each dose level was expanded
`as necessary to more precisely characterize the toxicity
`profile of this drug as given in this dosing schedule.
`Eight dose levels (400, 600, 800, 1000, 1200, 1440, 1730,
`
`and 2080 mg/m2 on Day 1) were explored. The initial
`cohort, at a dose of 400 mg/m2 on Day 1, received a
`total of 960 mg/m2 of suramin over the 1st month; the
`final cohort, at a dose of 2080 mg/m2 on Day 1, re-
`ceived a total dose of 4992 mg/m2 over the first month.
`Sixty-three patients were treated according to a sched-
`ule in which progressively decreasing doses of suramin
`were administered on Days 1, 2, 8, and 9 of each 28-
`day cycle. This regimen was designed to avoid gradu-
`ally increasing peak plasma levels over the treatment
`course. An additional 13 patients were treated using a
`modification of the original schedule (schedule A), in
`which the 2 doses in each week were consolidated into
`
`1 dose (i.e., Days 1 and 2 were administered on Day
`1, and Days 8 and 9 were administered on Day 8).
`Although this modification simplified the logistic as-
`pects of the dosing scheme, the total monthly amount
`of suramin was identical in both schedules. Written
`
`informed consent was obtained from all patients, in
`accordance with federal and institutional guidelines.
`Because the majority of patients enrolled in the
`Phase I study were men with hormone—refractory met-
`astatic prostate carcinoma, the baseline evaluations
`required by the protocol did not routinely include ab-
`dominal computed tomography (CT) scanning. How-
`ever, CT scans were performed as clinically indicated
`for those prostate carcinoma patients with known ex-
`traosseous disease and routinely for those patients
`with malignancies other than prostate carcinoma.
`Endocrinologic evaluations were not mandated by
`the protocol; however, posttreatment ACTH stimula-
`tion testing was discussed with patients as they discon-
`tinued treatment to assess their need for ongoing glu-
`cocorticoid replacement. At the time this study was
`initiated, all surviving patients who had discontinued
`suramin treatment were contacted and asked to un-
`
`dergo ACTH testing. Twenty of the 76 patients origi-
`nally enrolled agreed to undergo testing with the rapid
`ACTH test, assessing both cortisol and aldosterone re-
`sponse.2° The remaining 56 patients were either lost
`to follow—up, had died prior to being contacted, or
`refused endocrine testing. All 20 patients were ambu-
`latory outpatients who were able to present them-
`selves to either their local medical center or to the
`
`University of Chicago Hospitals for testing.
`Patients withheld their doses of hydrocortisone for
`24 hours prior to ACTH stimulation testing. After ob-
`taining baseline ACTH, aldosterone, and cortisol lev-
`els, 0.25 mg of cosyntropin was injected intravenously.
`Cortisol and aldosterone levels were measured at 30
`
`and 60 minutes after injection. ACTH, cortisol, and
`aldosterone levels were measured in the hospital clini-
`cal laboratories using commercially available radioim-
`munoassay kits (ACTH: Nichols Institute Diagnostics,
`
`

`
`San Juan Capistrano, CA; cortisol and aldosterone:
`Coat—A—Count, Diagnostic Products Corporation, Los
`Angeles, CA). The assay limits of sensitivity were 16
`pg/mL for aldosterone and 0.2 pg/dL for cortisol. A
`test was considered abnormal if the serum cortisol at
`60 minutes after administration of ACTH failed to rise
`
`by >7 pg/dL over the baseline and if it failed to reach
`a level of 18 ug/dL.2‘ Aldosterone stimulation testing
`was considered to be abnormal if the baseline level
`
`was below 5 ng/mL and if the level at 60 minutes failed
`to rise by 5 ng/mL over the baseline value.” The results
`of ACTH stimulation testing using a similar protocol in
`19 normal volunteers were kindly furnished by Nichols
`Institute Diagnostics (Jerrold Nelson, personal com-
`munication). ACTH, cortisol, and aldosterone concen-
`
`trations in these volunteers were measured using the
`same methods as already described.
`Suramin plasma concentrations were determined
`using a high—performance liquid chromatography
`assay method as previously described.”
`The design of the bioavailability study called for
`the oral administration of 20 mg of hydrocortisone on
`Day 1, followed by hourly collection of plasma samples
`over the next 8 hours. On Day 2, 20 mg of hydrocorti-
`sone was administered intravenously and hourly sam-
`pling was again obtained. Cortisol concentrations
`were determined using the same assay as described
`above.
`
`Examination of scatter plots relating absolute
`change and percentage change after ACTH stimulation
`to the baseline values of cortisol and aldosterone
`
`showed a strong dependence of the percentage change
`on baseline, whereas absolute change was relatively
`independent of the baseline. Accordingly, absolute
`change was used in the statistical analyses.” One—way
`analysis of variance with Bonferroni’s correction for
`post hoc analyses, Spearman’s rank order correlations
`(r), logistic regression analyses, nonparametric tests,
`chi—square, and Student's ttests were performed using
`SPSS for Windows, version 6.1.2.23 Two-compartment
`linear models were fit to the bioavailability data using
`the nonlinear least-squares fitting program PCNON—
`LIN, version 4.0 (Scientific Consultants, lnc., Lexing-
`ton, KY). Systemic bioavailability, quantified as the ra-
`tio of the area under the concentration x time curve
`
`(AUC) after oral administration to the AUC after intra-
`
`venous administration,24 was calculated using the lin-
`ear trapezoidal method with extrapolation to infinity.
`
`RESULTS
`
`ACTH Stimulation Testing
`Table 1 shows selected characteristics of the two pa-
`tient populations. The median age at the start of sura-
`min therapy differed significantly (P = 0.03), as did
`
`Suramin and Hypoaldosteronism/Kobayashi at al.
`
`2413
`
`the total received dose of suramin (P = 0.03), and the
`last recorded suramin concentration prior to ACTH
`stimulation testing (P = 0.002). This last result, how-
`ever, should be interpreted with caution, because the
`plasma concentrations could not be determined at a
`uniform time relative to the time of ACTH testing. The
`time from the last dose of suramin to the time of ACTH
`
`stimulation testing was not significantly different be-
`tween the two groups (P : 0.06).
`The ACTH stimulation test results of the 20 pa-
`tients are shown in Table 2. The occurrence of abnor-
`mal ACTH stimulation tests correlated with dose level
`
`(Spearman’s r = 0.67; P = 0.001) (Table 2). This is most
`strikingly illustrated by the fact that only 2 of 9 (22%)
`patients treated below the 1200 mg/m2 on Day 1 dose
`level who underwent ACTH stimulation testing after
`discontinuation of suramin had abnormal cortisol and
`
`aldosterone responses, whereas 9 of 11 (82%) patients
`treated at or above the 1200 mg/m2 dose level had
`abnormal responses. All patients who demonstrated
`impaired glucocorticoid responses to ACTH stimula-
`tion also demonstrated impaired or absent mineralo-
`corticoid responses, indicating defects in both miner-
`alocorticoid and glucocorticoid function.
`Table 3 compares the summary results of ACTH
`stimulation testing performed on 19 normal control
`subjects (J. Nelson, personal communication) with the
`results of the current study patients taken in their en-
`tirety. The extent of glucocorticoid responsiveness to
`ACTH stimulation, measured by the delta (A) cortisol,
`had a mean for the control subjects of 15 : 4.2 pg/
`dL, as compared with 6.4 : 6.8 pg/mL for all patients
`irrespective of dose level (P < 0.001). Considerable
`variability in suramin’s pharmacologic effect on the
`adrenal gland was noted among the patients; the effect
`of dose accounted for a large part of this variability,
`because A cortisol correlated well with dose cohort
`
`(Spearman’s r = -0.74; P < 0.001). When patients
`treated at doses from 400 mg/m2 on Day 1 to 800 mg/
`m2 on Day 1 were grouped together and compared
`against patients treated at higher dose levels and
`against control subjects (Table 3), statistically signifi-
`cant differences (P = 0.05) between patients and nor-
`mal controls were found in the aldosterone responses
`to ACTH stimulation.
`
`In comparing the cumulative received doses of
`those patients with and without abnormal responses
`to ACTH stimulation (Fig. 1), it appears that at doses
`below 9.6 g (4.8 g/mg), the incidence of adrenal insuf-
`ficiency is minimal, whereas at doses above 17 g (7.6
`g/m2) there is uniform occurrence of adrenal insuffi-
`ciency. At intermediate doses, other factors besides
`drug effect may be important. Logistic regression anal-
`ysis evaluating the effect of cumulative dose on the
`
`

`
`2414
`
`CANCER December 1, 1996 I Volume 78 I Number 11
`
`Glucocorticoid
`
`Mineralocorticoid
`
`Yes
`
`10
`60
`(34—76)
`4.1
`(3.8—4.7l
`1.2
`(1.l—?.)
`90
`180-1001
`13.3
`01.2-14.9)
`3.5
`(1-10)
`16,615
`(3994—2l,908)
`9134
`(1920-11,064)
`123.1
`(36.1—237.4l
`35
`[9—362l
`
`Yes
`
`10
`60
`(34-71)
`4.2
`(3.8—4.7)
`1.2
`(1-2)
`80
`(80-100)
`13.3
`(8.2—l4.9)
`
`3(
`
`2-4)
`15,586
`(9648—21,908)
`8349
`(4824-11,065)
`168.5
`(46.4—237.4)
`34
`(9— 145)
`
`TABLE 1
`Patient Characteristics
`
`Characteristic
`
`No. of patients
`Age (yrs)
`
`Albumin, mg/dL
`
`Creatinine, mg/dl.
`
`KPS, %
`
`Hemoglobin, g/dL
`
`No. of cycles
`
`Total dose, mg
`
`Total dose, mg/mi
`
`Last concentration, pg/dl.
`
`Time from last dose to endocrine testing, days
`
`KPS: Karnofsky performance status.
`“ Mann—Whitney U test.
`
`occurrence of an abnormal ACTH stimulation test
`
`received dose,
`confirmed the importance of total
`whether expressed as the raw total or as the total dose
`normalized to body surface area (Table 4). The odds
`of having an abnormal cortisol response to ACTH
`stimulation is increased 216% for each g/m2 of admin-
`istered suramin (P < 0.005). Although a significant
`proportion of the data on aldosterone response to
`ACTH stimulation were missing, similar analyses did
`not demonstrate a statistically significant effect of
`dose.
`
`No association of tumor response either for all
`patients or for only the prostate carcinoma patients
`with impaired glucocorticoid or mineralocorticoid re-
`sponse could be found (Table 2). Baseline CT scans
`were available in seven patients (one with breast carci-
`noma, two with sarcoma, and four with prostate carci-
`noma), including the two patients mentioned above.
`The adrenal glands in five patients were normal. Ab-
`normalities of the adrenal gland were noted in two
`patients, one of whom (Patient 416), demonstrated a
`normal cortisol ACTH stimulation test. Metastasis to
`
`tion and parenteral fluid and electrolyte support for
`addisonian crises and continued to have severe adre-
`
`nal insufficiency despite the use of increasing doses of
`glucocorticoid. The need for such intensive measures
`largely abated after the institution of mineralocorti-
`coid replacement with fludrocortisone.
`
`Bioavailability of Hydrocortisone
`A patient (Patient 456) with repeated episodes of addi-
`sonian crisis in spite of fludrocortisone and hydrocor-
`tisone replacement therapy underwent a bioavailabil-
`ity study to assess his ability to absorb orally adminis-
`tered hydrocortisone. As can be seen in Figure 2, after
`an oral dose of hydrocortisone (20 mg) and an intrave-
`nous dose of hydrocortisone (20 mg), the concentra-
`tion time profile is identical, demonstrating an unim-
`paired ability to absorb hydrocortisone. Consistent
`with this, the systemic bioavailability of hydrocorti-
`sone was 100%. The elimination half-life of hydrocorti-
`sone was also normal.
`
`DISCUSSION
`
`the adrenal glands developed over the course of the
`study in one other patient (Patient 476), who demon-
`strated impaired glucocorticoid responsiveness to
`ACTH stimulation. Two patients required hospitaliza-
`
`The occurrence of adrenal glucocorticoid insufficiency
`during treatment with suramin has been described
`previously” and is well known. However, mineralocor—
`ticoid insufficiency has been hitherto unsuspected,
`
`

`
`Suramin and Hypoaldosteronism/Kobayashi et al.
`
`241 5
`
`
`
`
`
`:23$23ea.:253%.::23E...:23E...2.53as=.5<2sans.3.ES23%$5.5:as8Name...%_§..:.5
`
`
`
`
`
`8.8Em3;3.2mmas338cs8»Ev
`
`3%gm982onE228%08ozEv
`
`2.2EN523.3mmmmNBESgmEm?
`
`<2<28.5«.3ES823.23-2VBENmm2Sm:8835._.§NS8.3fm«ZNIE85823.2:Wm8.23:Sis32o3
`
`3.33:2.:§EN825.23-S32NEE:88Ev
`
`B213.02Na:825?
`
`<2<28.2Emasme.33anozas
`
`.802.8?2$32%
`
`<2<2322V«N?at.
`
`_8.m-2V.03-Q2at2;
`
`_o2v.282asom:2%
`
`8.:3.2$2S:8523
`
`R.3-588oz9:»L.3893$2:8:5.o.<..oas33%om:3»ma.
`
`<233-E3%2.5ozmy
`
`
`
`
`
`_aE8_§8..can_:~E8:$9%258::mmv2:2528E85«E0528Samoa3235:3=<E
`
`
`
`
`
`.wsm=§«8:“<2..2§§o:m325.59.5:o.aE.8E8o=2E_:5<
`
`
`
`
`
`.3:E95ESE2:52:5:8uaeazEfiaam.£Eo=E8$50:2.333222338m.32E.88:52;.2538Soc5%.;958%2283E.3$3m8SEE.:55528mzamoaS.umcenmmmL
`
`
`
`
`
`
`
`
`
`.53mmumEo.Eo..S>wo_2_:uov:m=_6<.35.8Emsana2:E$25_EEoz.
`
`
`
`
`
`
`
`
`
`_e:oE§_gmozoanoooceumare$3588EswanesoE§B£8=.:£§__
`
`
`
`
`
`
`
`dambafimoee@3358so2%=22?_«E§E<9
`
`.33EEoE_<.
`
`
`
`
`
`ucE<<3£5.2::c3.2M__8_:eu<8_S_:ouc_8_:cou_.Fu<=c=~.:=8:8=ouE.:8=ou.3.E9:fineSouH85Eieu.5553:02am
`
`
`
`
`
`
`
`Se...35
`
`
`
`
`
`=_==:=m.233.35.3:33..8E53any=o_§=E=m=,5<Ea.=2E8.§Nm.__m<~
`
`

`
`2416
`
`CANCER December 1, 1996 I Volume 78 I Number 11
`
`TABLE 3
`
`Changes In Adrenal Function By Cohorta
`
`AAldo
`Aldo“ 0 mins
`Aldo“ so
`mins (pg/dL)
`(ngldL)
`(us/dL)
`mins (pig/dL)
`No.
`Patient ID
`
`(ngldL] Zj—._::_
`mins (ng/dL)
`
`Cortisol 60
`
`ACortisol
`
`Cortisol 0
`
`19
`
`20
`
`Normals
`
`All patients
`
`"Low dose cohorts"
`
`10.6 t 4.4
`25.7 : 7.2
`7.2 : 4.1
`11.8 : 5.3
`4.5 : 2.2
`15 : 4.2
`14.2—4l
`5.6—l9.4
`0-14
`5-20
`2-9
`75-25
`9.9 : 6.3
`16.3 : 11.4
`4.5 : 7.4
`10.7 : 11.7
`5.4 : 4.5
`5.4 : 6.8
`< l—2l.8
`15-33
`-3.5-20.7
`1-354
`< 1-143
`-0.8-19.7
`11.1 : 6.3‘
`20.6 : 10.8‘
`11.1 : 10.1“
`14.9 : 12.8“
`6.8 : 4.5“
`9.5 : 6.4"
`1—2l.B
`23-33
`0-207
`2.3-35.4
`1.3-14.8
`0-19.7
`8.1 : 6.2“
`9.9 : 9.69
`2 : 14.5‘
`5.3 : 8‘
`3.6 : 3.8‘
`18 : 45*
`“High dose cohorts"
`l.5~29.5
`I.3—l6.9
`
`
`
`< 1-115 1-232—O.8—I2.6 -3.5-11.5
`
`Aldo: aldosterone.
`“ Data expressed as mean : standard deviation; range is indicated in the second row.
`"Normal range at the University of Chicago Adult Endocrinology Laboratory.
`‘ Includes patients treated in cohorts receiving 600—I2t)t) mg/mi on Day I in : I2).
`" Includes patients treated in cohorts receiving 600-800 mg/m‘ on Day I (n = 5}.
`“Includes patients treated in cohorts receiving 1440-2080 mg/mg on Day I (n = 9).
`‘Includes patients treated in cohorts receiving I000—2080 mg/mi on Day I tn = 15].
`
`TABLE 4
`
`Results of Logistic Regression Analyses
`
`Outcome
`
`Covariate
`
`Abnormal cortisol response
`Abnormal cortisol response
`Abnormal aldosterone response
`Abnormal aldosterone response
`
`Total dose
`Normalized total dose
`Total dose
`Normalized total dose
`
`Coefticient
`
`1.4085
`2.1576
`1.1988
`1.4460
`
`and this report is the first to demonstrate the existence
`of a clinically significant mineralocorticoid defect after
`suramin therapy. In addition, the authors call atten-
`tion to the observation that some patients have intact
`glucocorticoid function after suramin therapy.
`The fact that two patients required hospitalization
`and parenteral fluid and electrolyte support for severe
`adrenal insufficiency despite taking adequate replace-
`ment doses of glucocorticoid demonstrates the clinical
`significance of the mineralocorticoid defect, because
`the need for hospitalization was at least in part caused
`by insufficient mineralocorticoid replacement. The
`need for combined mineralocorticoid and glucocorti-
`coid replacement in patients with Addison's disease
`has been emphasized by a report in which apparent
`glucocorticoid-resistant adrenal insufficiency in two
`patients appeared to stem, at least in part, from inade-
`quate mineralocorticoid replacement rather than from
`a state of primary glucocorticoid resistance.“ Although
`inherited glucocorticoid resistance in humans have
`been reported in certain kindreds,” it is rare and no
`instances of acquired glucocorticoid resistance have
`been reported. It is probable that the observed in-
`
`stances of apparent glucocorticoid insensitivity in the
`current study patients were in fact partly the result of
`mineralocorticoid insufficiency, which kept them in a
`state of partially compensated chronic hypovolemia
`without the physiologic reserve necessary to respond
`appropriately to otherwise minor stresses.
`The overall incidence of glucocorticoid insuffi-
`ciency in the current study group of 20 patients was
`55%. In the group of patients tested at more than 90
`days postdosing, the incidence was 20%. This data,
`taken together with that of other investigators,27'28
`demonstrates that suramin does not cause glucocorti-
`coid insufficiency in all patients. Eisenberger et al.27
`in their study of 73 men with hormone—refractory met-
`astatic prostate carcinoma treated with an intermittent
`infusion schedule of suramin observed a 41% inci-
`dence of clinical and biochemical adrenal
`insuffi-
`
`ciency in 17 patients undergoing ACTH stimulation
`testing. The current data most clearly show that, at low
`dose levels, the occurrence of long term glucocorticoid
`insufficiency is low. The implication is that if lifelong
`glucocorticoid replacement is continued on the as-
`sumption of irreversible adrenal toxicity, a significant
`
`

`
`suramin and Hypoaldosteronism/Kobayashi et al.
`
`2417
`
`CORTISOL
`
`ALDOSTERONE
`
`0
`
`0O®00000
`
`000000000
`
`
`
`[CORTISOL]ttg/ml
`
`Normal Abnormal
`
`Abnormal
`
`TIME (hours)
`
`FIGURE 2. Pharmacokinetic study of Case 2 (Patient 456). Measured
`plasma cortisol levels after the administration of 20 mg of hydrocortisone
`either intravenously (— — - -) or orally (~) are shown on the y axis. Time
`in hours after dosing is shown on the x axis. The arrow indicates the time
`at which hydrocortisone was administered.
`
`toxicity of suramin, there has been little formal study
`to fully characterize this effect in cancer patients. The
`instances of adrenal cortical insufficiency in the cur-
`rent study patients may have been caused by their
`underlying diseases, pharmacokinetic factors, non-
`compliance, glucocorticoid resistance, or exogenously
`administered steroids.
`
`It is unlikely that the adrenal insufficiency was
`caused by the underlying carcinomas. Two of the
`patients had either known or suspected adrenal
`involvement by their carcinoma prior to endocrino-
`logic testing. Abnormal ACTH stimulation tests oc-
`our in only 33% of patients with documented adrenal
`metastases by CT scan.” None of the remaining 18
`patients had manifested symptoms suggestive of ad-
`renal insufficiency. The only study to address the
`question of adrenal function in such patients, with
`no prior suspicion or knowledge of adrenal gland
`involvement, was performed in 14 patients with
`prostate carcinoma.” Although five patients had ab-
`normal ACTH stimulation tests, these results were
`
`not reproducible on repeat testing and it is difficult
`to interpret this finding.
`A pharmacokinetic basis for the adrenal insuffi-
`ciency, such as impaired absorption of hydrocorti-
`sone or accelerated clearance of hydrocortisone, is
`also unlikely based on the outcome of the hydrocor-
`tisone bioavailability study, although only one pa-
`tient was studied. Although noncompliance is also
`possible, all patients complied with other required
`
`
`
`
`
`TOTALSURAMINDOSE
`
`GRAMS/METER2
`
`Normal Abnormal
`
`Normal Abnormal
`
`FIGURE 1. Adrenocorticotropic hormone stimulation results as a func-
`tion of total received dose. For ease of comparison between the text and
`other published studies, figures depicting the relationship to both raw total
`dose (upper row) and total dose normalized to body surface area (lower
`row) are shown. Cumulative received dose is indicated on the y axis.
`Normal results are indicated on the left—hand side of each figure by shaded
`circles, whereas abnormal results are indicated on the right-hand side of
`each figure by open circles. Results of cortisol testing are shown in the
`two left figures, and results of aldosterone testing are shown in the two
`right figures. The dashed lines indicate a dose of 5.7 g, mentioned in the
`discussion.
`
`percentage of patients may be receiving long term ste-
`roid replacement unnecessarily. This is important not
`only because of the well known morbidities associated
`with chronic glucocorticoid therapy, but also because
`hydrocortisone is itself an active agent in treating hor-
`mone-refractory metastatic prostate carcinoma. Con-
`current use of glucocorticoid with suramin may thus
`affect the outcome and analysis of clinical trials involv-
`ing this combination.
`Although adrenal insufficiency is a well accepted
`
`

`
`2418
`
`CANCER December 1, 1996/ Volume 78 I Number 11
`
`therapy and monitoring procedures, and isolated
`noncompliance appears highly unlikely, especially
`because most patients developed some cushingoid
`features. Finally, given the substantial period of time
`for which some of the current study patients re-
`ceived hydrocortisone, it is possible that this exoge-
`nously administered glucocorticoid contributed to
`the adrenal insufficiency.
`Commonly used definitions of abnormal ACTH
`glucocorticoid stimulation tests are designed to de-
`tect clinically significant deficits in adrenal cortical
`function. They incorporate criteria measuring not
`only a change in the cortisol level, but also require
`that cortisol levels attain a certain level. Removing
`this latter constraint and focusing on the change in
`cortisol levels may allow detection of subtler deficits
`in organ function. Indeed, this appears to be the
`case with glucocorticoid production, because pa-
`tients treated at lower dose levels appear to have
`less impairment of adrenal functional reserve than
`those patients treated at the higher dose levels. Min-
`eralocorticoid functional reserve also exhibits a dose
`
`toxicity relationship, although it appears to be more
`sensitive to the effects of suramin than is glucocorti-
`coid functional reserve.
`
`Suramin may have more complex effects on adre-
`nal function than previously appreciated. In fact, there
`is a spectrum of effects ranging from subclinical im-
`pairment of functional reserve to severe anatomic and
`structural damage. In vitro, suramin depresses adrenal
`cortical function both by interference with hypotha-
`lamic-pituitary-adrenal regulatory mechanisms“ and
`by direct
`inhibition of adrenal steroidogenic en-
`zymes.32'36 The unexpectedly low ACTH concentra-
`tions shown in Table 2 may reflect a suramin—induced,
`rather than exogenous glucocorticoid—induced, sup-
`pression of ACTH secretion by the anterior pituitary
`gland, as has been shown to occur in cultured rat ante-
`rior pituitary cells.“ Adrenal cortical necrosis has been
`observed after suramin treatment
`in cynomolgus
`monkeys," other animals,” and in patients with bul-
`lous pemphigus.3""‘° The greater severity of adrenocor-
`tical impairment at the higher dose levels may reflect
`this type of process. Bearing in mind the retrospective
`nature of the current data and their limitations, it is
`
`tempting to speculate that these effects may be related
`to some pharmacokinetic quantity, such as close, con-
`centration, or systemic exposure.
`In support of this dose-response hypothesis are
`recent data from 17 patients receiving suramin therapy
`for human African trypanosomiasis.” The patients re-
`ceived suramin at doses up to a maximum of 5.75 g;
`no supplementation with glucocorticoid was reported.
`Serial testing of patients was conducted at baseline,
`
`during suramin therapy, and after discontinuation of
`treatment. Baseline adrenal insufficiency compared
`with normal Ugandan controls was noted prior to initi-
`ation of treatment, which improved over the course
`of therapy. Although no data were presented regarding
`the doses that were received, deteriorating adrenal
`function was not correlated with increasing serum sur-
`amin concentrations over time or with peak cortisol
`concentrations. The authors attributed the observed
`
`adrenal impairment to the effects of the underlying
`trypanosomiasis rather than to effects of suramin and
`concluded that the adrenotoxicity of suramin appears
`to be dose—dependent, with the toxic total dosage be-
`ing >5.75, and possibly >10, g. This conclusion is con-
`sistent with the current observations, because adrenal
`
`insufficiency was noted only in patients receiving cu-
`mulative doses of greater than 9 g.
`A dose-dependency of suramin’s adrenotoxic ef-
`fects has not been previously recognized in cancer
`clinical trials. This may be in part due to the design
`of other trials, in which a relatively limited dose range
`was administered to maintain fixed suramin plasma
`concentrations. Although the current study was able
`to test patients at multiple dose levels, the power of the
`current analysis regarding this point is limited because
`ACTH stimulation testing was conducted on a highly
`select group of patients. In addition, the sample size
`was small, there was some selection bias in that more
`
`patients were tested at the higher dose levels than at
`the lower levels, and finally, possibility that the defect
`in patients at lower dose levels was reversible and had
`resolved by the time of testing cannot be excluded.
`The current study patients also received varying
`amounts of suramin according to the response of their
`carcinoma, rather than according to the presence of
`adrenal insufficiency. This limits the authors ability to
`distinguish between dose, concentration, or systemic
`exposure as the crucial determinant of adrenal cortical
`toxicity. Thus, although the current data suggest that
`a dose—toxicity hypothesis may exist, this remains only
`a hypothesis that requires prospective testing for con-
`firmation.
`
`Nevertheless, the existence of a group of patients
`with intact adrenal function after administration of
`
`suramin, previously thought to uniformly cause de-
`struction of the adrenal gland, is significant and de-
`serves further investigation. Prospective endocrino-
`logic evaluations in future suramin trials are needed
`to fully characterize the endocrinologic effects of sura-
`min and to explore the validity of the dose-toxicity
`hypothesis. The authors have begun prospective endo-
`crinologic testing in a cohort of patients being treated
`with high dose suramin and can confirm that suramin
`treatment at a dose of 1730 mg/m2 in this schedule
`
`

`
`results in combined mineralocorticoi