THE HYPERCALCEMIA OF ADRENALINSUFFICIENCY

Mackenzie Walser, … , Brian H. B. Robinson, John W.Duckett Jr.

J Clin Invest. 1963;42(4):456-465. https://doi.org/10.1172/JCI104734.

Research Article

Find the latest version:

http://jci.me/104734/pdf

Journal of Clinical InvestigationVol. 42, No. 4, 1963

THE HYPERCALCEMIAOF ADRENALINSUFFICIENCY *

By MACKENZIEWALSER,BRIAN H. B. ROBINSON,t ANDJOHNW. DUCKETT, JR.t

(From the Department of Pharinacology and Experimental Therapeutics and the Departmnentof Medicine, The Johns Hopkins University School of Medicine and The Johns

Hopkins Hospital, Baltimore, Md.)

(Submitted for publication October 3, 1962; accepted December 6, 1962)

An interrelationship between the adrenal glandsand calcium metabolism was demonstrated in 1911by Guleke (1), who showed that adrenalectomyameliorates tetany in parathyroidectomized ani-mals. This observation has been repeatedly con-firmed (2).

Hypercalcemia after adrenalectomy was firstdemonstrated in 1924 by Kisch (3) in rabbits, andsoon thereafter by other workers in dogs (4-8),cats (5, 9), and in patients with Addison's disease(10-12). An early report (13) of hypercalcemiain adrenalectomized rats was based upon dubiousmethods of analysis; later reports indicate thatplasma calcium remains constant or falls in thisspecies after adrenalectomy (14-16).

Administration of adrenocortical steroids oftenameliorates spontaneous hypercalcemia (17), ex-cept when it is caused by hyperparathyroidism ( 18).Nevertheless, in rats cortisone modifies or preventsthe hypercalcemic response to parathyroid extract(16, 2). This discrepancy may be a matter ofdosage, since large amounts of steroids may some-times reduce plasma calcium even in hyperpara-thyroidism (19, 20). In normal subjects, adminis-tration of adrenocortical steroids has little effecton plasma calcium concentration, and Cushing'sdisease is not associated with hypocalcemia. Areport of hypercalcemia after cortisone adm-ninistra-tion to nephrectomized dogs (21) has not beenconfirmed (22).

Several hypotheses have been advanced to ex-plain these observations. The possibility thathemoconcentration might explain the hypercal-cemia of adrenal insufficiency was suggested by theearliest workers; however, no mneasurenments ofplasma protein concentration or ultrafiltrable cal-

* Supported by U. S. Public Health Service grantA-2306.

t Present address: Guy's Hospital, London, England.t Present address: Peter Bent Brigham Hospital,

Boston, Mass.

cium in this condition appear to have been re-ported. Increased parathyroid gland activity dur-ing adrenal insufficiency was suggested by the re-ported finding of parathyroid hyperplasia inadrenalectomized animals (23). Against thisview is the occurrence of hypercalcemia in adre-nalectomized parathyroidectomized dogs (22), aswell as the antagonistic effects of cortisone andparathyroid extract mentioned above.

Adrenal steroids block the excessive intestinalabsorption of calcium in sarcoidosis (24) and tendto inhibit the active transport of calcium by theisolated gut (25). Calcium absorption mighttherefore be excessive in adrenal insufficiency.Balance studies in an Addisonian subject (26) andin adrenalectomized rats (27) do not support thispossibility. The same objections apply to the hy-pothesis that cortisone acts by antagonizing the in-testinal action of vitamin D. The possibility thatthe skeletal actions of vitamin D are antagonizedby adrenal steroids has not been excluded, al-though the morphologic effects of vitamin D de-ficiency and steroid excess on bone are quite dif-ferent (28).

Steroid administration and hypercorticism areusually associated with increased urinary calciumexcretion (29, 30). After withdrawal of hormonetherapy in Addison's disease (26), urinary excre-tion of calcium may fall. Contrary results havealso been reported (31). In hypercalcemic statestreated with steroids, urinary calcium excretionfalls concomitantly with the reduction in plasmacalcium (22) ; clearly this effect on plasma calciumcannot be ascribed to a renal effect of the hormone.Nevertheless, a renal role in the hypercalcemia ofadrenal insufficiency remains a possibility.

Finally, evidence for a direct action of adrenalsteroids on the equilibrium between bone and ex-tracellular fluid has been presented (32). Ap-parently it has not been determined whetherchanges in ionic calcium concentration or some

456

THE HYPERCALCEMIAOF ADRENALINSUFFICIENCY

relevant ion product occur in hyper- and hypo-adrenal states as plasma calcium concentrationchanges.

In the present study, the incidence of hypercal-cemia in adrenocortical insufficiency has been ex-amined in three species, man, dog. and rat, andthe mechanism of this change has been investigatedin the dog.

In studying the incidence of hypercalcemia, itis necessary to establish some criterion for adrenalinsufficiency. The absence of adrenal function isclearly an inadequate criterion for this purpose.since the syndrome of adrenal insufficiency mayfail to occur in adrenalectomized animals main-tained on salt alone, whereas adrenal crisis maydevelop despite hormone therapy under conditionsof stress. One of the most characteristic featuresof the syndrome of adrenal insufficiency is hypo-natremia. Therefore the incidence of hypercalce-mia has been compared with the incidence of hy-ponatremia in patients with Addison's disease andin adrenalectomized dogs. In rats, hypercalcemiafailed to occur after adrenalectomy. so no furtherstudy of this species was made.

In examining the mechanism of this form ofhypercalcemia. observations in human subjectsare not readily available because patients with thesyndrome of adrenal insufficiency are uncommonand are treated as soon as the diagnosis is made.Therefore the mechanism of this disturbance hasbeen studied in dogs.

MATERIALS AND METHODS

SubjetctsPatients. Records of 31 patients with Addison's disease

have been collected in which hyponatremia, or hypercal-cemia, or both have been observed in a single bloodsample or in a single day. These include both treatedand untreated subjects. WN:e have arbitrarily selected 135mEq per L as the lowter limit of normal sodium concen-tration and 11 mg per 100 ml (2.75 mmoles per L) asthe upper limit of normal calcium concentration. Sixtycases of proven Addison's disease in the records of theJohns Hopkins Hospital have been examined, and 22 re-ports have been found of analyses of blood samples fromsixteen cases that meet these criteria. Apart from scat-tered reports of one or two cases, the only published ac-counts that include appreciable numbers of calcium de-terminations in Addisonian patients appear to be thoseof Loeb (10), Helve (11) and Leeksma, De Graeff, andDe Cock (12). The results of these authors are in-cluded in Figure 1.

Rats. Adrenalectomized, Sprague-Dawley, female rats

and litter-mate controls weighing 150 to 200 g were ob-tained the day after operation 1 and placed on diets ofeither milk or sodium-free milk (Lonalac) containing20% dextrose. Each day for 5 days, groups of animalswere anesthetized with ether and exsanguinated via theabdominal aorta. The adrenal regions were inspectedto confirm completeness of adrenalectomy.

Dogs. Male and female mongrel dogs weighing from7 to 16 kg were used. Twenty-five successful bilateraladrenalectomies were performed in one or two stagesthrough flank incisions under pentobarbital anesthesia.Cortisone,' desoxycorticosterone,3 and antibiotics wereprovided postoperatively and when required. From timeto time, hormone therapy was withdrawn and daily bloodsamples were obtained until frank adrenal insufficiencydeveloped. This required from 2 to 10 days (usually 4to 6) and was heralded by loss of appetite and decline inweight.

Three feeding regimens were employed: canned dogmeat supplemented with sodium chloride and sodium bi-carbonate, sweetened milk freed of calcium and magnes-ium, and ordinary dog chow. The milk diet was pre-pared by passing homogenized pasteurized cow's milkthrough a well-washed 10 X 50 cm column of Dowex Alresin in the sodium form, acidifying to pH 6 to 7 withHCl, adding 200 g dextrose per L, and freezing 400-mlsamples in plastic containers. Neither calcium nor mag-nesium could be detected in this material. Sodium con-centrations varied from 100 to 120 mEqper L. Activatedcharcoal was mixed with the resin in order to remove abitter-tasting material which this resin releases. Therate of onset of symptoms of adrenal insufficiency wasnot appreciably different with any of these three regi-mens.

Venous blood samples, usually from the jugular vein,were obtained in syringes containing heparin.

AtnalWtical methods

The methods for determination of plasma pH, proteinconcentration, calcium, magnesium, sodium, potassium,phosphate, and citrate were the same as those cited inprevious reports from this laboratory (33, 34). Ultra-filtrates were prepared from a number of samples, andthe concentration of free calcium ions and free magnesiumions was determined spectrophotometrically (33). It wasnot possible to perform all of the determinations on everysample. Consequently, in the calculation of individualcalcium complexes in plasma ultrafiltrate (34), valuesfor magnesium ion concentration have been assumedwhen not available

RESULTS

Incidence of hypercalcemtia in adrenal insufficiencyIn miani. In Figure 1, plasma calcium concentra-

tions have been plotted against plasma sodium1 Charles River Breeding Laboratories, Boston, Mass.2 17a,21-dihydroxy-4-pregnene-3,11,20-trione.321 -hydroxy-4-pregnene-3,20-dione.

457

MACKENZIEWALSER, BRIAN H. B. ROBINSON, AND JOHN W. DUCKETT, JR.

[Ca]p A0

,&HelveLoeb

2.8 ALeeksma ea/t

0

0

100 120 140 160

(Nao pmEq/L

FIG. 1. PLASMA CALCIUM IN RELATION TO PLASMA

SODIUM IN 31 PATIENTS WITH ADDISON'S DISEASE, EX-

CLUDING THOSE IN WHICH BOTH VALUES WERE NORMAL.

The normal limits were arbitrarily defined. Values pub-lished by Loeb (10), Helve (11), and Leeksma, DeGraeff, and De Cock (12) are included.

concentrations in thirty-one patients with Addi-son's disease, including cases from the Johns Hop-kins Hospital records and three published reports(10-12). In each case, the calcium and sodiumdeterminations were performed on the same bloodsamples, or at least on the same day. In theJohns Hopkins Hospital patients, hyponatremiaappears to be about twice as frequent as hyper-calcemia. Helve (11), however, found severalpatients with hypercalcemia without hyponatremia.These results obviously suffer from the lack ofcontrols, differing methods, and sampling bias.Considering all the data, however, it appears thathypercalcemia approaches hyponatremia both infrequency and in severity in Addison's disease.There is no correlation, positive or negative, be-tween hypercalcemia and hyponatremia.

In rats. Table I summarizes the results in 37adrenalectomized and 33 control rats. Day 0 re-

fers to the first day after adrenalectomy, when thediet was started. Plasma calcium concentrationsdid not change in controls or adrenalectomizedanimals, even when fed on sodium-free milk, de-spite a high mortality in the adrenalectomized ratson this diet.

In dogs. Figure 2 represents the incidence ofhypercalcemia as compared with hvponatremia inadrenalectomized dogs. As in the human subjects,any sample that revealed hypercalcemia, hyponia-tremia, or both has been included, whether theanimals were receiving therapy or not. Twentvobservations from one dog are included, and fromone to six observations from 24 others. The (lis-tribution of points from the one dog is similar tothat from the others. The frequtency of hvper-calcemia is seen to be nearly identical with thliat ofhyponatremia in these animals, even in those on a

calcium-free, magnesium-free diet. The severityof hypercalcemia, however, is much greater thanthat of hyponatremia. One-sixth of the 58 sam-ples exhibited calcium concentrations more than40% above the mean normal value and nearlyhalf were increased 20% or more. In only oneof the samples was the sodium concentration re-duced as much as 20%; half were reduced 10%,or more. There is no correlation, either positiveor negative, between hypercalcemia and hvpona-tremia in this group of observations. Twelvesamples exhibited hyponatremia without hyper-calcemia, 12 hypercalcemia without hyponatremia,and 34 showed both. The highest calcium con-centration observed was 4.18 miM, or 16.7 mg per100 ml.

There was no apparent difference between theclinical significance of hypercalcemlia and that ofhyponatremia in these dogs. In most cases ofeither disturbance, the animals were symptomatic.but a few appeared well and remained so for sev-

TABLE I

Mean plasma calcium concentration in adrenalectomized andcontrol rats fed milk and dextrose, and

sodium-free milk and dextrose*

Mean plasma calcium concentration

Sodium-free milk andMilk and dextrose diet dextrose diet

Days Adrenal- Adrenal-on ectomized ectomized

diet Controls rats Controls rats

mmoles/L mmoles/L0 2.70 [4] 2.57 [4] 2.70 [4] 2.57 [4]1 2.50 [1] 2.86 [4] 2.62 [4]2 2.52 [1] 2.50 [1] 2.68 [4] 2.58 [4]3 2.62 [1] 2.48 [1] 2.74 [4] 2.74 [4]4 2.64 [3] 2.63 [3] 2.66 [4] 2.59 [4]5 2.52 [3] 2.67 [4] 2.66 [4]

* Numbers in brackets refer to the number of observ-.l-tions in each group.

458

THE HYPERCALCEMIAOF ADRENALINSUFFICIENCY

- %O U) I' U) 00 V U)

00000000

0000 U) V V U)vi - C- %000-000000666666666

U U)o00 -0t U)i 00-0o - U UC4 ei U) U666666666

- d - V 0% V d' V) U

o o0 )o )o oo o

0. 0 V .- 0% .U . 00

666666666

0 - U-I U) V0- CU

Uo)ot- o UUU-o t-o

U)0-U) 00U 0% C-4

_n_n No r m

'U)U)1'U)

00 0o0 000

U)- -iU t- 00 t-0

ooo oU U - o

Vu U) U) 0- r- V0% U- 00% 0% 000000o-.-o.o.. .6

00 o 'o0 o 00o o -

u0 U-I U)U- C 0 UC

CU' CU ) U U) U )UU* .- .- .- .- .- .- U.

0v)d c)00 0 OC )--o°m°z°° 0 O

O; +

0 0O; +

r- U

O -H

U00000

00

0-i

VO

U) -

0 O

000

CU)

V 01

U) ---H

V 0

U) 0%

Vo o

Ol

U-I-H

4- 4- 4-4--4--

t-0 14 'n CU ,,4000%0 0UU

0-0-0-0000

4-4- 4-4-4-4--

% U) 00 U)~\ t- U) V- \

6666666666

LO to 0U) r) I- t- CU'4- 4-4-n

0000000000

66660t~h08 00000

00d00C 0000 00dr0000000000OO6666666666OO

4-. .-4-4-4-4-4-O % U) U) 000 - t- C U)

- CU'V---NO~

0 0

0-

00

0 _

U. -C,

dt CV1

C5 .

4+

0 r-

0 00 O6 .

0 0

C C

-H'- 0% 00C1 mt -I e#-,

In - C4- 4- 4-C O C4+95t n'-,d,1~4 t t +

U- % -U U- 0% CU U-I U-CUU- U'U0-. U) 0% U) N1 00s

0 -- C 00 c U,) en U-U- U) U- 00 U- 00O C0N U-

UN-t- 00

CU CCU CUC6 6

4- 4-4-

c0i _r4

CU U) U) U)

4- 4-4-- 4-40%00 0 00 % 00 U' U)

_U _0 V 00_t V_ _I 0

U''00 00 " "- -- U-- 0%.

00 -6666(V

0V 000 V - U-. U) U)5

U)U)U)U- U UNU)U)U)U

LnU 000000 'C 0 CIU)0000VdU-I T0oOO6 U-o6 r.0V .c;

4-4.- 4- 4-I-U) C 0- I t C

-.N

N U-) ,,

5

5 m-+ * < u u u

¢

4+00 V

0-H

V -

000

o-H

C oo

00

CU6

00Uo

CU' 00-. -

V 00

00 V

000V-11

ICU0

E0u

0

0.

H-,

;n

459

Kt.0

0

00

4.0

.0

CUu

.

CZ

2

o0

v

CUCZ10

~0

.00

24Z21

0C2E

022

C"4

CU0

Ira

VI

C,)

*U4

_U)

C,U)

0

Cb,)

CU

o

CZ)

CU

CCZ

od

CZ)

0

CU

0

C._

0

CU

._

Q

OU

-o

CU

U._

CU

C.)

CUQ11 I

0U

vOCU. O .r

CU

-XXU...CC. U .

11

U._C.

U.3

u

cou

ICI

CU

U

0

0I.

=0~

n

CUEUt

a.

WIll.

MACKENZIEWALSER, BRIAN H. B. ROBINSON, AND JOHN W. DUCKETT, JR.

mM/L I Mg%

3.5

15A

0

0 e

0

0

00

0 _0 * n

[Ca P3 - 13

2.5100 120 140

[Na]p , mEq/L

FIG. 2. PLASMACALCIUM IN RELATION TO PLASMA SO-DIUMI IN ADRENALECTOMIZEDDOGS, EXCLUDING SAMPLESIN XWHICH BOTH VALUES WXERENORMAL. Mean values ± 2SD for 20 normal dogs are shown. Triangles: dogs on

a calcium-free, magnesium-free diet. Open circles: dogson a regular diet. Closed circles: dogs on a milk diet.

eral days. No dogs with both disturbances re-mained well for more than a day.

Magnesium concentrations were subnormal(less than 0.81 mMi) in half of the 22 adrenallyinsufficient dogs on magnesium-containing diets.The lowest value was 0.33 mM. Low mag-nesium values were not correlated with hypercal-cemia, hyponatremia, or hyperkalemia. No highvalues were observed, contrary to an early re-port by Harrop, Soffer. Ellsworth, and Trescher(8).

Ionic composition of the plasma in the hvpercal-cemtia of adrenal insufficiency

In Table II are summarized the results of analy-sis of plasma in nine normal dogs and ten samplesfrom nine adrenalectomized dogs that were hyper-calcemic. The normal dogs were not the sameanimals as the operated dogs, except for dog 4J.The values in normal dogs are similar to thosepreviously reported in normal human subjects(34), with the following statistically significant

differences: total, ultrafiltrable, free ionic calcium,phosphate, and citrate were slightly higher; pro-tein concentration was lower. In the adrenalec-tomized animals, the major finding is that freecalcium ion concentration is normal in all but onedog, in which it is slightly elevated.4 Total plasmacalcium is increased 33%. on the average, whereasultrafiltrable calcium is increased 17% and freeionic calcium only 3 %1.

Complexed calcium, calculated as the differencebetween ultrafiltrable calcium and free ionic cal-cium, was not increased significantly in the groupas a whole, but was beyond the normal range inthree dogs. being markedly increased in one.Part of this increase is attributable to the calciumcitrate complex, CaCit-, which was abnormallyincreased in six out of ten dogs; a smaller part isattributable to increased amounts of the complexCaHPO,; a portion is unaccounted for, particu-larly in one sample. Figure 3 illustrates the dif-ference between normal dog plasma and thesehypercalcemic samples.

The cause of these changes in the plasma cal-cium fractions is revealed in part by the data inTable II. Hyperproteinemia was seen in all butone sample. In order to determine whether theincrease in protein concentration was adequate to

mM/L4 r-

3

2

0

mg%I 16

14

_12

I 0-Ca Prot-

-8

-4

2

NORMAL HYPERCALCEMICDOGS ADRENALECTOMIZED

DOGSFIG. 3. MEANVALUES FOR FRACTIONS OF PLASMA CAL-

CIUM IN 9 NORMALDOGS AND IN SAMPLES FROM 10 HY-PERCALCEMICADRENALECTOMIZEDDOGS. CaX = unidentifiedcalcium complexes.

4 This value lies 2.4 SD from the normal mean, a devi-ation to be expected by chance alone 1 out of 60 times.

460

THE HYPERCALCEMIAOF ADRENALINSUFFICIENCY

150 0 Loken et. a/.o Walser

KCaProt.%of

Value atum.16

100

50

0

IlL:05 .10 .15 .20

IONIC STRENGTHOF PLASMA

.I

.25

FIG. 4. EFFECT OF VARYING IONIC STRENGTHOF PLASMAON THE ABILITY OF PLASMA PROTEIN TO BIND CALCIUM.Results of Loken, Havel, Gordan, and Whittington (37)are from normal human plasma, and results of Walser(36), from normal Hog serum.

account for the increase in nonultrafiltrable calcium,the dissociation constant of calcium proteinate wascalculated by the following formula: KCaProt =[Ca++] (1.22 [Prot] -[CaProt] )/[CaProt], where[CaProt] = [Ca]p - [Ca] up (plasma and ultrafil-trate concentrations). This formula differs fromthat used in a previous report (35) in that freeionic calcium concentration replaces ultrafiltra-ble calcium. As stated previously, the choicebetween these two alternatives depends upon theextent of protein-binding of the filtrable calciumcomplexes. Since CaCit- accounts for a portionof the filtrable complexes in these samples and isnot appreciably bound to protein (36), the free ioniccalcium concentration was employed. As shownin Table II, five samples from four dogs exhibitabnormally low values for Kcaprot by this calcu-lation; that is, the elevation of protein concentra-tion is not sufficient to account for the increasein protein-bound calcium, and an abnormal affinityof plasma protein for calcium apparently exists.

One possible cause for this abnormal affinity ishyponatremia. When the ionic strength of plasmais varied, marked changes in the affinity of plasmaprotein for calcium occur (36, 37). Figure 4represents Kcaprot as the percentage of normal intwo reported studies in which a single plasmasample was diluted with water or salt solutionsand then subjected to ultrafiltration. Kcaprot, aspercentage of normal, can be approximated as[Ca]uF/( [Ca]p - [Ca]UF), expressed as the per-centage of the value obtained in plasma dilutedwith isotonic saline. The results in Figure 4 canbe represented approximately by a straight line.The results of Loken, Havel, Gordan, and Whit-

tington (37) are the mean of several determina-tions and have therefore been given greater weight.From this line, it is possible to calculate the changein plasma calcium concentration required to main-tain constant ultrafiltrable calcium when ionicstrength varies. If normal ionic strength is takenas u = 0.16 and normal sodium concentration as145 mEqper L, these results can be expressed interms of plasma sodium, as in Figure 5. Thisgraph shows that hyponatremia to the extent of110 mEqper L can be expected to increase plasmacalcium 2 mg per 100 ml if ultrafiltrable calciumis to remain constant.

When the values for KCaprot shown in Table IIare corrected for hyponatremia according to theline in Figure 4, three remain abnormally low.Thus an abnormally high affinity of plasma proteinfor calcium evidently exists in these three dogs.

Citrate concentrations were abnormally ele-vated in seven out of ten samples. In three, veryhigh values were seen. Obviously, free citrate ionconcentration was increased, as well as theamounts of NaCit--, MgCit-, and CaCit-. Thehypercitremia did not account, however, for morethan 20% of the increase in plasma calcium inany dog. In patients with Addison's disease,slight elevations of plasma citrate have been re-ported (38).

MM/L3.01

25

2.0-

[Ca] 1.5-

Mg70-12

-S0

- 8

- 6

'.0± 4

0.5+ 2

Protein- bound

Ultrof litrable

100 120 140 160[Na]opmEq/L

180 200

FIG. 5. TOTAL PLASMA CALCIUM VALUE REQUIRED TOMAINTAIN CONSTANTULTRAFILTRABLE CALCIUM VALUE INMAN (SOLID LINE) AND IN DOGS (DOTTED LINE) WHENIONIC STRENGTHOF PLASMA CHANGES, CALCULATED FROMDATA IN FIGURE 4.

461

MACKENZIEWALSER, BRIAN H. B. ROBINSON, AND JOHN W. DUCKETT, JR.

Plasma phosphate was elevated in six of tensamples and in the group as a whole. In sevensamples, the fraction of filtrable phosphate was re-duced, in four cases to a marked degree. Whenthese plasma samples were subjected to ultrafiltra-tion after the addition of 4 mmoles per L of thetrisodium salt of EDTA, the filtrability of phos-phate became normal. This same finding was re-ported in the other hypercalcemic states (35),and was attributed to the formation of so-called"colloidal calcium phosphate," either in vivo orin vitro. In the present samples, this explanationmay not be tenable because of the large amountsof nonultrafiltrable phosphate involved. In sample5I 18, for example, nearly 2 mmoles per L ofplasma phosphate were nonultrafiltrable untilEDTAwas added. Yet the total amount of non-ultrafiltrable calcium and magnesium was only2 mmoles per L. If the nonultrafiltrable phos-phate was associated with calcium and magnesium,as the results with EDTA indicate, the entireamounts of nonultrafiltrable calcium and mag-nesium must have been in this form. These con-siderations lead us to suspect that this form of non-ultrafiltrable phosphate is at least in part an arte-fact of some kind, as against the nonultrafiltrablephosphate in normal plasma, which is unaffectedby EDTA (39). Conceivably, there may be la-bile, nondialyzable, organic phosphate compoundsin plasma that are converted to inorganic phos-phate either by trichloracetic acid or by EDTA,but this seems improbable. This phenomenonmay account for some of the hypercalcemia ofadrenal insufficiency, since two of the animalswith large amounts of nonultrafiltrable phosphatewere those with abnormal values for K'Caprot 5

(5I18 and 5N3). In other words, the apparentlyexcessive affinity of protein for calcium in thesedogs may be due to the presence of a nonfiltrablecompound of calcium and phosphate. This com-pound must be present in vivo to contribute to hy-percalcemia, but in all probability the amount ofthis compound increased in vitro after the bloodwas drawn.

Magnesium concentrations were reduced in one-third to one-half of the samples in which measure-ments were made. None of these dogs was re-ceiving calcium-free, magnesium-free milk. Pro-

5 Calculated dissociation constant of calcium proteinatecomplex, corrected to ju = 0.16.

tein-binding of magnesium was normal in threesamples.

DISCUSSION

Hypercalcemia is evidently a common featureof adrenocortical insufficiency both in adrenalec-tomized dogs and in patients with Addison's dis-ease. For unknown reasons, it apparently doesnot occur in rats.

In contrast to the findings in other types of hy-percalcemia (35), the free ionic calcium concen-tration is not increased. Several other factors,singly or in combination, account for this hyper-calcemia: hemoconcentration, increased plasmaconcentration of citrate and other complexing an-ions, and an abnormal affinity of plasma proteinfor calcium, attributable in part to hyponatremiaitself.

Considering the dogs as a group, the averageincrease in plasma calcium was 0.85 mmole per L.Hyperproteinemia alone accounted for only afourth of this increment. A like amount is at-tributable to increased complexed calcium. Theother half of the increase is the result of an ap-parently increased affinity of plasma protein forcalcium. Part of this change in affinity is attribu-table to hyponatremia itself; the remainder is un-explained. Only three samples, however, ex-hibited this change and in two of the three, a non-filtrable form of calcium phosphate appears to beresponsible.

There is no reason to implicate excessive in-testinal absorption of calcium as a contributorycause of this type of hypercalcemia. The totalamount of calcium in the extracellular fluid isprobably close to normal. Although total calciumconcentration in plasma was increased an averageof 33%, plasma volume was probably considerablyreduced. Likewise, interstitial fluid volume wasprobably reduced, so that the 17%o increase inultrafiltrable calcium would not be inconsistentwith a normal amount of calcium in the interstitialfluid. Furthermore, hypercalcemia occurred withcomparable frequency in adrenalectomized dogsfed a calcium-free, magnesium-free diet.

A renal role in this hypercalcemia has not beenexcluded. As shown elsewhere (40), renal tu-bular reabsorption of calcium is *excessive inadrenally insufficient dogs. Daily urinary calciumexcretion was determined in a number of animals

462

THE HYPERCALCEMIAOF ADRENALINSUFFICIENCY

and was usually less than 0.3 mmole (12 mg), de-spite hypercalcemia. An increase in calcium ex-cretion could presumably restore a normal plasmacalcium concentration, but only at the cost of asubnormal ionic calcium concentration. If morecalcium were released from bone, hypercalcemiawould recur.

Since the concentration of free calcium ions wasnormal, there is no need to postulate an increasein parathyroid hormone secretion, an increasedsensitivity to vitamin D. or a direct effect of de-ficiency of adrenal hormones on the equilibriumbetween bone mineral and extracellular fluid.

There is no evidence that the hypercalcemia ofadrenal insufficiency is deleterious to the organism.Since the effective calcium concentration is nor-mal, the only possible effects are those of increasedprotein-bound or complexed calcium. These frac-tions are not known to exert any physiological ac-tions, with the possible exception of the complexCaHPO, (35).

In some respects, this form of hypercalcemiaresembles that produced locally by prolonged ve-nous obstruction, so-called "in vivo filtration"(41). In the venous blood from an arm whosevenous outflow has been occluded, protein con-centration and total calcium are increased, plasmavolume is reduced, and ultrafiltrable calcium con-centration is normal. It is uncertain whether hy-percalcemia is a common feature of the hemocon-centration seen in salt and water depletion.

The peculiar nonultrafiltrability of plasma phos-phate in half of these samples makes somewhat un-certain any conclusions about the free ion product[Ca'+] [HPOj--]. The observed values for thisproduct are proportional to the concentrations ofthe complex CaHPO4given in Table II, since thequotient of these quantities, [Ca++] [HPO4--]/[CaHPO,], is the dissociation constant of the com-plex. As indicated by the data in the table, in-creased products wvere observed in half of thesamples. Even higher products would have beenobtained if the filtrability of phosphate had beennormal. These high products may be a reflectionof renal failure, with a consequent accumulationof inhibitors of calcification in the plasma (35).There may also be some connection between non-ultrafiltrable phosphate and renal failure, sincethe four samples in which this phenomenon wasmost pronounced where also among the most hy-

perphosphatemic and acidotic ones. An alter-native possibility is that inhibitors of calcificationfail to occur in adrenal insufficiency, despite azo-temia, and that calcium phosphate microcrystalsconsequently form in plasma on cooling and stand-ing. The calcification of the external ears some-times found in Addison's disease might conceiv-ably have something to do with this phenomenon.

Just as slight hypercalcemia may be consideredas a normal response to hyponatremia (see Fig-ure 5), the hypocalcemia seen in hypernatremicstates (42) may be explained in part on this basis.It would be of interest to determine ultrafiltrablecalcium in this disorder.

The hypomagnesemia seen in adrenal insuffi-ciency in these dogs is apparently a new observa-tion, for which we have no explanation. It is in-teresting to note that hypomagnesemia is alsoseen occasionaly in hyperaldosteronism (43). Re-ports of plasma magnesium in patients with adre-nal insufficiency are apparently lacking.

SUMMARY

Hypercalcemia occurs as frequently as hypo-natremia in adrenalectomized dogs, and is con-siderably more pronounced. In patients with Ad-dison's disease, it appears to be somewhat lessfrequent, but adequate data are lacking. Inadrenalectomized rats, it does not occur at all, evenwhen they are fed a high-calcium, low-sodiumdiet.

In dogs, the hypercalcemia of adrenal insuffi-ciency may be severe (up to 4.18 mmoles per L.or 16.7 mg per 100 ml), but the concentration offree calcium ions was nevertheless found to benormal. Plasma magnesium was usually reduced.Three alterations in the plasma combined to pro-duce hypercalcemia: first, the elevated plasma pro-tein concentration associated with hemoconcentra-tion; second, an increase in filtrable calcium com-plexes. especially calcium citrate; and third, anincrease in the affinity of plasma protein for cal-cium. which could be explained in part as a con-sequence of hyponatremia and the resulting re-duction in ionic strength of plasma, and in part asa consequence of excessive amounts of a nonfiltra-ble compound of calcium and phosphate, fornedeither in vivo, or in vitro, or both.

The increased calcium concentration of the

463

MACKENZIEWALSER, BRIAN H. B. ROBINSON, AND JOHN W. DUCKETT, JR.

plasma is not dependent upon increased intestinalabsorption of calcium, since it also occurred on acalcium-free diet.

ACKNOWLEDGMENTS

WVe are greatly indebted to medical students J. Brown,P. Briscoe, J. Fries, M. Robson, and C. Van Way forpermission to include the observations on rats that weremade during a student research project as part of thecourse in pharmacology, and to Sylvia Butler, Mary JaneFord, and Edna Richardson for technical assistance.

REFERENCES

1. Guleke, P. Experimentelle Untersuchungen jiberTetanie. Langenbecks Arch. klin. Chir. 1911, 94,496.

2. Laron, Z., J. P. Muhlethaler, and R. Klein. The in-terrelationship between cortisone and parathyroidextract in rats. Arch. Path. 1958, 65, 125.

3. Kisch, B. Experimentelle Untersuchungen uber dieFunktion der Nebennieren Klin. Wschr. 1924, 3,1661.

4. Lucas, G. H. W. Blood and urine findings in desupra-renalized dogs. Amer. J. Physiol. 1926, 77, 114.

5. Taylor, N. B., and W. R. Caven. Observations uponthe serum calcium after adrenalectomy (abstract).Amer. J. Physiol. 1927, 81, 511.

6. Rogoff, J. M., and G. N. Stewart. Studies on adrenalinsufficiency. VII. Further blood studies (cho-lesterol and calcium) in control adrenalectomizeddogs. Amer. J. Physiol. 1928, 86, 25.

7. Hastings, A. B., and E. L. Compere. Effect of bi-lateral suprarenalectomy on certain constituents ofthe blood of dogs. Proc. Soc. exp. Biol. (N. Y.)1931, 28, 376.

8. Harrop, G. A., L. J. Soffer, R. Ellsworth, and J. H.Trescher. Studies on the suprarenal cortex. III.Plasma electrolytes and electrolyte excretion dur-ing suprarenal insufficiency in the dog. J. exp.Med. 1933, 58, 17.

9. Baumann, E. J., and S. Kurland. Changes in theinorganic constituents of blood in suprarenalecto-mized cats and rabbits. J. biol. Chem. 1926, 71,281.

10. Loeb, R. F. Chemical changes in the blood in Ad-dison's disease. Science 1932, 76, 420.

11. Helve, 0. A study of the metabolism in Addison'sdisease. II. On the metabolism of lipids, nitrogen,and minerals, and on the vitamin C household.Acta med. scand. 1947, 128, 1.

12. Leeksma, C. H. W., J. De Graeff, and J. De Cock.Hypercalcaemia in adrenal insufficiency. Actamed. scand. 1957, 156, 455.

13. Rohdenburg, G. L., and 0. F. Krehbiel. On the re-lation of certain endocrines to the salt content ofrat blood. J. Cancer Res. 1925, 9, 422.

14. Helve, 0. E. Studien uiber den Einfluss der Neben-nieren-exstirpation auf den tierischen Stoffweschsel.Biochem. Z. 1940, 306, 343.

15. Conway, E. J., and D. Hingerty. The influence ofadrenalectomy on muscle constituents. Biochem.J. 1946, 40, 561.

16. Kleeman, C. R., D. Bernstein, J. T. Dowling, J. M.Koplowitz, and M. H. Maxwell. The combinedinfluence of the adrenal and parathyroid glandsin the regulation of serum calcium (abstract).J. clin. Invest. 1959, 38, 1018.

17. Connor, T. B., T. R. Hopkins, W. C. Thomas, Jr.,R. A. Carey, and J. E. Howard. The use of cor-tisone and ACTH in hypercalcemic states (ab-stract). J. clin. Endocr. 1956, 16, 945.

18. Dent, C. E. Cortisone test for hyperparathyroidism.Brit. med. J. 1956, 1, 230.

19. Jackson, W. P. U. Lowering by prednisone of se-rum-calcium in hyperparathyroidism. Lancet 1961,1, 893.

20. Gordan, G. S. The Yearbook of Endocrinology,1960-61. Chicago, Year Book Medical Publishers,1961, p. 75.

21. Grollman, A. Effect of cortisone on serum calcium,magnesium and phosphate levels in nephrectomizeddogs. Proc. Soc. exp. Biol. (N. Y.) 1954, 85,582.

22. Myers, W. P. L. Studies of serum calcium regula-tion in Advances in Internal Medicine, I. Snapper,Ed. Chicago, Year Book Publishers, 1962.

23. Rogoff, J. M. Observations on functional interrela-tionship between the adrenal and parathyroidglands. Science 1934, 80, 319.

24. Anderson, J., C. E. Dent, C. Harper, and G. R.Philpot. Effect of cortisone metabolism in sar-coidosis with hypercalcaemia. Possibly antagonis-tic actions of cortisone and Vitamin D. Lancet1954, 2, 720.

25. Kimberg, D. V., D. Schachter, and H. Schenker.Active transport of calcium by intestine: effects ofdietary calcium. Amer. J. Physiol. 1961, 200,1256.

26. Tibbetts, D. M., and J. C. Aub. Magnesium metabo-lism in health and disease. III. In exophthalmicgoiter, basophilic adenoma, Addison's disease andsteatorrhea. J. clin. Invest. 1937, 16, 511.

27. Rubin, M. I., and E. T. Krick. The salt and watermetabolism of adrenal insufficiency and partialstarvation in rats. J. clin. Invest. 1936, 15, 685.

28. Laron, Z., B. D. Canlas, Jr., and J. D. Crawford.The interaction of cortisone and vitamin D onbone of rachitic rats. Arch. Path. 1958, 65, 403.

29. Molinatti, G. M., F. Camanni, and M. Olivetti. Astudy on the metabolism of calcium in the hyper-adrenocortical syndrome. Acta Endocrinol. 1960,34, 323.

30. Laake, H. The action of corticosteroids on the re-nal reabsorption of calcium. Acta Endocr. (Kbh.)1960, 34, 60.

464

THE HYPERCALCEMIAOF ADRENALINSUFFICIENCY

31. Tuttle, S. G., and WV. G. Figueroa. Hypercalciuriaassociated with reduction in corticoid therapy af-ter prolonged administration of prednisone and af-ter bilateral adrenalectomy for Cushing's syn-drom (abstract). J. clin. Invest. 1958, 37, 937.

32. Schartum, S., and G. Nichols, Jr. Influence ofadrenal glucocorticoids on distribution of calciumand phosphorus between bone and its surroundingfluids. Proc. Soc. exp. Biol. (N. Y.) 1961, 108,228.

33. Walser, M. Determination of free magnesium ionsin body fluids. Improved methods for free cal-cium ions, total calcium, and total magnesium.Analyt. Chem. 1960, 32, 711.

34. \Valser, M. Ion Association. VI. Interactions be-tween calcium. magnesium, inorganic phosphate,citrate and protein in normal human plasma. J.clin. Invest. 1961, 40, 723.

35. Walser, M. The separate effects of hyperparathy-roidism, hypercalcemia of malignancy, renal fail-ure, and acidosis on the state of calcium, phosphate,and other ions in plasma. J. clin. Invest. 1962, 41,1454.

36. Walser, M. Ion Association. I. The effect of multi-valent anions on the protein-bound and com-plexed calcium in serum. J. cell. comp. Physiol.1960, 55, 245.

37. Loken, H. F., R. J. Havel, G. S. Gordan, and S. L.Whittington. Ultracentrifugal analysis of pro-tein-bound and free calcium in human serum. J.biol. Chem. 1960, 235, 3654.

38. Mairtensson, J. Hypercitricemia in Addison's dis-ease. Acta med. scand. 1949, 134, 61.

39. Walser, M. Protein-binding of inorganic phosphatein plasma of normal subjects and patients with re-nal disease. J. clin. Invest. 1960, 39, 501.

40. Robinson, B. H. B., M. Walser, and E. B. Marsh, Jr.Unpublished observations.

41. Gerbrandy, J., A. M. van Leeuwen, H. B. A. Hellen-doorn, and L. A. de Vries. The binding betweenelectrolytes and serum proteins calculated from anin zivo filtration method. Clin. Sci. 1960, 19, 181.

42. Finberg, L. Experimental studies of the mecha-nisms producing hypocalcemia in hypernatremicstates. J. clin. Invest. 1957, 36, 434.

43. Milne, M. 0., R. C. Muehrcke, and I. Aird. Pri-mary aldosteronism. Quart. J. Med. 1957, 26, 317.

ANNOUNCEMENTOF MEETINGS

THE AMERICANFEDERATION FOR CLINICAL RESEARCHwillhold its Twentieth Annual Meeting in Atlantic City, N. J., at the Casino Theatreon the Steel Pier on Sunday, April 28, 1963, at 9:00 a.m. Joint sectional meetingswith The American Society for Clinical Investigation will be held on Sunday after-noon at Chalfonte-Haddon Hall, and additional meetings sponsored by The Ameri-can Federation for Clinical Research will be held there on Sunday evening.

THEAMERICANSOCIETY FORCLINICAL INVESTIGATION, INC.,will hold its Fifty-fifth Annual Meeting in Atlantic City, N. J., on Monday, April29, at 9 :00 a.m. at the Casino Theatre on the Steel Pier and in simultaneous pro-grams sponsored with The American Federation for Clinical Research on Sundayafternoon, April 28, in Chalfonte-Haddon Hall.

THE ASSOCIATION OF AMERICAN PHYSICIANS will hold itsSeventy-sixth Annual Meeting in Atlantic City, N. J., at the Casino Theatre onthe Steel Pier on Tuesday, April 30, at 9:30 a.m., and in the Vernon Room,Chalfonte-Haddon Hall, on Wednesday, May 1, at 9:30 a.m.

THEAMERICANSOCIETY FORCLINICAL NUTRITION will hold itsThird Annual Meeting in Atlantic City, N. J., at the Colton Manor Hotel on Satur-day, April 27, from 1:00 to 5:00 p.m.

465'