CLINICAL DETERMINATION OF THE SITES OF RED CELLSEQUESTRATION IN HEMOLYTIC ANEMIAS1

By JAMES H. JANDL, MORTIMER S. GREENBERG, ROBERT H. YONEMOTO, ANDWILLIAM B. CASTLE

(From the Thorndike Memorial Laboratory and Second and Fourth (Harvard), Medical ServicesBoston City Hospital, and the Department of Medicine, Harvard Medical

School, Boston, Mass.)

(Submitted for publication January 30, 1956; accepted April 3, 1956)

The widespread application of red cell survivaltechniques has revealed the importance of exces-sive red cell destruction in the pathologic physi-ology of many of the anemias. An increasing ar-ray of in vitro methods for detecting red cell orserum abnormalities has provided insight into thein vivo mechanisms underlying some of these proc-esses. In certain disease states the presence ofvisible or physically measurable alterations of thered cells has permitted detection of the sites andto some extent of the mechanisms of sequestrationof these cells. Such valuable observations havebeen made upon pathologic material from patientswith congenital hemolytic anemia (1-5) andsickle cell anemia (1, 5-7). The need has con-tinued to exist, however, for a clinical method bywhich to determine the sites of red cell destruction,both for an understanding of the relationship be-tween red cell abnormalities and the body's redcell clearing mechanisms, and for pragmatic aidin the often difficult decisions concerning splenec-tomy.The potential usefulness of Na2Cr5104-labelled

red cells in the clinical determination of the sitesof red cell destruction was suggested to us by ob-servations on postmortem tissues taken from apatient with the hemolytic anemia of liver disease,who succumbed during observation of the sur-vival of Cr5'-labelled autogenous red cells.2 TheCrp5 activity of the patient's splenic tissue was

1 This investigation was supported in part by a grantfrom the Helen Hay Whitney Foundation and by Re-search Grant RG3507 (C3) from the National Institutes ofHealth, Public Health Service.

2 Cr' either as Cr"OJ- (8) or as Cr5' "' (9) iscapable under appropriate conditions of firmly labellingred cells by chemical attachment to the hemoglobin andto the red cell membrane, respectively. For brevity, how-ever, red cells labelled with Na,Cr"04 will be describedhere as "Cr"-labelled red cells."

greater than that of other tissues even when cor-rection was made for the Cr5l activity of the re-sidual red cells. Moreover, the radioactivity ofthe packed red cells removed from the spleen ex-ceeded that of a comparable sample of packed redcells from the peripheral blood. In order to in-vestigate the possibility that Cr51-labelled red celldeposition could be determined by measuring bodysurface radioactivity, several questions requiredexploration: 1) Are the emanations of Cr5 suit-able for external body scanning at safe dosagelevels? 2) Does the site of tissue deposition ofCr65 following the intravenous injection of Cr51-labelled red cells necessarily indicate the site ofdeposition of whole red cells? 3) Is the turnoverof Cr5I deposited in tissues sufficiently slow toprovide a detectable accumulation? 4) Finally,do the results obtained confirm the informationalready established in certain specific instances,for example, in congenital hemolytic anemia andin sickle cell anemia? In order to investigatethese questions, observations were first made onthe distribution of Cr5' in normal subjects. Thefate of Cr51-labelled autogenous red cells was thenobserved in patients with known intracorpusculardefects and, lastly, in patients with certain diseasesgenerally characterized by excessive red cell de-struction and a preliminary report (10) was made.

METHODS

Cr' labelling technique. All red cell labelling was per-formed under sterile conditions with Na,CrMO43 dilutedwith normal saline to a concentration of about 30 micro-grams of chromium per cubic centimeter. Between 100and 150 microcuries of Cr5' in this form were added withimmediate mixing to 50 cc. of whole blood in a siliconizedflask containing 12 cc. of acid-citrate-dextrose solution.The flask was then gently agitated continuously for 45

3 "Rachromate," Abbott Laboratories, North Chicago,Illinois.

842

DETERMINATION OF THE SITES OF RED CELL SEQUESTRATION

FIGURE 1The gradual, almost linear, disappearance of Cr'-labelled normal autogenous red

cells from the blood stream shown in the upper portion of the figure is reflected by,a gradual accumulation of Cr' in the tissues. Little difference exists in the levelsof relative radioactivity over liver and spleen, as depicted in the lower portion ofthe figure.

minutes. Unbound or reduced chromium was removedby washing the red cells once in sterile normal saline,after which the red cells were suspended in 2 volumes ofnormal saline preparatory to injection. Over 99 percent of the Cr' injected was in the red cells.

Solutions of Cr'-labelled hemoglobin were preparedby similarly labelling red cells which were thereafterwashed 5 times in normal saline and then hemolyzed with4 volumes of distilled water.4 After adjusting the sodiumchloride concentration of the suspending fluid to 0.85 gm.per cent, the red cell "ghosts" were centrifuged down andthe supernatant solution was Seitz-filtered prior toinjection.

4 It is equally feasible to label the hemnoglobin in solu-

tion after hemolysis; this may be done either withNa2CrMO4 or Cr'Cl..

A solution of a reduction product of Na2CrkO,, Cr'G13,5was diluted with normal saline to a suitable volume; thiswas injected directly intravenously in one instance andafter preliminary incubation with serum in another.Specimen collection and Crt determination. Following

injection of Crc-containing preparations, venous bloodspecimens were collected at appropriate intervals inbottles containing dry "balanced" oxalate. Their radio-activity was determined after hemolysis by freezing andthawing. Specimens for the measurement of plasma ra-dioactivity were drawn into saline-wetted syringes andthe citrated plasma removed after centrifugation of thespecimens within one hour of their procurement. Theplasma samples were then frozen and thawed along with

5 Radioactive chromic chloride, Abbott Laboratories,North Chicago, Illinois.

Cr5I-LABELLED AUTOGENOUS RED CELLSSUBJECT NORMAL

90

I0-

zw 70

60

X.- soBLOOD

00

z30o _ _ _ ~~~~~~~~~~~~~TISSUE

z_ _ - ~~~~~~~~~ _ o ~~~URINE

o 4

31.-03

o,0

O * _ SPLEENP LIVER14

e 00 i'o l's 2o £ODAYS

843

J. H. JANDL, M. S. GREENBERG, R. H. YONEMOTO, AND W. B. CASTLE

the whole blood samples. Plasma Cru activity was ex-pressed as a percentage of the whole blood activity bycorrection for the plasma volume of a sample of wholeblood. When possible, all urine was collected as indi-vidual specimens for about 18 hours after the Cr' in-jection, and thereafter 24-hour urine collections weremade. In several subjects serial 2- or 4-day total stoolcollections were made. The stool specimens were di-gested with 10 per cent sodium hydroxide and mixed ina blender before sampling.The radioactivity of all types of materials was deter-

mined with a well-type scintillation counter. The "zero-time" peripheral blood radioactivity was estimated by ex-trapolation of a line drawn visually through points onsemilogarithmic graph paper on which the concentration

of whole blood Cr' was plotted against time. From thisvalue the initial blood volume was calculated by the dilu-tion principle and the CrM concentration of the circulatingblood was thereafter expressed as a percentage of thezero time value. The Cr'1 content of the circulating bloodwas calculated as the product of the blood Cr' concen-tration and the blood volume, by making the assumptionthat the latter was unchanged during the period of study.Red cell survival at any time was then taken as the per-centage of the injected Cr's which remained in the cir-culating blood. For graphic presentation data so ob-tained were plotted against time on rectilinear graphpaper without correction for Cr' "elution" from the redcells. The blood Cr'1 activity was followed in most sub-jects until its disappearance. However, for easier com-

FIGUE 2

Following intravenous injection of Crh-labelled hemoglobin solution, radioac-

tivity disappeared from the blood at a continuously slowing rate with a half-disap-pearance time of 3 hours, whereas by chemical determination plasma hemoglobinwas removed from the circulation "exponentially" with a half-time of from 1% to2 hours. With the removal of Cr' from the peripheral blood, there was a rapid,prominent rise in radioactivit-y over the liver and to a lesser extent over the spleen.

-if

844

Cr5l- LABELLED HEMOGLOBINSUBJECT NORMAL

I..

too

I-

SKS

o

'a.0SW -V

<

0 3w . . . .

DAYS

DETERMINATION OF THE SITES OF RED CELL SEQUESTRATION

FIGURE 3The upper portion of this figure indicates the great rapidity with which injected

cationic trivalent Crm in saline is cleared from the circulation. Most of this chrom-ium appeared in the urine, with only a slight accumulation by the liver and spleen.Preliminary incubation of Cr'Cl, and serum before injection enhanced the rela-tive Cr's uptake by the liver.

parison of data, a time axis of only 30 days was em-ployed in the presentation of the studies of individual pa-tients shown below in Figures 1 to 13. The urinaryexcretion of Cr' was expressed as the cumulative per-centage of the injected dose of Cr'. Since in no in-stance was Cr'1 excreted in the patients' stools, the chro-mium not accounted for in blood or urine was designatedas "tissue Cr"' in the figures and in Table I.Body surface counting. Body surface counting was

performed with a directional scintillation counter havinga conical collimator 4 inches in depth and 3 inches in in-ternal diameter at the open end. Twenty minutes afterthe injection of materials labelled with CrM, each patientwas scanned. In no instance among the patients re-ported here were high levels of activity found other thanover the heart, liver, or spleen. In each patient, there-

fore, surface projections of these organs were deter-mined by physical examination and their approximatecenters were marked with a skin pencil. In some patientsradioactivity levels were also followed over the thoracic,lumbar and sacral areas. In performing each count thescintillation counter was plumb-centered over the ap-propriate mark with the patient in a standardized position,permitting the body surface overlying the organ to ap-proximate parallelism with the collimator surface. Thecounter was then lowered to direct apposition with thebody surface and two or more measurements were madewith an automatic scaler.

In measuring the body surface radioactivity, it wasrecognized that a variable but important degree of radio-activity over each organ derived from the activity of theperfusing blood, as well as from stray radiation from one

Cr51 C13SUBJECT: NORMAL

loo

900so

o

0 r

O40s

z 30 - - - - - - - - - - - - - - - - - - -

0

.1-SO

go

.510

AL

o-i

I -

FL LIVTISSE

hi0~ ~ ~ ~ ~ ~ DY

845

J. H. JANDL, M. S. GREENBERG, R. H. YONEMOTO, AND W. B. CASTLE

organ to another. Since the geometric complexities in-volved prevented calculation of the absolute levels oforgan radioactivity, a relative expression was employed.The precordial Cr' activity grossly paralleled the activityof whole blood specimens and presumably emerged largelyfrom the perfusing blood.6 Body surface counts weretherefore expressed as the ratio of the radioactivity overa certain organ to that at the same time over the heart.This ratio represents a function of the quantity of Cr'accumulated in the organ concerned, whether in red cells

Is It is apparent, however, that precordial radioactivityreflects blood radioactivity most accurately when thereis the least stray irradiation from neighboring organssuch as the spleen.

or not, and-may be regarded-as a -crude measure of agradient between the organ and the circulating blood.Miscellaneous techniques. Plasma and urine hemo-

globin levels were measured by a photoelectric adaptation(11) of the benzidine method of Bing and Baker. Themean normal value for plasma hemoglobin concentrationin this laboratory is 2.6 + 1.2 mg. (2 S.D.) per cent.

Clinical material. The "normal" subjects studied wereelderly males hospitalized for minor illnesses and/or forsociologic reasons. Three of these were given autogenousCr'-labelled red cells. The 11 anemic patients displayedfindings characteristic of their diseases. A brief summaryof the vital statistics and of some of the hematologic datafrom these 14 subjects and patients is presented in Table I.

FIGURE 4A

The increased rate of Cr'-labelled red cell destruction in a patient, Case 1, withmoderately severe congenital hemolytic anemia (hereditary spherocytosis) was

attended by a predominant uptake of radioactivity over the spleen. Note both the

increased splenic activity at zero-time and the manner in which Cr' is progressivelydeposited in the spleen in proportion to its removal from the blood stream.

846

Cr5 - LABELLED AUTOGENOUS RED CELLSSUBJECT: CONGENITAL HEMOLYTIC ANEMIA

Go40

o!0 o

hI~~~~RN*10z?O~~~~~~~~~~~~~L

50

40

20

DAYS

m

DOD

DETERMINATION OF THE SITES OF RED CELL SEQUESTRATION

FIGURE 4BThe rate of red cell destruction in this patient, Case 2, with mild congenital he-

molytic anemia (hereditary spherocytosis) was less rapid than in Case 1. Therewas also a less pronounced but still predominant uptake of Cr' by the spleen.

RESULTS

General observations on red cell survival and cal-culated daily hemoglobin production

The half-survival of Cr51-labelled autogenous red

cells given 3 normal subjects was 33± 2 days(mean: 32.8). The red cell survival curves were

slightly concave when the uncorrected data were

charted on rectilinear graph paper. When thesedata were corrected for an in vivo "elution" of Cr51from Cr51-labelled red cells of 1 per cent daily as

suggested by Ebaugh, Emerson, and Ross (12),the normal red cell survival data formed a

straight line on rectilinear graph paper, indicating

red cell destruction by senescence resulting in a

mean cell life span of 112 days. On the other handin all but 1 of the 11 patients with hemolytic ane-

mias the blood Cr51 data uncorrected for elutionformed a straight line when plotted against time on

semilogarithmic graph paper. Red cell destructionin these subjects was thus "exponential" or ran-

dom in nature indicating that senescence did notdetectably participate in the process, presumably.since the rapidity of the random process (orprocesses) minimized cell death by aging. Therates of red cell destruction were thus calculateddirectly from the slopes of the red cell survivalcurves and are given in Table I. The correction

Cr51- LABELLED AUTOGENOUS RED CELLS

SUBJECT: CONGENITAL HEMOLYTIC ANEMIA

100

90a0--70 \

60 SQ~~0

0 BLOOD40-

zhi> 3030

ohiB. lo

0

6

0I1->- 5

40-4t 4

IL-SPLEEN(00 2

0

0 0 LIVER

E 00 5 X IS io i5 30DAYS

847

J. H. JANDL, M. S. GREENBERG, R. H. YONEMOTO, AND W. B. CASTLE

factor for the "elution" of Crl1 from normal redcells was again employed, although recognizingthat this rate may actually differ for diseased redcells. In one anemic patient, Case 9, despite stableperipheral blood levels during the observation pe-riod, the Cr51-labelled red cell survival data cor-rected for elution produced a straight line onneither rectilinear nor semilogarithmic graph pa-per. This survival curve presumably reflected amixture of accelerated senescence and of randomdestruction. The rate of red cell destruction inCase 9 was estimated by a simple method for de-termining the mean red cell life span based on theconcept of total red cell days (13).The rates of hemoglobin production shown in

Table I were determined in those 10 anemic sub-jects manifesting random red cell destruction asthe product of the calculated rate of red cell de-struction and the mean total circulating hemo-globin level during the period of observation. Inthe normal subjects and in one anemic subject,Case 9, with a mixed hemolytic process hemo-globin production was calculated on the basis ofthe estimated mean red cell lifespan.

Fate of Cr"l derived from labelled red cells, hemo-globin, and chromic chloride

Observations on one of the normal subjects whoreceived Cr5l-labelled autogenous red cells are pre-

FIGuRE 5AThe CrT-labelled red cells of a patient, Case 3, with moderately severe sickle

cell anemia were removed from the circulation in an "exponential" fashion andshowed a half-survival time of 11.2 days. There was a moderate increase in radio-activity over the liver but not over the splenic area.

CrSI- LABELLED AUTOGENOUS RED CELLSSUBJECT SICKLE CELL ANEMIA

I *,

a.~~~~~~~ah X

I-~~~~d

ffib - _ _ _ _ _ _ ~TSUg

to SLOOS

0 4s0

0g..~~ ~ ~ ~ ~ ~ A

00 lbIlb= * * * *^DAYDAYS

848

DETERMINATION OF THE SITES OF RED CELL SEQUESTRATION

FIGURE SB

The simultaneously observed survival of red cells from Case 3 in a normal com-patible recipient was appreciably shorter with a half-survival time of 6.0 days.This was associated with high levels of radioactivity over both spleen and liver.Note also the smaller excretion of Cr'1 in the urine of this subject.

sented in Figure 1. In the upper portion slow ac-

cumulation of Cr5l in the tissues is evident,amounting to about 60 per cent of the Cr51 re-

moved from the circulation at the red cell half-survival time. A high, transient Cr51 excretion inthe urine followed for 1 to 2 days the injection ofunwashed Cr'l-labelled red cells in subjects notpresented here, whether or not a reducing sub-stance such as ascorbic acid had been added to thesample of whole blood prior to injection; however,no such early hyperexcretion of Cr5' was evidentin the normal subjects who received washed Cr51-labelled red cells. The radioactivity over the liversas well as over the spleens of normal subjects 20minutes after the injection of Cr51-labelled red

cells ranged from 60 to 80 per cent of that overthe precordium; and even after 30 days this ratioincreased little. The lower portion of Figure 1depicts the slow minimal increase in relative radio-activity occurring almost equally over liver andspleen thereafter.Each of three other normal subjects (not listed

in Table I) was given 1 gm. of Cr51-labelled au-togenous hemoglobin injected intravenously overa period of 2 or 3 minutes. This produced nohemoglobinuria or hemosiderinuria and led tozero-time plasma hemoglobin levels which rangedfrom 30 to 40 mg. per cent. Hemoglobin, as de-termined colorimetrically, was cleared from theplasma in an exponential fashion with a half-disap-

849

J. H. JANDL, M. S. GREENBERG, R. H. YONEMOTO, AND W. B. CASTLE

pearance time of from 1.5 to 2 -hours. In con--trast, both the whole blood and the plasma Cr5'activity disappeared at a slower rate with a haf-disappearance time of about 3 hours and a con-tinually slowing rate of disappearance thereafter(Figure 2). Although this in vivo "separation"of chromium from hemoglobin began shortly afterinjection, Cr51-labelled hemoglobin incubated at370 C. in vitro in the presence of serum retainedits radioactivity for 24 hours, as determined byelectrophoresis and radioautography.7 As depicted

7 In a similar study Crk initially bound to gammaglobulin (Cohn fraction II-1, 2) migrated electropho-retically with the albumin and alpha globulin after 24hours' incubation of the labelled gamma globulin withnormal plasma.

I

a

1-

0

I-

z

IL0.

Cb0aw

I I

in Figure 2, after 24 hours in vivo 60 per cent, andby 8 days 75 per cent of the Cr5l removed fromthe blood stream was deposited in the body tis-sues. The highest level of body radioactivity ap-peared over the liver and approached a maximumwithin 6 hours; a parallel but lesser rise in radio-activity appeared over the spleen. Following theinjection of 1 gm. of Cr51-labelled hemoglobin intoa splenectomized patient the rate of disappearanceof Cr5' and of hemoglobin from the plasma wassimilar, but there was no increase in radioactivityover the "spleen" area. The maximal hepatic ac-tivity in this patient, as in normal subjects, ap-proximated 3 times that of the precordium.Twenty-four hours after the injection of Cr51-

FIGURE 6

The red cell half-survival time in this patient, Case 4, with Cooley's anemia(thalassemia major) was 16.5 days. The high initial relative radioactivity over

the patient's large spleen indicates a large vascular organ. The subsequent pro-gressive increase of radioactivity over the patient's spleen mirrored the Cr4-labelledred cell survival curve and indicates progressive red cell sequestration by the spleen.

Gr5l-LABELLED AUTOGENOUS RED CELLSSUBJECT: COOLEY'S ANEMIA

- LOOO

SPLEEN

IV_ __ _ _ __ _ __ _ _ __ _ __ _ _ LIVIER

7

850

DETERMINATION OF THE SITES OF RED CELL SEQUESTRATION

Cr5S - LABELLED AUTOGENOUS RED CELLSSUBJECT: PAROXYSMAL NOCTURNAL HEMOGLOBINURIA

0 7

la

04

I-

0

_s 7

4 O

1160

4 4

O 3m

0

aI.

44

- BLOOD

I I

0 SPLEEN

0---o LIVER

0

- 0, -o ~ IE

i 1o 15DAYS

FIGURE 7AThe early rapid phase of red cell destruction depicted in this patient, Case 5,

with paroxysmal nocturnal hemoglobinuria and splenomegaly presumably repre-sented destruction of the patient's own red cells with a half-survival estimated atabout 7.2 days. The late slow phase was presumed to represent disappearance ofthe normal previously transfused red cells which were estimated to constitute 30per cent of those in circulation at the time of labelling with chromium. An in-creased accumulation of Cr6' by the patient's enlarged spleen was observed. Thisuptake was about half that of a patient with hereditary spherocytic anemia, Case 1,whose red cell survival was similar (Figure 4A).

labelled autogenous hemoglobin solution into a

dog, the concentrations of radioactive chromium inthe liver and spleen of the sacrificed animal were

similar and greatly exceeded those of the lungs,lymph nodes and bone marrow. An intermediateconcentration of radioactivity was found in thekidney. The major amount of chromium was thuspresent in the heaviest organ, the liver.The fate of a reduction product of sodium

chromate, trivalent cationic chromium, was in-

vestigated in two normal subjects. In the firstsubject, Cr511CJ was injected in saline intrave-nously; and in a second subject after preliminaryincubation with the recipient's serum at 370 C. forone hour. In both, most of the Cr5' was promptlyremoved from the blood stream and in large partexcreted in the urine. The half-disappearancetime of Cr5' from the blood stream was about 15minutes in each instance. In the first subject, as

shown in Figure 3, a slight progressive hepatic

6.

6

4.

2-

Il

06

p

20 25 30

851

J. H. JANDL, M. S. GREENBERG, R. H. YONEMOTO, AND W. B. CASTLE

uptake of Cr5' occurred. In the second, a hepaticuptake of two times that of the precordium was ob-served at the end of 24 hours. In neither case wassignificant splenic radioactivity noted.

Patients utith intracorpuscular defects

Congenital hemolytic anemia. Studies of apatient, Case 1, with moderately severe congenitalhemolytic anemia (hereditary spherocytosis) dem-onstrated an abbreviated exponential red cell sur-vival curve in which half the cells were removed in8.0 days (Figure 4A). Radioactivity over themoderately enlarged spleen was 124 per cent ofthat of the precordium at zero-time and the rela-tive splenic radioactivity increased steadily in gen-eral proportion to the decline in blood activity to

reach about 400 per cent of that of the precordiumafter 25 days. Hepatic Crp5 activity remained low.No hemoglobin or Cr5l appeared in the patient'splasma.A second patient, Case 2, with hereditary

spherocytosis and anemia of tmild intensity,showed similar but less pronounced findings (Fig-ure 4B). The red cell half-survival in this patientwas 11.9 days. At that time the relative splenicradioactivity was over twice that of the pre-cordium. Again, the initial radioactivity over thepatient's enlarged spleen exceeded that of the nor-mal subjects and that over the liver remained low.A similar increase of Cr5' occurred over the spleenof a normal subject of compatible blood groupwho received labelled red cells from Case 2. The

FIGURE 7BInjection of the red cells of the same patient, Case 5, with paroxysmal nocturnal

hemoglobinuria into a normal compatible recipient resulted in a longer red cell

half-survival time of these cells, 16.0 days. As indicated above no definite Cr51 up-take was evident over either liver or spleen.

CfS-LADELLED PAROXYSMAL NOCTURNAL HEMOOLOBINURIA RED CELLSSUBJECT NORMAL

soo

soA

a 40

a .z BLOOD

10~~~~~.~.o0 URINE

IL t 00LvE"0

44

"3

0

OS.~~~~~~~DY

852

DETERMINATION OF THE SITES OF RED CELL SEQUESTRATION

FIGURE 8The moderate increase in autogenous red cell destruction in a patient, Case 6,

with homozygous hemoglobin C disease was associated with only a slight rise in

radioactivity over spleen and liver. The Cr' accumulation in the spleen probablyfalls within normal limits.

patient's red cells were destroyed more rapidly inthe normal recipient than in the patient himself.

Sickle cell anemia. The fate of the red cells ofa patient, Case 3, with sickle cell anemia is depictedin Figure 5A. Half of these red cells were re-

moved from the blood stream in 11.2 days at whichtime about 70 per cent of the derived Cr5' (35 per

cent of that injected) was retained in the body. Amoderate gradual increase in hepatic Cr5' uptakeaccompanied the disappearance of Cr5l from theblood, whereas "splenic" radioactivity was ini-tially abnormally low and increased little there-after. A second patient with sickle cell anemia,not reported here, showed almost identical find-

ings. Simultaneously with the study depicted inFigure 5A, a portion of the labelled red cells fromCase 3 were injected into a normal compatible re-

cipient. The rate of destruction of the sickle cellanemia red cells was distinctly greater than in thepatient herself. Their half-survival time was 6.0days, and at that time a larger fraction, over 80 per

cent, of the derived Crp5 was retained in the sub-ject's tissues. In this recipient a pronounced andprogressive splenic uptake of Cr5' accompanied a

moderately heavy increase in hepatic radioactivity.In neither subject did Cr5l appear in the plasma.

Cooley's anemia. The half-survival time of theautogenous labelled red cells of a patient, Case 4.

853

J. H. JANDL, M. S. GREENBERG, R. H. YONEMOTO, AND W. B. CASTLE

with Cooley's anemia- (thalassemia major> was transfused cells. Associated-with a period of rapid16.5 days (Figure 6). There was, accompanying red cell destrudtion during the first 15 days afterthis, a moderate increase in splenic radioactivity the injection of the labelled mixed red cell popula-beyond the initially elevated relative radioactivity tion, which presumably represented largely de-over the patient's moderately large spleen. struction of autogenous red cells, there was aParoxysmal nocturnal hemoglobinuria. The moderate increase in radioactivity over the pa-

half-survival time of the Cr51-labelled autogenous tient's enlarged spleen. Subsequently a slight de-red cells of a patient, Case 5, with paroxysmal-noc- cline in relative radioactivity over both spleen andturnal hemoglobinuria was 7.2 days (Figure 7A). liver attended a progressive slowing of the rate ofAt the time of labelling approximately 30 per cent Cr5l disappearance from the blood presumably as-of the circulating red cells were normal previously sociated with the slower destruction of the normal

FIGURE 9

In this patient, Case 7, with severe Laennec's cirrhosis and macrocytic anemiaassociated with; chronic alcoholism the increased rate of disappearance of Cr" fromthe blood indicated a moderately severe hemolytic process characteristic of thisdisease. This was associated with a distinct and progressive accumulation of Cr"over the patient's non-palpable spleen. Note also the initial and sustained abnor-mally low relative radioactivity over the patient's enlarged liver. This was ob-served in 3 other patients with cirrhosis, and presumably reflects a reduction inhepatic vascularity.

854

0

6

2

TISSUEBLOOD

URIHE

20 30

FIGURE 10

.The survival time of the red cells of a patient, Case 8, with untreated perniciousanemia was moderately reduced. The moderate increase in radioactivity over thepatient's non-palpable spleen indicates that the red cells were largely sequestered inthat organ.

transfused red cells. One hour after the injection

of the labelled red cells 0.6 per cent of the whole

blood radioactivity resided in the patient's plasma

indicating prompt release of labelled hemoglobin

from destroyed red cells. Plasma radioactivity in-

creased to 4.7 per cent of that in the whole blood 1

day later, remained at about 5 per cent for 4 more

days and then gradually diminished. The ratio of

Cr51 to hemogloliin in the plasma was several times

greater than that in t,he red, cells. No plasma Cr51

activityr remained at 14 days. In contrast to the

deposition of Cr51 from labelled red cells largely

sequestered in the spleen, the injection into this

patient of Cr51-labelled hemoglobin (not shown in

Figure 7A) led to an increase in relative radioac-

tivity over both the liver and spleen, with the

hepatic uptake predominating slightly. The half-

survival time of the Crpl derived from labelled

hemogloybin was about three and one-half hours.

Labelled red cells from Case 5 were destroyed

less rapidly when transfused into a normal com-

patible subject (Figure 7B). In this instance the

red cell half-survival time was 16.0 rather than 7.2

days, and there was no apparent sequestration by

eith,er sple,en or liver. Plasm,a radioactivity was

not measured. The possibility, implied by these

results, of an ex,aggerated red cell sequestering ac-

tion by the,patient's clinically somewhat enlarged

spleen superimposed upon the underlying intra-

vascular hemolytic process was borne out by the

855DETERMINATION OF THE SITES OF RE3D CELL SEQUESTRATION

Cr59 - LABELLED AUTOGENOUS RED CELLSSIJBJECT: PERNICIOUS ANEMIA

9

hi

o

I

J LIVER

S 10 15

DAYS

-

o

856 J. Hi. JANDL, M. 5. GREENBERG, R. H. YONEMOTO, AND W. B. CASTLE

subsequent beneficial effect of the surgical removal chronic alcoholism demonstrated the elevated retic-of a spleen weighing 660 gin. ulocyte levels (ranging from 8 to 24 per cent), hy-Hemoglobin C disease. A patient, Case 6, with peractive erythroid elements in the bone marrow,

a slight anemia, normal reticulocyte levels, many and increased urobilinogen excretion frequentlytarget cells, and a hemoglobin mobility on filter observed in this disease (14). The red cell half-paper electrophoresis characteristic of hemoglobin survival time was 6.0 days (Figure 9). ThereC disease, demonstrated a moderate reduction in was a moderate progressive increase of radioac-the survival of autogenous red cells labelled with tivity over the spleen. The Cr5' radioactivity overCrl (Figure 8). Associated with this was a small the liver one-half hour after injection of the la-gradual increase in radioactivity over the spleen belled red cells was abnormally low, 43 per centand somewhat less over the liver, of that over the precordium, and remained low.

In 3 other patients with cirrhosis this initial valuePatintswithextacopusclaror ixeddefcts ranged from 40 to 60 per cent whereas in 6 normal

subjects it ranged from 60 to 80 per cent of thatLiver disease. A patient, Case 7, with advanced over the precordium. Subsequently, Cr51-labelled

cirrhosis and macrocytic anemia associated with hemoglobin was injected intravenously into this

Cr1 -LABELLED AUTOBENOUS RED CELLS~SUBJECT: ANEMIA OF CHRONIC RENAL DISEASE

a1oh_

I-§t!w o$_-5s io " sz~~~~~~Dt

5~~~~~~~IU

Ths.ptet ae9 a arysvr,sal nmaascae ihuei0eutn rmcrncpdnprts esieaC"lWe e d afsrnaofI-r1. as oanra cuulaino r curdi ihrlvrospen

DETERMINATION OF THE SITES OF RED CELL SEQUESTRATION

Cr51-LABELLED AUTOGENOUS RED CELLSSUBJECT : ACQUIRED HEMOLYTIC ANEMIAlool'WITH AUTO-AGGLUTINATION

w ~~~~~~ ~TISSUE

F - /

S \ /

050 /

040,

~3020

URINE

so Ro Io/10o

SPLEEN

LIVER

10 ISDAYS

FIGURE 12A

The extremely rapid disappearance of Cr'-labelled red cells in a patient, Case 10,with a fulminating hemolytic process of unknown etiology was associated with anincreased tissue retention of Cr'1 and with high levels of radioactivity over bothspleen and liver. Autoagglutination of the red cells in the peripheral blood andhemoglobinemia were present.

20 ts

subject in divided daily doses calculated to ap-proximate the amount of hemoglobin which hadbeen released from the Cr51-labelled red cells eachday. The Cr51 injected in this manner accumu-lated largely in the patient's liver rather than inthe spleen.

Pernicious anemia. The survival time of redcells from patients with untreated pernicious ane-mia when transfused into normal subjects has beenobserved by the Ashby technique to be diminished(15, 16). A reduced survival of red cells la-be.lled with Crp5 was also observed here in a pa-tie:it, Case 8, with moderately severe untreatedpernicious anemia (Figure 10). A moderate ac-cumulation of Crll in the patient's non-palpable

spleen was also demonstrated whereas there wasno definite uptake of Cr5' by the liver.Anemia of chronic renal disease. A patient,

Case 9, with chronic "stationary" uremia had asustained, moderately severe, normocytic, normo-chromic anemia. The reticulocyte levels rangedfrom 1 to 4 per cent; there were about 5 per centof "burr cells" in the peripheral blood, and thebone marrow aspirate was normal in appearanceand cellular composition. The patient was foundto have a reduced survival time of Cr51-labelledautogenous red cells (Figure 11), a frequent find-ing in uremia (17-19). No increased Cr5' uptakeby either liver or spleen attended this process.Despite the shortened red cell survival, as in the

857

858 J. H. JANDL, M. S. GREENBERG, R. H. YONEMOTO, AND W. B. CASTLE

patient, Case 8, with pernicious--anemia, 'it-ism-i- cels were beingdee 'dailRy and that the dailyteresting that by calculation underproduction was hemoglobin production averaged 55 gm. This re-the greater factor in the production of this patient's markably rapid removal of Cr51-labelled red cellsanemia (Table I). from the blood stream was accompanied by an in-

Acquired hemolytic anemia. A patient, Case creased Crp5 uptake by the spleen and also, to only10, with severe, acute, idiopathic acquired hemo- a slightly lesser extent, by the liver. A repetitionlytic anemia of unknown etiology associated ini- of these studies was undertaken 3 weeks later,tially with hemoglobinemia and hemoglobinuria when auto-agglutination and hemoglobinemia hadwas studied on two occasions. During the first disappeared, and the signs of hemolysis had les-period of observation spherocytosis, mild auto- sened. At this time the patient's red cell half-agglutination of the red cells in the peripheral blood survival time had increased to 7.8 days, while theand a positive Coombs test were found, and the pa- spleen alone accumulated Cr51 (Figure 12B).tient's reticulocyte levels ranged from 17 to 47 per A second patient, Case 11, with idiopathic ac-cent. The half-survival time of the patient's Cr51- quired hemolytic anemia exhibited a remarkablylabelled autogenous red cells was only 2.3 days. constant degree of anemia for several months,It was estimated that about 29 per cent of her red with hemoglobin concentrations of from 6.0 to 7.0

FIGURE 12BRepetition 3 weeks later of the injection of Cr'-labelled autogenous red cells re-

vealed an improved red cell survival at a time when the anemia had diminished.At this time autoagglutination and hemoglobinemia had disappeared, but there was

a persistence of red cell sensitization, and less striking evidence of red cell seques-tration which was now confined to the spleen.

0r5l-LABELLED AUTOGENOUS RED CELLSSUBJECT ACQUIRED HEMOLYTIC ANEMIA

90

Io

zo60,> St \9 TISSUE

11640 04b //\tl:

///

10 / URINE BLOOD

O

>16

,F

SvPLEEN

OIL0

0 S510 1 f -' tDAYS

m

L

DETERMINATION OF THE SITES OF RED CELL SEQUESTRATION

FIGURE 13In this patient, Case 11, with moderately severe, stable idiopathic acquired hemo-

lytic anemia associated with evidence of red cell sensitization approximately 20 percent of the patient's Cr'-labelled red cells were removed from the circulation daily.A heavy accumulation of Cr' in the spleen mirrored the rapid disappearance ofCr' from the peripheral blood.

gm. per cent and reticulocyte levels ranging from19 to 34 per cent. The direct and indirect Coombstests were persistently positive and slight hemo-globinemia (10 to 15 mg. per cent) continued tobe present. An estimated 20 per cent of the pa-tient's Cr51-labelled red cells were destroyed dailyand high levels of radioactivity accumulated overthe slightly enlarged spleen (Figure 13).

DISCUSSION

Cr51 in the determination of the sites of red cellsequestration

The observations made on the normal subjectsindicate that the excretion and body distribution

of Cr5l derived from red cells labelled withNa2Cr51O4 differs characteristically from that ofCr5' derived from Cr51-labelled hemoglobin andfrom saline solutions of Cr51CJ3. The tissue con-centrations in the dog's liver and spleen of Cr5'derived from Cr51-labelled hemoglobin were simi-lar. Thus, the comparatively large liver containeda larger total amount of Cr51. This presumablyexplains the high radioactivity values found overthe livers of normal human subjects who were in-jected with labelled hemoglobin. A comparabledistribution of radioactivity between liver andspleen in rats was observed by Finch and hisassociates (20) with Fe 9-labelled hemoglobin.

859

Cr51-LABELLED AUTOGENOUS RED CELLSSUBJECT:ACQUIRED HEMOLYTIC ANEMIA

100

80TO

z

60.us

U.o 40

3..z30 010 20c:

20 <BLOOD

hi 13

O I00

-~6

SPLEEN

40

04

>W2LIVER

0

ox 0

0 10 IS 20 25 30DAYS

co C1 _ __m

C4 C I Iq 00 I )

_i I_ I ____ ___ Xw

I e I | C)

) |) e| 00 00 | UC4 e.s

C4oo is0 oV0 4'. 00F.oq0

co~~~~~~-coRC

0 o" 4"' U)oo eq '0+ 02]

oo eq WE) CO e o o00 e

U) U) '0o U)o VooX o

inC4 00E000' S

- U0eqo eqo g

U)U) 0%oUs .lX2W o

.U)00 00 coU) 44- eq et-c CN s.

-___- __ - ___ t_ _ _ _ - __ _ 6. 2

00 oo00I 0I r0o|o |o Uo 0% 0°|)

eq 0- t-eq UrE) .j U

lll l l l ~~0]}4)

01u>''-6e;c -

eq eq

d 0

0 U) VIO VDI V0 00 0_4.01

eq eq0o 00U4 r#0 CV)c 2 Wo.o4-. U') V 01 CoC~t.V

-O e eq - *u

ONV0 0~~~~~~~~~0

U)0U)0% U) 00eq ~~4 '~ ~eq ) e e-C;e F0.a I 3

u

Q4,.Ita

.1

.1Z4CI*

*'2

.%4414-I

1~

I131a

:

I

mA

14

14

*4,

4,E1Pe

I11IIA6

4DIAIRW U) ;ss in2 1 2 1 rz. 1 4. 1 4.

DETERMINATION OF THE SITES OF RED CELL SEQUESTRATION

The observed very abrupt clearance of Cr51CJ8from the circulation is consistent with the pro-pensity of cationic chromium to form colloidalmolecular aggregates above a pH of 6.0. Hereagain, the small fraction of Cr5' not excreted waspredominantly taken up by the normal liver as de-termined by body surface counts; the hepatic up-take of Cr5' was more striking when Cr51Cl hadbeen incubated with serum prior to injection.These results conform with studies of the chro-mium distribution in rat tissues following the in-jection of Cr51CJ. (21).

Because of the softness of the x-ray and gammaray emissions of Cr51, less than 20 per cent of itsradiations penetrate the trunk from a source placedbeneath a normal supine subject. Nevertheless,with use of the apparatus described, the quantityof Cr5' generally employed in red cell survivalstudies (12) is quite adequate for body surfacecounting. Indeed, the low energy and penetrationof these emissions reduces the factor of stray radi-ation from neighboring organs. An accumulationof local radiation obviously may arise from redcells in transit through or sequestered in a givenorgan, from released hemoglobin or from Crp5 insome other form deposited in the organ. Thegradual accumulation in the urine of Crp5 presum-ably results in part from its slow elimination fromthe tissues. On the basis both of direct bodycounts and of the urinary excretion of Cr51, therate of loss from the tissue was calculated in 4subjects who had abruptly sequestered labelledsensitized red cell suspensions (9) to averageabout 4 per cent per day; the half-life for tissueejlimination of Crp5 ranged from 14 to 21 days.Excretion of Cr5' was entirely renal. In no in-stance was there measurable Cr51 in the stool.Stool Cr5' measurements thus permit the readydetection of a potential error in red cell survivalstudies: occult intestinal bleeding.The findings in. normal subjects when com-

pared with those in the anemic patients presentedabove justify certain limited conclusions from bodysurface counting following injection of red cellslabelled with radioactive sodium chromate: (a)A progressive, high splenic uptake of Cr5' repre-sents the splenic sequestration of intact red cells.There appears to be no other way by which Cr5'from labelled red cells can be accumulated pre-ponderantly by the spleen. (b) A high hepatic

uptake of Cr5' in the presence of an intact spleensuggests the possibility, although not the certainty,that intravascular hemolysis of labelled red cellshas occurred.That the sequestration of whole red cells by the

spleen is indeed detectable by an attendant rise inradioactivity over the spleen was illustrated byCases 1 and 2. In these patients with congenitalhemolytic anemia the splenic Cr5' uptakes mir-rored the disappearance of chromium from theblood stream (Figures 4A and 4B). Further-more, the injection of Cr51-Iabelled red cells fromCase 2 into a normal compatible recipient also ledto a high relative radioactivity over the recipient'snormal-sized spleen. These observations cor-roborate the previously established evidence thatspherocytes may be selectively retained by eitherpathologic or normal spleens (1-5). The some-what longer survival of autogenous Cr51-labelledspherocytes in the patient, Case 2, with mild hered-itary spherocytosis than in a normal compatiblerecipient may have arisen from the competitionfor sequestration in the spleen of Case 2 by au-togenous unlabelled spherocytes. The moderatehepatic and comparatively low "splenic" uptake ofCr51-labelled autogenous red cells from a patient,Case 3, with sickle cell anemia (Figure 5A) wereinterpreted as reflecting the splenic atrophy char-acteristic of this disease (6, 7, 22), especially be-cause prominent progressive increases in splenicas well as hepatic radioactivity were observedwhen an aliquot of the sickle cell anemia patient'slabelled red cells was transfused into a compatiblenormal subject (Figure SB). That the sicklemiared cells survived only half as long in the normalsubject as in the patient herself presumably indi-cates in part the beneficial effect on red cell sur-vival of the "autosplenectomy" characteristic ofadult patients with sickle cell anemia. This is incontrast to the short survival time, even of normalred cells, in some children with sicklemia andlarge spleens observed by Lichtman, Shapiro,Ginsberg, and Watson (23). The considerablehepatic Crp5 uptake noted here also contrasts withthe preponderant uptake of Crp5 by the spleenalone in congenital hemolytic anemia. These ob-servations, again, are consistent with previousstudies (1, 5-7) revealing the hematologic basisfor the erythrostasis, thromboses and infarcts

861

J. H. JANDL, M. S. GREENBERG, R. H. YONEMOTO, AND W. B. CASTLE

which early cause severe damage to the spleen andwhich gradually injure the liver.

In addition to establishing the extent to whichCr5' is accumulated by the liver or spleen, it ispossible to discriminate by means of their respec-tive rates of increase between high relative radio-activity due on the one hand to sequestration oflabelled red cells and that on the other due tosimple organ enlargement with concomitant in-crease in vascularity. In the former situation localradioactivity necessarily builds up in an organ inapproximate proportion to its rate of disappear-

ance from the blood. In the latter situation thehigh initial relative activity over the organ (unlessfollowed at once by rapid accumulation) signifiesa large vascular bed rather than red cell sequestra-tion. Thus, the presence of a high initial ratioof spleen Cr5' to precordial Cr51, i.e., greater thanabout 90 per cent, merely indicates splenomegalywith an attendant increase in perfusing blood andunless the ratio subsequently increases, as for ex-ample in Cases 1, 2, 4, 5, 7, 10 and 11, does notnecessarily denote sequestration of red cells. Acase illustrating this point is presented in Figure

FIGURE 14This 48-year old Jewish female had splenomegaly and thrombocytopenia on the

basis of Gaucher's disease. The patient was not anemic, however, and her red cellsurvival, as indicated above, was almost normal. Although a high relative radio-activity appeared in the enlarged spleen shortly after the injection of Cr'-labelledautogenous red cells, there was no subsequent accumulation of Cr' by the spleen.This pattern of radioactivity reflects the vascular size of the spleen rather than thefiltration of red cells.

862

Cr51-LABELLED AUTOGENOUS RED GELL'S

SUBJECT: GAUCHER S DISEASE00

c90 <w

70Xw

W6 IBLOOD0 40

Z 30w

~ 0 o

CL10

>56

s-I-

maD0.j445

OD 1< --4 @ * * * ~~~SPLEEN0I

0LIVER

0 S 10l 15 2O 5 D0DAYS

DETERMINATION OF THE SITES OF RED CELL SEQUESTRATION

14. This patient had splenomegaly on the basisof Gaucher's disease and showed no evidence ofhemolysis. Her red cell half-survival time wasabout 25 days. Within a few hours of the injec-tion of Cr51-labelled autogenous red cells the radio-activity over the spleen was 125 per cent of thatover the heart; as blood Crp5 activity declined,however, there was no further uptake of radioac-tivity by the spleen, suggesting that this largespleen possessed a large vascular bed but did notsequester red cells. Accordingly, Crp5 may some-times be employed to determine the nature ofmasses in the left upper quadrant when this is nototherwise possible. It must be pointed out, how-ever, that splenomegaly may exist with little orno increase in relative radioactivity over the spleen.Thus, whereas the initial Cr5' activity over thespleen was elevated in all observed cases of sple-nomegaly associated with a hemolytic anemia orusually with cirrhosis, this was not so, despite thepresence of marked splenomegaly, in one case ofchronic lymphatic leukemia; and the Cr5' activitywas only slightly increased in one case of agno-genic myeloid metaplasia with massiye spleno-megaly. Presumably in these 2 patients spleno-megaly reflected non-vascular cellular hyperplasia.The use of Cr51-labelled normal red cells in theanatomic identification of the spleen is thus reli-able only in a positive sense, that is, when a highinitial relative radioactivity occurs over the sus-pected area. Conceivably sufficiently large vas-cular tumors, such as hemangiomata of the liver,might be similarly identified.

Observations (9) reported elsewhere upon theorgan distribution of Cr5' and the appearance ofhemoglobin in the plasma following injection ofCr51-labelled red cells after exposure to completeor to incomplete antibodies indicate the experi-mental association of red cell agglutination anddeposition in the lung and liver with hemoglobi-nemia. On the other hand, the association of redcell sensitization, without manifest agglutination,was with splenic sequestration and the absence ofhemoglobinemia. On the basis of the plasma he-moglobin levels these processes have classicallybeen distinguished as "intravascular" and "ex-travascular" hemolysis. The former state has of-ten been taken to indicate the presence of a cir-culating hemolysin despite the rarity with whichhemolysins are actually demonstrable in hemolytic

diseases, even in patients (such as Case 10) withfairly striking increases in plasma hemoglobinlevels. These experimental observations and thoseclinical observations reported or referred to hereindicate the general tendency of mild or of discretered cell alterations (as encountered in spherocytes,sensitized cells, and the morphologically peculiarcells encountered in pernicious anemia and inCooley's anemia) to lead to rather specific seques-tration by the spleen presumably due to the re-fined filtering function of that organ. This is at-tended by little or no release of hemoglobin intothe circulation and may be interpreted as an"extravascular" process. More severe or grossred cell changes, such as agglutination, or discretecellular changes which may grossly affect bloodviscosity, as in sickle cell anemia, lead to a moreindiscriminate filtration by the capillaries, espe-cially in the several organs with very large vascularbeds: the lungs and liver, as well as the spleen ifpresent and functioning. Such widespread redcell trapping, especially when not in close relationto concentrations of reticulo-endothelial cells, mayfavor the release of hemoglobin into the plasmaand thus simulate the effects of a circulatinghemolysin.

It is interesting to note the coexistence of redcell agglutination, a large hepatic uptake of Cr5',and high plasma hemoglobin levels at the time ofthe first observation on Case 10, a patient at thattime with fulminating acquired hemolytic anemia.In Case 11, with an acquired hemolytic anemia ofslightly less severity, the red cells were sensitizedbut not agglutinated, the preponderance of Cr51 ac-tivity in the spleen was striking, and there waslittle rise in the plasma hemoglobin level. Of fur-ther interest was the observation that Case 5, thepatient with paroxysmal nocturnal hemoglobinuria,manifested an increase in hemoglobinuria follow-ing splenectomy despite improvement in the redcell hemoglobin concentration. Later still irondeficiency ensued as evidenced by a fall in meancorpuscular hemoglobin concentration and in theserum iron level. These observations upon Case5 suggest that following removal of the spleen agreater proportion of the patient's cells were de-stroyed intravascularly rather than in the spleenwith consequent increase in the urinary excretionof iron.

863

J. H. JANDL, M. S. GREENBERG, R. H. YONEMOTO, AND W. B. CASTLE

Technical hazards in the use of Cr1In the clinical application of the present method,

the Crl5-labelling techniques described by others(12) are generally suitable. The uptake of chro-mate by red cells is highly pH-dependent, diminish-ing sharply as the pH exceeds 7.0. Thus, acidanticoagulant solutions, such as ACD solution, aresuitable for red cell labelling with sodium chro-mate.8 In addition, trivalent chromium arisingfrom reduction of the chromate by dextrose or byadded ascorbic acid is bound by citrate, which inturn inhibits binding by the plasma proteins andby the red cell membranes (9). During the initialmoments of the labelling process, chromate en-ters the red cells so rapidly that a disproportionate"loading" of a small fraction of the cells occursunless precautions are taken to employ a dilutesolution of sodium chromate and to secure its im-mediate and thorough mixture with the bloodsample. Chromate is an oxidant relative to hemo-globin, and methemoglobin appears in detectablequantities when the molecular ratio of chromateto hemoglobin exceeds 0.1, i.e., above about 22micrograms of Crp5 or about 70 micrograms ofNa2Cr5104 per cubic centimeter of packed redcells.9 The glycolysis of red cells, as measured inthe Warburg apparatus, was found to be de-pressed in arithmetic proportion to the chromateconcentration of the suspension. A significant de-pression of red cell glycolysis was demonstrablefollowing the addition of as little as 10 micro-grams of Cr5' in the form of sodium chromateper cubic centimeter of packed red cells. Chro-mated hemoglobin was observed to migrate morerapidly than does normal hemoglobin toward the

8An exception to this was made in studying the redcells of the patient with, paroxysmal nocturnal hemoglobi-nuria. In order to avoid the acid hemolysis characteristicof this patient's red cells, heparin was employed as ananticoagulant during the labelling procedure.

9 This reaction appears to be a direct chemical oxidationof hemoglobin progressing in proportion to dosage and totime equally well with hemoglobin solutions or with in-tact red cells. That inactivation of the enzyme normallyconcerned with the reduction of methemoglobin is notinvolved was suggested by study of the effect of chro-mate upon the red cells of a patient with congenitalmethemoglobinemia, a disease in which this enzyme isbelieved to be absent or deficient: the rate and extent ofconversion of oxyhemoglobin in the patient's red cellsto methemoglobin by chromate was the same as innormal red cells.

anode of a filter paper electrophoresis cell main-tained at a pH of 8.6 by barbital buffer. Underthese conditions, the greater negative charge ofthe Cr51-labelled hemoglobin induced an increasein the electrophoretic mobility of hemoglobin ofover 15 per cent. All of these observations stressthe chemical alterations brought about by the Cr51-labelling procedure, and urge that care be taken tolabel the red cell aliquots evenly and withNa2Cr51O4 of high specific radioactivity.

Ordinarily there is no apparent necessity forwashing the patient's red cells prior to re-injectionor for measuring the urinary excretion of Cr.51.The chief advantage of washing the red cells priorto injection is to obviate an initial urinary hyper-excretion of Cr5l which is not derived from la-belled red cells but from the plasma. The in-jection of a small amount of Cr51 unattached tored cells will also impair the accuracy of the ini-tial blood volume determinations if whole bloodspecimen radioactivity is measured, but it shouldnot interfere with the assessment of splenic se-questering activity.The extensive accumulation of Cr5' observed

to occur in the spleens and livers of patients withhemolytic anemias necessitates a re-appraisal ofthe radiation hazard involved. The radiation dos-age contributed by the injection of 100 ,uC of Cr5'into a 70 Kg. subject has been estimated to beabout 0.1 rep, with the assumption that the iso-tope is evenly distributed within the circulationand that no excretion has occurred (24). In apatient with a hemolytic process in which the in-jected Cr5' is accumulated entirely by the liverand spleen, with combined organ weights of 1600grams, the exposure to these tissues approximates2.0 rep. If a 100 !C dose of Cr5' is accumulatedentirely by a 200-gram spleen, the exposure tothat organ will approximate 16.1 rep, an exces-sive level. Although total clearance of Cr5' by thespleen is unlikely, it is suggested that the dose ofCr51 employed in subjects with presumed hemo-lytic anemias be reduced to a maximum of 50 ,uC.

Cr5l in predicting the need for splenectomy

After each injection of labelled red cells deter-mination of the body surface Cr5' activity by meansof three or four appropriately spaced sets of ob-servations over precordium, liver, spleen and

864

DETERMINATION OF THE SITES OF RED CELL SEQUESTRATION

thorax should suffice to establish the sites of redcell sequestration. For correlation with the ratesof disappearance of Cr51-labelled red cells from theperipheral blood it is desirable to perform bodysurface counts for at least half the red cell sur-

vival period. A suggested procedure is to makea set of observations 20 or 30 minutes after the in-jection of labelled red cells and again one day later.Two or three more observations of body surfaceradioactivity (for example, after an estimated 25per cent, 50 per cent, and 90 per cent, respectively,of the labelled cells have been destroyed) shouldprovide the desired information. Employment ofa uniform, geometric positioning of patients andof counting apparatus at the time of each observa-tion is, of course, essential to obtaining reproduci-ble data. Nevertheless, a portable counting unitat the bedside should be sufficiently accurate todetermine the approximate distribution of radio-activity between heart, liver, and spleen.The in vivo determination of the site of red cell

sequestration should prove of greatest practicalvalue in predicting the need for splenectomy in ac-

quired hemolytic anemias, wherein the greatestprognostic uncertainty presently exists. Inter-pretation of the data obtained by the methods de-scribed may be simplified by using a standardizedexpression such as the difference between the rela-tive radioactivity (expressed as a percentage ofthat over the precordium) for liver or spleen atthe red cell half-survival time (R50) and suchvalues at zero time (R0). A simple formula maythen be invoked as an index of red cell sequestra-tion as follows:

Index of Sequestration = R50-Ro

In 3 normal subjects, these values for the liverranged from 23 to 35 and for the spleen from 38to 48 with average values of 27 and 43, respec-

tively, as recorded in Table I. A tentative gradingof the intensity of splenic sequestration on the ba-sis of the index of sequestration values recordedin Table I for the 11 anemic subjects might be as

follows: less than 30-subnormal; 30 to 60-nor-mal; 60 to 100-mild to moderate; and greaterthan 100-moderate to severe. The employmentof even such crude numerical assessments as guidesin deciding the need for splenectomy will as-

suredly require many case studies. However,

Korst, Clatanoff, and Schilling (25) have studied,with methods resembling ours, 3 patients with ac-quired hemolytic anemia, on whom splenectomieswere later performed. In 2 of the patients highsplenic accumulations of Cr51 occurred and splen-ectomy was beneficial; in 1 patient, in whom therewas little build-up of radioactivity over the spleen,removal of the spleen was ineffectual. There are,in addition to this category of patients, other con-ditions sometimes characterized by splenomegaly(23, 26-28) in which, although splenectomy isusually not indicated, in individual instances re-moval of the spleen has benefitted the patient bydecreasing the rate of red cell destruction (26-28).An unexpected example eventuated from the ob-servation that radioactivity accumulated in the en-larged spleen of a patient, Case 5, with paroxysmalnocturnal hemoglobinuria following injection ofCr51-labelled red cells. Removal of the patient's660-gram spleen produced temporarily a moderatediminution of the anemia, and thereafter the patienthas required fewer transfusions to maintain agiven hemoglobin concentration. It is likely thatdetectable instances of excessive sequestration bypathologic spleens will arise in a variety of diseasestates. The surprising predilection of Cr51-labelledpernicious anemia red cells for sequestration by thespleen, observed in Case 8, may explain the thera-peutic value of splenectomy reported in this dis-ease prior to the use of liver extract (29). Thefact that the defectively formed red cells charac-teristic of certain diseases of red cell underproduc-tion may survive poorly, and, as in the case ofpernicious anemia, despite a calculated rate ofhemoglobin production that was subnormal, maybe detectably sequestered by the spleen, enjoins theneed to characterize the fundamental balance be-tween red cell production and destruction in eachcase as clearly as possible. The anemia of chronicliver disease, for example, has been shown to beprimarily hemolytic in most instances (14). Thefinding that all of 4 patients with the anemia ofliver disease demonstrated mild to moderate splenicuptakes of Cr51, as illustrated by Case 7, indicatesthe potential therapeutic value of splenectomy insuch patients although only rarely is such an op-eration warranted. The occasional patients withcongenital non-spherocytic hemolytic anemia whomay benefit from splenectomy should also be iden-tifiable by the Cr51 distribution technique.

865

J. H. JANDL, M. S. GREENBERG, R. H. YONEMOTO, AND W. B. CASTLE

SUMMARY

The sites of deposition of Cr5l in the humanbody were determined by body surface countingfollowing the injection into normal subjects ofCr51-labelled red cells, hemoglobin, and chromicchloride in saline. Differences characteristic ofeach were observed and served as a basis for inter-pretation of a study of a series of 11 patients withhemolytic anemias of diverse kinds. Measure-ments made on these patients following injection ofCr51-labelled autologous red cells suggested thatprogressive accumulation of Cr51 in the spleen ac-

companying the disappearance of Cr51-labelled redcells from the blood stream indicates active redcell sequestration. In determining the extent ofthis sequestration a simple expression may be em-

ployed (index of sequestration) which deductsthat radioactivity initially present and due simplyto the size of the splenic vascular bed from latervalues obtained over that organ. The methodsdescribed should in practice constitute a simpleclinical device for determining the need for splen-ectomy. In certain hemolytic anemias the shortsurvival of labelled red cells in the circulation isaccompanied by a progressive increase of radio-activity over the spleen and the anemia is relievedby splenectomy.

ACKNOWLEDGMENTS

We are indebted to Miss Martha Hoge, Miss PatriciaLambert, Miss Nancy Sheldon, and Miss Ann Tomlinsonfor technical assistance.

REFERENCES

1. Ham, T. H., and Castle, W. B., Relation of increasedhypotonic fragility and of erythrostasis to themechanism of hemolysis in certain anemias. Tr.A. Am. Physicians, 1940, 55, 127.

2. Emerson, C. P., Jr., Shen, S. C., and Castle, W. B.,The osmotic fragility of the red cells of the periph-eral and splenic blood in patients with congenitalhemolytic jaundice transfused with normal redcells. J. Clin. Invest., 1946, 25, 922.

3. Emerson, C. P., Jr., Shen, S. C., Ham, T. H., andCastle, W. B., The mechanism of blood destruc-tion in congenital hemolytic jaundice. J. Gin. In-vest., 1947, 26, 1180.

4. Young, L. E., Platzer, R. F., Ervin, D. M., and Izzo,M. J., Hereditary spherocytosis. II. Observationson the role of the spleen. Blood, 1951, 6, 1099.

5. Weisman, R., Jr., Hurley, T. H., Harris, J. W., andHam, T. H., Studies of the function of the spleenin the hemolysis of red cells in hereditary sphero-cytosis and sickle cell disorders. J. Lab. & Clin.Med., 1953, 42, 965.

6. Diggs, L. W., and Ching, R. E., Pathology of sicklecell anemia. South. M. J., 1934, 27, 839.

7. Bauer, J., Sickle cell disease. Pathogenic, clinicaland therapeutic considerations. Arch. Surg., 1940,41, 1344.

8. Gray, S. J., and Sterling, K., The tagging of red cellsand plasma proteins with radioactive chromium.J. Clin. Invest., 1950, 29, 1604.

9. Jandi, J. H., Sequestration by the spleen of red cellssensitized with incomplete antibody and withmetallo-protein complexes. J. Clin. Invest., 1955,34, 912.

10. Jandl, J. H., Greenberg, M. S., Yonemoto, R. H., andCastle, W. B., Clinical determination of the sitesof red cell destruction. Clin. Res. Proc., 1955, 3,95.

11. Ham, T. H., Ed., A syllabus of laboratory examina-tions in clinical diagnosis. Critical evaluation oflaboratory procedures in the study of the patient.Cambridge, Harvard University Press, 1950.

12. Ebaugh, F. G., Jr., Emerson, C. P., and Ross, J. F.,The use of radioactive chromiumi as an erythro-cyte tagging agent for the determination of redcell survival in zivo. J. Clin. Invest., 1953, 32,1260.

13. Jandl, J. H., To be published.14. Jandl, J. H., The anemia of liver disease: Observa-

tions on its mechanism. J. Clin. Invest., 1955, 34,390.

15. Loutit, J. F., Discussion on the life and death of thered blood corpuscle. Proc. Roy. Soc. Med., 1946,39, 757.

16. Singer, K., King, J. C., and Robin, S., The life spanof the megalocyte and the hemolytic syndrome ofpernicious anemia. J. Lab. & Clin. Med., 1948, 33,1068.

17. Emerson, C. P., Jr., and Burrows, B. A., The mecha-nism of anemia and its influence on renal functionin chronic uremia. J. Clin. Invest., 1949, 28, 779.

18. Loge, J. P., Lange, R. D., and Moore, C. V., Charac-terization of the anemia of chronic renal insuffi-ciency. J. Clin. Invest., 1950, 29, 830.

19. Chaplin, H., Jr., and Mollison, P. L., Red cell life-span in nephritis and in hepatic cirrhosis. Clin.Sc,, 1953, 12, 351.

20. Finch, C. A., Hegsted, M., Kinney, T. D., Thomas,E. D., Rath, C. E., Haskins, D., Finch, S., andFluharty, R. G., Iron metabolism. The patho-physiology of iron storage. Blood, 1950, 5, 983.

21. Visek, W. J., Whitney, I. B., Kuhn, U. S. G., III,and Comar, C. L., Metabolism of Crt by animalsas influenced by chemical state. Proc. Soc. Exper.Biol. & Med., 1953, 84, 610.

866

DETERMINATION OF THE SITES OF RED CELL SEQUESTRATION

22. Diggs, L. W., Siderofibrosis of the spleen in sicklecell anemia. J.A.M.A., 1935, 104, 538.

23. Lichtman, H., Shapiro, H., Ginsberg, V., and Watson,J., Splenic hyperfunction in sickle cell anemia

Am. J. Med., 1953, 14, 516.24. Frank, H., and Gray, S. J., The determination of

plasma volume in man with radioactive chromicchloride. J. Clin. Invest., 1953, 32, 991.

25. Korst, D. R., Clatanoff, D. V., and Schilling, R. F.,External scintillation counting over the liver andspleen after the transfusion of radioactive erythro-cytes. Clin. Res. Proc., 1955, 3, 195.

26. Hahn, E. V., and Gillespie, E. B., Sickle cell anemia.

Report of a case greatly improved by splenectomy.

Experimental study of sicide cell formation. Arch.Int. Med., 1927, 39, 233.

27. Shotton, D., Crockett, C. L., Jr., and Leavell, B. S.,Splenectomy in sickle cell anaemia: Report of acase and review of the literature. Blood, 1951, 6,365.

28. Lichtman, H. C., Watson, R. J., Feldman, F., Gins-berg, V., and Robinson, J., Studies on thalassemiaPart II. The effects of splenectomy in thalassemiamajor with an associated acquired hemolytic ane-

mia. J. Clin. Invest., 1953, 32, 1233.29. Pearce, R. M., Krumbhaar, E. B., and Frazier, C. H.,

The Spleen and Anemia. Experimental and Clini-cal Studies. Phila. and London, J. B. Lippincott,1918.

867