From www.bloodjournal.org by guest on August 14, 2018. For personal use only.

The Pathophysiology of Pure Red Cell Aplasia: Implications for Therapy

By Robert J. Charles, Kathleen M. Sabo, Pamela G. Kidd, and Janis L. Abkowitz

To determine the utility of marrow culture in defining the
natural history and therapeutic response of pure red cell
aplasia we have studied 37 patients. Patients were evaluated
a t the University of Washington before specific therapies (n
= 21) or at the time of treatment failure (n = 16). Evaluation
included a medical and drug exposure history, a physical
examination, a chest x-ray or computed tomography t o rule
out thymoma, lymphocyte immunophenotype studies, anti-
nuclear antibody and rheumatoid factor determinations,
marrow cytogenetics, and marrow progenitor cell cultures.
Retrospective Southern analyses t o detect human parvovi-
rus B19 was performed in the 27 patients for whom sera
was stored. Clinical follow-up was obtained t o document
therapeutic responses. Normal burst forming unit-erythroid
(BFU-E) growth (>30 bursts/105 marrow mononuclear cells
[MMNCI) in culture proved an outstanding predictor of clini-
cal response, as 27 of 29 individuals with normal frequencies
of erythroid bursts in culture responded t o immunomodulat-
ing therapies (sensitivity 96%. specificity 7% predictive
value 93%. P = .OW1 with two-tailed chi square analysis).
Overall, 28 patients responded t o either immunomodulating
therapies or drug withdrawal. Twenty-four patients ob-
tained a normal hematocrit (complete response [CRI) and 4
additional patients became transfusion independent (partial
response).Although responding patients often required sev-
eral therapies, 20 of 24 (83%) patients who obtained a CR
have sustained a normal hematocrit without maintenance
therapy at the time of last follow-up (median 5 years). In
contrast, of 8 patients with poor in vitro BFU-E growth ( e 6
bursts/105 MMNC), 7 failed t o respond t o any therapy and
all died (median survival time 17 months). Our data suggest
that in individuals, from whom BFU-E mature appropriately
in culture, immunosuppressive drugs should be used se-
quentially until a CR is obtained and a durable remission is
the expected outcome.
0 1996 b y The American Society of Hematology.

PURE RED CELL aplasia (PRCA) is a clinical syndrome

defined by the absence of mature erythroid precursors
in an otherwise normocellular bone marrow (BM). Patients
(n = 3) with trilineage dysmorphic abnormalities allowing the defin-
itive diagnosis of myelodysplasia were specifically excluded. Studies
were approved by the Human Subjects Committee, University of
have severe anemia, a low reticulocyte count and normal Washington, Seattle.
Initial clinical evaluation. The initial clinical evaluation in-
platelet and granulocyte counts.'
cluded a medical and drug exposure history, a physical examination,
The physiology of PRCA is likely heterogenous. It has and a computer-assisted tomographic scan or x-ray to rule out thy-
been associated with autoimmune, viral, and neoplastic dis- moma. Blood was obtained to immunophenotype lymphocytes to
eases including t h y m ~ m a ,rheumatoid
~.~ arthriti~,~'~
systemic detect excess natural killer (NK) cells or diagnose CLL; antinuclear
lupus erythematosus,"-" hepatitis,I2 mononucle~sis,'~~'~antibody (ANA), rheumatoid factor and other rheumatologic studies;
l y m p h ~ m a , ' ~ -B-cell
'~ chronic lymphocytic leukemia and a complete blood count and morphology. A retrospective analy-
(CLL),I9 T-cell CLL," and large granular lymphocytic sis for HPV B 19 was performed on the 27 patients for whom serum
(LGL) l e ~ k e m i a . ~In' . these
~ ~ settings, semm antibodies with was stored and has been reported separately.I6The presence or ab-
selective cytotoxicity for marrow erythroid cells or erythro- sence of proerythroblasts on the marrow aspirate was specifically
poietin have been r e p ~ r t e d . ~In~ . other
' ~ studies T-cells from noted. Marrow cytogenetics were obtained from all patients.
Marrow culture studies. To quantitate the frequencies of ery-
patients with PRCA associated with t h y m ~ m a , chronic ~~.~~
throid progenitors, 105 marrow mononuclear cells (MMNC) were
Epstein-Barr virus (EBV),14.28 CLL,20,30-33and suspended in 1 mL a-medium containing final concentrations of
LGL l e ~ k e m i ahave ~ ~ ,been
~ ~ shown to suppress erythropoie- 1.2% methylcellulose, 1% bovine serum albumin, mol& beta
sis in vitro. PRCA may also occur in association with myelo- mercaptoethanol, penicillidstreptomycin, 30% fetal calf serum
dysplasia or as a consequence to chronic human parvovirus (FCS). 1 U/mL recombinant human erythropoietin and 2.5% human
B 19 (HPV B19) i n f e ~ t i o n . ~ ~ * ~ ~ lymphocyte conditioned medium!" Cultures were incubated at 37°C
The optimal treatment of PRCA remains uncertain. Be- in 5% CO2 and erythroid colonies (from colony forming unit-ery-
cause clinical and morphologic presentations of PRCA are throid [CFU-E] and erythroid bursts (from burst forming unit-ery-
overlapping, management decisions are difficult. In this re- throid [BFU-E] were enumerated on days 7 and 14, respectively,
port, we have prospectively studied 37 PRCA patients to with inverted microscopy. Granulocyte macrophage (GM) colonies
(from CFU-GM) were quantified on day 14 on the same plates
evaluate the roles of clinical presentations and marrow cul-
ture studies as prognostic indicators of long-term response.
To gain further insight into the pathophysiology and natural From the Departments of Medicine and Laboratory Medicine Uni-
history of PRCA, we have separately analyzed those patients versity of Washington, Seattle.
(n = 21) evaluated before any pharmacologic intervention. Submitted August 22, 1995; accepted January 16, 1996.
Supported by Grants No. ROI HL 31823 (J.L.A.)and MO1 RR
MATERIALS AND METHODS
00037 from the National Institutes of Health. J.L.A. is the recipient
Criteria for admission. This series includes patients with PRCA of a Faculty Research Award from the American Cancer Society.
evaluated at the University of Washington between 1982-1992. All Address reprint requests to Janis L. Abkowitz. MO, Department
patients fulfilled standard criteria for PRCA and had anemia with of Medicine, Division of Hematology, University of Washington, Box
reticulocytopenia (reticulocyte count < l%), normal white cell 357710, Seattle, WA 98195.
counts, normal differential counts, apd normal platelet counts.'.24BM The publication costs of this article were defrayed in part by page
biopsies revealed no decrease in overall cellularity. Hemoglobinized charge payment. This article must therefore be hereby marked
cells comprised less than 3% of nucleated marrow cells on marrow "advertisement" in accordance with 18 U.S.C. section 1734 solely to
aspirate. All patients but two (patients 9 and 19) met the stricter indicate this fact.
criteria described by Clark et (hemoglobinized cells comprising 0 1996 by The American Society of Hematology.
less than 0.5% of nucleated marrow cells). For this analysis, patients 0006-4971/96/8711-0022$3.00/0

Blood, Vol 87, No 11 (June 1). 1996: pp 4831-4838 483 1

 From www.bloodjournal.org by guest on August 14, 2018. For personal use only.
4832
to control for any technical issues. All studies were performed in
triplicate.
To determine if aberrant T cells or a serum antibody suppressed
erythroid differentiation, coculture studies were performed. Periph-
eral blood (PB) E-rosette + (ER+) cells were added to l@ ER-
MMNC in ratios of 1:1, 2: 1, and 3: 1 T-cell mediated inhibition
of erythropoiesis was defined as a 240% decrease in the numbers
of detectable erythroid bursts or erythroid colonies (compared to
baseline ER- cultures) and with no change in the numbers of CFU-
GM colonies. To investigate antibody mediated inhibition, marrow
cells were incubated for 30 minutes at room temperature with autolo-
gous serum or normal AB serum, then for 30 minutes at room
temperature with an equal volume of rabbit complement before cul-
ture.40As an additional assay, 15% autologous or normal AB serum
was added directly to marrow cultures (replacing 15% of the FCS).40
Antibody mediated inhibition was considered present if the numbers
of erythroid bursts or erythroid colonies decreased by 35%, and if
no change in GM colony numbers was detected, when MMNC were
incubated with autologous serum then complement, or were cultured
in the presence of autologous serum.
Follow-up studies. Follow-up has been acquired from contact
with patients or their physicians, and ranged from 1 month to 12
years. Patients were classified to have a complete response (CR) if
their hemoglobin or hematocrit became normal and a partial response
(PR) if their need for red cell transfusions ended. These results were
correlated with results from marrow culture and coculture studies to
determine significant prognostic indicators of therapeutic success.
Among patients with a CR or PR, median follow-up was 5 years.
Relapses occurred in several patients. Relapses were defined as a
transfusion need which reoccurred after a patient obtained a thera-
peutic remission. Marrow aspirates and biopsies were not generally
performed at this time, and therefore it cannot be stated that these
patients fulfilled all criteria for PRCA at relapse.
Statistical analysis. Results were analyzed by two-tailed chi
square analysis employing the Statview 512 plus program (Abacus
Concepts, Inc, Berkeley, CA).
A ESULTS
Patient characteristics. Clinical data are summarized in
Table 1. Patients ranged in age from 4 to 83 years with a
median of 57 years. Twenty-two patients were men and fif-
teen women. Twenty-one of the patients were evaluated be-
fore specific therapies for PRCA. The other 16 patients were
referred following treatment failure. Fifty-one percent (19
of 37) of patients had other conditions known to be associ-
ated with PRCA. Immunologic disorders (eg, rheumatoid
arthritis, aortitis, Sweet's syndrome) were present in 14% (5
of 37) of patients. In addition, a history of previous autoim-
mune disease (eg, idiopathic thrombocytopenic purpura,
Hashimoto' s thyroiditis) or of nonspecific rheumatologic
complaints was reported in 11% (4 of 37) of patients. Human
immunodeficiencyvirus (HIV) or EBV infection was present
in 5% (2 of 37), CLL or lymphoma in 8% (3 of 37) and
previous or recurrent thymoma in 8% (3 of 37) of patients.
Two patients or 5% had evidence of LGL leukemia and
one patient developed LGL leukemia during the follow-up
period. Three patients (8%) were exposed to a new drug 1
to 6 months before the onset of PRCA. Data from patients 1,
2,3,7,9, 17, 19,20,21, and 26 were reported previou~ly?~"'
Several additional studies were performed to further char-
acterize the PRCA. Inverted ratios of CD4+ to CD8+ cells
were present in nine patients and did not correlate with dis-
CHARLES ET AL
ease process (P > .OS). HPV B19 infection was found retro-
spectively in 4 (1 1 %) patients using Southern blot analysis
of sera. One of these patients had concurrent HIV infection,
1 had Sweet's syndrome, 1 admitted to marijuana abuse that
may have contributed to immune incompetence, and 1 had
no immunologic abnormality.
As previously stated, LGL leukemia was diagnosed in
three patients. These patients fulfilled the diagnostic criteria
of Loughran, and had both an absolute increase in large
granular lymphocytes (>700/pL) and either an absolute in-
crease in CD8+ lymphocytes (>l,OOO/pL) or NK (CD 56+
or CD 57+) cells ( 1,000/pL).21In addition, 1 patient (patient
25) developed LGL leukemia during the follow-up period.
T-cell receptor gene rearrangement studies were performed
in 2 of these patients and confirmed a clonal abnormality.
Seven additional patients had elevated percentages of NK
cells (ranging from 21% to 5 1%) and greater than 200 NK
cells but failed to meet the diagnostic criteria for LGL leuke-
mia. None of these 7 patients were evaluated with T-cell
receptor gene rearrangement studies.
Although no patient had trilineage morphologic abnormal-
ities to allow the clinical diagnosis of myelodysplasia, micro-
megakaryocytes were present in the marrow aspirate and a
single unilobed neutrophil was seen in the PB of patient 3 1.
Marrow from patient 7 could not be aspirated and biopsy
revealed myelofibrosis, but no additional abnormality. Cyto-
genetic abnormalities were present in three patients. These
included 5q- deletions in patients 17 and 20 and trisomy 8
in patient 31.
Responses to therapy. Individual responses to therapy
are summarized in Table 1. Twenty-eight patients (76%)
obtained hematocrits adequate to end transfusion require-
ments (24 CR, 4 PR). Twenty-five of these patients had
remissions secondary to immunomodulating therapies. Three
patients had a CR following drug withdrawal.
Although not specified in the study design, 33 of 37 pa-
tients received prednisone as their initial therapy. Nine of 33
(27%) patients obtained a complete remission. No additional
therapy was required in 6 patients.
Patients often received more than one therapy. Conse-
quently, 1 patient may contribute to the response rate of
more than 1 agent. Overall, response rates (CR and PR)
to cyclophosphamide, antithymocyte globulin (ATG), and
cyclosporine were 6 of 15 (40%), 8 of 13 (62%), and 2 of
3 (67%), respectively. Two patients responded to IgG ther-
apy (one was viremic with HPV B19 and one was not).
Sustained low-dose methotrexate (10 to 15 kg/week) resulted
in remissions in 2 patients. One patient had a CR following
thymectomy and another had a CR following thymectomy
and prednisone therapy. Other forms of therapy including
vincristine, splenectomy, and androgens were attempted
without benefit in a few patients failing prednisone, ATG,
or cytotoxic agents.
To consider possible selection bias, response rates to vari-
ous agents were independently analyzed in the group evalu-
ated before specific therapies for PRCA. Of these 21 patients,
17 received prednisone alone as their initial treatment and 8
patients (47%) obtained a complete remission. In the patients
failing prednisone, 78% responded to either ATG, cyclo-
 From www.bloodjournal.org by guest on August 14, 2018. For personal use only.

PURE RED CELL APLASIA 4833

Table 1. Characteristics of PRCA Patients

Patient No./ Initial Excess T-cell/
AoeEex Studv Associated Conditions and Remarks BFU-E CFU-E Pro-E Antb Theraoeutic ReSDOnSe Clinical Follow-UD

- --
1/57/F Hx of Hashimoto's thyroiditis and Y N N P NR, ANDR NR, ATG PR
-+ -+ -+ CR -+ 12 yr
sarcoidosis, LGL leukemia Relapse after 3 wk, CY CR
U 12/F
3/4F
P HPV B19 Y
Y
Y
N
Y
N -
P - NR, ATG PR CR
P NR, spontaneous
remission, relapse after 1 mo,
CR
CR
-+

-+
12 yr
1 yr

ATG CR -+

4/34/F
5/63/M
6/77/M
P
P
Y
N
Y
Y
N
Y
N
N
N
P NR, ATG CR
-+ -+

P NR, Azathioprine NR
-+

P NR, IgG NR
-+ -+
- CR
NR
NR
-
-+

-+
5 yr
2 yr, died of ANLL
9 mo, died of MI
Azathioprine NR -+

7/72/M Myelofibrosis N N N ANDR + NR, P NR, ATG -+ NR 5 yr. died of ANLL

-
-+ +

NR
8/46/M P Hx of ITP and recurrent Y N N P NR, CY CR
-+ CR+5yr
pericarditis

9/50/M
ANA 1:160
Chronic EBV infection Y Y N P NR, Acyclovir + NR, CY - - PR 10 yr, died of

-
-+

10/38/F Rheumatoid arthritis Y N N

NR
- -
Plasmaphresis NR, ATG PR
P CR, relapse after taper,
cardiovascular DX

CR 5 yr on

- -
-+

ATG + P CR, relapse after

-+ methotrexate
taper, AZA + P CR, MTX

-- -
CR
11/58/M P Thymoma Y Y N P NR, CY NR, VIN NR,
-+ CR-+9yr
AZA NR, thymectomy -+

12/60/M P Y Y N
NR, P CR
-+ -
-+

P NR, CY PR (d/c because

of neutropenia)
PR 3 mo; transfusion
-+

supported, died 1 yr
later of UGlB & COPD
13/42/M P Marijuana x 6 mo Y Y N D/c marijuana -+ CR CR-7yr
HPV B19
14/59/M Sweet's syndrome, HPV B19 N N N P 81 Danazol -+ NR NR 9 mo, died of
-+

aortic dissection
15/83/M P Y N N P-+CR CR-5mo
16/62/F
17/49/F
Rheumatoid arthritis
5q-cytogenetics
Y
N
Y
N
Y
N
T,A
-
P+NR,CY-+CR
P NR, CY NR, ATG NR
-+ - - CR+8yr
NR 5 yr, died of ANLL
18/57/F
19/63/M
P
P
Thymoma
B-cell CLL
Y
Y
ND
Y
N
N
Thymectomy + CR
-
P & Chlorambucil + NR, CY
NR, splenectomy NR, ATG
-+
CR 1 mo, died of MI
-+

PR 1 yr. died of CLL

- - -- -
-+

PR
-+

20/75/M P 5q-cytogenetics N N N P NR, CY NR, ATG NR NR 7 yr. died of ANLL

- -
-+ +

21/25/F ANA 1:40 Y Y Y P NR, ANDR NR, CY NR

-+ CR 5 yrs
plasmaphresis NR, ATG
CR
22/69/M P Y Y Y P-+CR CR 10 yr
-+

23/37/M P Aortitis Y Y N P+CR CR+1 mo

24/35/M
ZM~/M
P
P
Dilantin x 1 mo Y
Y
Y
ND
Y
N
D/c Dilantin CA
- -
-+

P NR, CY NR, IgG PR

-+

relapse 1 yr later IgG NR,

-+

-+
CR-8yr
CR 2 yr on
-+

methotrexate, LGL

26/26/M HIV
HPV B19
N N N -CyA PR, MTX + CR
-+

P NR, IgG CR, relapse, IgG

CR
-+
-+ -
leukemia diagnosed
CR 6 mo, died of
mycobacterial

27/68/F LGL leukemia Y N N P-NR,CY-+NR

ATG + NR, IgG NR, CyA- -
infection & AIDS
CR 5 yr

-
-+

--
CR
28/56lM P Thymoma resection 3 mo before Y N N P -+ CR, CY CR, relapse with CR 5 yr on

29/63/M
onset of PRCA
Rheumatoid arthritis Y N N

(Continued on followina Page)

taper, P CR
P PR, IgG NR, CY PR
-+ -+ -+ -
prednisone
PR 1 yr. died of COPD
 From www.bloodjournal.org by guest on August 14, 2018. For personal use only.

4834 CHARLES ET AL

Table 1 (Cont'd). Characteristics of PRCA Patients

Patient No./ Initial Excess T-cell/
AgelSex

30/60/M
Study Associated Conditions and Remarks

Stage IV non-Hodgkin's
BFU-E

Y
CFU-E

N
Pro-E

N
Antb

-
Therapeutic Response Clinical Follow-Up

M BACOD NR, P + NR, IgG + NR + 2 yr. died of

-- -
lymphoma NR lymphoma
CyA NR, ATG NR, CAPBOP

31/78/F Trisomy 8 cytogenetics,

micromegakaryocytes a single
N N N
NR
P + NR, Chlorambucil NR - -
NR 4 mo, died of
ANLL

32/76/F
33/30/M
P
P
unilobed neutrophil in smear
Y
Y
N
Y
N
N
--
P CR
P CR
CR
CR
-
+
1 yr. died of CHF
2 yr
34/57/F P Isoniazid x 1 mo Y Y N D/c isoniazid + CR CR + 2 yr

-- -
Rheumatoid factor 1:2,560
35/66/M B-cell CLL N N N CVP NR, P NR, CY NR
+ + NR 2 mo, died of MI
36/34/F
37/46/F
P
P
Arthralgias
Arthralgias, Hx of Hashimoto's
thyroiditis
Y
Y
Y
Y
N
Y
P CR
P CR, relapse with taper, P
+

CR
- CR + 6 yr
CR 3 yr o n
+

prednisone
Twenty-one patients were studied prior (P)to any drug therapies. In 16 patients, the evaluation was within 3 weeks of presentation with
anemia. In 4 patients. evaluation was within 14 weeks. Patient 25 was transfused for 2 years before diagnosis and study. Fifteen patients were
referred after previous treatment failure. Inverted populations of CD4- to CD8' cells were shown in 9 patients (8, 11, 12, 14, 17, 25, 27, 28, and
29). Excess (Exc) proerythroblasts (greater than 7% of nucleated marrow cells) were present in 6 patients (2, 16,21, 22,24, and 37). LGL leukemia
was diagnosed in two patients (1 and 27). These patients had both an absolute increase in large granular lymphocytes (>700/pL) and either an
absolute increase in CD8' lymphocytes (>l,OOO/pL) or NK (CD 56' or CD 57') cells (l,OOO/pL). Seven additional patients had elevated percentages
of NK cells (ranging from 21% to 51%) and greater than 200 NK cells but failed to meet the diagnostic criteria for LGL leukemia (patients 10,
11, 12, 14, 16, 17, and 22). The presence of cytogenetic abnormalities was statistically associated with poor therapeutic response ( P = ,001 with
chi square analysis). Although the sensitivity was 100%. the specificity was only 33%. The presence or absence of other associated conditions,
when tested individually or when grouped together, did not correlate with clinical response (all P values > ,051.
Marrow culture and coculture studies: Normal in vitro growth was defined as greater than 30 BFU-E and CFU-GM and greater than 40 CFU-
E frequencies per l o 5 MMNC. Eight patients had poor BFU-E and CFU-E growth (5, 7, 14, 17, 20, 26, 31, and 35). Ten patients had poor CFU-E
growth only (1, 3, 8, 10, 15, 27, 28, 29, 30, and 32). CFU-E frequencies were not determined in 2 patients (18 and 25). In all other patients, BFU-
E and CFU-E frequencies were normal. All patients had adequate CFU-GM growth, except patient 14. Blood cells were used for studies in patient
7 whose marrow as inaspirable. T-cell mediated inhibition (T) was shown in 10 patients (1, 2, 3,4, 9, 10, 11, 16, 21, and 22). Antibody mediated
inhibition (A) was detected in 3 patients (15, 16, and 19).
Abbreviations: F, female; M, male; Hx, history; LGL leukemia, large granular lymphocytic leukemia; ITP, idiopathic thrombocytopenic purpura;
HPV 619, human parvovirus 819; NR, no response; d/c, discontinued; P, prednisone; ANDR, androgens; ATG, anti-thymocyte globulin; CY,
cyclophosphamide; M A , azathioprine; IgG, intravenous immunoglobulin; MTX, methotrexate; CyA, cyclosporine; ANLL, acute nonlymphocytic
leukemia; MI, myocardial infarction; UGIB, upper gastro-intestinal bleed; COPD, chronic obstructive pulmonary disease; CHF, congestive heart
failure; Dx, disease. Patients 1 to 3, 7 to 10, and 12 to 21 are JA#1 to 3, JA#7 to 10, and JA#12 to 21, respectively. Patients 22 to 24 correspond
to JA#26 to 28, patients 25 and 26 to JA#31 and 32, patients 27 to 30 to JA#34 to 37, patients 31 to 34 to JA#41 to 44, and patient 35 to JA#46.

phosphamide, or methotrexate. Response rates (CR + PR)
to ATG, cyclophosphamide or methotrexate were (3 of 4),
(3 of 7) and (1 of l), respectively. Taken together, 81% of
patients evaluated before therapy for PRCA obtained a CR
compared to only 44%of those referred after prior treatment
failure (significant at P = .02).
Clinical follow-up. Clinical follow-up showed sustained
remissions (PR or CR) ranging from 1 month to 12 years,
with a median of 5 years (see Table 1). When last follow-
up was available, of the 24 patients who obtained a CR,
20 (83%) patients had a sustained CR without maintenance
therapy. Only 4 (17%) patients had a CR requiring continued
therapy. Although 7 (29%)patients relapsed after obtaining
an initial CR, subsequent therapies achieved a complete re-
sponse in all cases. Thus, once a CR has been obtained, no
patient has developed refractory disease.
Of the 4 patients with a PR, all died (a median of 1 year
after initial response) because of related (CLL) or unrelated
causes (chronic obstructive pulmonary disease; upper gastro-
intestinal bleed, cardiovascular disease).
All 9 patients without a therapeutic remission died. Me-
dian survival time was 17 months. Causes of death included
leukemia, lymphoma, and acquired immunodeficiency syn-
drome (AIDSj.
Marrow culture results and treatment implications. BFU-
E frequencies were normal or increased in 29 of 37 patients
(>30/105 MMNC). CFU-E were detected in 17 patients (>40/
105 MMNC) and excess proerythroblasts were present in 6
patients. CFU-E frequencies were not determined for 2 patients
(18 and 25). Thus, the level of differentiationat which erythro-
poiesis was blocked could be determined in 27 patients and
was between BFU-E and CFU-E in 10 patients, between CFU-
E and proerythroblasts in 11 patients and between proerythro-
blasts and hemoglobinized cells in 6 patients (Fig 1). The level
of block in erythropoietic differentiation failed to predict the
specific drug to which a patient might respond when evaluated
with chi square analysis. CFU-GM frequencies were normal
(>30/1@ MMNC) in all patients except patient 14 whose CFU-
GM frequency was decreased. This intemal control showed
that cell culture methods were adequate.
 From www.bloodjournal.org by guest on August 14, 2018. For personal use only.
PURE RED CELL APLASIA
-
I I
BFUE CFU-E ---+ W E -Hemoglobin
Containing
cell
Fig 1. Marrow culture shows the level of erythropoieticblock. The
level of differentiation at which erythropoiesis was blocked could be
determined in 27 patients. In all 27 patients, BFU-E growth was nor-
mal. CFU-E growth was normal in 17 of these patients. Excess pro-
erythroblasts were present in 6 patients. Thus, the level of block
was between BFU-E and CFU-E in 10 patients, between CFU-E and
proerythroblasts in 11 patients, and between proerythroblasts and
hemoglobinized cells in 6 patients.
The relationship between the presence of erythroid bursts
in culture and whether the patient obtained a therapeutic
remission is central to this study (Fig 2). Of 29 patients
who had normal numbers of erythroid bursts in culture, 27
obtained a remission. In contrast. of 8 patients from whom
BFU-E failed to mature in culture (erythroid bursts <6/1OS
MMNC), 7 failed to respond to any treatment. Therefore,
BFU-E maturation in vitro was a superb predictor of clinical
response. Its sensitivity was 96%, its specificity was 78%
and its predictive value was 93% (P = .0001 with 2-tailed
chi-square analysis).
Coculture studies showed T-cell mediated suppression of
erythropoiesis in IO patients and antibody mediated inhibi-
tion in 3 patients (one patient had evidence of both T-cell and
antibody mediated inhibition). Although positive coculture
studies were an excellent predictor of ATG response (CR or
PR) (P = .OOOl), it was not predictive of responses to other
agents (P > .6 cyclophosphamide or prednisone). All 12
patients had remissions ( I O CR, 2 PR) following therapy,
with transfusion free survivals extending 7 years at present.
Among the 7 patients with poor BFU-E growth in vitro
and no response to therapies, patient 14 had chronic HPV
B19 infection, and patient 35 had extensive B-cell CLL. Five
patients had or developed some evidence of myelodysplasia.
Cytogenetic abnormalities and a unilobed neutrophil were
seen in patient 3 I , cytogenetic abnormalities alone were seen
in patients 17 and 20, and myelofibrosis in patient 7. Patient
5 had marrow morphology indistinguishable from the other
PRCA patients and normal cytogenetics but later died of
acute nonlymphocytic leukemia (ANLL). Patients 7, 17, 20,
and 31 also died with ANLL.
Instructive cases. Several other cases were instructive
clinically. Patient 27, a 68-year-old female presented with a
hematocrit of 19% and reticulocyte index of 0. Immunophe-
notype studies showed 63% of lymphocytes were CD8' (ab-
solute CD8 count of 3,995 cells/pL) and 35% were NK cells
(absolute count 1,645 NK cells/pL). T-cell receptor gene
rearrangement studies confirmed the diagnosis of LGL leu-
4835
kemia. Her marrow aspirate showed no excess proerythro-
blasts. Neither T-cell nor antibody mediated inhibition was
shown with coculture studies. Cell culture results revealed
normal BFU-E and CFU-GM growth. Only 6 CFU-E colo-
nies were detected.
The patient failed prednisone, cyclophosphamide, ATG
and IgG therapy. At the time when cyclosporine was initi-
ated, her transfused hematocrit was 26% and reticulocyte
index was 0. Two months after beginning cyclosporine, her
hematocrit normalized at 41.6% (hgb 14.3 g/dL). She has
remained in complete remission without relapse or mainte-
nance therapy 5 years at present. This case illustrates that
in patients with normal BFU-E growth, one should persist
with trials of drug therapies until a CR is obtained.
Patient 25 presented with a hematocrit of 17% and reticu-
locyte index of 0.1. His age of 79 years and high MCV of
107 raised the concern of myelodysplasia. Marrow culture
studies were obtained when he was referred to the University
of Washington Medical Center after 2 years of red cell trans-
fusions and showed normal BFU-E maturation suggesting
immunologically mediated PRCA. For this reason he was
started on prednisone and then cyclophosphamide, both
without changing his transfusion requirement. He next re-
ceived a 5 day course of IgG therapy with a subsequent
reticulocytosis and a rising hematocrit to 32. This partial
response continued for one year while he remained transfu-
sion and symptom free (Hct = 29 to 32). However, following
a viral illness, his PRCA relapsed. He did not respond to
repeated IgG therapy. Cyclosporine then resulted in a partial
remission (Hct = 26 to 29). At this time, 4 years after his
initial diagnosis of PRCA, a large granular lymphocytosis
(LGL's > 1,2OO/pL) developed. Repeated T-cell immuno-
phenotype studies revealed increased populations of CD8-'
(3,265/pL) and N K (826/pL) cells and gene rearrangement
studies revealed a clonal T-cell population confirming LGL
leukemia. This patient has subsequently responded to metho-
trexate with a complete response (stable Hct = 35) currently
extending 24 months on maintenance therapy of 5 mg metho-
trexate orally per week. This case shows that individual pa-
tients can respond to different therapies at different times.
BFU-E MATURATION
NO
YES I
REMISSION
OBTAINED
NO
Fig 2. The clinical value of in vitro culture. The relationship of
normal BFU-E growth (>30 bursts/105 MMNC) and remission (CR
and PR) is shown. BFU-E maturation in vitro was a superb predictor
of clinical response. Its sensitivity was 96%. its specificity was 78%
and its predictive value was 9396 ( P = ,0001 with 2-tailed chi-square
analysis).
 From www.bloodjournal.org by guest on August 14, 2018. For personal use only.
4836
It also shows that associated diseases can become clinically
appearant years after the presentation of PRCA.
DISCUSSION
In this study, we prospectively followed 37 PRCA patients
at the University of Washington. Overall, the clinical disor-
ders associated with their PRCA were compatible to those
previously described by Clark et al.39We also report a low
incidence of thymoma (8%) compared to case reports in
the literature of up to ?io%.’-’ This discrepancy probably
represents the reporting partiality of thymoma associated
PRCA in the earlier literature. Our results are consistent
with the thymoma incidences of 5 to 13% reported more
reCently.4.38.39.42One of the patients in our study developed
PRCA 3 months after thymoma resection; similar patients
have been reported p r e v i ~ u s l y . ~ ~
Recently, considerable attention has been focused on the
association between various lymphoproliferative disorders,
particularly LGL leukemia, and PRCA.2’-’3.34,38.42 Evidence
suggests that NK cells may directly inhibit erythropoie-
sis.21.W Although 9 patients had elevated percentages
(>20%) of NK cells (see Table I), only 2 had an increase
in their absolute numbers of NK cells and thus fulfilled
the criteria” for LGL leukemia at the time of their initial
evaluation. A third patient (patient 25) developed LGL leu-
kemia 4 years later. As T-cell gene rearrangement studies
were performed infrequently, the actual incidence of a clonal
T-cell or NK-cell process may have been higher than the
8% we report.
To define physiologies of PRCA and correlate this with
therapeutic responses, marrow BFU-E, CFU-E, and CFU-
GM frequencies were determined and coculture studies were
obtained. Our results indicate that PRCA involves three
mechanisms: immunologically mediated disease including
T-cell or antibody mediated inhibition of erythropoiesis,
HPV B19 infection, and an intrinsic stem or multipotent
progenitor cell defect (ie, myelodysplasia).
Immunologically mediated PRCA. In patients with im-
munologically mediated PRCA, BFU-E differentiate nor-
mally in culture. We interpret this data to imply that BFU-E
are present in vivo, and when moved from the immunologic
milieu of the patient, are able to fully differentiate to hemo-
globinized cells. Presumably within the patient there is
T-cell or antibody mediated inhibition of erythropoiesis, al-
though the in vitro coculture studies are significantly sensi-
tive to confirm these mechanisms in only a fraction of
studies.
Normal frequencies of BFU-E in vitro were associated
with therapeutic remission in 27 of 29 patients (sensitivity
0.93, specificity 0.78, predictive value 0.96, P = .Owl).
There were only 2 patients with normal BFU-E growth but
no clinical response (patients 6 and 30). Patient 6, without
any associated conditions, was referred following prednisone
failure. After failing subsequent azathioprine treatment, fur-
ther agents were not tried because of his poor health status
resulting from severe congestive heart failure. It is possible
that further immunomodulating therapy may have been suc-
cessful. Patient 30 presented with PRCA secondary to stage
IVA non-Hodgkin’s lymphoma of intermediate grade that
CHARLES ET AL
was unresponsive to chemotherapy. Response of PRCA in
this setting is reported to depend on a remission of the
lymph~ma.’~
These results are compatible with data previously reported
by Lacombe et a14’ who retrospectively analyzed 22 PRCA
patients and found a positive correlation between in vitro
erythroid growth and patient response rates to immunosup-
pressive therapies.
Human parvovirus B19 associated PRCA. PRCA can
result from chronic parvovirus infection in patients with im-
munodeficiency. HPV B19 specifically infects and lyses ery-
throid progenitor cells. When immunocompromised patients
are unable to develop neutralizing antibodies, persistent in-
fection results in PRCA. Therapy with intravenous IgG
(which contains high titers of antibody to HPV B 19) at 400
mg/kg/d X 5 d is indicated and universally effe~tive.~‘.~’ In
a recent retrospective study analyzing sera from 57 PRCA
patients, HPV B19 DNA was detected in 14%, suggesting
chronic parvovirus infection may be a relatively common
cause of PRCA.’~
Marrow culture of HPV B 19 infected patients showed two
growth patterns. Two patients had normal in vitro erythroid
differentiation (2 and 13) and obtained a CR to ATG or drug
(marijuana) withdrawal, currently extending 12 and 6 years,
respectively. BFU-E failed to mature in marrow cultures
from two additional patients. Patient 26 had a CR with IgG
but died 6 months later of a mycobacterial infection second-
ary to AIDS. Patient 14 failed prednisone therapy and subse-
quently died of an aortic dissection. As the anemias re-
sponded to ATG as well as IgG, HPV B 19-associated PRCA
may have an immunologic component to its pathogenesis in
certain circumstances.”
Intrinsic stem cell defect (myelodysplasia). An important
subset was 5 of the 8 patients whose BFU-E failed to grow
in culture. Four of these patients had some evidence of my-
elodysplasia (eg, a cytogenetic abnormality, myelofibrosis,
a single unilobed neutrophil). None of these 5 patients re-
sponded to immunosuppressive therapies and all died sec-
ondary to ANLL. We assume that these individuals had a
neoplastic hematopoietic stem multipotent progenitor cell
that was unable to fully differentiate along the erythroid
pathway, resulting in the morphology of PRCA. Erythroid
bursts were not seen because BFU-E were absent or were
unable to mature in vitro because of this genetic defect.
Interestingly, the two patients with 5q- abnormality (pt.
numbers 17 and 20) had transient improvement in erythroid
production concurrent with acute hepatitis (associated with
red cell transf~sion).~’This may imply that early in its
course, myelodysplastic erythroid cells have some capability
for erythropoiesis, but that reactive T cells may inhibit their
differentiation. Later in its evolution, immunosuppressive
therapies are no longer helpful because the myelodysplastic
stendprogenitor cell lacks any intrinsic capacity to differenti-
ate.
Three patients meeting standard criteria for PRCA2‘.”
were excluded from our study because of trilineage abnor-
malities (thrombocytopenia, circulating blast cells or Pelger-
Huet cells in peripheral smear, small and dysmorphic
megakaryocytes, megaloblastic erythropoiesis, and ringed
 From www.bloodjournal.org by guest on August 14, 2018. For personal use only.
PURE RED CELL APLASIA
sideroblasts) allowing the definitive morphologic diagnosis
of myelodysplasia. Erythroid bursts were similarly absent
from marrow cultures and in all three the anemia was refrac-
tory to treatment. It appears that in vitro culture is a powerful
clinical tool in predicting myelodysplasia when clinical and
morphologic findings are diagnostically indistinct.
Management of PRCA. An initial work-up should in-
clude a drug and medical history, a physical examination,
and a complete blood count and blood smear. Chest x-ray
or computed tomography to rule out thymoma, lymphocyte
immunophenotype, andor T-cell gene rearrangement stud-
ies, marrow cytogenetic studies, and serum Southem analysis
for presence of HPV B 19 are generally indicated. Specific
therapies (ie, drug withdrawal, IgG for chronic HPV B19,
thymectomy) should be pursued where appropriate. If the
work-up for concomitant disease is unhelpful, immunosup-
pressive therapy should be employed. Because many PRCA
patients respond to corticosteroids, we suggest initial therapy
with prednisone. If corticosteroids do not produce a remis-
sion within 6 weeks, a second line agent such as ATG,
cyclosporine, or cyclophosphamide should be initiated de-
pending on their relative contraindications. Low dose metho-
trexate therapy may be of specific benefit in patients with
coexistent LGL leukemia.
We suggest in vitro erythroid culture be used in patients
refractory to prednisone therapy and a second-line agent. If
BFU-E mature poorly in vitro no further immunologic ther-
apy is warranted, and the clinical diagnosis of myelodys-
plasia should be strongly considered. If in vitro culture shows
adequate BFU-E growth, then individual therapies should be
used consecutively until a sustained remission is achieved.
Our results show that PRCA is amenable to therapy, and
that durable remissions can be obtained in most patients.
ACKNOWLEDGMENT
The authors thank Drs Norbert Frickhofen and Neal Young, Na-
tional Institutes of Health, Bethesda, MD, for performing the HPV
B19 Southem analyses and Zeny Sisk and Allan Dimaunahan for
their assistance in preparation of the manuscript.
REFERENCES
1. Krantz SB: Acquired pure red cell aplasia, in Hoffman R, Benz
El Jr, Shattil SJ, Furie B, Cohen HJ, Silberstein LE (eds): Biology
of Stem Cells and Disorders of Hematopoiesis, New York, NY,
Churchill Livingstone, 1995, p 172
2. Hirst E, Robertson TI: The syndrome of thymoma and erythro-
blastopenic anemia. Medicine 46:225, 1967
3. Schmid JR, Kiely JM, Harrison EG Jr, Bayrd ED, Pease GL:
Thymoma associated with pure red-cell agenesis. Cancer 18:216,
1965
4. Wong KF, Chau KF, Chan JKC, Chu YC, Li CS: Pure red cell
aplasia associated with thymic lymphoid hyperplasia and secondary
erythropoietin resistance. Am J Clin Pathol 103:346, 1995
5. Morgenthaier TI, Brown LR, Colby TV,Harper CM Jr, Coles
DT: Thymoma. Mayo Clin Proc 68:1110, 1993
6. Souadjian JV, Enriquez P, Silverstein MN, Wpin J-M: The
spectrum of diseases associated with thymoma. Arch Intem Med
134:374, 1974
7. Roland AS: The syndrome of benign thymoma and primary
aregenerative anemia. Am J Med Sci 247:719, 1964
4837
8. Dessypris EN, Baer MR, Sergent JS, Krantz SB: Rheumatoid
arthritis and pure red cell aplasia. Ann Intem Med 100:202, 1984
9. Rodrigues JF, Harth M, Barr RM: Pure red cell aplasia in
rheumatoid arthritis. J Rheumatol 15:1159, 1988
10. Dainiak N, Hardin J, Floyd V, Callahan M, Hoffman R: Hu-
moral suppression of erythropoiesis in systemic lupus erythematosus
(SLE) and rheumatoid arthritis. Am J Med 69537, 1980
11. Cassileth PA, Myers AR: Erythroid aplasia in systemic lupus
erythematosus. Am J Med 55:706, 1973
12. Wilson HA, McLaren GD, Dworken HJ, Tebbi K: Transient
pure red-cell aplasia: Cell-mediated suppression of erythropoiesis
associated with hepatitis. AM Int Med 92:196, 1980
13. Purtilo DT, Zelkowitz L, Harada S, Brooks CD, Bechtold T,
Lipscomb H, Yetz J, Rogers G: Delayed onset of infectious mononu-
cleosis associated with acquired agammaglobulinemia and red cell
aplasia. Ann Int Med 101:180, 1984
14. Williams ML, Loughran TP Jr, Kidd PG, Starkebaum GA:
Polyclonal proliferation of activated suppressor/cytotoxic T cells
with transient depression of natural killer cell function in acute infec-
tious mononucleosis. Clin Exp Immunol 77:71, 1989
15. Suzuki A, Takahashi T, Taniguchi A, Kotake C, Seo T, Toda
T, Kobayashi K, Tsukamoto N, Fukumoto M: Pure red cell aplasia
associated with non-Hodgkin’s lymphoma and hemolytic anemia.
Jpn J Clin Oncol 21:384, 1991
16. Carloss HW,Saab CA, Tavassoli M: Pure red cell aplasia
and lymphoma. J Am Med Assoc 242:67, 1979
17. Morgan E, Pang KM, Goldwasser E: Hodgkin disease and
red cell aplasia. Am J Hematol 5:71, 1978
18. Alter R, Joshi SS, Verdirame JD, Weisenburger DD: Pure
red cell aplasia associated with B cell lymphoma: Demonstration of
bone marrow colony inhibition by serum immunoglobulin. Leuk Res
14:279, 1990
19. Mangan KF, Chikkappa G, Farley PC: T gamma (Ty) cells
suppress growth of erythroid colony-forming units in vitro in the
pure red cell aplasia of B-cell chronic lymphocytic leukemia. J Clin
Invest 70: 1148, 1982
20. Chikkappa G, Zarrabi MH, Tsan M-F: Pure red-cell aplasia
in patients with chronic lymphocytic leukemia. Medicine 65:339,
1986
21. Loughran TP Jr: Clonal diseases of large granular lympho-
cytes. Blood 82:1, 1993
22. Oshimi K, Yamada 0, Kaneko T, Nishinarita S, Lizuka Y,
Urabe A, Inamori T, Asano S, Takahashi S, Hattori M, Naohara T,
Ohira Y, Togawa A, Masuda Y, Okubo Y, Furusawa S, Sakamoto
S, Omine M, Mori M, Tatsumi E, Mizoguchi H: Laboratory findings
and clinical courses of 33 patients with granular lymphocyte-prolif-
erative disorders. Leukemia 7:782, 1993
23. Kwong YL, Wong KF, Chan LC, Liang RHS. Chan JKC,
Lin CK, Chan TK: Large granular lymphocyte leukemia: A study
of nine cases in a Chinese population. Am J Clin Pathol 103:76,
1995
24. Dessypris EN: Pure Red Cell Aplasia. Baltimore, MD, The
Johns Hopkins University Press, 1988
25. Takahashi M, Nikkuni K, Tanaka I, Furukawa T, Moriyama
Y, Shibata A, Satoh S, Kuwabara Y, Maruyama Y,Matsunaga Y:
Serum erythropoietic inhibitors in patients with pure red cell aplasia.
Ann Hematol 63:9, 1991
26. Peschle C, Marmont AM, Marone G, Genovese A, Sasso GF,
Condorelli M. Pure red cell aplasia: Studies on an IgG serum inhibi-
tor neutralizing erythropoietin. Br J Haematol 30:411, 1975
27. Mangan KF, Volkin R, Winkelstein A: Autoreactive erythroid
progenitor-T suppressor cells in the pure red cell aplasia associated
with thymoma and panhypogammaglobulinemia. Am J Hematol
23:167, 1986
28. Socinski MA, Ershler WB, Tosato G , Blaese RM: Pure red
 From www.bloodjournal.org by guest on August 14, 2018. For personal use only.
4838
blood cell aplasia associated with chronic Epstein-Barr virus infec-
tion: evidence for T cell-mediated suppression of erythroid colony
forming units. J Lab Clin Med 104:995, 1984
29. Reid TJ 111, Mullaney M, Burrell LM, Redmond J 111, Mangan
KF: Pure red cell aplasia after chemotherapy for Hodgkin’s
lymphoma: In vitro evidence for T cell mediated suppression of
erythropoiesis and response to sequential cyclosporin and erythro-
poietin. Am J Hematol 46:48, 1994
30. Akard LP, Brandt J, Lu L, Jansen J, Hoffman R: Chronic T
cell lymphoproliferative disorder and pure red cell aplasia. Am J
Med 83:1069, 1987
31. Mangan KF, Chikkappa G, Scharfman WB, Desforges JF:
Evidence for reduced erythroid burst (BFU,) promoting function of
T lymphocytes in the pure red cell aplasia of chronic lymphocytic
leukemia. Exp Hematol 9:489, 1981
32. Nagasawa T, Abe T, Nakagawa T: Pure red cell aplasia and
hypogammaglobulinemia associated with Tr-cell chronic lympho-
cytic leukemia. Blood 57: 1025, 1981
33. Hoffman R, Kopel S, Hsu SD, Dainiak N, Zanjani ED: T
cell chronic lymphocytic leukemia: Presence in bone marrow and
peripheral blood of cells that suppress erythropoiesis in vitro. Blood
52255, 1978
34. Abkowitz JL, Kadin ME, Powell JS, Adamson JW: Pure red
cell aplasia: Lymphocyte inhibition of erythropoiesis. Br J Haematol
6359, 1986
35. Levitt LJ, Reyes GR, Moonka DK, Bensch K, Miller RA,
Engleman EG: Human T cell leukemia virus-I-associated T-suppres-
sor cell inhibition of erythropoiesis in a patient with pure red cell
CHARLES ET AL
aplasia and chronic Ty-lymphoproliferative disease. J Clin Invest
81538, 1988
36. Frickhofen N, Chen U ,Young NS, Cohen BJ, Heimpel H,
Abkowitz JL: Parvovirus B19 as a cause of acquired chronic pure
red cell aplasia. Br J Haematol 87:818, 1994
37. Dessypris EN, Fog0 A, Russell M, Engel E, Krantz SB: Stud-
ies on pure red cell aplasia. X. association with acute leukemia and
significance of bone marrow karyotype abnormalities. Blood 56:421,
1980
38. Dessypris EN: The biology of pure red cell aplasia. Semin
Hematol 28:275, 1991
39. Clark DA, Dessypris EN, Krantz SB: Studies on pure red cell
aplasia. XI. results of immunosuppressive treatment of 37 patients.
Blood 63:277, 1984
40. Abkowitz JL, Powell JS, Nakamura JM, Kadin ME, Adamson
JW: Pure red cell aplasia: Response to therapy with anti-thymocyte
globulin. Am J Hematol 23:363, 1986
41. Frickhofen N, Abkowitz JL, Safford M, Berry JM, Antunez-
de-Mayolo J, Astrow A, Cohen R, Halperin I, King L, Mintzer D,
Cohen B, Young NS: Persistent B19 parvovirus infection in patients
infected with human immunodeficiency virus type 1 (HIV-I): A
treatable cause of anemia in AIDS. Ann Int Med 113:926, 1990
42. Lacey MO, Kurtin PJ, Tefferi A: Pure red cell aplasia
(PRCA): Association with large granular lymphocytic leukemia
(LGL) and the prognostic value of cytogenetic abnormalities. Blood
81:216a, 1995 (abstr, suppl 1)
43. Lacombe C, Casadevall N, Muller 0, Varet B: Erythroid
progenitors in adult chronic pure red cell aplasia: Relationship of in
vitro erythroid colonies to therapeutic response. Blood 64:7l, I984
 From www.bloodjournal.org by guest on August 14, 2018. For personal use only.

1996 87: 4831-4838

The pathophysiology of pure red cell aplasia: implications for therapy

RJ Charles, KM Sabo, PG Kidd and JL Abkowitz

Updated information and services can be found at:

http://www.bloodjournal.org/content/87/11/4831.full.html
Articles on similar topics can be found in the following Blood collections

Information about reproducing this article in parts or in its entirety may be found online at:
http://www.bloodjournal.org/site/misc/rights.xhtml#repub_requests

Information about ordering reprints may be found online at:

http://www.bloodjournal.org/site/misc/rights.xhtml#reprints

Information about subscriptions and ASH membership may be found online at:
http://www.bloodjournal.org/site/subscriptions/index.xhtml

Blood (print ISSN 0006-4971, online ISSN 1528-0020), is published weekly by the American
Society of Hematology, 2021 L St, NW, Suite 900, Washington DC 20036.
Copyright 2011 by The American Society of Hematology; all rights reserved.