Hematological profile in pyrexia of unknownorigin: role of bone marrow trephine biopsyvis-a-vis aspiration

Ruchika Gupta, Namrata Setia, Prerna Arora, Sompal Singh and Tejinder Singh

Department of Pathology, Maulana Azad Medical College, New Delhi, India

Background: Bone marrow examination, by aspiration and/or trephine biopsy, is an important

procedure in arriving at a diagnosis for long-duration febrile illness. The role of trephine biopsy in

immunocompromised host, especially HIV-positive patients, has been well studied in the

literature. However, its utility in immunocompetent patients is still shrouded by controversy. Thus,

the authors attempted to evaluate the utility of marrow aspirate vis-a-vis trephine biopsy in

establishing a diagnosis in cases of pyrexia of unknown origin in immunocompetent individuals,

along with an analysis of haematological alterations in these patients.

Materials and methods: Over a period of 8 years, 121 patients with pyrexia of unknown origin

underwent both bone marrow aspiration and trephine biopsy as a part of diagnostic work-up.

These cases were reviewed for their clinical data and hematological findings, including detailed

morphological features in aspiration smears and trephine biopsies. Bone marrow aspiration and

biopsy were compared for their diagnostic efficacy in these patients.

Results: A wide age range (2–65 years) was noted with a slight male predominance (2 : 1).

Anemia was the most common feature in peripheral blood findings, seen in 97.5% of patients.

Bone marrow aspiration was diagnostic in only 16.5% of cases, which revealed leishmaniasis or

pure red cell aplasia. Granulomas were infrequent in marrow aspiration smears, as only two cases

(1.6%) showed ill defined epithelioid cell collections. Compared to this, trephine biopsy offered a

diagnosis in 76% of the cases. Granulomas were a frequent finding in the trephine biopsy, being

present in 70% of the cases included. Additional cases diagnosed on biopsy (over those

diagnosed with aspiration smears) included lymphoma, tuberculosis, fungal infection, sarcoidosis

and hypocellular marrow.

Conclusion: Bone marrow trephine biopsy is an important adjunct to aspiration in arriving at an

aetiological diagnosis of patient with long-duration febrile illness, and should be routinely

performed in such cases. The presence of granulomas in trephine biopsy increases the likelihood

of an etiologic diagnosis in these patients.

Keywords: Pyrexia of unknown origin, bone marrow, biopsy, diagnosis

Introduction

Bone marrow examination has become an important

diagnostic procedure in the evaluation of patients

with pyrexia of unknown origin (PUO), especially in

Correspondence to: Dr Ruchika Gupta, Department of Pathology,Maulana Azad Medical College, 162, Pocket-B, Sarita Vihar, New Delhi110076, IndiaE-mail: ruchika257@yahoo.com

� W. S. Maney & Son Ltd 2008Received 8 February 2008; accepted 15 May 2008DOI 10.1179/102453308X343446 Hematology 2008 VOL 13 NO 5 307

cases with prolonged history of fever.1 However, the

exact role of bone marrow trephine biopsy vis-a-vis

aspiration in patients with prolonged fever is not well

established in the literature. Most of the studies

investigating the role of bone marrow trephine biopsy

in PUO have been limited to HIV-positive patients.2,3

These studies have concluded that bone marrow

biopsy still plays a role in investigation of PUO in

HIV-positive patients.2,3 However, one study ques-

tioned the utility of bone marrow trephine biopsy in

immunocompetent individuals with long-standing

fever.4 A systematic study of trephine biopsy in

immunocompetent patients with PUO is still lacking

in the available literature.

Thus, the aim of this study was to evaluate the

diagnostic role of bone marrow biopsy in immuno-

competent patients.

Materials and methods

Over a period of 8 years (1999–2007), 265 patients

with pyrexia of unknown origin of more than two

weeks’ duration underwent bone marrow aspiration

at our hospital. Of these 265 patients, 121 also

underwent bone marrow trephine biopsy for further

evaluation. Their detailed clinical history, physical

examination and various relevant investigations were

noted from the case records. Bone marrow aspiration

and biopsies had been done from posterior superior

iliac spine under local anesthesia. The aspirate smears

were stained with Giemsa stain whereas the histolo-

gical sections of biopsy were stained with hematox-

ylin-eosin stain. The parameters studied were

cellularity of marrow particles; status of hematopoie-

tic cells; increase in plasma cells; presence of fibrosis;

presence, site and type of granulomas; presence of

necrosis and any other associated features seen

including hemophagocytosis. Biopsy sections were

also stained by reticulin stain for degree of fibrosis,

Ziehl-Neelsen (ZN) stain for acid fast bacilli and

periodic acid Schiff (PAS) stain for fungi.

Results

The age of the study patients ranged from 2 to 65 years

with a mean age of 31 years. There was a slight male

preponderance (male: female ratio of 2 : 1). All the

patients had prolonged fever, duration ranging from

three weeks to two months. Nearly half the cases gave

history of weight loss (47.1%). Hepatosplenomegaly

was seen in 57 patients (47.1%). The clinical features

are summarized in Table 1.

The clinical differential diagnoses considered ran-

ged from infectious conditions (such as tuberculosis,

malaria, leishmaniasis and typhoid fever) to haema-

tological malignancies.

Hematological features

Anaemia was the most common finding, seen in 118

patients (97.5%). It was normocytic normochromic

in 101 cases, macrocytic in 15 and microcytic

hypochromic in two cases. The median hemoglobin

in this study was 8.02 g/dl (inter-quartile range 6.8–

9.4 g/dl). Similarly, the median total leukocyte

count was 5930 mm23 (inter-quartile range being

3355–6950 mm23) and median platelet count

was 1.62 lac mm23 (inter-quartile range 1.0–

2.1 lac mm23). Peripheral smear was negative for

malarial parasite in all cases. The details of hemato-

logical profile of these cases are given in Table 2.

Bone marrow aspiration findings

Bone marrow aspiration smears were normocellular

in 64 cases (52.8%), hypercellular in 40 cases (33%)

and hypocellular in eight cases (6.6%). In the rest nine

cases, it was diluted with peripheral blood. Erythroid

hyperplasia was observed in 20 aspirates (16.5%)

while others showed myeloid preponderance.

Erythropoeisis was normoblastic in 101 cases

(83.4%), megaloblastic in 15 cases (12.4%) and

micronormoblastic in two cases. There was a reactive

plasmacytosis and increase in marrow histiocytes in

48 cases each (39.6%). Additional associated findings

included increase in eosinophilic precursors in 24

cases (19.8%), hemophagocytosis in 15 cases (12.4%)

and pseudo-gaucher cells in two cases. Intracellular

and extra cellular amastigote forms of Leishmania

donovani (L.D. bodies) were seen in 18 cases (14.8%)

Table 1 Summary of the clinical presentation of thecases of PUO who underwent marrowexamination

Signs and symptoms Number of cases n5121 (%)

Fever 121 (100)Weight loss 57 (47.1)Diarrhea/vomiting 10 (8.3)Amenorrhea 6 (4.9)Epistaxis 4 (3.3)Malena 4 (3.3)Rashes/Jaundice 1 each (0.8)Hepatosplenomegaly 57 (47.1)Lymphadenopathy 24 (19.8)

Table 2 Common hematological abnormality in the casesof PUO

Hematological features Number of cases (%)

Anemia 118 (97.5)Pancytopenia 27 (22.3)Leucopenia 18 (14.8)Thrombocytopenia 10 (8.2)Leucocytosis 6 (4.9)

Gupta et al. Hematological profile in pyrexia of unknown origin

308 Hematology 2008 VOL 13 NO 5

(Fig. 1). In two cases, a marked myeloid preponder-

ance (myeloid: erythroid ratio of 25 : 1) with few giant

proerythroblasts (GPE) was seen, thus suggesting

pure erythroid cell aplasia (PRCA). An ill-defined

collection of epithelioid cells was observed in only

two cases (1.6%) in this series. An etiologic diagnosis

could be ascertained in 20 cases based on aspiration

smears alone (or in combination with peripheral

smears).

Bone marrow biopsy findings

Trephine biopsies were adequate in all 121 cases. The

cellularity and pattern of hematopoiesis in biopsy was

similar to that seen in corresponding aspirate smears.

In nine cases with aspiration smears diluted with

blood, trephine biopsy revealed hypoplastic marrow

with reduced hematopoiesis.

In the 18 cases showing L.D. bodies in aspirate

smears, the trephine biopsy showed an increase in

histiocytes with few L.D. bodies seen intracellularly

in histiocytes (Fig. 2a). There was also an associated

Figure 1 Bone marrow aspirate (Giemsa stain) showing

numerous Leishmania donovani (L.D.) bodies

intracellularly within the macrophages (Giemsa

6400)

Figure 2 Photomicrographs showing L.D. bodies in a section of trephine biopsy (a, H&E 6400). A case of tuberculosis

showing extensive caseation necrosis surrounded by epithelioid cell granulomas with giant cells (b, c, H&E

6100, 6200 respectively) and positive staining for acid fast bacilli (d, Ziehl-Neelsen stain 61000)

Gupta et al. Hematological profile in pyrexia of unknown origin

Hematology 2008 VOL 13 NO 5 309

plasmacytosis in the marrow. In two patients

diagnosed as PRCA on aspiration smears, trephine

biopsy showed similar features.

Granulomas were identified in trephine biopsy in

85 of the 121 cases studied (70.2%). There was ill-

defined collection of epithelioid cell in 32 of these 85

cases (37.6%). The granulomas were variable in their

location, being paratrabecular or interstitial.

Caseation necrosis was observed in 19 (22.3%) cases.

There was marrow fibrosis in 13 biopsies of our study

cases, with three showing myelofibrosis and second-

ary myeloid metaplasia. Other associated features

seen in trephine biopsies with granulomas were

increase in eosinophilic precursors in 24 cases

(28.2%) and increase in histiocytes in 48 cases

(56.4%) with evidence of hemophagocytosis in 15

(17.6%) of these. Special stain for iron revealed

increase in iron stores in 26 of 121 (21.5%) biopsies,

decreased in 40 (33%) and normal stores in the rest.

An etiology for granuloma could be established in

63 of the 85 cases, in which they were observed. Out

of 85 cases, 54 were diagnosed as tuberculosis (63.5%).

Of these 54 cases, acid fast bacilli was demonstrable in

only eight cases, while in 28 patients, a subsequent

lymph node aspirate showed acid-fast bacilli, confirm-

ing a diagnosis of tuberculosis. In 18 patients, bone

marrow biopsy as well as a subsequent lymph node

aspirate revealed epithelioid cell granulomas with

caseation and these were diagnosed as tuberculosis

based on these features (Fig. 2b–d.). All these 54

patients responded to anti-tubercular therapy (ATT).

Bone marrow biopsy in four cases (4.7%) showed

non-caseating granulomas with associated fibrosis,

increase in eosinophilic precursors and plasma cells.

Occasional large mononuclear as well as Reed-

Sternberg like cells were also seen (Fig. 3). These

cases were diagnosed as Hodgkin’s involvement of

marrow based on marrow findings. Lymph node

biopsy was suggested, which showed features of

Hodgkin’s lymphoma.

In one case (1.1%), bone marrow biopsy showed

myeloid preponderance, ill-defined granulomas and

Figure 3 Photomicrographs from a case of Hodgkin’s involvement of marrow showing a granuloma (a, H&E 6100). A

polymorphic infiltrate with atypical mononuclear, binucleate and occasional giant cells (b, c, H&E 6200).

Marrow fibrosis is also seen in this case (d, H&E 6100)

Gupta et al. Hematological profile in pyrexia of unknown origin

310 Hematology 2008 VOL 13 NO 5

few paratrabecular nodular deposits of atypical

lymphoid cells, immunoreactive for B-cell marker

(CD20) and negative for T-cell marker (CD3). A

diagnosis of non-Hodgkin’s lymphoma, B-cell immu-

nophenotype, involving the bone marrow was made.

In two patients (2.3%), bone marrow trephine

biopsy showed ill-defined granulomas with foamy

macrophages. Periodic acid Schiff (PAS) stain

revealed budding yeast forms surrounded by a clear

space, which did not stain with mucicarmine. The

morphology was compatible with histoplasma. In

another two cases (2.3%), marrow showed non-

caseating epithelioid cell granulomas with rare cell

showing asteroid body. These cases were diagnosed

presumptively as sarcoidosis. Chest radiograph in

these two cases showed hilar and mediastinal

lymphadenopathy. Fine needle aspiration from med-

iastinal lymph node showed non-caseating granulo-

mas and stain for acid fast bacilli were negative.

All patients were seronegative for human immu-

nodeficiency virus by enzyme linked immunosorbent

assay for HIV-1 and HIV-2. A summary of etiolo-

gical diagnosis reached at in our study is given in

Table 3.

Thus, bone marrow trephine biopsy suggested a

diagnosis in 92 (76%) of the 121 cases. Of these, 18

were leishmaniasis, 54 had tuberculosis, nine patients

showed hypoplastic marrow and four had Hodgkin’s

involvement of the marrow. Two cases each showed

fungal infection and sarcoidosis. Other diagnoses in

bone marrow trephine biopsy included non-

Hodgkin’s lymphoma (1) and pure red cell aplasia

(2). In comparison, aspiration smears offered an

etiologic diagnosis in only 20 (16.5%) cases, 18 with

leishmaniasis and two with pure red cell aplasia.

Discussion

Pyrexia of unknown origin (PUO) is defined as

unexplained fever for more than 2 weeks, where no

etiology could be found after extensive routine

investigations. Hematological alterations are com-

mon in these patients as seen in the present study.

Anemia was most common (97.5%), followed by

pancytopenia, leucopenia and thrombocytopenia.

This is similar to the findings revealed in earlier

studies.5,6 The high frequency of anemia, in our view,

may be attributable to a mechanism similar to anemia

of chronic disease coupled with some degree of

nutritional deficiencies.

A variety of morphologic changes in the bone

marrow have been described in various infectious and

systemic diseases resulting in PUO.7 These changes

may be features of acute inflammation (interstitial

edema, vascular congestion, hemorrhage, ischemic

necrosis or suppurative necrosis) or chronic inflam-

mation with granuloma formation, reactive lymphoid

hyperplasia, plasmacytosis, histiocytosis or fibrosis.7

Bone marrow aspiration, in this study, showed

variable features, most commonly myeloid prepon-

derance. Reactive increase in histiocytes and plasma

cells was observed in 39.6% cases each. Aspiration

was diagnostic in 20 (16.5%) cases, with presence of

amastigote form of Leishmania donovani in 18 and

PRCA in two. No definite granulomas were seen in

any of the cases on aspiration smears, though an ill-

defined collection of epithelioid cells was seen in two

cases (1.6%), which were subsequently diagnosed as

tuberculosis on trephine biopsy. This has been

ascribed to the associated fibrosis in relation to

granulomas and hence trephine biopsy is considered

to be a better tool for demonstration of granuloma.8

The utility of trephine biopsy in immunocompetent

patients with PUO has been questioned.4 In a study

by Riley et al., an evaluation of bone marrow biopsy

in detection of mycobacteria in patients with PUO

was conducted. They concluded that bone marrow

culture for mycobacteria should be reserved for

severely immunosuppressed patients and BMB has

a role in investigation of PUO in HIV-positive

patients.3 The findings of the present study are in

contrast to these published results.

Trephine biopsy in the authors’ series concluded or

strongly suggested an etiological diagnosis in 92 cases

(76%) as compared to 20 cases (16.5%) diagnosed by

marrow aspiration, as shown in Table 3. Of these 92

cases, 18 were leishmaniasis and two PRCA diag-

nosed on aspirate smears. Nine cases were diagnosed

as hypoplastic anemia leading to pancytopenia and

recurrent infections. Of the cases showing granulo-

mas in trephine biopsy, 54 were diagnosed as

tuberculosis, four as Hodgkin’s disease, two cases

each as fungal infections and sarcoidosis, and one as

Table 3 Summary of the diagnoses offered on bonemarrow aspirate and biopsy

Bone marrowaspirate

Bone marrowbiopsy

Tuberculosis 0 54Leishmaniasis 18 18Hypoplastic marrow 0 9Hodgkin’s disease 0 4Fungal infection 0 2Sarcoidosis 0 2Non-Hodgkin’s lymphoma 0 1Pure red cell aplasia 2 2No diagnosis 101 29

Gupta et al. Hematological profile in pyrexia of unknown origin

Hematology 2008 VOL 13 NO 5 311

non-Hodgkin’s lymphoma involvement of the mar-

row. Earlier studies in literature have reported an

etiological diagnosis in 78–87% of cases with granu-

lomas in bone marrow biopsy.1,5 In the present series,

a definite etiology was noted in 74% of bone marrow

biopsies in which granuloma was noted.

Tuberculosis was the most common etiological

diagnosis in the present study, similar to another

study from South-east Asia.5 However, acid-fast

bacilli could be demonstrated in only eight cases

(14.8% of tuberculous cases). It has to be remembered

that the failure to detect acid-fast bacilli in histolo-

gical sections does not exclude a diagnosis of

tuberculosis since organisms are identified in only

about 25% of the marrow biopsies in documented

disease.5 A remarkable observation in the present

study was the presence of caseous necrosis in 19 of

the 54 cases of tuberculosis (35.2%). This is in

contrast with the reports by Bodem et al. and

Vilalta-Castel et al.1,6 Few studies report caseation

in bone marrow biopsies with granuloma in cases of

tuberculosis.5 The presence of caseating necrosis in

granulomas helps in suggesting a diagnosis of

tuberculosis, especially in endemic areas like the

authors’. Contrary to the findings by Vilalta-Castel

et al.,6 performance of ZN and PAS stain helped in

identification of eight cases of tuberculosis and two

cases of fungal infection respectively in the present

study. Studies by Akpek et al. and Riley et al.

concluded that bone marrow culture does not play a

significant role in immunocompetent individuals with

suspected tuberculosis. However, both these studies

also agreed that histopathologic examination of bone

marrow has a high diagnostic specificity when

granulomas are detected.2,3 The present study lends

support to the importance of histologic examination

of bone marrow in PUO.

Thus, it can be concluded that bone marrow biopsy

is a complementary and in some respects, better

diagnostic tool than aspiration in patients with

pyrexia of unknown origin. Presence of granulomas

in trephine biopsy helps to narrow down the

diagnostic possibilities and further guide the required

special stains. Thus, bone marrow trephine biopsies

should always be performed in conjunction with

aspiration as a routine diagnostic procedure in

patients with long-duration febrile illness.

References1 Bodem CR, Hamory BH, Taylor HM, Kleopfer L. Granulomatous

bone marrow disease. Medicine 1983; 62: 372–383.

2 Akpek G, Lee SM, Gagnon DR, Cooley TP, Wright DG. Bone

marrow aspiration, biopsy and culture in the evaluation of HIV-

infected patients for invasive mycobacteria and histoplasma infec-

tions. Am J Haematol 2001; 67: 100–106.

3 Riley UBG, Crawford S, Barrett SP, Abdalla SH. Detection of

mycobacteria in bone marrow biopsy specimens taken to investigate

pyrexia of unknown origin. J Clin Pathol 1995; 48: 706–709.

4 Prego V, Glatt AE, Roy V, Thelmo W, Dincsoy H, Raufman JP.

Comparative yield of fungi and mycobacteria in liver biopsy and

bone marrow biopsy in human immunodeficiency virus-infected

patients. Arch Intern Med 1990; 50: 333–336.

5 Basu D, Saravana R, Purushotham B, Ghotekar LH. Granulomas

in bone marrow – a study of fourteen cases. Indian J Pathol

Microbiol 2005; 48: 13–16.

6 Vilalta-Castel E, Valdes-Sanchez MD, Guerra-Vales JM et al.

Significance of granulomas in bone marrow: a study of 40 cases. Eur

J Haematol 1988; 41: 12–16.

7 Diebold J, Molina T, Camilleri-Broet S, le Tourneau A, Audouin J.

Bone marrow manifestations of infections and systemic diseases

observed in bone marrow trephine biopsy. Histopathology 2000; 37:

199–211.

8 Bain BJ. Bone marrow biopsy. J Clin Pathol 2001; 54: 737–742.

Gupta et al. Hematological profile in pyrexia of unknown origin

312 Hematology 2008 VOL 13 NO 5