Q J Med 2001; 94: 101-105
© 2001 Association of Physicians
The utility of bone-marrow examination in HIV-infected adults in South Africa
1 From the Department of Medicine, Chris Hani Baragwanath Hospital 2 Department of Anatomical Pathology, South African Institute for Medical Research, and University of the Witwatersrand, Johannesburg, South Africa
Received 16 February 2000 and in revised form 27 November 2000
 |
Summary |
|---|
 Top Summary IntroductionMethodsResultsDiscussionReferences |
|---|
We retrospectively reviewed a consecutive case series of 257 adults with HIV infection who had undergone a bone-marrow examination with trephine bone biopsy, to assess the diagnostic usefulness of bone-marrow examination and evaluate possible predictors of a diagnostic examination. Bone-marrow examination was positive in 97 (38%) patients and gave a unique diagnosis in 61 (24%). The diseases were tuberculosis (83 patients), Mycobacterium avium complex infection and cryptococcosis (four patients each), and haematological malignancies (eight patients). The yield of the examination was significantly increased, by univariate analysis, in patients with wasting, oral thrush, leukopenia, CD4
100/mm3, and granuloma formation on histopathology. Granulomata were present in 113 (44%) patients, of whom 28 (25%) had no specific cause identified. Granulomata occurred in 22 (25%) of 89 patients with CD4 >100/mm3 compared to 36 (51%) of 70 patients with CD4100/mm3 (OR 0.3; 95%CI 0.15–0.62). Of 48 patients with CD4 <50/mm3, 25 had granulomata, including 15 with caseation necrosis. The yield of bone-marrow examination was considerable in our setting. Expanding access to modern blood culture techniques for mycobacteria to primary care level could limit the number of bone-marrow examinations required. |
Introduction |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
Opportunistic infections and haematological malignancies can involve the bone marrow in people with human immunodeficiency virus (HIV) infection. In the industrialized world, bone-marrow examination (BME) can identify such disease in 22–42% of patients, but this diagnosis is unique to the bone marrow in only 8–10% of cases.1–12 Blood culture for mycobacterial infections, the commonest diagnosis in most series, is as effective as BME with bone-marrow culture.2,3,12
There have been no large series examining the utility of BME in Southern Africa nor in sub-Saharan Africa in general. Due to limited financial resources in such countries, lysis-centrifugation blood culture systems for mycobacteria tend only to be available in some academic hospitals in major cities, and are not available in the primary care settings which manage the majority of HIV-infected individuals. We have studied retrospectively a consecutive series of BME in HIV-infected adults at Chris Hani Baragwanath Hospital, Soweto, South Africa, to assess the diagnostic yield of this investigation, and evaluate possible predictors of a diagnostic BME.
|
Methods |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
Chris Hani Baragwanath Hospital is a 3300-bed public sector university hospital serving the people of Soweto. There are approximately 2500 adult medical admissions per month, 26% of whom were HIV-seropositive in March, 1997. The study was approved by the Committee for Research on Human Subjects of the University of the Witwatersrand.
All adults (
18 years) with HIV infection who had undergone a BME with trephine bone biopsy over the 2 years from January 1996 to December 1997 were identified. The medical and laboratory records of these patients were retrospectively reviewed. A standardized data collection sheet was used to retrieve relevant information, including patient characteristics, indication for performing the BME and final diagnosis; clinical findings such as oral thrush, hepatomegaly and splenomegaly; laboratory results of CD4 lymphocyte count, blood count, liver function tests and culture results of bone marrow aspirates; pathological results of BME, including the presence of acid-fast bacilli, fungal elements, granulomata, and evidence of malignancy.
The BME was considered positive for disseminated infection or malignancy if the culture of the aspirate grew a significant pathogen, if there were fungal elements, acid-fast bacilli, well-formed granulomata with caseation necrosis, or evidence of lymphoma or other malignancy on histopathological examination.
Marrow core biopsies were fixed in formalin and decalcified briefly in weak acid. Two micrometer sections were stained with haematoxylin and eosin, Ziehl-Neelsen and Grocotts methenamine silver stain, and other special stains where appropriate. Bone-marrow aspirates were stained routinely with Wright-Giemsa stain.
We regarded any collection of epithelioid histiocytes as a granuloma. Poorly formed granulomata had indistinct margins, while well-formed granulomata were sharply demarcated. Well-formed granulomata with caseation necrosis formed a third group.
For mycobacterial cultures, marrow aspirate material was inoculated at the bedside into Lowenstein-Jensen agar slopes and incubated at 35°C under 50% CO2 for 6 weeks before being designated negative. Acid-fast organisms cultured were identified as Mycobacterium tuberculosis (MTB) or Mycobacterium avium complex (MAC) using nucleic acid hybridization (Gen-Probe).
2 or Fisher's exact test was used to evaluate differences between groups.
|
Results |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
During the two-year period, 257 HIV-seropositive patients had a BME with trephine bone biopsy. There were 116 (45.1%) males and 141 (54.9%) females. The mean age was 34 years (range 18–65 years).
A disease was identified by BME in 97 cases giving a diagnostic yield of 37.7% (TB in 32.2%, haematological malignancy in 3.1%, two with dual disease with TB, and MAC and cryptococcosis in 1.6% each). In 36 of these patients, the same disease was diagnosed by one or more alternative modalities outside of the bone marrow. Thus BME provided a unique diagnosis in 61 (23.7%) patients and in 62.9% of cases with a positive BME. In a further three patients, the diagnosis on BME was more rapid than that obtained by other means, so that BME was unique or rapid in 64 (24.9%) patients. Any diagnosis of disease was obtained in 140 (54.4%) patients, so BME provided the diagnosis in 69.3% of such patients. Of 123 patients with mycobacterial or fungal (other than Pneumocystis carinii) disease, BME revealed the diagnosis in 91 (74.0%).
The yield and diagnosis according to indication for BME is shown in Table 1
. Of the patients with peripheral blood cytopenia alone as an indication, 18 had pure thrombocytopenia, and in none of these cases did BME provide a diagnosis. The other 35 patients had bicytopenia or pancytopenia. All patients with cryptococcosis of the bone marrow already had cryptococcal meningitis proven on cerebrospinal fluid culture and had a BME performed for cytopenia. Pneumocystis carinii pneumonia (PCP) was diagnosed only after the BME. The BME was uniquely diagnostic in one patient with Hodgkin's lymphoma and one with non-Hodgkin's lymphoma. There was no significant difference in yield by indication for BME.
|
Subjects with diagnoses made by investigations other than BME included another 31 patients with TB (yield of BME 83 of 114 (72.8%)), five with non-Hodgkin's lymphoma, nine with PCP and one each with cryptococcal meningitis, Kaposi's sarcoma and haemophagocytic syndrome. Concomitant infection with TB was found in four patients with PCP and one each with cryptococcal meningitis, non-Hodgkin's lymphoma and acute myelogenous leukaemia. TB was diagnosed on sputum in 42 (37.1%) TB patients. Very few patients had blood cultures performed for mycobacterial infection.
Microbiological evidence of infection in the BME was found in 53 (58.2%) of 91 cases. The mycobacterial culture was positive in 29 patients (25 with TB and four with MAC) giving a yield of 29 (33.3%) out of 87 with mycobacterial disease on BME and a yield of 11.3% over the whole cohort. Thirteen of the 87 did not have a mycobacterial culture performed. A further 12 patients with a negative BME (including three with TB) did not have a mycobacterial culture. None of the cohort had fungal cultures performed on bone marrow aspirate. Organisms were seen in 31 patients, 16 on histology and 15 on aspirate smear (24 TB, three MAC, four cryptococcus).
Granuloma formation was present in 113 (44.0%) patients (Table 2), of whom 85 (75.2%) had a positive BME and 28 (24.8%) had no specific cause identified. Of patients with infections found on BME, 77 (92.8%) of 83 with TB had granulomata present, while 3/4 each with MAC and cryptococcosis had granulomas identified. Overall, 83 (67.4%) of 123 patients with mycobacterial or cryptococcal infection had granulomata on BME. Twenty (64.5%) patients with organisms seen on BME had granulomata present, while 11 did not. Caseation necrosis was not exclusive to TB, but occurred in one case each of MAC infection and Hodgkin's disease.
|
The clinical, laboratory, and pathological markers of a positive BME, based on univariate analysis, are shown in Table 3
. Predictors identified were a history of wasting, oral thrush, leukopenia and low CD4 count. There were only 32 patients without a predictor and seven (21.9%) of them had a positive BME, compared to 90 (40%) of 225 with at least one predictor (p=0.048). CD4 cell counts were available for 159 (61.9%) patients with a median count of 118/mm3 (range 1–1756/mm3). Granuloma formation occurred in 22 (24.7%) of 89 patients with CD4 >100/mm3 compared to 36 (51.4%) of 70 patients with CD4 
100/mm3 (OR 0.3; 95%CI 0.15–0.62). Of the 48 patients with CD4 <50/mm3, 34 (70.8%) had a diagnosis made (28 (58%) with TB, two (4%) with MAC, four (8%) with PCP, and one each with cryptococcosis, multiple myeloma and Kaposi's sarcoma), and 25 patients had granulomata (15 with caseation necrosis, two well-formed and eight poorly-formed). There were no significant differences between males and females regarding any new diagnosis (57.7% vs. 55.3%; p=0.694) or diagnosis with infection (50.0% vs. 46.1%; p=0.533). Fifty-nine (22.9%) patients died in hospital, 30 (30.9%) of those with a positive BME, compared to 29 (18.1%) with a negative BME (p=0.018).
|
|
Discussion |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
In this series of 257 symptomatic HIV-infected patients, the yield of BME was considerable (38%). This result is comparable to the yield of other published studies of 25–42%,1,2,4–6,8–12 although in some of these other series patients were more carefully selected. In our series a high proportion (24%) had a unique diagnosis on BME, compared to other studies where the diagnosis was unique in 8–10% of patients.3,6,9–10,12 The likely explanations for this are two-fold: firstly, that other invasive investigations were often withheld pending the result of the BME; and secondly, that blood culture for mycobacterial disease by lysis centrifugation technique was not available. Since blood culture for mycobacteria by modern techniques has been shown to be equivalent to BME and bone-marrow culture,3,12 blood culture has become the preferred early investigation, with BME being recommended when other investigations have proved negative, when a patient is critically ill, or in geographic areas where diseases such as histoplasmosis3,10 or leishmaniasis11 are prevalent or suspected. Many of the patient cohort in this study were critically ill, as evidenced by the high (23%) in-hospital mortality in an acute care facility and would therefore have merited BME. Information was lacking on the extent of investigations performed prior to hospital admission.
The spectrum of HIV-related disease in Southern Africa has been considered to be inadequately reported.13 This study provides evidence that South Africa is a tuberculosis-predominant geographical area comparable to other sub-Saharan African countries, since 44% of the patient cohort had active TB diagnosed. MAC infection does occur in South Africa. In patients with CD4 <50/mm3, the 4% of patients with MAC and 58% with TB is similar to that in a Spanish study of patients with pyrexia of unknown origin, most of whom had advanced HIV disease, in whom 48% of the patients had TB and 6% MAC.8 Lymphoma was present in 4% of patients, which is similar to the 6% found in London, UK.9 The patients with acute myelogenous leukaemia and multiple myeloma probably represented coincidental HIV infection in a high seroprevalence area. The indications for BME all proved equally useful, with the exception of peripheral blood thrombocytopenia alone, where the yield was neligible, a result found in similar studies.9
In none of the similar studies reported have there been substantial numbers of women,1,12 but there were in our series, which was performed in an area where heterosexual sex is the main route of acquisition of HIV infection. There were no significant differences in yield of BME or in disease process by gender.
Granulomata are formed as part of a non-specific response to persistent antigen or as a hypersensitivity reaction. Non-immunological or humoral mechanisms are presumably operative in those with low CD4 cell counts who had a significantly higher frequency of granuloma formation, probably associated with higher rates of opportunistic infection. Limitations on this relationship between granuloma formation and CD4 lymphocyte count include the lack of a final diagnosis in 25% of such patients, and the fact that not all subjects had CD4 counts performed. The high number of patients with CD4 <50/mm3 and caseation necrosis was particularly unexpected. This finding may be analogous to people with severe neutropenia who can still form abscesses. The 44% rate of granulomata in this cohort exceeds the 20–30% reported elsewhere.4,7,10,12 Granulomata alert the pathologist and the clinician to the high probability of an opportunistic process ranging from 80 to 100%,4,9,10 and 75% in our study, but they may be insensitive and non-specific for mycobacterial disease.3,12 This is an important issue in a TB-predominant geographical region, as empirical antituberculosis therapy is frequently recommended on the basis of the presence of non-caseating granulomata. This approach would have been microbiologically appropriate in only 24 (41%) of 58 patients with non-caseating granulomata in this study (Table 2
). Follow-up of such patients on empirical treatment may contribute to clarification of this issue.
The highest yield of BME in other studies has been found in patients with extremely advanced HIV disease, fever and cytopenia,7,10,12 which is consistent with our findings (Table 3). These predictors would not have had much impact on the numbers of BME performed, as only 32 subjects had no predictor of a positive BME.
The aim of diagnostic testing for an opportunistic infection is to identify the responsible organism in a timely and cost-effective way, and to use non-invasive tests where possible. Although BME has been shown to be a valuable and frequently unique test in our setting, this was in a hospital without modern blood culture techniques for mycobacteria. Lysis centrifugation blood cultures have been shown to be of value in Tanzania in sub-Saharan Africa, where 20% of febrile HIV seropositive hospitalized patients were culture positive for M. tuberculosis.14 It was not difficult in Tanzania to expand the microbiological capabilities of a university hospital department. Most national TB control programmes in Africa focus on sputum-smear-positive patients, to the detriment of many patients such as those in our study, since these patients would not be definitively diagnosed with TB, and empirical anti-tuberculous therapy is discouraged. The reality of the situation in South Africa is that the majority of HIV-infected patients would not have easy access to BME or blood cultures for mycobacteria, nor to CD4 counts. If we wish to impact positively and significantly on the lives of people living with advanced HIV infection in countries in Southern Africa, then TB blood culture facilities will need to be made widely available at primary-care level for sputum-smear-negative people with unexplained fevers, weight loss and cytopenias. This would both enable a diagnosis of mycobacterial disease to be made before the patients are extremely debilitated, and also limit the necessity for hospital referral for investigation to those in whom cultures are negative and who have persistent symptoms. Cash-strapped national departments of health are unlikely to be able to introduce such testing. It will therefore be up to the international community to provide technical, logistic and financial assistance in order to expand the microbiological capabilities of countries in which the majority of HIV-infected people in the world reside.
|
Notes |
|---|
Address correspondence to Dr A.S. Karstaedt, Department of Medicine, Chris Hani Baragwanath Hospital, P.O. Bertsham, 2013, Johannesburg, South Africa. e-mail: 014alan{at}chiron.wits.ac.za
|
References |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
1. Bishburg E, Eng RH, Smith SM, Kapila R. Yield of bone marrow culture in the diagnosis of infectious diseases in patients with acquired immunodeficiency syndrome. J Clin Microbiol1986; 24:312–14.
2.
Prego V, Glatt AE, Roy V, Thelmo W, Dincsoy H, Raufman JP. Comparative yield of blood culture for fungi and mycobacteria, liver biopsy, and bone marrow biopsy in the diagnosis of fever of undetermined origin in human immunodeficiency virus-infected patients. Arch Intern Med1990; 150:333–6.
3. Northfelt DW, Mayer A, Kaplan LD, Abrams DI, Hadley WK, Yaiko DM, Herndier BG. The usefulness of diagnostic bone marrow examination in patients with human immunodeficiency virus (HIV) infection. J Acquir Immune Defic Syndr1991; 4:659–66.
4. Nichols L, Florentine B, Lewis W, Sattler F, Rarick MU, Brynes RK. Bone marrow examination for the diagnosis of mycobacterial and fungal infections in the acquired immunodeficiency syndrome. Arch Pathol Lab Med1991; 115:1125–32.[Web of Science][Medline]
5. Ciaudo M, Doco-Lecompte T, Guettier C, d'Agay MF, David F, Rional N, Buu-Hoia A, Molina JM, Kazatchkine M, Modai J, Lecompte T. Revisited indications for bone marrow examination in HIV-infected patients. Eur J Haematol1994; 53:168–74.[Web of Science][Medline]
6. Engels E, Marks PW, Kazanjian P. Usefulness of bone marrow examination in the evaluation of unexplained fevers in patients infected with human immunodeficiency virus. Clin Infect Dis1995; 21:427–8.[Web of Science][Medline]
7.
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 Pathol1995; 48:706–9.
8. Lozano F, Torre-Cisneros J, Bascunana A, Polo J, Viciana P, Garcia-Ordonez MA, Hernandez-Quero J, Marquez M, Vergara A, Diez F, Pujol B, Torres-Tortosa M, Pasquan J, Hernandex-Burruezo JJ, Suarez I. Prospective evaluation of fever of unknown origin in patients infected with the human immunodeficiency virus. Eur J Clin Microbiol Infect Dis1996; 15:705–11.[Web of Science][Medline]
9. Brook MG, Ayles H, Harrison C, Rowntree C, Miller RF. Diagnostic utility of bone marrow sampling in HIV positive patients. Genitourin Med1997; 73:117–21.[Web of Science][Medline]
10. Keiser P, Rademacher S, Smith JW. Utility of bone marrow culture and biopsy in the diagnosis of disseminated infections in AIDS. Am J Hematol1997; 56:1–4.[Web of Science][Medline]
11.
Benito N, Nunez A, de Gorgolas M, Esteban J, Calabuig T, del Carmen Rivas M, Fernandez Guerrero ML. Bone marrow biopsy in the diagnosis of fever of unknown origin in patients with acquired immunodeficiency syndrome. Arch Intern Med1997; 157:1577–80.
12. Kilby JM, Marques MB, Jaye DL, Tabereaux PB, Reddy VB, Waites KB. The yield of bone marrow biopsy and culture compared with blood culture in the evaluation of HIV-infected patients for mycobacterial and fungal infections. Am J Med1998; 104:123–8.[Web of Science][Medline]
13. Grant AD, Djomand G, de Cock KM. Natural history and spectrum of disease in adults with HIV/AIDS in Africa. AIDS1997; 11 (Suppl B):S43–54.
14. Archibald LK, den Dulk MO, Pallangyo KJ, Reller LB. Fatal Mycobacterium tuberculosis bloodstream infections in febrile hospitalized adults in Dar es Salaam, Tanzania. Clin Infect Dis1998; 26:290–6.[Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||