Changing Pattern of AIDS A Bone Marrow Study

Transcription

1 Hematopathology / BONE MARROW CHANGES OF AIDS Changing Pattern of AIDS A Bone Marrow Study Xiaohui Zhao, MD, PhD, 1* Nora C.J. Sun, MD, 1 Mallory D. Witt, MD, 2 Margaret Keller, MD, 3 and Yutaka Niihara, MD 4 Key Words: Bone marrow; AIDS; AIDS-related malignant neoplasms; CD4+ cell count; Parvovirus DOI: /K8FU6KW125UTJE7B Abstract We compared bone marrow findings in 2 groups of patients with AIDS during 2 different periods: group 1, n = 20; male/female ratio, 19/1; and group 2, n = 120; male/female ratio, 6/1. Bone marrow iron stores were decreased significantly in group 2 (P <.01), and the incidence of AIDS-related lymphomas was higher, with frequent bone marrow involvement. Two group 1 patients had Kaposi sarcoma, and a 21-month-old girl with transfusion-transmitted AIDS had Burkitt-like lymphoma. In group 2, 44 patients had a history of malignant neoplasms, including Kaposi sarcoma (10 cases), hematologic neoplasms (33 cases), and metastatic leiomyosarcoma (1 case). Of the 120 patients, 15 (12.5%) had bone marrow involvement by malignant neoplasms. The majority of the non-hodgkin lymphomas were high-grade lymphomas. Patients with AIDS-related malignant neoplasms had higher CD4+ cell counts and viral loads than patients without malignant neoplasms (P <.01, P <.05, respectively). The finding of decreased iron stores in patients with AIDS might aid clinical management of their anemia. The increased incidence of malignant neoplasms, especially lymphomas, in recent years might be related to prolonged survival but with incomplete reconstitution of immune function after antiretroviral therapy. The year 2001 marked the 20th anniversary of AIDS. An estimated 60 million people have been infected worldwide, and about 40 million people live with HIV or AIDS. 1 Since the recognition of AIDS as a distinct disorder of the immune system, the epidemiologic, clinical, immunologic, hematologic, and immunopathologic aspects of the disease have been studied widely. 1-5 Clinical manifestations of HIV disease include hematologic abnormalities (isolated anemia, leukopenia or thrombocytopenia, bicytopenia or pancytopenia), opportunistic infections, unexplained fever or a fever of unknown origin (FUO), and, less commonly, certain types of neoplasms. 6-8 Bone marrow examination frequently is undertaken, especially in the evaluation of unexplained or persistent fever, 8 in suspected opportunistic infection with bone marrow involvement, 9 or in staging for malignancy. 10 Findings of bone marrow examination in patients with AIDS have been described in the literature since the 1980s, 11,12 and a spectrum of morphologic characteristics have been described, ranging from normal hematopoiesis to bone marrow dysplasia, lymphocytic infiltration, plasmacytosis, reticulin fibrosis, granulomatous myelitis, and other conditions. 12 The introduction of highly active antiretroviral therapy (HAART) into clinical practice in industrialized countries dramatically changed the course of HIV disease. 13,14 The number of deaths due to AIDS-related opportunistic infection has been reduced by 75% 15,16 since the advent of protease inhibitors and combination antiretroviral therapy in To study the changing pattern of bone marrow involvement after the introduction of HAART, we studied the bone marrow findings in patients from the same institution who had HIV disease during 2 portions of the period. Am J Clin Pathol 2004;121: DOI: /K8FU6KW125UTJE7B 393

2 Zhao et al / BONE MARROW CHANGES OF AIDS Materials and Methods All patients who fulfilled the Centers for Disease Control and Prevention defined criteria for AIDS 17 and underwent bone marrow examination during the period September 1982 through August 1987 (group 1) or January 1996 through May 2001 (group 2) were entered into the study. Indications for bone marrow examinations included FUO, one or more cytopenias, opportunistic infections, and staging for malignant neoplasms. These indications have been consistent in the Division of HIV Medicine (Harbor/UCLA Medical Center, Torrance, CA) for decades. Group 1 included 20 patients, and group 2 included 120 patients. The peripheral blood smear data, CBC count data, bone marrow biopsy specimens, bone marrow aspiration smears, and histories were retrieved. Bone marrow biopsy specimens and aspirates were taken from the posterior iliac crest using a Jamshidi needle, except in pediatric cases in which the anterior tibia was the sampling site of choice. Smears prepared from bone marrow aspirates were stained with the modified Romanowsky method. Biopsy specimens were fixed in B-5 fixative and embedded in paraffin. Sections were cut at 4 µm and stained routinely with H&E, Giemsa, periodic acid Schiff, or the Prussian blue method for iron. Wilder stain was used for reticulin fibers, Fite-Faraco stain for the identification of Mycobacteria (acidfast bacilli [AFB]), and Gomori methenamine silver stain for fungi; immunohistochemical stains also were applied on bone marrow sections whenever needed. The histochemical stains that included Sudan black B, α-naphthol acetate esterase, and α-naphthyl butyrate esterase stains were used on bone marrow aspiration smears when indicated. Immunohistochemical staining was performed by using a streptavidinbiotin-peroxidase complex method (streptavidin-biotinylated horseradish peroxidase, DAKO, Carpinteria, CA) enhanced by microwave oven antigen retrieval. The primary antibodies used included the following: BLA36 (dilution 1:100; DAKO), bcl-1 (cocktail, Novocastra Laboratories, Newcastle upon Tyne, England), bcl-2 (dilution 1:200; DAKO), bcl-6 (dilution 1:20; DAKO), CAM 5.2 (prediluted, Becton Dickinson, San Jose, CA), CD3 (dilution 1:200; DAKO), CD5 (dilution 1:40; DAKO), CD10 (dilution 1:50; Novocastra), CD15 (dilution 1:100; Becton Dickinson), CD20 (dilution 1:800; DAKO), CD21 (dilution 1:20; DAKO), CD30 (dilution 1:40; DAKO), CD43 (dilution 1:1,000; Becton Dickinson), CD45 (leukocyte common antigen; dilution 1:800; DAKO), CD45RO (dilution 1:800; DAKO), CD68 (dilution 1:3,200; DAKO), CD74 (dilution 1:4; ICN Biomedicals, Aurora, OH), CD79a (dilution 1:100; DAKO), CDw75 (dilution 1:4; Immunotech, Marseille, France), epithelial membrane antigen (dilution 1:800; DAKO), hemoglobin A (dilution 1:2,000; DAKO), hemoglobin F (dilution 1:6,400; Perkin Elmer Wallace, Norton, OH), κ light chain (dilution 1:51,200; DAKO), λ light chain (dilution 1:51,200; DAKO), latent membrane protein-1 of Epstein-Barr virus (dilution 1:800; DAKO), lysozyme (dilution 1:16,000; DAKO), myeloperoxidase (dilution 1:6,400; DAKO), parvovirus B19 (prediluted, Cell Marque, Austin, TX), Pneumocystis carinii (dilution 1:200; DAKO), S-100 (dilution 1:4,000; DAKO), Toxoplasma (dilution 1:200; DAKO), Ulex europaeus agglutinin type 1 (UEA-1) (dilution 1:1,600; DAKO), and vimentin (dilution 1:3,200; DAKO). Antibodies were applied as required by the case. Anemia was defined as a hemoglobin concentration of less than 10 g/dl (100 g/l), leukopenia as a WBC count less than 4,000/µL ( /L), and thrombocytopenia as a platelet count less than /µl ( /L). Bone marrow cellularity was evaluated according to the patients age, as described by Hartsock and associates. 18 Plasmacytosis was noted as present when the plasma cells in the marrow constituted 5% or more of all nucleated cells and lymphocytosis when the lymphocytes in the bone marrow equaled or exceeded 20% of all nucleated cells. Nonnumeric values were compared by using the χ 2 test and numeric values by using the Student t test. Results Patient Populations Group 1 included 19 males and 1 female. The female was a 21-month-old girl who contracted HIV infection from a blood transfusion. Group 2 included 103 males and 17 females. Compared with group 1, the patient population in group 2 was significantly different Table 1. The percentage of white patients declined significantly (P <.001), whereas the percentage of Hispanic patients increased significantly (P <.05). The percentage of African American patients showed a slight but insignificant increase. Manifestations Anemia and leukopenia were the most common manifestations in group 1, while FUO and pancytopenia were most common in group 2. Peripheral Blood Cell Counts Hemoglobin concentrations and WBC and platelet counts from the 2 groups are given in Table 1. No statistically significant differences were found between the groups in hemoglobin concentration, WBC count, or platelet count, although group 1 patients tended to have lower WBC counts than group 2 patients. 394 Am J Clin Pathol 2004;121: DOI: /K8FU6KW125UTJE7B

3 Hematopathology / ORIGINAL ARTICLE Bone Marrow Findings The bone marrow findings are summarized in Table 2. Hypercellularity Image 1 was seen in 5 (25%) of 20 patients in group 1 and 42 (35.6%) of 118 patients in group 2. Panhyperplasia frequently was responsible for the increased cellularity. One or more lymphoid aggregates of various sizes were present in 7 (35%) of 20 patients in group 1 and in 26 (22.0%) of 118 patients in group 2. These lymphoid aggregates were composed mainly of small lymphocytes and some plasma cells and were considered reactive. Increased numbers of plasma cells were observed in 7 patients (35%) in group 1 and 44 patients (37.3%) in group 2. Plasma cell cuffing around the blood vessels was found in 7 patients (35%) in group 1 and 53 patients (44.9%) in group 2. Granulomas were observed in 7 (35%) of 20 patients in group 1 and 31 (26.3%) of 118 patients in group 2, although the difference was not statistically significant. In group 1, special stains in 7 cases of granulomatous changes in the bone marrow revealed fungal microorganisms in 2 cases and AFB in 2 cases. In the other 3 cases, no microorganisms were shown by special stains, and bone marrow cultures (bacterial) on these cases were sterile. In group 2, 31 bone marrow specimens revealed granulomas; microorganisms were demonstrated in 15 cases by special stains, including 10 specimens revealing fungi, 4 specimens exhibiting AFB Image 2, and 1 specimen revealing P carinii. Seventeen samples of granulomas showed negative stains. A bone marrow sample from 1 patient had a positive fungal stain but no granuloma was identified. Human parvovirus B19 was studied by immunohistochemical analysis in 43 group 2 cases. Of 43 cases, 5 (12%) had positive results Image 3. Of the 5 cases with positive immunohistochemical stains for parvovirus B19, 2 showed erythroid hypoplasia, and 3 cases with positive immunohistochemical results had normal myeloid/erythroid ratios. Decreased iron stores (<1+) were found in 58 (49.2%) of 118 cases in group 2 and in only 1 (5%) of 20 cases in group 1 (P <.01). Two patients had a history of Kaposi sarcoma (KS), and an abnormal lymphocytic infiltrate was found in the bone marrow specimen in a 21-month-old girl in group 1. At the time of autopsy, dissemination of Burkitt-like lymphoma involved the terminal ileum, stomach, appendix, lymph nodes, spleen, liver, lungs, kidneys, and thymus. No malignancy was detected in the bone marrow samples from the other 2 patients with KS. In group 2, 44 (36.7%) of 120 patients had a history of malignant neoplasm Table 3. Of the 44 patients, 33 had a hematologic malignancy, including 28 with non-hodgkin lymphoma (NHL), 3 with Hodgkin lymphoma (HL), 1 with multiple myeloma, and 1 with acute myeloid leukemia. Table 1 Patient Characteristics and Blood Cell Counts * Group 1 (n = 20) Group 2 (n = 120) P Sex.2572 Male 19 (95) 103 (85.8) Female 1 (5) 17 (14.2) Ratio 19:1 6:1 Age Range 21 mo-50 y y Median Mean Race African American 4 (20) 33 (27.5).4799 White 12 (60) 24 (20.0) <.001 Hispanic 4 (20) 60 (50.0) <.05 Asian 0 (0) 2 (1.7) Samoan 0 (0) 1 (0.8) Hemoglobin concentration, g/dl (g/l).5729 <10.0 (<100) 11 (55) 74 (61.7) 10.0 ( 100) 9 (45) 46 (38.3) WBC count, /µl ( 10 9 /L).0814 <4,000 (<4.0) 15 (75) 65 (54.2) 4,000 ( 4.0) 5 (25) 55 (45.8) Platelet count, 10 3 /µl ( 10 9 /L) (15) 18 (15.0) (85) 102 (85.0) * Data are given as number (percentage) unless otherwise indicated. For a description of the groups, see the Materials and Methods section. Table 2 Bone Marrow Findings * Group 1 (n = 20) Group 2 (n = 118) Cellularity Hypercellularity 5 (25) 42 (35.6).3557 Normocellularity 8 (40) 38 (32.2).4951 Hypocellularity 7 (35) 38 (32.2).8044 Lymphoid aggregate 7 (35) 26 (22.0).2081 Plasmacytosis 7 (35) 44 (37.3).8451 Plasma cell cuffing 7 (35) 53 (44.9).4070 Granuloma 7 (35) 31 (26.3).4200 Positive staining result GMS 2 10 AFB 2 4 PCP 0 1 Negative staining result Iron stores Normal 17 (85) 56 (47.5) <.01 Increased 2 (10) 4 (3.4).1804 Decreased 1 (5) 58 (49.2) <.01 Malignancy involvement 1 (5) 15 (12.5).3278 AFB, acid-fast bacilli; GMS, Gomori methenamine silver; PCP, Pneumocystis carinii pneumonia. * Data are given as number (percentage) except for data for the granuloma subheadings, which are numbers of cases. For a description of the groups, see the Materials and Methods section. By immunohistochemical stain. Iron store was quantified from 0 to 4+ and was judged as decreased (<2), normal (2+), and increased (>2+) according to Beutler et al (Ann Intern Med. 1958;48:60-82). Group 2 was divided further into 2 subgroups based on bone marrow involvement by malignant neoplasm. Of 120 patients, 15 (12.5%) had a malignant neoplasm involving the bone marrow; this group consisted entirely of men, with a median age of 34 years (range, years). Of these 15 P Am J Clin Pathol 2004;121: DOI: /K8FU6KW125UTJE7B 395

4 Zhao et al / BONE MARROW CHANGES OF AIDS Image 1 Hypercellular bone marrow with megakaryocytic hyperplasia and dysplasia such as micromegakaryocytes (inset, H&E, 1,000) and plasmacytosis in a patient with AIDS (H&E, 400). A Image 2 Granuloma in the bone marrow with a positive stain for acid-fast bacilli (H&E, 400; inset, Fite-Faraco stain, 1,000). Image 3 A, Bone marrow samples from a patient with parvovirus infection shows erythroid precursors containing intranuclear inclusions (arrows) (H&E, 1,000; inset, aspirate smear, Romanowsky stain, 1,000). B, Immunohistochemical stain for parvovirus B19 (mouse, R92F6) on the same patient s bone marrow section displaying positive intranuclear stains in many erythroid precursors (anti parvovirus B19, 400). B patients, 11 had NHL, 2 had HL, 1 had multiple myeloma, and 1 had acute myeloid leukemia. Of 31 patients with AIDS-related lymphoma (ARL, ie, NHL and HL), 13 (42%) had bone marrow involvement Image 4 and Image 5. In the subgroup with bone marrow involvement, high-grade lymphomas (8/11 [73%]) were most common, with Burkitt or Burkitt-like being the most common subtypes, whereas in the group without marrow involvement, high-grade lymphomas constituted 7 (41%) of 17 cases (Table 3). In the uninvolved group, 1 primary central nervous system lymphoma, 1 primary effusion lymphoma, 10 cases of KS, and 1 leiomyosarcoma metastatic to the liver were found. It is interesting to note that, taken together, more patients had diffuse large B-cell lymphoma (13 patients) than Burkitt or Burkitt-like lymphoma (8 patients). A comparison of the subgroups of bone marrow specimens with and without involvement by malignant neoplasms in group 2 revealed no significant differences in 396 Am J Clin Pathol 2004;121: DOI: /K8FU6KW125UTJE7B

5 Hematopathology / ORIGINAL ARTICLE Table 3 Histopathologic Findings for Patients With a History of Malignant Neoplasms * Bone Marrow Involvement Group 1 (n = 3) Group 2 (n = 44) Present (n = 1) Absent (n = 2) Present (n = 15) Absent (n = 29) Non-Hodgkin lymphoma High-grade Burkitt Immunoblastic T-cell NOS Intermediate-grade DLBCL Hodgkin lymphoma Multiple myeloma Acute myeloid leukemia Kaposi sarcoma Leiomyosarcoma metastatic to liver DLBCL, diffuse large B-cell lymphoma; NOS, not otherwise specified. * For a description of the groups, see the Materials and Methods section. Image 4 Burkitt lymphoma involving the bone marrow of a patient with AIDS (H&E, 630). age, sex, and race (data not shown). CD4+ cell counts and HIV RNA copies per milliliter were compared between patients with and without malignant neoplasms, between patients with and without bone marrow involvement by malignant neoplasms, and between patients with and without granulomas involving the bone marrow Table 4. The CD4+ cell counts were significantly higher in patients with malignant neoplasms and in patients with bone marrow involvement by malignant neoplasms. In group 2 patients with malignant neoplasms, the NHL and hematopoietic Image 5 Classic Hodgkin lymphoma involving the bone marrow of a patient with AIDS (H&E, 630). neoplasm subgroups had significantly higher CD4+ cell counts than did the subgroups without malignant involvement. There was no significant difference in CD4+ cell counts for the KS subgroups. However, when comparing the NHL and hematopoietic neoplasm subgroups with the KS group, the CD4+ cell count in the KS group was significantly lower (P <.05; data not shown). In addition, the CD4+ cell count was substantially lower in patients with granulomas than in patients without granulomas. In contrast, the HIV RNA level was significantly lower in Am J Clin Pathol 2004;121: DOI: /K8FU6KW125UTJE7B 397

6 Zhao et al / BONE MARROW CHANGES OF AIDS Table 4 Comparison of CD4+ Cell Count and Viral Load for Subgroups in Group 2 * CD4+ Cell Count, /µl ( 10 6 /L) Viral Load, HIV RNA Copies per Milliliter With Without P With Without P Malignant neoplasm n = 40 n = 74 n = 24 n = 48 Median 81.5 (82) 21.5 (22) < , ,411 <.05 Range (4-319) (0-252) , ,680 Non-Hodgkin lymphoma n = 25 n = 89 n = 15 n = 57 Median 99.5 (100) 24 (24) <.01 51,655 35, Range (11-319) (0-252) , ,077 Hematopoietic neoplasm n = 30 n = 84 n = 18 n = 53 Median 94.5 (95) 21 < , , Range (11-319) (0-252) , ,077 Kaposi sarcoma n = 9 n = 105 n = 5 n = 67 Median 19 (19) 34 (34) ,754 35,889 <.05 Range (5-104) (0-319) , ,680 Bone marrow involvement n = 13 n = 27 n = 6 n = 18 Median 98 (98) 63 (63) < , , Range (32-319) (4-179) 2, , ,077 Granuloma n = 32 n = 84 n = 17 n = 55 Median 18 (18) 44 (44) <.01 38,043 40, Range (0-101) (2-319) , ,077 NS, not significant. * For a description of group 2, see the Materials and Methods section. With and Without refer to the categories listed in column 1, eg, Malignant neoplasm. P <.05 between Kaposi sarcoma and hematopoietic neoplasms and/or non-hodgkin lymphoma subgroups. patients without malignant neoplasms or KS in comparison with patients with malignant neoplasms or KS. Discussion Hematologic abnormalities and FUO are common in patients with HIV infection, often prompting examination of the bone marrow. We and others have found a spectrum of morphologic characteristics, ranging from normal hematopoiesis to marrow dysplasia, lymphocytic infiltration, plasmacytosis, reticulin fibrosis, granulomatous myelitis, and bone marrow involvement by malignant neoplasms. 9,10,12 During the past 10 years, the introduction of HAART and prophylactic medications for opportunistic infections have led to dramatic reductions in AIDS-associated deaths in industrialized countries. 13,15 Therefore, a comparison of bone marrow findings in HIV-infected patients during the period of the early AIDS epidemic and currently in the same institution might provide corroboration of the changing spectrum of disease affecting patients with HIV. Changes in Demographics We noted a relative but significant decrease in the white population and an increase in Hispanic population in group 2 compared with group 1 (Table 1). A definite but statistically insignificant increase in HIV-positive women also was observed in group 2. Similar trends were found in another institution in the Los Angeles area. 19 Changes in Blood Values The degree and incidence of anemia and thrombocytopenia among patients with HIV infection in group 2 did not change significantly compared with group 1. Leukopenia, however, was less common and less severe in group 2, although the difference in the 2 groups was not significant. The introduction of cytokines in recent years, such as granulocyte colony-stimulating factor for patients with AIDS, might have altered these results. 20 Changes in Bone Marrow Findings Granuloma Granulomatous inflammatory responses in patients with AIDS have been a subject of great interest because of the loss of primary function of cell-mediated immunity in patients with AIDS. In group 1, 7 (35%) of 20 bone marrow specimens showed granulomas with stainable fungi in 2, stainable AFB in 2, and negative stains in 3. In 1 of the negatively stained cases, the patient had a history of pulmonary tuberculosis and had been treated with antituberculosis drugs. The results of bone marrow cultures in all 3 patients showed no bacterial growth. In group 2, only 26.3% of patients had granulomas in their bone marrow, which was not a statistically significant difference compared with group 1. However, bone marrow involvement by P carinii and nongranulomatous infection such as human parvovirus B19 occurred in rare cases in group 2. A Spanish study of 1,312 AIDS bone marrow specimens revealed the most common infection in bone marrow 398 Am J Clin Pathol 2004;121: DOI: /K8FU6KW125UTJE7B

7 Hematopathology / ORIGINAL ARTICLE to be mycobacterial infection (58%) with fungal infection constituting 6% of infections. 8 In our studies of 2 different periods, mycobacterial and fungal infections were equally common in group 1, but fungal infections were the predominant cause of granulomas in group 2 (Table 2). The reasons for the discrepancy likely are multiple and include the following: (1) There might have been differences in the prophylactic regimen for these infections in the Spanish study and ours. (2) AFB usually cannot be demonstrated on histologic sections once the patient has been treated with antimycobacterial therapy. 9 (3) Patients with advanced HIV disease might be unable to form granulomas in response to mycobacterial infection of the bone marrow. 11 (4) The mechanisms of granuloma formation are complex; for example, hypersensitivity might contribute to the maturation and augmentation of granuloma formation. 21 In the present study, a significantly lower CD4+ cell count was observed in patients with granulomas than in patients without granulomas (Table 4). This finding indicates that patients with and without granuloma might have different immune profiles. Parvovirus Infection Up to 80% of HIV-infected persons are anemic during the course of their disease, and the percentage increases with advancing HIV infection. 7 There have been reports on the possible role of human parvovirus B19 infection in such conditions. 22,23 Human parvovirus has been studied using serum samples from patients with AIDS to detect antibodies and nucleic acid by polymerase chain reaction. 22 Immunohistochemical analysis and in situ hybridization also have been applied to tissue samples to detect viral antigen and DNA, respectively. 23 Owing to a high incidence (30%-60%) of parvovirus IgG positivity in the healthy adult population 22 and the inadequate or absent antibody response in patients with AIDS, 24 the serologic study of IgM and IgG might not have diagnostic significance in these patients. In the present study, 5 (12%) of 43 patients showed positive results for parvovirus B19 by immunohistochemical analysis. However, Liu and associates 23 failed to identify any viral antigens in the bone marrow specimens of 73 patients with AIDS; they found nucleic acid of parvovirus B19 in 7 (10%) of 73 patients by an in situ hybridization method. 23 The detection of parvovirus DNA by polymerase chain reaction in serum samples and by in situ hybridization of bone marrow samples might be superior to serologic study and immunohistochemical analysis for the study of parvovirus B19 infection in patients with AIDS. Iron Stores The most reliable early detection of iron deficiency is the study of bone marrow samples for iron stores. Increased and decreased iron stores in patients with AIDS have been reported. 11,25 Iron stores were normal in the majority of patients in group 1, but approximately half (49.2%) of the patients in group 2 had decreased iron stores. The significant changes in iron stores between these 2 groups raise a practical question about the presence of iron deficiency in patients with AIDS. It is possible that the demographic change might have contributed to the difference. The predominantly white male patients in group 1 might be more affluent and were less likely to have nutritional anemia and iron deficiency anemia. The anemia of patients with AIDS might be due to a combination of factors, such as chronic infections, neoplasms, bleeding, and iron deficiency. Malignancy Involvement Bone marrow involvement has been reported in approximately 20% to 30% of patients with AIDS-related NHL at initial diagnosis and in approximately 20% of HIV-negative patients with diffuse large-cell NHL. 26 A study of 291 cases of ARL by Seneviratne et al 10 revealed 19% bone marrow involvement. In the present study, the incidence of bone marrow involvement in ARL was 1 of 1 in group 1 and 13 (42%) of 31 in group 2 (Table 3). The higher incidence of bone marrow involvement in ARL in our study might be because we identified all AIDS cases from the bone marrow registry instead of from HIV services. Among the 11 cases of bone marrow involvement by NHL in group 2, 8 (73%) with high-grade NHLs had bone marrow involvement, which is comparable to the findings of 78% by Seneviratne et al. 10 An interesting observation is that 2 of 3 cases of HL in our study had bone marrow involvement (Table 3), which suggests that patients with AIDSassociated HL usually have advanced disease at the time of diagnosis. One study of 362 cases of HL showed 9.9% bone marrow involvement in patients with non HIV-related disease. 27 Although HL does not represent an AIDS-defining condition, recent evidence consistently indicates that HIVinfected people have a significantly increased risk of developing HL. 28 HIV-related HL is characterized by the preponderance of aggressive histologic subtypes, advanced stage at diagnosis, and malignant clinical course. 28 To study whether bone marrow involvement by ARL is related to immune function, we analyzed the CD4+ cell count and viral load in group 2 cases. The CD4+ cell counts were significantly higher in patients with a malignant neoplasm or bone marrow involvement than in patients without a malignant neoplasm or bone marrow involvement (Table 4); however, the CD4+ cell counts in patients with granulomas were significantly lower than in patients without granulomas. On the other hand, HIV RNA values were significantly higher in patients with a malignant neoplasm or with KS than in patients without these conditions. Patients Am J Clin Pathol 2004;121: DOI: /K8FU6KW125UTJE7B 399

8 Zhao et al / BONE MARROW CHANGES OF AIDS with high viral loads possibly are at higher risk for developing opportunistic infection, including KS caused by human herpesvirus 8. A higher viral load might indicate increased apoptosis of virally infected cells or might reflect deficient immune status. The CD4+ cell count alone might not be a good reflection of T-cell immune function. Also, immune mechanisms other than those associated with helper T-cell functions might be impaired by a high viral burden. Combination therapies with protease inhibitors and nucleoside and nonnucleoside reverse transcriptase inhibitors (HAART) dramatically suppress plasma HIV-1 RNA levels and delay progression to AIDS, but the impact on HIV-related malignant neoplasms remains to be established. AIDS-related KS and NHL are the 2 most common malignant neoplasms among patients with AIDS and affected 20% to 40% of these patients during the pre-haart era. 29 Observational and timetrend data indicate that the incidence of KS and primary brain lymphoma have decreased, but the same study suggests that current therapies have not had a proportionate effect on systemic NHL. 30 It has been postulated that patients treated with HAART might survive longer with continued B-cell stimulation and dysregulated T-cell function resulting in an increased incidence of lymphoma over time. Grulich 31 concluded that the incidence of AIDS-related NHL has decreased with the use of HAART, but the magnitude of the decrease seems to be less than that for AIDS-associated opportunistic infections and KS. Tam and colleagues 32 demonstrated that the use of HAART prolongs overall survival among HIV-infected men diagnosed with KS and NHL. HAART seems to be effective in improving survival even when initiated after the diagnosis of NHL and KS. 32 Although a majority of our patients in group 2 were receiving HAART, a more careful and thorough study is needed to evaluate the impact of HAART on ARL. Conclusions The present report studied the changes in the bone marrow of 2 groups of patients with HIV infection at 2 different times during the period. An increase in the number of women and Hispanic patients in group 2 was noted. Significantly decreased iron stores were found in group 2, suggesting that iron deficiency might be an important cause of anemia in the setting of AIDS in recent years. A higher incidence of AIDS-related malignant neoplasms, especially intermediate- and high-grade NHLs, is an important finding of this study. Interestingly, patients with AIDS-related malignant neoplasms had higher CD4+ cell counts and higher viral loads than those without malignant neoplasms. However, a comparison of the NHL and hematopoietic neoplasm subgroups with the KS group revealed a significantly lower CD4+ cell count in the KS group. These findings demonstrate that different immune profiles are present in patients with AIDS with or without AIDS-related malignant neoplasms and in patients with AIDS-related KS. Whether these different immune profiles are involved in the development of the ARL and whether HAART causes an incomplete immune reconstitute in ARL are unknown. Further studies are needed to elucidate the impact of HAART on ARL. From the 1 Department of Pathology, 2 Department of Medicine, Division of HIV Medicine, 3 Department of Pediatrics, and 4 Department of Medicine, Division of Hematology, Harbor/UCLA Medical Center, Torrance, CA. Address reprint requests to Dr Sun: Dept of Pathology, Harbor/UCLA Medical Center, 1000 W Carson St, Box 12, Torrance, CA * Dr Zhao is now at UCI Medical Center, Orange, CA. References 1. Stover J, Walker N, Garnett GP, et al. Can we reverse the HIV/AIDS pandemic with an expanded response? Lancet. 2002;360: Yeni PG, Hammer SM, Carpenter CCJ, et al. Antiretroviral treatment for adult HIV infection in JAMA. 2002;288: Tirelli U, Bernardi D. Impact of HAART on the clinical management of AIDS-related cancer. Eur J Cancer. 2001;37: Bain BJ. The haematological features of HIV infection. Br J Haematol. 1997;99: Moses A, Nelson J, Bagby GC Jr. The influence of human immunodeficiency virus-1 on hematopoiesis. Blood. 1998;91: Brizzi MF, Porcu P, Porteri A, et al. Haematologic abnormalities in the acquired immunodeficiency syndrome. Haematologica. 1990;75: Zon LI, Arkin C, Groopman JE. Haematologic manifestations of the human immune deficiency virus (HIV). Br J Haematol. 1987;66: Mayo J, Collazos J, Martinez E. Fever of unknown origin in the HIV-infected patient: new scenario for an old problem. Scand J Infect Dis. 1997;29: Akpek G, Lee SM, Gagnon DR, et al. Bone marrow aspiration, biopsy, and culture in the evaluation of HIVinfected patients for invasive mycobacteria and Histoplasma infections. Am J Hematol. 2001;67: Seneviratne L, Espina BM, Nathwani BN, et al. Clinical, immunologic, and pathologic correlations of bone marrow involvement in 291 patients with acquired immunodeficiency syndrome related lymphoma. Blood. 2001;98: Castella A, Croxson TS, Mildvan D, et al. The bone marrow in AIDS. Am J Clin Pathol. 1985;84: Sun NCJ, Shapshak P, Lachant NA, et al. Bone marrow examination in patients with AIDS and AIDS-related complex (ARC). Am J Clin Pathol. 1989;92: Hammer SM, Squires KE, Hughes MD, et al. A controlled trial of two nucleoside analogues plus indinavir in persons with human immunodeficiency virus infection and CD4 cell counts of 200 per cubic millimeter or less. N Engl J Med. 1997;337: Am J Clin Pathol 2004;121: DOI: /K8FU6KW125UTJE7B

9 Hematopathology / ORIGINAL ARTICLE 14. Hogg RS, Heath KV, Yip B, et al. Improved survival among HIV-infected individuals following initiation of antiretroviral therapy. JAMA. 1998;279: Palella FJ, Delaney KM, Moorman AC, et al. Declining morbidity and mortality among patients with advanced human immunodeficiency virus infection. N Engl J Med. 1998;338: Scadden DT. AIDS-related malignancies. Curr Opin Oncol. 1998;10: Hollander H, Katz MH. HIV infection. In: Tierney LM, McPhee SJ, Papadakis MA, eds. Current Medical Diagnosis and Treatment. 38th ed. London, England: Appleton & Lange; 1999: Hartsock RJ, Smith EB, Petty CS. Normal variations with aging of the amount of hematopoietic tissue in bone marrow from the anterior iliac crest: a study made from 177 cases of sudden death examined by necropsy. Am J Clin Pathol. 1965;43: Levine AM, Seneviratne L, Espina BM, et al. Evolving characteristics of AIDS-related lymphoma. Blood. 2000;96: Kuritzkes DR. Neutropenia, neutrophil dysfunction, and bacterial infection in patients with human immunodeficiency virus disease: the role of granulocyte colony-stimulating factor. Clin Infect Dis. 2000;30: Jagadha V, Andavolu RH, Huang CT. Granulomatous inflammation in the acquired immune deficiency syndrome. Am J Clin Pathol. 1985;84: van Elsacker-Niele AMW, Kroon FP, van der Ende ME, et al. Prevalence of parvovirus B19 infection in patients infected with human immunodeficiency virus. Clin Infect Dis. 1996;23: Liu W, Ittmann M, Liu J, et al. Human parvovirus B19 in bone marrows from adults with acquired immunodeficiency syndrome: a comparative study using in situ hybridization and immunohistochemistry. Hum Pathol. 1997;28: Frickhofen N, Young NS. Persistent parvovirus B19 infections in humans. Microb Pathog. 1989;7: Shenoy CM, Lin JH. Bone marrow findings in acquired immunodeficiency syndrome (AIDS). Am J Med Sci. 1986;292: Levine AM, Sullivan-Halley J, Gill PS, et al. HIV-related lymphoma: prognostic factors predictive of survival. Cancer. 1991;68: Wang J, Weiss LA, Chang KL, et al. Diagnostic utility of bilateral bone marrow examination. Cancer. 2002;94: Dolcetti R, Boiocchi M, Gloghini A, et al. Pathogenetic and histogenetic features of HIV-associated Hodgkin s disease. Eur J Cancer. 2001;37: Krown SE. AIDS-associated malignancies. Cancer Chemother Biol Response Modif. 1996;16: Rabkin CS. AIDS and cancer in the era of highly active antiretroviral therapy (HAART). Eur J Cancer. 2001;37: Grulich AE. AIDS-associated non-hodgkin s lymphoma in the era of highly active antiretroviral therapy. J Acquir Immune Defic Syndr. 1999;21(suppl 1):S27-S Tam HK, Zhang ZF, Jacobson LP, et al. Effect of highly active antiretroviral therapy on survival among HIV-infected men with Kaposi sarcoma or non-hodgkin lymphoma. Int J Cancer. 2002;98: Am J Clin Pathol 2004;121: DOI: /K8FU6KW125UTJE7B 401