A Low Incidence of Nontuberculous Mycobacterial Infections in Pediatric Hematopoietic Stem Cell Transplantation Recipients

Biology of Blood and Marrow Transplantation 12:1188-1197 (2006) 䊚 2006 American Society for Blood and Marrow Transplantation 1083-8791/06/1211-0001$32.00/0 doi:10.1016/j.bbmt.2006.07.006

A Low Incidence of Nontuberculous Mycobacterial Infections in Pediatric Hematopoietic Stem Cell Transplantation Recipients Elif Unal,1 Catherine Yen,1 Lisa Saiman,1 Diane George,1 Phyllis Della-Latta,2 Carmella van de Ven,1 Erin Morris,1 M. Brigid Bradley,1 Gustavo Del Toro,1 James Garvin,1 Monica Bhatia,1 Joseph Schwartz,2 Prakash Satwani,1 Elizabeth Roman,1 Erin Cooney,1 Karen Wolownik,1 Ria Hawks,1 Sandra Foley,1 Mitchell S. Cairo1,2,3 Departments of 1Pediatrics, 2Pathology, and 3Medicine, Columbia University, New York, New York Correspondence and reprint requests: Mitchell S. Cairo, MD, Professor of Pediatrics, Medicine and Pathology, Chief, Division of Pediatric Hematology and Blood and Marrow Transplantation, Morgan Stanley Children’s Hospital of New York-Presbyterian, Columbia University, 180 Fort Washington, HP-506, New York, NY 10032 (e-mail: [email protected]). Received April 25, 2006; accepted July 10, 2006 Presented in part at the Société Internationale d’Oncologie Pédiatrique (SIOP) 2005 Annual Meeting; Vancouver, Canada; September 2005.

ABSTRACT Hematopoietic stem cell transplantation (HSCT) is being used to treat a wide spectrum of clinical disorders but opportunistic infection remains an important factor determining outcomes for these patients. Nontuberculous mycobacterial (NTM) infections are being reported more frequently in HSCT recipients and the incidence of NTM infections in adult recipients is reported to be 0.4%-4.9%. However, the incidence and severity of NTM infections are less well described in pediatric HSCT recipients. Centers for Disease Control and Prevention guidelines were used to define definite and probable NTM infection among 132 children undergoing 169 HSCT between January 2000 and December 2004 at our institution. NTM infection was diagnosed in 5 of 132 pediatric recipients (3.8%). There were no NTM infections diagnosed in the autologous HSCT recipients and the incidence of NTM in allogeneic HSCT recipients was 6.4% (95% confidence interval, 0.8-11.9). The mean age of the HSCT recipients who developed NTM infections was 8 years (range, 2-19 years); 3 were male and 2 were female. Four conditioning regimens included alemtuzumab and 3 had antithymocyte globulin. Of the 5 patients with NTM infections, 2 met the criteria for definite infection and 3 for probable infection. Of the 2 patients with definite NTM infection, 1 had disseminated disease with Mycobacterium avium complex and the other had Mycobacterium chelonae catheter-related bloodstream infection. The probable NTM infections were 1 skin infection with Mycobacterium kansasii and 2 lower respiratory tract infections with M avium complex. Median time to NTM infection was 115 days (range, 14-269 days) after HSCT. Two patients had graft-versus-host disease at the time of NTM infection. All 5 patients received 3-4 antimycobacterial drugs and all NTM infections resolved. In summary, the incidence of NTM infection in pediatric HSCT recipients appears similar to that described in adult HSCT recipients and the outcome appears to be excellent with the proper antibiotic therapy. The increased use of anti-T cell antibodies appears to be associated with an increased risk of NTM infections in pediatric HSCT recipients. Multicenter studies are needed to identify the risk factors, early diagnostic criteria, and optimal therapy. © 2006 American Society for Blood and Marrow Transplantation

KEY WORDS Atypical mycobacteria zumab 1188

●

Nontuberculous mycobacteria

●

Transplant

●

Pediatric

●

Alemtu-

1189

NTM Infections in Pediatric Hematopoietic Stem Cell Transplantation

INTRODUCTION Hematopoietic stem cell transplantation (HSCT) offers a potential cure for an increasingly wide spectrum of malignant and nonmalignant disorders but opportunistic infections remain an important limiting factor in determining outcome. During the past few years, in addition to bacterial, viral, and fungal infections in HSCT recipients, nontuberculous mycobacterial (NTM) infections are being reported more frequently [1-4]. Nontuberculous mycobacteria are ubiquitous pathogens that can cause a broad spectrum of diseases in immunocompetent and immunocompromised hosts. Approximately 50 different mycobacterial species are currently considered etiological agents of human disease [5]. Nontuberculous mycobacteria have been reported to cause pneumonia, lymphadenitis, skin and soft tissue disease, skeletal infections, foreign body and central venous catheter infections, and disseminated infection [6]. Although contact with environmental mycobacteria is regular and frequent, overt disease is uncommon because of the low virulence of NTM organisms [6]. Protective immunity against infection with mycobacteria seems to depend primarily on T cells, which are the main producers of interferon-␥, together with natural killer cells in response to interleukin-12, which is produced by dendritic cells and macrophages upon mycobacterial invasion [7]. NTM disease has been reported more commonly in immunocompromised hosts including patients with acquired immunodeficiency syndrome, primary immunodeficiencies, and after chemotherapy or bone marrow and solid organ transplantation [1-4,8-17]. In a review of NTM infections in immunocompromised hosts, Doucette et al [13] reported incidences of NTM infection among renal, heart, and lung transplant recipients to be 0.16%-0.38%, 0.24%2.8%, and 0.46%-2.3%, respectively. In the same review, the incidence of NTM infections among HSCT recipients was reported as 0.4%-4.9%. Most previous reports of NTM in HSCT recipients have been described in adult recipients and the highest incidence of NTM infection among adult HSCT recipients was reported by Weinstock et al [2] who described 65% of patients receiving T cell-depleted HSCT. HSCT recipients have severely impaired cell-mediated immunity because of their underlying disease, pretransplantation chemotherapy and radiation, conditioning regimen, graft-versus-host disease (GVHD), and GVHD prophylaxis and treatment. T cell-depleted HSCT is not being used in our institution. With the increased use of unrelated cord blood transplantation with delayed engraftment [18] and increased use of alemtuzumab and antithymocyte globulin (ATG) in the conditioning regimens, we hypothesized that the

incidence of NTM infections in our pediatric population would be higher than reported by Weinstock et al. To test our hypothesis we investigated the incidence of NTM infections and associated risk factors in a large cohort of pediatric HSCT recipients treated between January 2000 and December 2004 at Morgan Stanley Children’s Hospital of New York-Presbyterian at Columbia University Medical Center.

METHODS Patients

Between January 2000 and December 2004, 132 children underwent 169 HSCT at Morgan Stanley Children’s Hospital of New York-Presbyterian at Columbia University Medical Center. One hundred twenty-nine of these HSCT patients had malignant disease and 40 did not. Patient characteristics are listed in Table 1. All patients were enrolled in protocols approved by the institutional review board in compliance with the Declaration of Helsinki. Allogeneic HSCT recipients were defined as having highrisk disease with ⱖ1 of the following criteria: leukemia in relapse, leukemia with induction failure, leukemia in third complete remission or beyond, malignant progressive disease, and/or second allogeneic HSCT. Human Leukocyte Antigen Typing and Allogeneic Donors

Human leukocyte antigen (HLA) typing for patients receiving stem cells from matched family donors or umbilical cord blood (UCB) donors was performed by serology for HLA-A and -B and by high-resolution DNA typing for DRB1. Patients receiving unrelated adult donor peripheral blood stem cells (PBSCs) also had a 6/6 match with typing at class I (HLA-A and -B) by serology and class II (HLA-DRB1) at high-resolution DNA typing. Matched family donors were required to be a 6/6 or 5/6 HLA match, and unrelated UCB donor units had to be a 6/6, 5/6 or 4/6 HLA match. UCB donor units were also required to have a minimum of 1.5 ⫻ 107 nucleated cells of cryopreserved UCB unit/kg of recipient. Unrelated adult donors were required to have a minimum of 5 ⫻ 106 CD34 cells/kg and to be an 8/10, 9/10, or 10/10 HLA match (A, B, C, DRB1, DQB1). For autologous HSCT, PBSCs were collected using apheresis. Endpoint for successful PBSC harvesting was a collection of ⱖ5 ⫻ 106 CD34⫹ cells/kg. For neuroblastoma patients, in case of detectable tumor cells in the product by immunocytology, CD34⫹ cells were selected using the Isolex 300i system (Baxter Immunotherapy, Round Lake, Ill) or the CliniMacs system (Miltenyi Biotec, Auburn, Calif).

1190

E. Unal et al.

Conditioning Regimens

Table 1. Patient Demographics (n ⫽ 132) Age (y), median (range) Sex, n (%) Male Female Diagnosis, n (%) Malignant ALL (3 CR1, 10 CR2, 4 CR3, 1 PD, 1 IF) AML (7 CR1, 4 CR2, 1 CR3, 4 PD, 1 IF) CML CP HD (1 CR1, 6 CR2, 1 PR, 1 PR1, 2 PR2, 2 SD, 3 PD) NHL (1 CR1, 1 CR2, 3 PR, 1 PR1, 3 SD, 1 PD) Neuroblastoma (6 CR1, 2 CR, 13 PR, 4 PD) Rhabdomyosarcoma (CR2) Wilms tumor (5 CR2, 2 PR1, 3 PD, 1 SD) Brain tumor Medulloblastoma (CR1) Ependymoma (5 SD, 2 PR) Choroid plexus (1 PR, 3 SD) Pineoblastoma (CR1) Others Nonmalignant Aplastic anemia Fanconi anemia HLH/LCH Sickle cell disease ␤-Thalassemia WAS SCID Hurler syndrome MDS Risk, n (%) Average (all nonmalignant) High Transplant type, n (%) Autologous Allogeneic Transplant source, n (%) Family donor 6/6 HLA match 5/6 HLA match Unrelated Cord blood 4/6 HLA match 5/6 HLA match 6/6 HLA match PB/BM 8/10 HLA match 9/10 HLA match 10/10 HLA match Conditioning regimen, n (%) Myeloablative Reduced intensity/nonmyeloablative GVHD prophylaxis, n (%) Tacrolimus/MMF CSA/steroids CSA/MTX CSA/MTX/steroids

5.25 (0.25-21.75) 78 (59) 54 (41) 129 (76.3) 19 17 6 16 9 25 2 11 2 7 4 3 8 40 (23.7) 11 2 7 3 5 2 2 2 6 112 (66.3) 57 (33.7) 64 (37.9) 105 (62.1) 42 (40) 32 10 63 (60) 59 34 18 7 4 1 2 1 122 (72.2) 47 (27.8) 94 5 5 1

(89.5) (4.8) (4.8) (0.9)

ALL indicates acute lymphoblastic leukemia; AML, acute myelogenous leukemia; CML, chronic myelogenous leukemia; CP, chronic phase; HD, Hodgkin lymphoma; NHL, non-Hodgkin lymphoma; CR, complete remission; CR1, first complete remission; CR2, second complete remission; CR3, third complete remission; PR, partial remission; PR1, first partial remission; PR2, second partial remission; PR3, third partial remission; PD, progressive disease; IF, induction failure; SD, stable disease; HLH, hemophagocytic lymphohistiocytosis; LCH, Langerhans cell histiocytosis; WAS, Wiskott-Aldrich syndrome; SCID, severe combined immunodeficiency; MDS, myelodysplastic syndrome; HLA, human leukocyte antigen; PB, peripheral blood; BM, bone marrow; GVHD, graftversus-host disease; MMF, mycophenolate mofetil; CSA, cyclosporine; MTX, methotrexate.

Conditioning regimens were myeloablative in 72.2% of patients, including all 64 autologous HSCT recipients, and nonmyeloablative (reduced intensity ⫹ allogeneic HSCT) in 27.8%. Alemtuzumab-containing conditioning regimens were used in 9.5% of recipients. GVHD Prophylaxis

Tacrolimus and mycophenolate mofetil (MMF) were administered as GVHD prophylaxis in 89.5% of allogeneic HSCT recipients, as we previously described [19]. Four patients who received matched unrelated donor HSCT received methotrexate (15 mg/m2 on day ⫹1, 10 mg/m2 on days ⫹3, ⫹6, and ⫹11) in addition to this regimen. Tacrolimus was administered intravenously (IV) at 0.03 mg/kg per day as a continuous infusion or orally at 0.12 mg/kg per day in 2 divided doses starting on day ⫺1 or on the first day of conditioning (protocol dependent). As we previously described, tacrolimus dosing was adjusted to maintain tacrolimus steady-state or trough concentrations of 5-20 ng/mL (whole-blood enzyme-linked immunosorbent assay) [19]. MMF was administered daily starting on day ⫹1 at a dose of 15 mg/kg per dose IV or orally twice daily, as we previously described [19]. Beginning in January 2002, MMF doses were adjusted to maintain mycophenolic acid trough concentrations within a reference range of 1-3.5 ␮g/mL. MMF was tapered or discontinued as per individual protocol, but not before day ⫹28. Other GVHD prophylaxis regimens included cyclosporine/methylprednisolone (n ⫽ 5, 4.8%), cyclosporine/methotrexate (n ⫽ 5, 4.8%), and cyclosporine/ methylprednisolone/methotrexate (n ⫽ 1, 0.9%). GVHD was graded according to the Glucksberg and Seattle consensus criteria [20,21]. The “rule of 9” or burn chart was used to estimate extent of skin rash. If a clinical diagnosis of acute GVHD (aGVHD) or chronic GVHD (cGVHD) was made, histologic confirmation was obtained whenever possible. Supportive Care

All patients were hospitalized in protective isolation, defined as single hospital rooms with high-efficiency particulate air filtration system and reverse isolation requiring strict hand degerming and mask use by staff for unrelated HSCT patients. Beginning in January 2001, liposomal amphotericin B (3 mg/kg per day IV over 2 hours) was administered to all patients who received allogeneic HSCT starting on day 0 until day ⫹100, as we previously described [22]. Fluconazole was used as antifungal prophylaxis for the remaining patients. Additional anti-infective prophylaxis against Pneumocystis carinii pneumonia, herpes simplex virus, and cytomegalovirus [23], if indicated, was ad-

1191

NTM Infections in Pediatric Hematopoietic Stem Cell Transplantation

ministered. None of the patients received NTM prophylaxis except 1 patient with severe combined immunodeficiency who had been on azithromycin since diagnosis. All patients received hematopoietic growth factors starting with sargramostim (granulocyte-macrophage colony-stimulating factor, GM-CSF; 250 ␮g/m2 per day IV daily) on day 0 within 3 hours of stem cell infusion and continued until white blood cell count was ⱖ0.3 ⫻ 109 cells/l for 2 consecutive days. Patients were then switched to filgrastim (granulocyte colony-stimulating factor; 10 ␮g/kg per day IV daily) and tapered by 50% when the absolute neutrophil count reached ⱖ2.5 ⫻ 109 cells/L for 2 consecutive days, as we previously described [24]. Definition of NTM Disease

Definitions from the Centers for Disease Control and Prevention (CDC) were used as described in a report by Gaviria et al [1] to diagnose NTM disease (Table 2). The day of diagnosis of NTM infection was defined as the day on which the initial NTM isolate was recovered or a diagnostic histopathologic specimen was obtained.

MGIT 960 System. In addition, samples were plated onto Lowenstein-Jensen slants and 7H11 selective and nonselective agar plates (Becton Dickinson) and incubated (35°C, 5%-10% CO2) for 6-8 weeks. After growth was observed, the mycobacteria were identified to species level by using commercially available DNA probe assays (AccuProbe, Gen-Probe, San Diego, Calif) for identification of Mycobacterium avium complex (MAC), Mycobacterium kansasii, and Mycobacterium gordonae and standard biochemical tests for identification of other species. NTM Susceptibility Testing

Non-MAC NTM mycobacterial isolates were referred to the University of Texas (Tyler, Tex) for testing with the following anti-NTM drug susceptibility panel: amikacin, cefoxitin, ciprofloxacin, clarithromycin, doxycycline, gatifloxacin, imipenem, kanamycin, levofloxacin, linezolid, meropenem, minocycline, sulfamethoxazole, and tobramycin. MAC isolates were referred to the National Jewish Medical and Research Center (Denver, Colo), where the following drugs were tested: amikacin, ciprofloxacin, clarithromycin, clofazimine, ethambutol, levofloxacin, rifabutin, rifampin, and streptomycin.

NTM Isolation and Identification

All specimens for culture of mycobacteria were obtained for suspected opportunistic infection and initially digested and decontaminated with 2% NaLCNaOH (Mycoprep BBL, Becton Dickinson, Sparks, MD), stained with AFB auramine-O fluorescent stain (ENG Scientific Inc, Clifton, NJ), and inoculated onto appropriate media, according to established protocol [25]. Samples were inoculated onto broth media (Mycobacterial Growth Indicator Tube, Becton Dickinson) and incubated in the automated BACTEC

Statistical Analysis

Results are presented as medians with specified ranges of datasets. The Kaplan-Meier method was used to calculate the probabilities of incidences of NTM, aGVHD, cGVHD, and survival in children and adolescents who underwent HSCT from January 2000 to 2004. Chronic GVHD was evaluated in the group of children and adolescents who underwent HSCT and were followed for ⱖ18 months in the same period, and incidence of cGVHD was calculated using

Table 2. Center for Disease Control and Prevention Definitions for Mycobacterial Disease [1] 1. Lower respiratory tract disease a. Definite—Respiratory symptoms compatible with pneumonia, pulmonary infiltrates on chest radiograph, >1 positive culture for NTM from lower respiratory samples (ie, bronchoalveolar lavage or deep tracheal sputum), and a concomitant blood culture for NTM or histologic evidence of tissue invasion by NTM. b. Probable—Same criteria as definite but without concomitant blood culture for NTM or histologic evidence of tissue invasion by NTM. c. Possible—Same as probable and identification of a copathogen likely to account for the clinical picture or improvement of respiratory symptoms and pulmonary infiltrates after treatment of an alternative diagnosis without specific treatment against NTM. 2. Catheter-related infection* a. Definite—Typical symptoms (eg, exit site or tunnel erythema, purulence, fever) and positive blood culture for NTM. b. Probable—Typical symptoms and isolation of NTM from the catheter site, tunnel, or tip. 3. Other sites a. Definite—Compatible symptoms, a positive culture for NTM from a normally sterile site, and histologic evidence of atypical mycobacterial tissue invasion or a positive blood culture. b. Probable—Compatible symptoms, a positive culture for NTM from a normally sterile site, but no histologic evidence of atypical mycobacterial tissue invasion or positive blood culture.

NTM indicates nontuberculous mycobacteria. *Infections were further categorized as exit site infection (⬍2 cm of inflammation, ie, erythema, tenderness, or edema around exit site), tunnel infection (⬎2 cm of inflammation from the catheter exit site extending along the subcutaneous tract of tunneled catheter), or catheterrelated bacteremia (without exit or tunnel infection).

1192

E. Unal et al.

Graft-versus-Host Disease

A

Probability of

G r a de I I I aG V H D

100

The probabilities of aGVHD (at least grade II) and cGVHD (evaluated in children surviving ⱖ18 months) in the allogeneic HSCT recipients were 48.5% (95% confidence interval, 38.3-58.7) and 20.7% (95% confidence interval, 11.3-30.1), respectively (Figure 1A,B).

80

60

NTM Infections 40

20

0 0

300

600

900

1200

1500

1800

Day

B 100

Pro bability of cGVH D

80

60

40

20

0 0

100

200

300

400

500

Day Figure 1. A, Probability of developing grade III-IV aGVHD was determined by the Kaplan-Meier methods in patients after HSCT. B, Probability of developing cGVHD in patients followed ⱖ18 months after HSCT determined by the Kaplan-Meier method.

During this 5-year period, NTM infection was diagnosed in 5 (3.8%) of 132 pediatric recipients of HSCT. None of the NTM infections was diagnosed in patients after myeloablative plus autologous HSCT. The incidence of NTM in allogeneic HSCT was 6.4% (95% confidence interval, 0.8-11.9; Figure 2). Two had definite NTM infection and 3 had probable NTM infection. Characteristics of patients with NTM infection are listed in Table 3. Three of these patients (504-003, 505-002, and 509-008) were partly described by Nicholson et al [26]. Three males and 2 females were diagnosed with NTM infection. Mean age was 8 years (median, 3 years; range, 2-19 years). Diagnoses were chronic myelogenous leukemia (CML) in chronic phase (CP) in 2 patients, neuroblastoma/post-SCT myelodysplastic syndrome in 1, hemophagocytic lymphohistiocytosis in 1, and severe aplastic anemia in 1. These 5 patients underwent 8 HSCT procedures (7 allogeneic and 1 autologous) and the donor sources for allogeneic HSCT were 6/6 sibling in 2, 5/6 sibling in 1, 6/6 UCB in 2, and 5/6 UCB in 2. Two of 5 patients received ⬎1 HSCT and for both patients NTM infection was diagnosed after the second HSCT. Although 4 of 5 patients were diagnosed with GVHD at some point during their transplantation course, only 2 had active GVHD (1 with acute gut/skin GVHD and 1 with extensive cGVHD involving gut and skin) at the time

100

NTM-all HPSCT recipients NTM-allo HPSCT recipients

80

the Kaplan-Meier method. All group comparisons were made using the log-rank test. Statistical analysis was performed with the Prism statistical program (GraphPad, San Diego, Calif).

60

40

RESULTS

20

Demographic and Clinical Characteristics

From January 2000 to December 2004, 132 children and adolescents with malignant and nonmalignant disease underwent 169 HSCT procedures at our institution (Table 1). Median follow-up was 827.5 days (range, 170-1958 days).

0 0

500

1000

1500

2000

Day Figure 2. Incidence of NTM in all patients (solid line) receiving HSCT and in those after allogeneic HSCT (dashed line).

Patient Number

Age (y)

Donor Source

Sex

Disease

504-003

3

F

NBL/MDS

UCB

505-002

19

M

CML-CP

509-010

13

M

505-004

3

509-008

2

NTM Culture/Biopsy Sites

NTM

NTM Infection

Days from HSCT to NTM Infection

Antibiotic Combinations

Days of NTM Tx

Outcome

MAC

Definite/disseminated

14

Rifampin, levofloxacin, amikacin, azithromycin

119

AlloSib

Gut biopsy, bone marrow culture, CSF culture Blood culture

M chelonae

133

AlloSib

Sputum culture

MAC

191

Alive, NED

F

CML-CP

AlloSib

BAL culture

MAC

Imipenem, amikacin, clarithromycin Rifabutin, azithromycin, ethambutol, levofloxacin Rifabutin, ethambutol, azithromycin

468

SAA

Definite/catheterrelated bacteremia Probable/lower respiratory tract infection

Dead on day ⴙ119, VOD Alive, NED

40

Alive, NED

M

HLH

UCB

Skin nodule culture

M kansasii

153

Alive, NED

Probable/lower respiratory tract infection Probable/skin and soft tissue infection

269

127

35

Rifampin, amikacin, ethambutol, levofloxacin, azithromycin

NTM Infections in Pediatric Hematopoietic Stem Cell Transplantation

Table 3. Characteristics of Patients with NTM Infection (n ⫽ 5)

NTM indicates nontuberculous mycobacterium; HSCT, hematopoietic stem cell transplantation; NBL, neuroblastoma; MDS, myelodysplastic syndrome; CML, chronic myelogenous leukemia; CP, chronic phase; SAA, severe aplastic anemia; HLH, hemophagocytic lymphohistiocytosis; UCB, umbilical cord blood; AlloSib, allogeneic sibling; CSF, cerebrospinal fluid; BAL, bronchoalveolar lavage; MAC, Mycobacterium avium complex; M chelonae, Mycobacterium chelonae; M kansasii, Mycobacterium kansasii; VOD, veno-occlusive disease; NED, no evidence of NTM disease/infection.

1193

1194

of NTM diagnosis. For the 2 patients who had active GVHD at the time of NTM diagnosis, 1 had been treated with prednisone and 1 had been treated with prednisone, infliximab, denileukin diftitox, rituximab, and etanercept. The median time from HSCT to diagnosis of NTM infection was 115 days (range, 14-269 days). The median times from the submission of the culture to a positive result for MAC, M kansasii, and M chelonae were 26, 6 and 5.5 days, respectively. Definite NTM Infections

Disseminated MAC was diagnosed in a 3-year-old female after her second HSCT (Table 3). While on treatment for NTM infection, the patient received a second allogeneic HSCT for graft failure and persistent myelodysplastic syndrome 110 days after NTM diagnosis. She died secondary to veno-occlusive disease of the liver and subsequent multiorgan failure 13 days after this second allogeneic HSCT without evidence of an NTM infection. Catheter-related NTM bacteremia (M chelonae) was diagnosed in a 19-year-old male patient with CML in first CP (Table 3). He received a 6/6 matched allogeneic HSCT from his sibling with a conditioning regimen consisting of busulfan (1.6 mg/kg ⫻ 4 days), fludarabine (30 mg/m2 ⫻ 5 days), and alemtuzumab (54 mg/m2 ⫻ 5 days). He is currently alive without evidence of an active NTM infection. Probable NTM Infections

Two patients were diagnosed with probable lower respiratory tract NTM infections. A 13-year-old male with severe aplastic anemia received a 6/6 allogeneic HSCT from his brother (Table 3). The conditioning regimen was cyclophosphamide (50 mg/kg ⫻ 4 days), fludarabine (30 mg/m2 ⫻ 6 days), and ATG (2 mg/kg ⫻ 4 days). He developed aGVHD of the gut and skin that responded to prednisone. Ground-glass opacities were reported on a chest computed tomogram and sputum culture grew MAC 269 days after SCT. Bronchoalveolar lavage was negative and blood cultures remained negative for NTM. He currently is alive and free of NTM infection. A 3-year-old female received a 5/6 allogeneic HSCT from her sister for CML in CP (Table 3). Conditioning regimen was busulfan (4 mg/kg ⫻ 4 days), fludarabine (30 mg/m2 ⫻ 5 days), and alemtuzumab (54 mg/m2 ⫻ 5 days). Her course was complicated by acute skin and gut GVHD followed by extensive cGVHD involving the skin and gut that was resistant to steroid treatment. She had received combination GVHD treatment with infliximab and denileukin diftitox followed by rituximab and etanercept for refractory GVHD. Computed tomogram of the chest demonstrated punctate nodular densities involving both lung fields. Bronchoalveolar lavage cultures

E. Unal et al.

grew MAC and M gordonae 127 days after HSCT. She is currently alive 9 months after anti-NTM antibiotic treatment with disappearance of the chest computed tomographic findings and no evidence of active NTM infection. A 2-year-old male patient with hemophagocytic lymphohistiocytosis developed graft failure and received a second allogeneic HSCT from a 5/6 UCB donor with a conditioning regimen consisting of busulfan (4 mg/kg ⫻ 4 days), cyclophosphamide (60 mg/kg ⫻ 2 days), VP-16 (40 mg/kg ⫻ 1 day), and alemtuzumab (54 mg/m2 ⫻ 5 days; Table 3). A right thigh skin nodule was noted on examination before engraftment. Biopsy of a right thigh skin nodule was positive for M kansasii 35 days after second allogeneic HSCT. He is currently alive and free of NTM infection after anti-NTM antibiotic treatment. Survival

Kaplan-Meier survival curves for patients with NTM infection and all patients who underwent allogeneic HSCT are included in Figure 3. Probabilities of overall survival in all patients (Figure 3A) and those in average- and high-risk groups (Figure 3B) were 58% (95% confidence interval, 49.1-100), 68.4% (95% confidence interval, 58.6-78.3), and 33.7% (95% confidence interval, 17.5-49.9), respectively. Survival among patients with NTM (Figure 3C) was 80.0% (95% confidence interval, 44.9-100). There were no deaths secondary to NTM infection.

DISCUSSION This is the largest pediatric series of HSCT recipients to report the incidence and outcome of NTM disease. In this series, 5 of 132 (3.8%) children were diagnosed with an NTM infection and, as expected, the risk was higher (6.4%) among allogeneic HSCT recipients. The incidence after autologous HSCT was negligible. As evidence of the diagnostic difficulties inherent in NTM infections, definite infection was diagnosed in only 2 patients, and probable infection was diagnosed in 3. Outcome in the patients with NTM infections appeared to be excellent because all infections resolved with medical treatment with antiNTM antimicrobial agents. The incidence of NTM infection in our allogeneic HSCT recipients (6.4%) is higher than that in the adult allogeneic HSCT recipients reported by Weinstock et al [2] (4.9%). This higher incidence may have been secondary to the use of alemtuzumab and/or rabbit ATG, a large number of unrelated donors, and/or active GVHD treatment with systemic steroids. GM-CSF is an earlier and broader-acting CSF, which also induces more T helper 1 and dendritic cell immune subsets preferentially in addition to mono-

NTM Infections in Pediatric Hematopoietic Stem Cell Transplantation

A 100

80

60

40

20

0 0

500

1000

1500

2000

Day

B 100

Probability of Survival

80

60

40

20

Average risk HSCT recipients High risk HSCT recipients

0 0

500

1000

1500

2000

Day

C 100

80

60

40

20

0 0

300

600

900

1200

Day

Figure 3. A, Probability of overall survival determined by the Kaplan-Meier method in patients after HSCT. B, Probability of overall survival stratified by risk was determined by the KaplanMeier method in patients with average risk (solid line) versus those with high risk (dashed line) after HSCT. Allogeneic HSCT recipients were defined as having high-risk disease with ⱖ1 of the following criteria: leukemia in relapse, leukemia with induction failure, leukemia in third complete remission or beyond, malignant progressive disease, and/or second allogeneic HSCT. All other patients were considered to have average risk. C, Probability of overall survival in patients with NTM from time of HSCT by the Kaplan-Meier method in patients with NTM after HSCT.

1195

cytes and macrophages compared with granulocyte colony-stimulating factor [24] , which might play an important role in the host defense against NTM infections. Theoretically, the use of GM-CSF may have prevented some of the early NTM infections. An increase in NTM infections in recent years has been reported [13] and is thought to be due to increased patient survival, increased awareness of NTM, and enhanced diagnostic capabilities. In a recent review of NTM infections in HSCT recipients, the incidence ranged from 0.4% to 4.9% [13]. In the case series reported by Roy and Wisdorf [3], the incidence was only 0.4%, but 46.8% of their patients were ⬍18 years of age, which potentially explained the lower incidence of infections. In contrast, Weinstock et al [2] reported that 4.9% of allogeneic HSCT recipients were diagnosed with a definite or probable NTM infection. The increased incidence was potentially explained by the use of T cell depletion, an increased awareness of NTM, and increased diagnostic testing and the frequency of infections caused by M haemophilum [2,9]. Three large series in the literature have reported a higher incidence of NTM infections in unrelated allogeneic HSCT recipients: related versus unrelated, 0.22% versus 1.0% [1], 0.36% versus 0.54% [3], and 4.7% versus 5.3% [2]. The increased use of unrelated mismatched donors and immunosuppressive medications in recent years may also contribute to this increase in NTM. Three of our patients with NTM infection had received matched related sibling transplants. In addition, the somewhat higher incidence in our population could be secondary to the immunosuppressive treatment regimens our patients received. Although the limited number of patients with NTM at our institution did not allow an analysis of risk factors, the use of alemtuzumab in the conditioning regimen before the diagnosis of NTM in 3 of 5 patients is striking. Alemtuzumab is a humanized monoclonal antibody directed against the CD52 antigen that is expressed on virtually all lymphocytes (T lymphocytes more prominently than B lymphocytes) and monocytes [27]. It is used for in vivo and in vitro T cell depletion and for the treatment of lymphoma, autoimmune disorders, and GVHD because of its profound effect on T cells [28]. Reactivation of M xenopi pulmonary infection, systemic tuberculosis, and disseminated disease with M bovis immediately after administration of alemtuzumab have been reported [2931]. Thus, alemtuzumab may be an additional risk factor for developing NTM in HSCT recipients. NTM infections of the lower respiratory tract can be very difficult to diagnose. The diverse clinical spectrum, coinfections, the delay until a positive culture because of the slow growth of many of these organisms, and the difficulty distinguishing colonization and contamination from true infection are factors

1196

complicating the diagnosis of pulmonary NTM infections. Although the CDC criteria were used in this study [1], it is possible that some cases were misclassified because any positive culture from lower respiratory tract secretions in combination with radiologic and clinical symptoms is sufficient for the diagnosis of probable or possible disease. Diagnostic criteria defining “infection” could also be a reason why statistics vary among studies. Thus, it is possible that 2 of our cases were only colonized and not truly infected, although no copathogens were identified. The characteristics of NTM infections in adult HSCT recipients are somewhat different when compared with NTM infections in our series of pediatric HSCT recipients. Doucette and Fishman [13] reviewed 93 previously reported NTM cases in adult HSCT recipients. The median time between transplantation and presentation was 4.6 months, median age was 32 years, and the incidence was equal for male and female patients. GVHD was present in 46% of cases. The most common presentation was catheterrelated infection (37%) and the most frequent species isolated was rapidly growing NTM, with M fortuitum being the most common followed by MAC. Sixty-two percent of the cases were reported as “resolved” or “cured” with medical therapy. In 7.5% of patients, death was attributed to NTM disease. In our series, 60% of our patients with NTM were male, the median time to infection was 3.8 months, and 40% of patients had GVHD at the time of infection. The most frequent species detected in our series was MAC. Bacteremia or catheter-related infections, which are caused more by rapidly growing NTM, were less in our series. In our series, no deaths were attributed to NTM infection. The optimal treatment for NTM infections in HSCT recipients remains undefined. Four of our patients were started on 3 or 4 anti-NTM antibiotics and the fifth patient with M chelonae infection received 2 antibiotics after catheter removal. The mean duration of treatment for the 3 patients who completed the course was 270 days. Thus, treatment regimens for NTM are often complex because multidrug regimens are required, with numerous toxicities and the potential to affect the metabolism of other treatments. Empiric treatment is usually begun before susceptibility testing results and the role of susceptibility testing is still evolving. The duration of treatment is largely derived from other patient populations and reflects the species, site of infection, and clinical, microbiologic, and radiologic responses to treatment [13]. Depending on the severity of the infection, decreasing the immunosuppressive regimens might be necessary. In summary, NTM infections do occur in pediatric HSCT recipients and the incidence is similar to that reported in the adult population. Although there was a higher incidence in allogeneic HSCT recipients,

E. Unal et al.

the incidence after autologous HSCT was negligible. In this small series, the outcome was excellent, although it is possible that some patients were misclassified with disease when they had only colonization. Susceptibility to NTM infection is most likely multifactorial and identifying the risk factors will be important to optimize management in this complex patient population. Because of the relative rarity of NTM infections, multicenter studies are needed to develop strategies to identify risk factors and optimize management of these NTM infections in HSCT recipients.

ACKNOWLEDGMENTS This study was supported in part by the Pediatric Cancer Research Foundation, Brittany Barron Fund, Sonia Scaramella Fund, Pediatric Cancer Foundation, Marisa Fund, and Bevanmar Foundation. The authors thank Linda Rahl for expert assistance on the preparation of this report. The authors also thank the BMT nursing staff on 5 Tower and IP-7 for expert care of these patients.

REFERENCES 1. Gaviria JM, Garcia PJ, Garrido SM, Corey L, Boeckh M. Nontuberculous mycobacterial infections in hematopoietic stem cell transplant recipients: characteristics of respiratory and catheter-related infections. Biol Blood Marrow Transplant. 2000; 6:361-369. 2. Weinstock DM, Feinstein MB, Sepkowitz KA, Jakubowski A. High rates of infection and colonization by nontuberculous mycobacteria after allogeneic hematopoietic stem cell transplantation. Bone Marrow Transplant. 2003;31:1015-1021. 3. Roy V, Weisdorf D. Mycobacterial infections following bone marrow transplantation: a 20 year retrospective review. Bone Marrow Transplant. 1997;19:467-470. 4. Cordonnier C, Martino R, Trabasso P, et al. Mycobacterial infection: a difficult and late diagnosis in stem cell transplant recipients. Clin Infect Dis. 2004;38:1229-1236. 5. Brown-Elliott BA, Griffith DE, Wallace RJ Jr. Newly described or emerging human species of nontuberculous mycobacteria. Infect Dis Clin North Am. 2002;16:187-220. 6. Wagner D, Young LS. Nontuberculous mycobacterial infections: a clinical review. Infection. 2004;32:257-270. 7. Reichenbach J, Rosenzweig S, Doffinger R, Dupuis S, Holland SM, Casanova JL. Mycobacterial diseases in primary immunodeficiencies. Curr Opin Allergy Clin Immunol. 2001;1:503-511. 8. Levendoglu-Tugal O, Munoz J, Brudnicki A, Fevzi Ozkaynak M, Sandoval C, Jayabose S. Infections due to nontuberculous mycobacteria in children with leukemia. Clin Infect Dis. 1998;27:12271230. 9. White MH, Papadopoulos EB, Small TN, Kiehn TE, Armstrong D. Mycobacterium haemophilum infections in bone marrow transplant recipients. Transplantation. 1995;60:957-960. 10. Jacobson KL, Teira R, Libshitz HI, et al. Mycobacterium kansasii infections in patients with cancer. Clin Infect Dis. 2000; 30:965-969.

NTM Infections in Pediatric Hematopoietic Stem Cell Transplantation

11. Kurzrock R, Zander A, Vellekoop L, Kanojia M, Luna M, Dicke K. Mycobacterial pulmonary infections after allogeneic bone marrow transplantation. Am J Med. 1984;77:35-40. 12. Shah MK, Sebti A, Kiehn TE, Massarella SA, Sepkowitz KA. Mycobacterium haemophilum in immunocompromised patients. Clin Infect Dis. 2001;33:330-337. 13. Doucette K, Fishman JA. Nontuberculous mycobacterial infection in hematopoietic stem cell and solid organ transplant recipients. Clin Infect Dis. 2004;38:1428-1439. 14. Patel R, Roberts GD, Keating MR, Paya CV. Infections due to nontuberculous mycobacteria in kidney, heart, and liver transplant recipients. Clin Infect Dis. 1994;19:263-273. 15. Hermida G, Richard C, Baro J, et al. Allogeneic BMT in a patient with CML and prior disseminated infection by mycobacterium avium complex. Bone Marrow Transplant. 1995;16:183-185. 16. Au WY, Cheng VC, Ho PL, et al. Nontuberculous mycobacterial infections in Chinese hematopoietic stem cell transplantation recipients. Bone Marrow Transplant. 2003;32:709-714. 17. Mohite U, Das M, Saikia T, et al. Mycobacterial pulmonary infection post allogeneic bone marrow transplantation. Leuk Lymphoma. 2001;40:675-678. 18. Bradley MB, Cairo MS. Cord blood immunology and stem cell transplantation. Hum Immunol. 2005;66:431-446. 19. Osunkwo I, Bessmertny O, Harrison L, et al. A pilot study of tacrolimus and mycophenolate mofetil graft-versus-host disease prophylaxis in childhood and adolescent allogeneic stem cell transplant recipients. Biol Blood Marrow Transplant. 2004;10: 246-258. 20. Przepiorka D, Weisdorf D, Martin P, et al. 1994 Consensus conference on acute GVHD grading. Bone Marrow Transplant. 1995;15:825-828. 21. Glucksberg H, Storb R, Fefer A, et al. Clinical manifestations of graft-versus-host disease in human recipients of marrow from HL-A-matched sibling donors. Transplantation. 1974;18:295-304. 22. Roman E, Osunkwo I, Militano O, et al. Liposomal amphotericin B (AMB)prophylaxis is effective in prevention of invasive mold infections (IMI) following allogeneic stem cell transplantation in pediatric recipients. Pediatr Blood Cancer. 2005;45:137.

1197

23. Shereck EB, Bessmertny O, Cooney EM, et al. Cytomegalovirus (CMV) prophylaxis with alternate day ganciclovir/foscarnet in at risk pediatric allogeneic stem cell transplant (AlloSCT) recipients is 100% effective in preventing CMV infections and reducing ganciclovir hematopoietic toxicity. Blood. 2002;100: 2470. 24. Militano O, Boskoff B, Del Toro G, et al. Sequential administration of sargramostim (GM-CSF) and filgrastim (G-CSF) in pediatric allogeneic stem cell transplant (AlloSCT) recipients undergoing myeloablative (MA) conditioning: cost-effective and more rapid platelet recovery in UCB recipients. Biol Blood Marrow Transplant. 2006;12. Abstract 116. 25. Pfyffer G. Mycobacterium: General characteristics, isolation and straining procedures. In: Murray P, Baron EJ, Jorgensen JH, Pfaller MA, Yolken RH, eds. Manual of Clinical Microbiology. Volume 1. 8th ed. Washington, DC: ASM Press; 2003:532559. 26. Nicholson O, Feja K, Larussa P, et al. Nontuberculous mycobacterial infections in pediatric hematopoietic stem cell transplant recipients: case report and review of the literature. Pediatr Infect Dis J. 2006;25:263-267. 27. Hale G, Xia MQ, Tighe HP, Dyer MJ, Waldmann H. The CAMPATH-1 antigen (CDw52). Tissue Antigens. 1990;35:118127. 28. Flynn JM, Byrd JC. Campath-1H monoclonal antibody therapy. Curr Opin Oncol. 2000;12:574-581. 29. Willemze R, Richel DJ, Falkenburg JH, et al. In vivo use of Campath-1G to prevent graft-versus-host disease and graft rejection after bone marrow transplantation. Bone Marrow Transplant. 1992;9:255-261. 30. Weinblatt ME, Maddison PJ, Bulpitt KJ, et al. CAMPATH1H, a humanized monoclonal antibody, in refractory rheumatoid arthritis. An intravenous dose-escalation study. Arthritis Rheum. 1995;38:1589-1594. 31. Abad S, Gyan E, Moachon L, et al. Tuberculosis due to Mycobacterium bovis after alemtuzumab administration. Clin Infect Dis. 2003;37:e27-e28.