Extrapulmonary tuberculosis active infection misdiagnosed as cancer

Mycobacterium tuberculosis disease in patients at a comprehensive cancer center (2001–2005)

Authors

  • Gabriel M. Aisenberg M.D.,

    1. Department of Infectious Diseases, Infection Control, and Employee Health, The University of Texas M. D. Anderson Cancer Center, Houston, Texas
    Search for more papers by this author
  • Kalen Jacobson M.D.,

    1. Department of Infectious Diseases, Infection Control, and Employee Health, The University of Texas M. D. Anderson Cancer Center, Houston, Texas
    Search for more papers by this author
  • Roy F. Chemaly M.D.,

    1. Department of Infectious Diseases, Infection Control, and Employee Health, The University of Texas M. D. Anderson Cancer Center, Houston, Texas
    Search for more papers by this author
  • Kenneth V. Rolston M.D.,

    1. Department of Infectious Diseases, Infection Control, and Employee Health, The University of Texas M. D. Anderson Cancer Center, Houston, Texas
    Search for more papers by this author
  • Isaam I. Raad M.D.,

    1. Department of Infectious Diseases, Infection Control, and Employee Health, The University of Texas M. D. Anderson Cancer Center, Houston, Texas
    Search for more papers by this author
  • Amar Safdar M.D.

    Corresponding author
    1. Department of Infectious Diseases, Infection Control, and Employee Health, The University of Texas M. D. Anderson Cancer Center, Houston, Texas
    • Department of Infectious Diseases, Infection Control, and Employee Health, Unit 402, The University of Texas M. D. Anderson Cancer Center, 1515 Holcombe Blvd., Houston, TX 77030===

    Search for more papers by this author
    • Fax: (713) 745-6839


  • A portion of this study was presented at the 43rd Annual Meeting of Infectious Diseases Society of America, October 6–9, 2005.

Abstract

BACKGROUND

Extrapulmonary tuberculosis is an uncommon disease in the U.S., even in immunosuppressed cancer patients. This study evaluated characteristics and frequency of extrapulmonary tuberculosis in patients at a tertiary care referral cancer center.

METHODS

The records of all consecutive patients with Mycobacterium tuberculosis diagnosed during January 2001 through April 2005 at the M. D. Anderson Cancer Center were reviewed after obtaining institutional review board approval.

RESULTS

There were 26 patients with active tuberculosis during the period studied; 18 of them were cancer patients and the others had been referred for a presumed cancer but did not have cancer. The overall rate of active tuberculosis during this period was 0.2 in 1000 new cancer diagnoses. There were 18 men (69%), the median age was 54 years (range, 3–84 yrs), and 16 patients (62%) were born in the U.S. Thirteen (72%) of the 18 cancer patients had solid-organ tumors; 3 of the 5 patients with a hematologic malignancy had non-Hodgkin lymphoma. Three patients (12%) had diabetes mellitus, and 2 patients (8%) had received high-dose (>1 mg/kg of prednisone daily) corticosteroids in the previous week. No patient had a recent history (within the past 4 wks) of chemotherapy; 4 patients had neutropenia. Cough was a prominent symptom (31%), followed by bone pain (19%), dyspnea (15%), and fever (12%). Fifteen patients (58%) had extrapulmonary infection, including 5 patients with concurrent pulmonary involvement; 7 noncancer patients (88%) and 8 cancer patients (44%, P = 0.22) had extrapulmonary disease. In 11 patients (42%), the lungs were the only site of active tuberculosis. Cavitary pneumonia was seen radiographically in 3 of 16 patients (19%) with pulmonary tuberculosis. All M. tuberculosis isolates were susceptible to isoniazid, rifampin, ethambutol, and pyrazinamide; streptomycin resistance was noted in 1 of 22 (5%) isolates tested. Twenty-two patients (85%) received appropriate antituberculosis treatment; all had a clinical and radiographic response. In 3 patients (12%) the cause of death was attributed to M. tuberculosis disease; 2 of 18 cancer patients (11%) died of progressive M. tuberculosis, and they had advanced solid-organ cancer, whereas 1 of 8 patients (13%) without cancer died and the tuberculosis diagnosis was made only on postmortem examination. Univariate analysis showed no significant differences in patients or disease characteristics between non–U.S.-born and U.S.-born patients, whereas noncancer patients (age 52 yrs) and those with extrapulmonary tuberculosis (age 53 yrs) were younger compared with cancer patients (63 yrs; P < 0.007) and those with pulmonary disease (age 60 yrs; P = 0.09).

CONCLUSIONS

Extrapulmonary tuberculosis was relatively common in younger patients with active M. tuberculosis infection, and was often initially misdiagnosed as cancer. Cancer 2005. © 2005 American Cancer Society.

Active Mycobacterium tuberculosis infection can occasionally lead to serious disease in immunosuppressed cancer patients.1, 2 Mycobacterium tuberculosis has a profound influence on global public health,3, 4 although in the U.S. new M. tuberculosis cases, which increased during the emergence of the human immunodeficiency virus (HIV) epidemic, has declined since 1997.5, 6 However, given that most M. tuberculosis infections in the U.S. now occur in individuals born in other countries,4 it was expected that most active tuberculosis infection cases would occur in non–U.S.-born patients who were receiving antineoplastic therapy or stem cell transplantation in the U.S.6

The clinical spectrum of extrapulmonary tuberculosis disease, including cases that were initially suspected as cancer, remains uncertain. In a preceding study, we found that extrapulmonary body sites were commonly involved in cancer patients with active M. tuberculosis infection.1 The relative frequency of extrapulmonary tuberculosis in the U.S. has increased since 1991, and nearly 20% of documented M. tuberculosis cases now involve extrapulmonary sites.7 Women, non-Hispanic blacks, and patients with HIV-AIDS (acquired immunodeficiency syndrome) have a significantly higher risk of extrapulmonary tuberculosis; however, the presence of an underlying malignancy has not been considered as a risk factor for extrapulmonary infection.7

The frequency of extrapulmonary tuberculosis in patients that were initially misdiagnosed as cancer and require referral to a tertiary care cancer center is not known. Therefore, we sought to evaluate the patient and disease characteristics of active pulmonary and extrapulmonary M. tuberculosis infection in patients receiving care at our comprehensive cancer center during the past 4 years.

MATERIALS AND METHODS

The records of all consecutive patients with active M. tuberculosis infection diagnosed at The University of Texas M. D. Anderson Cancer Center (MDACC) between January 2001 and April 2005 were evaluated retrospectively after obtaining institutional review board approval; the requirement for patient consent was waived. The patients were identified from a microbiology database previously established for patients with M. tuberculosis isolated from any body site. Patient information was obtained from electronic medical records and paper charts were also reviewed. The cases of new cancer diagnosis at MDACC were obtained from institutional data register.

Organism Identification

Mycobacterium tuberculosis was identified by DNA probing and confirmed by standard biochemical tests described previously.1 Antimicrobial susceptibility was determined according to National Committee for Clinical Laboratory Standards guidelines.8

Definitions

Patients were considered to have an active infection if all of the following criteria were met: 1) M. tuberculosis was isolated in cultures from respiratory, blood, biopsy, or postmortem specimens; 2) the clinical signs and symptoms were compatible with a diagnosis of tuberculosis; and 3) radiographic features, especially those relating to lung infection, were consistent with tuberculosis. Patients with a recent history of a reactive tuberculin skin test had pulmonary infiltrates that were compatible with tuberculosis and responded to combination antituberculosis therapy were also considered to have active tuberculosis. Disseminated disease was evidence of infection involving two or more noncontiguous body sites.

Neutropenia was defined as an absolute neutrophil count lower than 500/μL at the time when samples that grew M. tuberculosis were obtained. Lymphopenia was defined as an absolute lymphocyte count lower than 500/μL at the time when samples that grew M. tuberculosis were obtained. Systemic steroid use was defined as more than 1 mg/kg/day of prednisone or an equivalent dose of another corticosteroid given during the immediate 7 days before diagnostic samples were obtained.

Patients were considered to have advanced cancer if any of the following occurred: 1) AJCC Stage IV disease in patients with solid-organ cancer; 2) first cancer recurrence after hematopoietic stem cell transplantation; or 3) second recurrence in patients with hematologic malignancies. Deaths were attributed to M. tuberculosis if 1) the patient died with progressive tuberculosis within 4 weeks of infection diagnosis; 2) no other known terminal events such as intracranial hemorrhage, myocardial infarction, or pulmonary embolism occurred (the presence of sepsis and septic shock-related multiorgan failure including adult respiratory distress syndrome were considered sequelae of progressive systemic infection); or 3) the postmortem examination revealed that disseminated M. tuberculosis was a cause of death.

Statistical Analysis

Study data are presented as mean values ± standard deviations. Differences in categorical variables between extrapulmonary and pulmonary active tuberculosis, foreign-born versus patients born in the U.S., and patients with cancer were compared with those in whom tuberculosis was initially misdiagnosed as cancer were analyzed using the Fisher exact test and differences in discrete variables were analyzed by the Wilcoxon test for unpaired samples. A two-sided P-value < 0.05 was considered statistically significant. MedCalc 7.0 computer software (MedCalc, Mariakerke, Belgium) was used for the statistical analysis.

RESULTS

During the period studied, 81,899 patients were registered with a new cancer diagnosis at MDACC. Active tuberculosis was diagnosed in 18 cancer patients, for a frequency of 0.2 in 1000 new cancer diagnoses. In addition, eight patients who were referred for a suspicious mass presumed to be cancer but who did not have cancer were determined to have tuberculosis. The distribution of active tuberculosis cases over the period studied included seven cases in 2001, three cases in 2002, four cases in 2003, eight cases in 2004, and four cases in 2005.

Patient and Disease Characteristics

Table 1 summarizes the baseline patient characteristics. Of the 10 non–U.S.-born patients, 4 were from Vietnam, 3 from Mexico, and 1 each were from China, the Philippines, and Spain. Among the 18 patients with an underlying malignancy, 1 patient had undergone an allogeneic hematopoietic stem cell transplantation for non-Hodgkin lymphoma, and M. tuberculosis granulomatous hepatitis was diagnosed 104 days after the transplant.

Table 1. Patient and Disease Characteristics of 26 Patients with Active Tuberculosis Disease
CharacteristicNo. of patients (%)
  • a

    Within 30 days.

  • b

    Within 2 years.

Median age in yrs (range)54 (3-84)
Male/female ratio18 (69)/8 (31)
Born in the U.S.16 (62)
Neutropenia 4 (15)
Lymphocytopenia 3 (12)
Systemic steroid use 2 (8)
Recent chemotherapya 0 (0)
Previous radiationb 1 (4)
Diabetes mellitus 3 (12)
No underlying cancer 8 (31)
Underlying cancer18 (69)
Hematologic malignancy 5 (19)
Non-Hodgkin lymphoma 3
Hodgkin disease 1
Chronic myelogenous leukemia 1
Solid tumor13 (50)
Lung 3
Prostate 2
Esophagus 1
Larynx 1
Liver 1
Oral 1
Stomach 1
Testicle 1
Thyroid 1
Metastasis of unknown origin 1
Advanced cancer 4
Clinical presentation 
Cough 8 (31)
Bone pain 5 (19)
Dyspnea 4 (15)
Fever 3 (12)
Night sweats 2 (8)
Abdominal pain 2 (8)
Neck mass 2 (8)
Superior vena caval syndrome 1 (4)
Sputum production 1 (4)
Chest pain 1 (4)
Hemoptysis 1 (4)
Weight loss 1 (4)
Headaches 1 (4)
Jaundice 1 (4)
Hand arthritis 1 (4)
Site of infection 
Pulmonary only11 (42)
Extrapulmonary only10 (38)
Lymph nodes 2
Hip 2
Spine 2
Cerebellum 1
Liver 1
Pleura 1
Hand 1
Extrapulmonary and pulmonary 5 (19)
Disseminated 4
Disease outcome 
Death attributed to tuberculosis disease 3 (12)
Treated for tuberculosis22 (85)
Death attributed to tuberculosis disease among treated 0 (0)
New culture negative (n = 8) 6 (75)
Radiographic improvement (n = 13) 9 (69)

Seven of 10 patients with known tuberculin skin test results had a positive reaction; 2 of 5 patients (40%) with cancer had negative reaction, whereas 1 of 5 patients (20%) without cancer did not react to intradermal tuberculin injection. In 17 patients the diagnosis of M. tuberculosis required fine-needle aspiration (n = 12) or surgical biopsy (n = 5); 13 of these samples grew M. tuberculosis in cultures, whereas the other 4 were acid-fast stain-positive but did not grow any organism.

Disease characteristics are also presented in Table 1. Eleven patients (42%) had pulmonary tuberculosis only, 10 patients (38%) had extrapulmonary tuberculosis only, and 5 patients (19%) had both pulmonary and extrapulmonary disease. In the radiographic presentation of 15 patients with pulmonary tuberculosis, 6 had bilateral consolidation, 3 had cavitary lesions, 2 had nodules, 2 had lobar infiltrates, 1 had interstitial plus nodular infiltrates, and 1 had tuberculous pleural effusion. Seven of eight patients (88%) in whom tuberculosis was initially misdiagnosed as cancer had extrapulmonary disease, whereas one patient with pulmonary tuberculosis was referred to our institution for possible lung cancer. Disseminated tuberculosis was noted in 1 of 18 cancer patients (6%) and 3 of 8 patients (38%) who did not have cancer (P = 0.07).

The antimicrobial susceptibility of 24 M. tuberculosis isolates was tested and all strains were susceptible to isoniazid, rifampin, pyrazinamide, and ethambutol. Twenty of 21 isolates were also susceptible to streptomycin; 1 was resistant.

Univariate analysis (Table 2) failed to demonstrate significant differences in patient or disease characteristics between non–U.S.-born and U.S.-born patients. Patients with cancer and those with pulmonary tuberculosis were older compared with patients in whom tuberculosis was initially misdiagnosed as cancer and those with extrapulmonary tuberculosis. Similarly, active tuberculosis infection in cancer patients and those with pulmonary disease were frequently seen in men (P = 0.06). All other characteristics were not significantly different in any of these groups.

Table 2. Univariate Analysis Comparing Patient and Disease Characteristics in Tuberculosis Patients by Place of Birth, Cancer vs. Noncancer Status, and Site of Tuberculosis
CharacteristicU.S. born (n = 16)Non–U.S. born (n = 10)P valueCancer (n = 18)Noncancer (n = 14)P valuePulmonary (n = 11)Extrapulmonary (n = 10)P value
  • NA: not applicable; TST: tuberculin skin test; ANC: absolute neutrophil count; ALC: absolute lymphocyte count.

  • a

    Radiographic improvement on 6-month follow-up after antituberculosis therapy began.

Median age in yrs (range)56 (3-84)55 (14-74)0.8263 (24-84)52 (3-72)0.00760 (46-84)53 (3-72)0.09
Gender, male11 (69) 7 (70)115 (83) 3 (38)0.0610 (91) 5 (50)0.06
U.S. bornNANANA13 (72) 3 (38)0.19 7 (64) 6 (60)1.00
Underlying cancer13 (81) 5 (50)0.19NANANA10 (91) 6 (60)0.15
Extrapulmonary tuberculosis 9 (56) 5 (50)1 8 (44) 7 (88)0.22NANANA
Cavitary lung lesions 2 (13) 1 (10)1 2 (11) 1 (13)0.46NANANA
Positive TST (n = 7)4 of 5 (80)3 of 5 (60)13 of 5 (60)4 of 5 (80)0.393 of 3 (100)1 of 4 (25)0.14
ANC >10,000/mL 3 (19) 1 (10)1 3 (17) 1 (13)1 2 (18) 1 (10)1.00
ALC < 1000/mL 7 (44) 3 (30)0.68 9 (50) 1 (13)0.15 4 (36) 4 (40)1.00
ALC < 500/mL 1 (6) 2 (20)0.54 3 (17) 0 (0)0.53 0 (0) 3 (30)0.09
Tobacco use 8 (50) 3 (30)0.43 9 (50) 2 (25)0.39 6 (55) 5 (50)1.00
Systemic corticosteroid use 2 (13) 1 (10)1 2 (11) 1 (13)1 1 (9) 2 (20)0.59
Anti-M. tuberculosis therapy12 (75) 9 (90)0.6214 (78) 7 (88)1 9 (82) 9 (90)1.00
Death 2 (13) 1 (10)1 2 (11) 1 (13)0.53 0 (0) 1 (10)1.00
Radiographic improvementa (n = 13)7 of 8 (88)2 of 5 (40)0.228 of 10 (80)1 of 3 (33)0.545 of 7 (71)2 of 4 (50)1.00

Outcome

In three patients death was attributed to M. tuberculosis diseases and in all three the diagnoses was made only after death. In these three patients(two of whom [11%] had advanced solid-organ cancer and 1 of whom [13%] was without cancer) the cause of death was progressive and/or disseminated tuberculosis. The population was too small to calculate mortality-associated risk factors.

Twenty-two patients received appropriate antituberculosis treatment and all had a clinical and radiographic response. Twenty patients had been followed for more than 180 days at the time of this report and no tuberculosis disease recurrence occurred. In 6 of 8 patients in whom repeat mycobacterial cultures were obtained within 3 months after antituberculosis treatment began a negative culture was noted in a median of 43 days (range, 13–91 days). Radiographic resolution of lung signs was noted in 9 of 13 patients by 6 months.

DISCUSSION

Extrapulmonary tuberculosis was a frequent complication in our patients with active M. tuberculosis infection and a substantial number of patients with extrapulmonary disease were initially misdiagnosed to have cancer. It was reassuring that the overall frequency of active tuberculosis infection remained unchanged (0.2 in 1000 new cancer diagnoses in 1990–2000 and in 2001–2005) in our cancer patients during these 4 years.1 Extrapulmonary tuberculosis was noted in 44% of patients with cancer, whereas nearly all patients who were referred for a suspicious mass presumed to be cancer but who did not have cancer and were determined to have tuberculosis had extrapulmonary M. tuberculosis disease. This finding most likely reflects an overrepresentation of extrapulmonary active tuberculosis in the noncancer population seen at our referral cancer center.

The rate of infections due to multidrug-resistant (MDR) tuberculosis have declined in the US.9 In keeping with the current trend, none of our M. tuberculosis isolates were resistant to the first-line antituberculosis agents, even though nearly one-third of patients in this study were non–U.S.-born. This finding was in contrast to reports that MDR tuberculosis in the U.S. is observed mostly in non–U.S.-born individuals.10 The lack of MDR tuberculosis in our non–U.S.-born cancer patients may reflect a lack of risk factors, which are frequently associated with increased probability of drug resistance, such as AIDS and low socioeconomic status, or both.11

Active M. tuberculosis is a concern in cancer patients with cellular immune defects; these patients are also at an increased risk of disseminated infection.5, 12, 13 It was interesting that in our study all four cases of disseminated infection occurred in patients with no known T–cell-mediated immune dysfunction. Tuberculosis poses a relatively frequent problem in patients undergoing hematopoietic stem cell transplantation in regions in which the disease is endemic,13–16 whereas our observation that approximately 5% of cancer patients with active tuberculosis had an underlying hematopoietic stem cell transplant was in agreement with other reports from the U.S. and northern Europe,17, 18 indicating a low frequency of active tuberculosis even in severely immunosuppressed stem cell transplant recipients. Therefore, based on the current information, antituberculosis prophylaxis is not recommended even in non–U.S.-born patients with cancer receiving antineoplastic therapy in the U.S.; however, in patients with a history of M. tuberculosis infection immunosuppressive therapy should be approached with caution, as these patients may experience infection recrudescence.

The diagnosis of active tuberculosis requires a high level of suspicion because these infections are uncommon in cancer patients in the U.S. and the clinical and radiographic presentation may be quite variable.19 In our study, even patients with pulmonary tuberculosis presented with a wide variety of radiographic features and the associated constitutional symptoms were nonspecific. Furthermore, extrapulmonary tuberculosis20–25 and occasionally pulmonary tuberculosis26, 27 may be initially mistaken as a tumor, as occurred in eight patients in our study. The diagnostic workup for an abdominopelvic and spinal mass and lymphadenopathy of indeterminate etiology should include evaluation for extrapulmonary foci of M. tuberculosis infection. Patients with extrapulmonary tuberculosis and those with disseminated miliary tuberculosis have a high rate of treatment failure and a high risk of infection-associated death, which is often due to delayed diagnosis and the institution of appropriate antituberculosis therapy.

Tuberculin skin tests were reactive in 7 of 10 patients in whom test results were available. This lack of immunogenic anergy to the tuberculin skin test in these patients with active tuberculosis infection was not limited to patients who did not have cancer. In our cancer patients, 60% had a positive tuberculin reaction, which most likely reflects an underrepresentation of patients who are often immunologically anergic, such as those with T-cell senescence due to advanced age, HIV infection, high-dose systemic corticosteroid use, and allogeneic hematopoietic stem cell transplantation.28, 29

The limitations of this study include a retrospective study design and low infection-associated deaths. In this report we have described the features of extrapulmonary active tuberculosis in patients who were presumed to have cancer and referred to our cancer center for antineoplastic therapy.

Active M. tuberculosis infection remains uncommon in patients receiving care at our comprehensive cancer center, and nearly one-third of the tuberculosis cases identified were in patients with no underlying malignancy, in whom extrapulmonary tuberculosis was initially mistaken for a possible cancer. A high level of suspicion in patients with lymphadenopathy and spinal or abdominopelvic mass of indeterminate etiology may help in early diagnosis and institution of appropriate antituberculosis therapy.

Acknowledgements

The authors thank Melissa Burkett of the Department of Scientific Publications for editorial assistance.

Ancillary