| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Guest Access | Sign In via User Name/Password
|
|
|
|||||||
Guest Access | Sign In via User Name/Password |
|||||||||
* From the University of Mississippi Medical Center (Drs. Baumann, Nolan, and Petrini), Jackson, MS; Oxford University (Dr. Lee), Oxford, UK; Vanderbilt University (Dr. Light), Nashville, TN; and the Centers for Disease Control and Prevention (Dr. Schneider), Atlanta, GA.
Correspondence to: Michael H. Baumann, MD, MS, FCCP, Professor of Medicine, 2500 North State St, Division of Pulmonary and Critical Care Medicine, University of Mississippi Medical Center, Jackson, MS 39216-4505; e-mail: mbaumann{at}medicine.umsmed.edu
Abstract
Background: Pleural tuberculosis (TB) should be considered in any patient with a lymphocytic pleural effusion. The diagnostic approach is under debate. Knowledge of pleural TB epidemiology would be beneficial. To help clarify pleural TB epidemiology, we analyzed US national TB surveillance data for 1993 to 2003.
Methods: We compared pleural TB to pulmonary TB (where each was reported as the major site of TB disease, and no additional sites of disease were reported). Applicable statistical tests were performed; p < 0.05 was considered to be significant.
Results: From 1993 through 2003, 7,549 cases of pleural TB and 156,779 cases of pulmonary TB were reported (in 2003: pleural TB, 536 cases; pulmonary TB, 10,551 cases). The annual proportion of pleural TB was relatively stable (median rate, 3.6%; range, 3.3 to 4.0%) compared to that for pulmonary TB, which steadily decreased (average annual decrease, 0.9%; p < 0.01). Pleural TB occurred significantly more often than pulmonary TB among persons 
65 years old (30.4% vs 23.3%, respectively; p < 0.01), and it occurred significantly less often among children < 15 years old (1.8% vs 6.1%, respectively; p < 0.01) and persons 45 to 64 years old (22.9% vs 27.9%, respectively; p < 0.01). Pleural TB patients (63.4%) were born slightly more often in the United States than were pulmonary TB patients (60.9%; p < 0.01). Drug-resistance patterns of pleural TB broadly reflected those of pulmonary TB. However, isolates from pleural TB patients were less often resistant to at least isoniazid (6.0% vs 7.8%, respectively; p < 0.01) and to at least one first-line TB drug (9.9% vs 11.9%, respectively; p < 0.01) compared with pulmonary TB patients.
Conclusions: Knowledge of pleural TB demographic, clinical, and drug-resistance patterns may assist clinicians in making diagnostic and therapeutic decisions.
Key Words: pleura • resistance • tuberculosis
Pleural tuberculosis (TB) should be considered in any patient with an exudative effusion, particularly a lymphocyte-predominant exudative pleural effusion.123 To date, there have been no in-depth national pleural TB epidemiologic studies. In 2003, 14,874 TB cases were reported in the United States, of which 3,029 cases (20.4%) were exclusively extrapulmonary. Of patients with extrapulmonary TB, the pleura was reported as the major site of disease in 559 patients (18.5%).4 The precise incidence of pleural TB can be difficult to determine, since it may not be routinely distinguished from other forms of respiratory TB when reported.5 Pleural TB incidence may also be impacted by coinfection with HIV.567
The best diagnostic approach for pleural TB continues to be debated.12 Central to the debate is the role of pleural biopsy vs the use of pleural fluid level measurements, including interferon-
, nucleic acid amplification, including polymerase chain reaction, tuberculous proteins, and antibodies, and lysozyme level measurements,38 without biopsy. Measurement of pleural fluid adenosine deaminase (ADA) levels appears to be one of the most promising new tests.8 Sahn9 has noted that by combining pleural fluid (ie, culture and acid-fast staining) and pleural biopsy tissue studies (ie, culture and assessment for pleural granulomas) a diagnosis can be established in 90 to 95% of cases. Alternatively, a 2003 metaanalysis10 demonstrated the summary measure of the test characteristics derived from the receiver operating characteristic curve to be 92.2% for both sensitivity and specificity for pleural fluid ADA in patients with pleural TB. Not addressed in this metaanalysis10 was the potential role of the assessment of pleural fluid ADA isoenzyme levels in increasing the diagnostic sensitivity and specificity of the test.11 A central shortcoming of the use of any pleural fluid measurement alone to diagnose pleural TB is the failure to provide antituberculous chemotherapeutic susceptibility data. This is particularly problematic in an era of increasing concerns for drug-resistant TB.
Given the uncertainty of the current incidence and patterns of drug resistance of pleural TB in the United States, we analyzed the national TB surveillance system database to describe the US epidemiology and clinical characteristics of pleural TB compared with those of pulmonary TB.
Materials and Methods
The national TB surveillance system uses a standardized case report form (Report of Verified Case of Tuberculosis) to collect information on newly diagnosed persons with TB from 50 states and the District of Columbia.1213 Demographic, clinical, laboratory, and sociodemographic data, including country of origin, date of arrival in the United States, HIV seropositivity status, and anti-TB drug resistance, were collected.
We analyzed data from the national TB surveillance database, which is maintained by the Centers for Disease Control and Prevention, from January 1, 1993, through December 31, 2003, using reports that were complete as of April 14, 2005. For the purpose of this analysis, pulmonary only was defined when "pulmonary" was reported as the major site of disease without any additional sites of disease reported. Pleural only was defined when "pleural" was reported as the major site of disease without any additional sites of disease reported. Unless specified, pleural-only and pulmonary-only cases will be termed pleural and pulmonary, respectively. We compared pleural TB cases to pulmonary TB cases in our analysis.
Statistical Analysis
Tests of trend were performed on ordinal data (regression, t test for slope) to assess for any differences in the incidence of cases of pleural and pulmonary TB. 
2 analysis for two sample proportions was used to assess differences between pleural and pulmonary TB cases; p < 0.05 was considered to be significant. Bonferroni correction for multiple comparisons was performed. For purposes of statistical analysis using the 2 test, the numbers of cases within the pulmonary subgroups were used as expected frequencies. Within-group comparisons (ie, pleural vs pleural TB cases) were made using a contingency table and the Fisher exact test. In all tests in which there were unknown results, analysis was performed using only the known values.
Results
Demographic Characteristics
From 1993 through 2003, a total of 210,978 TB cases were reported in the United States. The major sites of TB disease were pulmonary (n = 167,261; 79.3%), pleural (n = 8,354; 3.9%), and other (eg, lymph node, bone/joint, genitourinary, meningeal, and peritoneal) [n = 35,363; 16.8%]. For the 167,261 reports of pulmonary TB, 156,779 (93.7%) gave no secondary sites of disease, 3,392 (2.0%) listed "pleural" as a secondary site, and 7,090 (4.3%) listed "other" as a secondary site. For the 8,354 reports of pleural TB, 7,549 (90.4%) gave no secondary sites of disease, 572 (6.8%) listed "pulmonary" as a secondary site, and 233 (2.8%) listed "other" as a secondary site. For this analysis, we compared these 7,549 pleural TB patients having no secondary sites of disease to these 156,779 pulmonary TB patients having no secondary sites of disease.
From 1993 through 2003, the number of cases of pleural and pulmonary TB (p < 0.001), and the proportion of pulmonary TB cases relative to the total number of reported cases in the United States decreased annually (p < 0.001) [Fig 1 ]. In contrast, the proportion of pleural cases remained relatively stable (p = 0.6). In 2003, there were 536 pleural TB cases and 10,551 pulmonary TB cases reported in the United States.
|
65 years of age, with a smaller proportion among pediatric patients (ie, < 15 years old) and patients 45 to 64 years old, compared to cases of pulmonary TB (p < 0.01) [Table 1
]. The distribution of race/ethnicity was similar for pleural and pulmonary patients, except that pleural TB patients were slightly more often black non-Hispanic (35.7%) than were pulmonary TB patients (32.1%; p < 0.01) [Table 1].
|
15 years old), pleural TB patients had resided within the United States significantly longer prior to diagnosis than pulmonary TB patients (median duration, 7.8 vs 5.6 years, respectively; p < 0.01) [Table 1].
Diagnostic Characteristics
Unlike pulmonary TB patients, diagnosis was made in pleural TB patients through culture and microscopic examination of tissues/fluids other than sputum. Of the pleural TB patients, 15.8% had a positive result from microscopic examination of tissues/fluids for TB (p < 0.001) [compared with 8.8% of pulmonary TB patients), and 62.8% had a culture of tissues/fluids that was positive for Mycobacterium tuberculosis (p < 0.001) [compared with 16.7% of pulmonary TB patients] (Table 2
). Almost all of these positive specimens (> 97%) were pleural fluid or pleural tissue (Table 2). Approximately half of patients with pleural TB either did not have a sputum smear for acid-fast bacilli (AFB) or sputum culture performed for M tuberculosis, or the results were unknown. Despite this, 1.7% of pleural patients were sputum smear positive and 3.5% were sputum culture positive (Table 2). During 1993 to 2003, 64.3% of pleural TB patients were culture-positive for M tuberculosis and 22.7% met the clinical case definition for TB (Table 2).12 The majority of both pleural TB patients (94.2%) and pulmonary TB patients (94.5%) had an abnormal chest radiograph. More than half of pleural TB patients (60.6%) and pulmonary TB patients (58.3%) reported tuberculin skin test results that were positive, although many (25.0% and 28.3%, respectively) did not have a tuberculin skin test performed or the results were unknown. From 1993 to 2003, HIV test results were not reported for a large proportion of TB patients 25 to 44 years old (pleural TB patients, 46.9%; pulmonary TB patients, 43.7%). Pulmonary TB patients, however, were almost twice as likely to be coinfected with HIV (20.0%) as pleural TB patients (11.7%; p < 0.001).
|
|
Outcomes
From 1993 through 2001, 82.0% of pleural TB patients and 81.4% of pulmonary TB patients completed therapy. There was no difference (p = 0.99) in all-cause mortality during TB treatment between pleural TB patients (9.0%) and pulmonary TB patients (9.8%). No information regarding relapse rates was available.
Discussion
This study is the most recent and comprehensive assessment of the epidemiology of pleural TB in the United States. Although the total number of pleural TB cases decreased between 1993 and 2003, the proportion of pleural TB cases compared to the total number of cases in the United States remained relatively stable (median proportion, 3.6%) [Fig 1]. This confirms earlier estimates and counters thoughts that those estimates were low.36 The total number and proportion of pulmonary TB cases have been steadily decreasing.
However, concerns about the underreporting of pleural TB owing to frequently negative mycobacterial culture results3 may continue to raise concerns that our pleural TB incidence findings were underestimated. Pleural TB is thought to arise primarily as the sequela of primary pulmonary TB infection or as the reactivation of pulmonary TB disease.3 The decline in the numbers of both pleural and pulmonary TB cases reflects the overall decline in TB in the United States since the resurgence (ie, 1986 to 1992). This can be attributed to the aggressive efforts of the public health community and health-care providers, which resulted in improved diagnostic, therapeutic, and prevention strategies.141516
Several demographic findings regarding pleural TB deserve emphasis (Table 1). A greater proportion of pleural TB cases occurred in patients 65 years of age, and a smaller proportion occurred among pediatric patients and persons aged 45 to 64 years compared with pulmonary TB cases. No clear explanation can be offered for the disparity in the pediatric population. However, the diagnosis of pediatric TB is often achieved by documenting an epidemiologic link with a known TB patient and by clinical findings (eg, perihilar adenopathy on seen on a chest radiograph), rather than laboratory confirmation (eg, sputum).17 The greater proportion in the older age group may be related to the occurrence of pleural TB as a sequela of TB reactivation in this age group. A study from Spain18 noted that pleural TB patients without chest radiographic signs of reactivation were younger (mean age, 28 vs 40 years, respectively; p < 0.01) than patients with radiographic findings and presumed reactivation-derived pleural TB. Such data reinforce the notion that pleural TB occurs in an older age group, as found in our study, and may be owing to the reactivation of disease. Malignancy should also be considered as a cause in patients with a lymphocyte-predominant pleural effusion,3 particularly in patients who are > 60 years of age.9
Not surprisingly, foreign-born TB patients comprised an increasing proportion of both pleural and pulmonary cases. Mexico, the Philippines, Viet Nam, China, and India were the most common countries of origin for pleural TB or pulmonary TB occurring in foreign-born persons. A larger proportion of foreign-born pleural TB patients were from India compared to foreign-born pulmonary TB patients. Although we have no explanation for this pattern, this may be important to consider when assessing a patient with a pleural effusion.
The differences observed in the types of diagnostic studies performed to diagnose pleural vs pulmonary disease were not unexpected (Table 2). Sputum analysis (ie, AFB smear and culture) is the mainstay for diagnosing pulmonary TB, while pleural fluid analysis and pleural biopsy remain the mainstay for the diagnosis of pleural TB.114 The greater use of microscopic examination and cultures of tissues/fluids in pleural cases (Table 2) reflects the standard diagnostic assessment of pleural fluid and pleural biopsy specimens. Relying only on pleural fluid analysis may lead to the underdiagnosis of pleural TB. Pleural fluid smears for AFB have a yield of approximately 10%, and the culture of fluid has a yield of between 25% and 75%.1 Pleural biopsy histopathology (ie, the visualization of granulomas) or a positive culture result substantiates a diagnosis of TB in 55 to 80% of cases. Combined culture and histopathology of tissue obtained by pleural biopsy yields a diagnosis in as many as 87% of cases.1 Combined analysis of pleural fluid and pleural biopsy specimens may increase the diagnostic yield to 95%.1 The method of obtaining pleural tissue was not available in the data we reviewed. Thoracoscopy has yields equivalent to or better than the combination of pleural fluid and pleural biopsy tissue analysis.1 Data on the frequency of use of ADA assessment of pleural fluid were not available in the database.
Sputum smears were positive in only 1.7% of pleural TB cases, and sputum culture findings were positive in only 3.5% of pleural TB cases (Table 2). Almost half of the patients with pleural TB either did not have a sputum smear or culture performed, or the results were unknown. However, in a study from Brazil,19 induced sputum samples may yield a diagnosis in 55% of patients whose chest radiographs show only an effusion. The yield of induced sputum samples in patients with chest radiographs suggesting pulmonary parenchymal disease in addition to effusions was 45%.19 The accompanying editorial20 notes that pursuing induced sputum samples for the diagnosis of pleural TB challenges "traditional dogma" that patients with tuberculous pleural effusions without concomitant pulmonary disease are not contagious. Importantly, the applicability of the findings of the study in low-incidence TB locations needs to be addressed by "independent replication."20 Meantime, these data suggest that induced sputum samples should be used more frequently in the setting of suspected tuberculous pleural effusions given the relative ease and safety of obtaining sputum samples.1920
Of greatest import, our data demonstrate that pleural and pulmonary TB show broadly similar patterns of antimicrobial resistance (Table 3). A small but statistically significant difference was found in resistance to at least INH and in resistance to at least one first-line drug. The similarity of resistance patterns has implications for both treatment and diagnosis. Given these patterns of drug resistance, the initiation of four-drug therapy15 in cases of pleural TB appears to be appropriate while awaiting culture and drug-susceptibility results. The measurement of pleural fluid ADA levels, including ADA isoenzymes, has been advocated3811 as a valuable tool for the diagnosis of pleural TB. It appears to be one of the most promising pleural fluid measurements for the diagnosis of pleural TB.3811 In common with all other tests of pleural fluid for TB diagnosis, the measurement of ADA levels does not provide antimicrobial susceptibility data.1 As a result, one must still attempt to culture the organism to detect antimicrobial resistance. Obtaining antibiotic susceptibility data is particularly important in cases of pleural TB occurring in foreign-born patients, who have higher rates of drug resistance. The pursuit of culture confirmation and drug-susceptibility testing remains a mainstay of Centers for Disease Control and Prevention recommendations.15 By increasing the awareness of pleural TB, we hope to increase the appropriate use of specific diagnostic tests, such as pleural fluid ADA measurement, thereby increasing the diagnostic accuracy and reporting of pleural TB.
Although the national TB surveillance system is the most comprehensive TB database in the United States (reporting completeness, > 95%21), the following several potential limitations may exist with this analysis: (1) mortality is reported as being all-cause, rather than specific cause, while the patient is receiving TB treatment; (2) although the reporting of HIV test results has increased, it is not complete, in part owing to state reporting directives, and therefore one must be aware of the proportion of missing or unknown test results when interpreting results; and (3) our analysis compared pleural-only TB to pulmonary-only TB and therefore the results may not be generalizable to those patients with mixed pleural and pulmonary TB.
In conclusion, the incidences of both pleural TB and pulmonary TB are decreasing in the United States, but are decreasing less for pleural TB. Difficulties in diagnosis may hinder the identification of pleural TB patients. Overall, pleural TB drug resistance patterns parallel those for pulmonary TB, with foreign-born patients being at particular risk for drug resistance. Knowledge of the demographic, clinical, and drug-resistance characteristics of pleural TB, as well as specific diagnostic test options (eg, ADA measurement) can assist the clinician in making both diagnostic and therapeutic decisions, and thus can improve outcomes.
Footnotes
Abbreviations: ADA = adenosine deaminase; AFB = acid-fast bacilli; INH = isoniazid; TB = tuberculosis
The findings and conclusions in this report are those of the authors and do not necessarily represent the views of the Centers for Disease Control and Prevention.
The authors have reported to the ACCP that no significant conflicts of interest exist with any companies/organizations whose products or services may be discussed in this article.
Received for publication September 23, 2006. Accepted for publication November 22, 2006.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
