© 2003 American Thoracic Society
Sputum InductionSimpler, Cheaper, and Safer—But Is It Better?Montreal Chest Institute McGill University Montreal, Quebec, Canada More than 3 decades ago, sputum induction was widely used for the diagnosis of tuberculosis (TB) as it had superior yield to gastric lavage and, not surprisingly, was preferred by patients (1). However, with the advent of fiberoptic bronchoscopy, sputum induction was largely abandoned. It has enjoyed a modest revival recently because of lower risk of nosocomial TB transmission (2), and much lower costs (3, 4). A number of studies over the past decade have demonstrated that diagnostic yield of a single induced sputum is as good as fiberoptic bronchoscopy (3, 5), and yield of repeated inductions is better (6). Patient tolerability and safety are excellent (3–8), and the procedure is feasible in developing country settings (4, 7), and for young children (8) In this issue of the AJRCCM (pp. 723–725), Conde and colleagues (9) report a remarkably high yield of mycobacterial culture in a single specimen of induced sputum in 84 patients with tuberculous pleural effusions. Mycobacterial cultures were positive in induced sputum specimens from 44 (52%) patients compared with 10 (12%) pleural fluid specimens and 52 (62%) pleural biopsies (9). Induced sputum specimens provided bacteriologic confirmation in 55% of patients with normal lung parenchyma on chest X-ray, and the yield of sputum induction was independent of HIV status. This carefully performed and highly original study has a number of implications for public health and clinical practice. Traditional dogma has been that patients with tuberculous pleural effusions without concomitant pulmonary disease were not contagious. The finding in this study, that induced sputa were smear-positive in 12% of all patients with tuberculous pleurisy and culture-positive in more than half of the patients with normal lung parenchyma (9), challenges this traditional thinking. These patients should be considered contagious and, at minimum, their household contacts should be investigated. Until recently, patients with smear-negative pulmonary disease had also been considered non-contagious. This, however, was convincingly refuted by a molecular epidemiologic study showing that patients with smear-negative pulmonary TB contributed importantly to TB transmission in San Francisco (10). A molecular epidemiologic study to define the contagiousness of patients with pleural TB, and the population impact of transmission from them, would be of great interest. Could these striking findings have resulted from bias or unusual patient characteristics, and are the results applicable in low-incidence settings? Bias seems most unlikely, as Conde and colleagues (9) performed a thorough, prospective study. All patients underwent a battery of investigations, providing complete data on all patients, and diagnoses were confirmed microbiologically or from independent radiographic review. Higher yield could have occurred if many study subjects had recent contact, HIV infection, or primary TB—all associated with positive sputum cultures, but normal chest X-rays. HIV status was known in all patients, and yield was high even in HIV sero-negative individuals. Approximately 40% of HIV-negative patients had a history of contact in the preceding 3 years, suggesting that primary infection could have contributed to the higher yield, at least in some patients. The bacteriologic yield from pleural fluid and biopsies and the histologic yield from pleural biopsies were similar to earlier reports (11), suggesting their patients' severity of illness was similar. The applicability of these results in low-incidence countries is best addressed by independent replication in such settings—an important and needed study. In the meantime, what are the clinical implications? In the investigation of a pleural effusion, the usual first step is thoracentesis, followed by pleural biopsy if analysis of the fluid is nondiagnostic (11). Bacteriologic confirmation of TB effusions from pleural fluid is relatively low, both in Conde's experience (9) and elsewhere (11). On the other hand, pleural biopsy will provide histologic proof in 60–80% of cases, and bacteriologic confirmation in a similar proportion (9, 11). This approach is appropriate in low-incidence settings where TB is much less likely than other causes that may also be diagnosed with these techniques. These invasive techniques, however, can have major complications. Although estimates vary, pneumothorax requiring chest tube can occur in 1–5% (11, 12), major bleeding in a similar proportion (11), and mortality can be as high as 1% (11). A very important advantage of sputum induction is its excellent record of safety. In this (9) and six other reports published since 1995 (3–8), sputum induction was well tolerated without major complications by a total of 2,844 patients. The only potential adverse effect is bronchospasm in patients with asthma or COPD, which can be prevented by routine administration of a bronchodilator before induction, and/or use of a less concentrated saline solution. In resource-poor countries, sputum induction may be particularly helpful because TB is a common cause of pleural effusions, and invasive diagnostic procedures, such as pleural biopsy, are less accessible. But will clinicians in industrialized countries choose sputum induction over more invasive techniques? Selection of treatment for many respiratory disorders is increasingly guided by evidence-based medicine. If these criteria were applied for diagnostic procedures, sputum induction would be strongly favored over bronchoscopy for initial investigation of pulmonary TB, as yield of sputum induction has been consistently equivalent or better (3, 5, 6). Sputum induction may also be favored for lung cancer, where yield was found to be equivalent to bronchoscopy in one series (13). Despite these advantages, the idea that a less complex procedure has better diagnostic yield may be slow to gain acceptance. Although unpopular to suggest, another disincentive for sputum induction may be remuneration, which has been shown to influence performance of bronchoscopy for diagnosis of Pneumocystis carinii pneumonia (14). Now, in light of Conde's article (9), perhaps it is time to review and reassess our diagnostic approach to suspected tuberculous pleural effusion. Yield of induced sputum is high, but a major drawback is the delay in confirmation. This must be balanced against the superior patient comfort and safety, particularly when risks of invasive techniques are increased in patients with small or loculated effusions, or with bleeding diatheses (11). In such patients, and those in whom TB is strongly suspected, prudent clinicians should consider sputum induction for the initial investigation of undiagnosed pleural effusions. REFERENCES
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
