Search tips
Search criteria 


Logo of jcmPermissionsJournals.ASM.orgJournalJCM ArticleJournal InfoAuthorsReviewers
J Clin Microbiol. 2011 March; 49(3): 772–776.
PMCID: PMC3067741

Role of the Clinical Mycobacteriology Laboratory in Diagnosis and Management of Tuberculosis in Low-Prevalence Settings[down-pointing small open triangle]


Tuberculosis (TB) remains a global epidemic, despite a significant decline in reported cases in the United States between 2008 and 2009. While the exact nature of this decline is unclear, one thing remains certain: TB, including multidrug-resistant (MDR) and extensively drug-resistant (XDR) TB, is no longer restricted to developing regions of the globe. It is of vital importance that both public and private mycobacteriology laboratories maintain the ability to detect and identify Mycobacterium tuberculosis from patient specimens, as well as correctly determine the presence of antibiotic resistance. To do this effectively requires careful attention to preanalytical, analytical, and postanalytical aspects of testing. Respiratory specimens require digestion and concentration followed by fluorescence microscopy. The Centers for Disease Control and Prevention (CDC) recommends the performance of a direct nucleic acid amplification method, regardless of smear results, on specimens from patients in whom the suspicion of tuberculosis is high. Liquid-based technologies are more rapid and sensitive for the detection of M. tuberculosis in culture and nucleic acid probes, but biochemicals are preferred for identification once growth is detected. Susceptibility testing is most often done using either the agar proportion method or a commercial broth system. New genotypic and phenotypic methods of susceptibility testing include first- and second-line agents and are promising, though not yet widely available. Finally, gamma interferon release assays are preferred to the tuberculin skin test for screening certain at-risk populations, and new CDC guidelines are available that assist clinicians in their use.

Tuberculosis (TB) remains a global scourge, despite a significant recent decline in the number of reported cases in the United States (4). Between 2008 and 2009, the TB case rate in the United States decreased 11.4%, from 4.2 to 3.8 cases per 100,000 population (4). This decline is notable in that it represents the largest decrease in a single year and the lowest rate since the advent of a national surveillance program in 1953 (4). So “unusual and unexpectedly large” is this decline that an investigation has been initiated by the CDC and the National Tuberculosis Controllers Association (4) to determine whether the decrease represents an actual reduction in TB cases due to better control or is an artifact created by population/demographic shifts related to the economic downturn or underdiagnosis or underreporting of cases. While the cause of this decline has yet to be elucidated, one thing is certain. This age-old disease, including multidrug-resistant (MDR) and extensively drug-resistant (XDR) TB, is no longer restricted to developing regions of the globe. People and the disease are mobile. In addition, roughly one-third of the world's population is latently infected with Mycobacterium tuberculosis, an estimated 14 million of whom reside in the United States (1, 7). These latently infected individuals serve as a repository for future reactivation cases. Thus, it is of vital importance that both public and private mycobacteriology laboratories maintain the ability to detect and identify M. tuberculosis from patient specimens, as well as to determine the presence of antibiotic resistance. Recently, these efforts have been affected to various degrees by budget cuts, declining proficiency in the wake of the decreased numbers of clinical specimens, and staffing shortages resulting from the closure of training programs and the retirement of experienced personnel. And yet, appropriate TB treatment and prevention of transmission rests on rapid and dependable laboratory results. This minireview is designed to offer some general testing guidelines for mycobacteriology laboratories using currently available technologies in settings in which the prevalence of TB is low relative to the increasing numbers of clinically relevant nontuberculous Mycobacterium species (NTM).


The confirmation of a single case of TB often involves a complex network of laboratories performing a wide range of tests. This is because multiple methods are required to recover, identify, and determine drug resistance for mycobacteria, including M. tuberculosis, not all of which are available in every laboratory. Existing methods for detection range from simple smear microscopy and slow culture methods to advanced, costly, or technically demanding molecular assays. Average turnaround times for reporting results, even in state-of-the-art laboratories, are often measured in days to weeks. In the United States, roughly 2,000 TB laboratories exist at the local and state levels (3). These laboratories are within both the public and private sectors. Communication between these laboratories is an essential component of TB control. Of the 50 state public health laboratories, most serve as the primary reference center for culture, identification, and susceptibility testing. This function is also carried out by a limited number of medical centers and commercial laboratories (22, 24). It is estimated that more than 80% of initial laboratory testing for TB is conducted in the private sector. Initial laboratory testing in this case refers only to microscopic smear and inoculation of culture. Subsequent species identification and susceptibility testing is predominantly (≥50%) done by public health laboratories (22). This complicated network of specimen and culture referrals to larger laboratories often leads to delays in reporting that can be detrimental for both patients and TB control programs. However, the very existence of this complex, interconnected web is necessary to maintain the critical services and range of tests required for the diagnosis of TB (20, 21). These services/tests are summarized in Fig. 1 and will be covered briefly in the subsequent sections. They include the following: (i) prompt delivery of specimens to the laboratory, preferably in less than 24 h; (ii) use of the most rapid currently available methods for detection and identification; (iii) reporting of smear results within 24 h; (iv) culture identification of the M. tuberculosis complex within 21 days; (v) drug susceptibility results within 30 days; and (vi) reporting of all positive results within 24 h from the date of report (20, 21).

Fig. 1.
Recommended testing algorithm for TB laboratories. Please note, MB/BacT ALERT 3D is not FDA approved for susceptibility testing of 1st-line drugs.


Specimen procurement, transport, and processing.

Proper specimen collection and transport are essential for the submission of an adequate and appropriate specimen for the diagnosis of TB. In general, inappropriate specimens include those which (i) have insufficient volume, (ii) consist of pooled samples, (iii) arrive in broken containers, (iv) exceed the optimal time between collection and processing (>7 days), (v) consist of saliva rather than sputum, or (vi) have been sent on swabs (17). Swabs are of particular concern since they present a problem in digestion/decontamination and the hydrophobic components of the mycobacterial cell wall interfere with transfer of the bacilli to culture media. In addition, care should be taken to prevent specimen contamination with sources of environmental mycobacteria (e.g., tap water), which can confound both smear and culture results. All specimens should be transported to the laboratory in less than 24 h and should be maintained at 4°C if transport takes more than an hour. Decontamination is required for all specimens from nonsterile body sites due to the slow growth of the mycobacteria, which are easily overgrown by other microorganisms. The most widely used digestion-decontamination method consists of N-acetyl-l-cysteine (NALC) and 2% sodium hydroxide. (17). Although all decontamination methods are somewhat toxic to mycobacteria, this method is mild enough not to kill all of the acid-fast bacilli (AFB) present while providing for adequate inhibition of contaminating organisms. In addition, this method is compatible with several commercially available automated broth culture detection systems. With respect to sputum, after initial processing, concentration by centrifugation or sedimentation has been shown to increase the sensitivity of smear microscopy by as much as 33%, resulting in a recommendation that this be adapted as a global standard (19).


Microscopic examination of clinical sputum specimens has been the mainstay of TB case detection for over 100 years. As such, smear microscopy is the most widely utilized test for active TB and is central to most control strategies. Microscopy is rapid, fairly inexpensive, and less labor intensive than other technologies (18). In addition, in the case of TB, a positive smear is indicative of a high risk of transmitting infection to others. Recent technological improvements have increased both the speed and sensitivity of the standard microscopic smear and include changes in microscope design and microbial stains. Although traditional light microscopy has been used extensively for the detection of pulmonary TB, fluorescence microscopy provides a 10%-more-sensitive option (18). In the past, widespread use of fluorescence microscopy has been hindered by the costs associated with purchase of the required microscopes and the need for frequent replacement of UV lamps. This is changing with the development of fluorescence microscopes which require light-emitting diodes rather than UV lamps, resulting in significantly reduced equipment and maintenance costs (15, 16, 18). In addition, a more rapid fluorescent stain, the auramine-O (AO) stain, has been developed which minimizes the staining of background debris while maximizing the staining of AFB. This stain (Scientific Device Laboratories, Des Plaines, IL) can be completed in 2 minutes, versus 22 minutes for the standard AO stain, thus reducing the time of the technologist in preparation and improving turnaround time in reporting results (8).


Culture remains the most sensitive method for the detection of M. tuberculosis in clinical specimens. Isolation in culture provides for all subsequent testing, including identification and antibiotic susceptibility testing. Culture is capable of detecting as few as 101 to 102 organisms/ml of specimen, surpassing the sensitivity of smear (≤104 CFU/ml). While a number of culture options exist for clinical mycobacteriology laboratories, broth-based detection methods are preferable for the initial isolation of organisms from specimens. Recovery in broth is significantly more rapid (average time, 10 to 14 days), and isolation is better than on solid media (6). Though manual broth culture systems exist, automated systems have been developed that are continuously monitored, such as the Bactec 9000MB (Becton Dickinson, Sparks, MD), the Bactec mycobacterial growth indicator tube 960 (MGIT960; Becton Dickinson), the ESP culture system II (Trek Diagnostic Systems; Cleveland, OH), and the MB/BacT Alert 3D (bioMérieux, France) (6). Although the performance and operational characteristics of these systems are similar, all are subject to higher bacterial contamination rates than the semiautomated Bactec 460 TB (Becton Dickinson), which is being phased out, or conventional agar-based methods. None of these methods can provide a definitive identification of the mycobacterial species present or indication of a mixed mycobacterial culture. Thus, parallel cultures on solid media will provide confirmation of a single colonial morphology. Ancillary tests, such as nucleic acid probes, DNA sequencing, and high-performance liquid chromatography (HPLC) analysis of mycolic acids are required to differentiate M. tuberculosis from other, nontuberculous species. These methods shorten the time to identification but add additional complexity and cost to the testing algorithm.

Molecular detection.

Molecular detection of M. tuberculosis continues to change the landscape of TB diagnostics. Two nucleic acid amplification techniques (NAATs) have been approved by the U.S. Food and Drug Administration (FDA) for use in the United States: the Gen-Probe Amplified MTD (M. tuberculosis direct) test (Gen-Probe, San Diego, California) and the Roche Amplicor MTB test (Roche Molecular Systems, Branchburg, NJ). Both tests are approved for use with smear-positive respiratory specimens; the Gen-Probe assay is also approved for use with smear-negative pulmonary specimens. Other currently available commercial kits include the BD-ProbeTec direct (Becton Dickinson, Sparks, MD) and the COBAS Amplicor (Roche Diagnostic Systems, Mannheim, Germany) (13). None of these methods has been approved for direct detection of M. tuberculosis from extrapulmonary specimens. Although all of these current technologies are rapid and have demonstrated excellent specificity, their performance characteristics can vary and their sensitivity still does not equal that of culture-based methods, especially for smear-negative samples. A recent meta-analysis was conducted which examined over 125 studies looking at primarily smear-positive and -negative samples. Although this study reported pooled sensitivities and specificities of 85% and 97%, respectively, significant variability in sensitivity (range, 36% to 100%) and specificity (range, 54% to 100%) was found between studies (11, 13). NAATs are costly to perform, often require expensive instrumentation which must be maintained, and require technical expertise and a relatively high test volume to be cost effective. Currently, the Centers for Disease Control and prevention (CDC) recommends that a NAAT be performed on the first sputum sample of all TB suspects regardless of smear status (13). While this recommendation was made in the interest of public health, clinical laboratories are often unaware of which patients are clinically at high risk for TB. Routine implementation of NAAT testing on all first specimens submitted for mycobacterial culture without consideration of the clinical history seems impractical and costly. Clinical microbiology laboratories will need to work with public health laboratories and their medical staff to ensure that the above recommendations are carried out in a fashion that guarantees testing for patients at risk of having TB while being sensitive to both the incurred costs and potential poor performance of these assays were these tests to be implemented routinely in low-prevalence populations.

Susceptibility testing.

Most commercially available automated broth culture detection systems (Bactec MGIT960 [Becton Dickinson] and ESP culture system II [Trek Diagnostic Systems]) have been FDA approved for susceptibility testing of first-line antibiotics (isoniazid, rifampin, ethambutol, and pyrazinamide), with an average of 2 to 4 weeks from time of inoculation to reporting of susceptibility results (14). Most often, the “indirect susceptibility method” is used, in which growth is first isolated in pure culture from clinical specimens, with subsequent inoculation of drug-containing broth. The extension of these commercial systems for “direct susceptibility testing” of smear-positive specimens would potentially reduce reporting turnaround times to 1 to 3 weeks (14). However, direct testing of clinical specimens is problematic due to the potential for bacterial contaminants and other nontuberculous species, resulting in assay failure rates of ~15%. For this reason, most laboratories rely on the indirect method for susceptibility testing. The use of these systems has not been cleared by the FDA for susceptibility testing of second-line drugs, and thus, most laboratories rely on agar proportion as the reference standard. This method is labor intensive and slow, requiring several weeks for completion. Recently, in a multicenter study, Lin and coworkers demonstrated that susceptibility testing of second-line drugs with the Bactec MGIT960 system could be standardized, producing reliable results in a fraction of the time required for agar proportion, possibly paving the way for an alternative, more rapid method (10).


Newer technologies currently in development will undoubtedly redefine the way in which TB is diagnosed, providing the future essentials for the diagnostic laboratory. The development of these technologies has been fueled by the increasing prevalence of MDR and XDR TB, which threaten global TB control. MDR TB is defined as resistance to two key first-line agents, rifampin (RIF) and isoniazid (INH). Rifampin resistance is associated with resistance to other drugs, such as INH, especially in high-burden regions of the globe (13). For this reason, development efforts by multiple companies have centered on the rapid detection of RIF and/or INH resistance from clinical samples or culture. Such advances include the XpertMTB/RIF (GeneXpert system; Cepheid, Sunnyvale, California), which is an automated molecular test for TB case detection and RIF resistance. This test uses a heminested real-time PCR to amplify a specific region within the RIF-resistance-determining region (RRDR) of the rpoB gene (2). In a recent study funded through a public-private partnership consisting of Cepheid, the Foundation for Innovative New Diagnostics (FIND), the National Institutes of Health, and the Bill and Melinda Gates Foundation, the XpertMTB/RIF demonstrated sensitivities of 98.2% (551/561) and 72.5% (124/171) with smear-positive and -negative TB, respectively. The test's specificity was 99.2% (604/609) in patients without TB. This assay requires a single manual step with minimal sample manipulation. The remaining analysis is completely performed by the GeneXpert instrument in as little as 2 h (2).

Other commercially available molecular tests for the rapid detection of RIF and INH resistance include the INNO-LiPARif.TB (Innogenetics, Ghent, Belgium) and the GenoTypeMTBDR (Hain LifeScience, Nehren, Germany). Both tests are line probe assays which utilize nucleic acid amplification technology to detect M. tuberculosis and drug resistance simultaneously (13). The sensitivity and specificity of the INNO-LipARif.TB assay vary depending on whether clinical specimens or cultured isolates are tested. With cultured isolates, the test's sensitivity ranges from 82% to 100%, but it decreases slightly (80% to 100%) with direct testing of clinical specimens (12). The specificity remains 100% with both types of samples (12). The GenoTypeMTBDRplus assay detects both rifampin and isoniazid resistance. This assay demonstrated a pooled sensitivity of 98.4% and a specificity of 98.9% for rifampin. However, the pooled sensitivity for isoniazid ranged from 84% to 88.7% with clinical specimens and cultured isolates, respectively (11). In a study conducted by Lacoma and coworkers, the sensitivity for isoniazid was 73%, indicating that the detection of resistance to this particular drug may be more problematic, due to the involvement of multiple genes/mutations (9). Although the World Health Organization (WHO) has endorsed these tests for use with smear-positive clinical specimens and culture, neither test is currently FDA approved for use in the United States (13).

Advances in phenotypic susceptibility methods have also been made. In a recent multicenter study, the new Sensititre MycoTB MIC plate (Trek Diagnostic Systems, Cleveland, OH) was evaluated for susceptibility testing of M. tuberculosis (23). This study compared the MycoTB plate, a broth microdilution method, to modified agar proportion (APM), which was used as the reference standard. As a phenotypic test, the MycoTB plate is unique in a number of ways: (i) it includes both first- and second-line drugs in a single platform, with the exception of pyrazinamide (INH, RIF, streptomycin, ethambutol, para-aminosalicylic acid, cycloserine, ofloxacin, moxifloxacin, rifabutin, amikacin, kanamycin, and ethionamide); (ii) it incorporates the critical concentration of each drug currently used to determine susceptibility or resistance, along with a range of concentrations for each drug to permit determination of an MIC (MIC); and (iii) it can be completed in an average of 7 to 10 days, compared with 14 to 21 days for APM. In this initial study, the overall agreement between the MycoTB and APM was 96.7%, indicating that this test is capable of producing reliable results compared with those of the reference standard and may provide for a standardized, single testing platform for the determination of phenotypic drug resistance in M. tuberculosis (23).

Other technologies exist which are currently in development and, therefore, beyond the scope of this review (15, 16). However, these tests may provide for a new wave of diagnostic improvements for TB in the future (15, 16).


Three gamma interferon release assays (IGRAs) have been approved by the FDA and are commercially available in the United States: the QuantiFERON-TB gold (QFT-G; Cellestis, Victoria, Australia), the QuantiFERON-TB gold in-tube test (QFT-GIT; Cellestis), and the T-Spot.TB test (T-Spot; Oxford Immunotec, Oxford, United Kingdom). The CDC recently endorsed the use of IGRAs “in place of (but not in addition to) a TST (tuberculin skin test)” in situations for which the TST is currently indicated, with some important exceptions, such as severely immunocompromised patients or the very young (5, 13). When considering the use of these tests, it must be remembered that the results may not be directly comparable between methods, since different antigens and interpretive criteria are utilized (5). IGRAs are especially useful in Mycobacterium bovis BCG-vaccinated individuals or those with a low probability of returning to have a TST read. The specific recommendations established by the CDC take into consideration a number of situations which can be used to guide decisions about when to use IGRAs versus TST. These include: (i) situations in which an IGRA is preferred but a TST is acceptable (i.e., those unlikely to return to have a TST read or BCG-vaccinated persons); (ii) situations in which a TST is preferred but an IGRA is acceptable (i.e., children <5 years old); (iii) situations in which either a TST or IGRA may be used without preference (i.e., recent contacts of active TB cases or periodic screening of health care workers); and (iv) situations in which both an IGRA and a TST may be considered. Although the latter is not generally recommended, situations may arise in which the initial test result (either IGRA or TST) is inconclusive or inconsistent with the overall clinical scenario (5). In such situations, the use of the alternative test may be warranted.

Despite these recommendations, one key disadvantage of IGRAs exists, the requirement to process blood samples within hours of collection to ensure the viability of the white cells present (5). This requirement may limit the implementation of these tests to laboratories in close proximity to patient blood drawing services. Laboratories whose location is too far removed from such services would necessarily be excluded from such testing (5).


Unfortunately, no laboratory diagnostic is one hundred percent foolproof one hundred percent of the time. Phenotypic, molecular, and immunologic tests are no exception. For instance, molecular detection of resistance depends ultimately on the presence of the resistance-conferring mutation. Alternate mechanisms of resistance may develop or mutations may appear which the test was not designed to detect. In such a case, a phenotypic test is required to identify drug resistance. And yet, phenotypic detection of resistance is subject to contamination and is very slow. In both cases, assay failure can result. In the end, given the current state of TB diagnostics, a combination of multiple tests that include smear using fluorescence microscopy, liquid culture methods, and molecular detection to rapidly distinguish M. tuberculosis from NTM is necessary to provide reliable and accurate results for patient care.


[down-pointing small open triangle]Published ahead of print on 22 December 2010.


1. Bennett D. E., et al. 2008. Prevalence of tuberculosis infection in the United States population: the national health and nutrition examination survey, 1999-2000. Am. J. Respir. Crit. Care Med. 177:348–355 [PubMed]
2. Boehme C. C., et al. 2010. Rapid molecular detection of tuberculosis and rifampin resistance. N. Engl. J. Med. 363:1005–1015 [PMC free article] [PubMed]
3. Centers for Disease Control and Prevention 2005. National plan for reliable tuberculosis laboratory services using a systems approach. MMWR Recommend. Rep. 54(RR-6):1–12 [PubMed]
4. Centers for Disease Control and Prevention 2010. Decrease in reported tuberculosis cases—United States, 2009. MMWR Morb. Mortal. Wkly. Rep. 59:289–294 [PubMed]
5. Centers for Disease Control and Prevention 2010. Updated guidelines for using interferon gamma release assays to detect Mycobacterium tuberculosis infection—United States, 2010. MMWR Recommend. Rep. 59(RR-5):1–13 [PubMed]
6. Dinnes J., et al. 2007. A systematic review of rapid diagnostic tests for the detection of tuberculosis infection. Health Technol. Assess. 11:1–196 [PubMed]
7. Dye C., Scheele S., Dolin P., Pathania V., Raviglione M. C. 1999. Consensus statement: global burden of tuberculosis—estimated incidence, prevalence, and mortality by country. WHO Global Surveillance and Monitoring Project. JAMA 282:677–686 [PubMed]
8. Hendry C., Dionne K., Hedgepeth A., Carroll K., Parrish N. 2009. Evaluation of a rapid fluorescent staining method for detection of Mycobacteria in clinical specimens. J. Clin. Microbiol. 47:1206–1208 [PMC free article] [PubMed]
9. Lacoma A., et al. 2008. GenoType MTBDRplus assay for molecular detection of rifampin and isoniazid resistance in Mycobacterium tuberculosis strains and clinical samples. J. Clin. Microbiol. 46:3660–3667 [PMC free article] [PubMed]
10. Lin S. Y., Desmond E., Bonato D., Gross W., Siddiqi S. 2009. Multicenter evaluation of Bactec MGIT 960 system for second-line drug susceptibility testing of Mycobacterium tuberculosis complex. J. Clin. Microbiol. 47:3630–3634 [PMC free article] [PubMed]
11. Ling D. I., Flores L. L., Riley L. W., Pai M. 2008. Commercial nucleic-acid amplification tests for diagnosis of pulmonary tuberculosis in respiratory specimens: meta-analysis and meta-regression. PLoS One 3:e1536. [PMC free article] [PubMed]
12. Morgan M., Kalantri S., Flores L., Pai M. 2005. A commercial line probe assay for the rapid detection of rifampin resistance in Mycobacterium tuberculosis: a systematic review and meta-analysis. BMC Infect. Dis. 5:62. [PMC free article] [PubMed]
13. Nyendak M. R., Lewinsohn D. A., Lewinsohn D. M. 2009. New diagnostic methods for tuberculosis. Curr. Opin. Infect. Dis. 22:174–182 [PubMed]
14. Parrish N. M., Carroll K. 2008. Importance of improved TB diagnostics in addressing the extensively drug-resistant TB crisis. Future Microbiol. 3:405–413 [PubMed]
15. Perkins M. D., Cunningham J. 2007. Facing the crisis: improving the diagnosis of tuberculosis in the HIV era. J. Infect. Dis. 196:S15–S27 [PubMed]
16. Palomino J. C. 2006. Newer diagnostics for tuberculosis and multi-drug resistant tuberculosis. Curr. Opin. Pulm. Med. 12:172–178 [PubMed]
17. Pfyffer G. E. 2007. Mycobacterium: general characteristics, laboratory detection, and staining procedures, p. 543–572 In Murray P. R., Baron E. J., Jorgensen J. H., Landry M. L., Pfaller M. A., editors. (ed.), Manual of clinical microbiology, 9th ed ASM Press, Washington, DC
18. Steingart K. R., et al. 2006. Fluorescence versus conventional sputum smear microscopy for tuberculosis: a systematic review. Lancet Infect. Dis. 6:570–581 [PubMed]
19. Steingart K. R., et al. 2006. Sputum processing methods to improve the sensitivity of smear microscopy for tuberculosis: a systematic review. Lancet Infect. Dis. 6:664–674 [PubMed]
20. Styrt B. A., Shinnick T. M., Ridderhof J. C., Crawford J. T., Tenover F. C. 1997. Turnaround times for mycobacterial cultures. J. Clin. Microbiol. 35:1041–1042 [PMC free article] [PubMed]
21. Tenover F. C., et al. 1993. The resurgence of tuberculosis: is your laboratory ready? J. Clin. Microbiol. 31:767–770 [PMC free article] [PubMed]
22. Tokars J. I., et al. 1996. U.S. hospital mycobacteriology laboratories: status and comparison with state public health department laboratories. J. Clin. Microbiol. 34:680–685 [PMC free article] [PubMed]
23. Wengenack N., Hall L., LaBombardi V., Parrish N. 2010. Evaluation of the new Sensititre (MYCOTB) MIC plate for the susceptibility testing of Mycobacterium tuberculosis (Mtb) to first and second line drugs, abstr. D-1508/169. Abstr. 50th Intersci. Conf. Antimicrob. Agents Chemother. (ICAAC), American Society for Microbiology, Washington, DC
24. Woods G. L., Long T. A., Witebsky F. G. 1996. Mycobacterial testing in clinical laboratories that participate in the College of American Pathologists Mycobacteriology Surveys: changes in practices based on responses to 1992, 1993, and 1995 questionnaires. Arch. Pathol. Lab. Med. 120:429–435 [PubMed]

Articles from Journal of Clinical Microbiology are provided here courtesy of American Society for Microbiology (ASM)