American Journal of Respiratory and Critical Care Medicine

Rationale: The World Health Organization recently revised its recommendations for tuberculosis (TB) diagnosis in people with HIV. Most studies cited to support these policies involved HIV-uninfected patients and only evaluated sputum specimens.

Objectives: To evaluate the performance of acid-fast bacilli smear and mycobacterial culture on sputum and nonsputum specimens for TB diagnosis in a cross-sectional survey of HIV-infected patients.

Methods: In Thailand and Vietnam, we enrolled people with HIV regardless of signs or symptoms. Enrolled patients provided three sputum, one urine, one stool, one blood, and, for patients with palpable peripheral adenopathy, one lymph node aspirate specimen for acid-fast bacilli microscopy and mycobacterial culture on solid and broth-based media. We classified any patient with at least one specimen culture positive for Mycobacterium tuberculosis as having TB.

Measurements and Main Results: Of 1,060 patients enrolled, 147 (14%) had TB. Of 126 with pulmonary TB, the incremental yield of performing a third sputum smear over two smears was 2% (95% confidence interval, 0–6), 90 (71%) patients were detected on broth-based culture of the first sputum specimen, and an additional 21 (17%) and 12 (10%) patients were diagnosed with the second and third specimens cultured. Of 82 lymph nodes cultured, 34 (42%) grew M. tuberculosis. In patients with two negative sputum smears, broth-based culture of three sputum specimens had the highest yield of any testing strategy.

Conclusions: In people with HIV living in settings where mycobacterial culture is not routinely available to all patients, a third sputum smear adds little to the diagnosis of TB. Broth-based culture of three sputum specimens diagnoses most TB cases, and lymph node aspiration provides the highest incremental yield of any nonpulmonary specimen test for TB.

Scientific Knowledge on the Subject

The World Health Organization recently revised its recommendations for tuberculosis (TB) diagnosis in people with HIV. Most studies cited to support these policies involved HIV-uninfected patients and only evaluated sputum specimens.

What This Study Adds to the Field

In people with HIV living in settings where mycobacterial culture is not routinely available to all patients, a third sputum smear adds little to the diagnosis of TB. Broth-based culture of three sputum specimens diagnoses most TB cases, and lymph node aspiration provides the highest incremental yield of any nonpulmonary specimen test for TB.

Tuberculosis (TB) is the leading cause of death of persons with HIV infection, resulting in an estimated 700,000 illnesses and 200,000 deaths worldwide in 2008 (1). Although over 25 years have passed since the TB/HIV syndemic first emerged, basic components of TB disease control, such as diagnosis, remain challenging (2). The most widely used TB diagnostic test, microscopic examination of sputum for acid-fast bacilli (AFB), is highly insensitive in HIV-infected patients, and modern broth-based culture methods are not routinely available in most high-burden TB countries (3). Extrapulmonary TB occurs frequently and causes high mortality, but definitive diagnosis requires culture or pathological examination of specimens that may be difficult to obtain (4).

To improve diagnosis and accelerate the development of TB laboratory diagnostic services for people with HIV, the World Health Organization (WHO) released new guidelines for TB diagnosis in 2006 and 2007. These guidelines called for mycobacterial sputum culture to be performed in HIV-infected patients who are sputum smear negative and have a clinical suspicion of TB, reduced from three to two the number of sputum specimens that need to be examined before calling a TB suspect “smear negative,” and provided simplified guidelines for diagnosing common types of extrapulmonary TB (5, 6). Although strongly supported by expert opinion, the WHO's recent recommendations have important evidence gaps. The recommendation to perform two, rather than three, sputum smears was based on a systematic review of studies that involved very few patients with HIV infection (7). Three studies that included HIV-infected patients either did not use modern broth-based culture methods or evaluated small numbers of cases (810). The WHO's recommendation to perform culture in HIV-infected, smear-negative TB suspects did not specify how many specimens to culture and what method to use, both of which could greatly affect yield (11). The WHO's guidelines for diagnosing extrapulmonary TB recommend using symptoms or signs to guide testing of extrapulmonary sites, even though multiple case series have documented the yield of lymph node aspiration (1218) or blood culture (1924) for diagnosing TB, regardless of the primary site of disease.

To evaluate current international recommendations, we comprehensively evaluated the TB diagnostic performance of AFB smear and culture on sputum and nonsputum specimens in a large cross-sectional survey of HIV-infected patients from Southeast Asia, a region greatly affected by the TB/HIV syndemic (1). We sought to answer the following questions: (1) In HIV-infected patients, how many sputum specimens should be examined for AFB to diagnose TB, and how many specimens should be cultured to diagnose TB? (2) Which mycobacterial culture technique provides the best recovery of Mycobacterium tuberculosis (MTB) from sputum? (3) When sputum smear and culture are negative, what is the yield of testing nonsputum specimens?

Enrollment and Specimen Collection

The WHO guidelines recommend screening HIV-infected patients for TB during encounters with the health system; therefore, we enrolled patients from outpatient health facilities that perform HIV counseling, testing, and clinical care. One clinic in Bangkok, Thailand, and three clinics in Ho Chi Minh City, Vietnam, participated. In Bangkok, the study was conducted at a large, outpatient HIV counseling and testing center run by a nongovernmental organization; health facility staff sequentially recruited patients who were first diagnosed with HIV or who presented to the laboratory for CD4+ T-lymphocyte (CD4) testing. In Ho Chi Minh City, the study was conducted at three HIV outpatient clinics run by the city government; health facility staff sequentially recruited patients first presenting to the clinics for HIV care, most of whom were recently diagnosed with HIV and a small proportion of whom were transferring their care from a previous facility. In all facilities, patients were recruited regardless of the presence or absence of symptoms or prior suspicion of TB. Patients were eligible for the study if they were documented to be HIV infected, were more than 6 years old, were not currently being treated for TB, had not received treatment for TB disease or infection in the past year, had not undergone TB screening with chest radiography or sputum smears in the previous 3 months, and had taken no medications with anti-TB activity within the past month.

After providing written informed consent, patients underwent a standardized interview and physical examination, chest radiography, and blood testing for complete blood cell count and CD4. Patients also provided three sputum specimens over 2 days, including a “spot” specimen collected on the first day (first sputum) and a morning (second sputum) and “spot” specimen (third sputum) collected on the second day. Patients also provided one specimen each of urine, blood, and stool. If patients had enlarged peripheral lymph nodes (>1 cm for noninguinal nodes; >2 cm for inguinal nodes), the largest palpable node was aspirated. Specimens of all types underwent microscopy and culture regardless of the findings from previous specimens examined. Specimens were transported to one reference laboratory in each country within 48 hours of collection.

Sputum Specimen Processing

Sputum was processed using methods recommended by the United States Centers for Disease Control and Prevention (USCDC), including specimen volume adjusted to 10 ml; equal volume of N-acetyl-l-cysteine w/NaOH-citrate (final concentration of NaOH 1%) added; specimen mixed well and incubated for 15 to 20 minutes at room temperature; PBS (pH 6.8) added up to the 45-ml mark; solution mixed well and centrifuged (3,000 × g at 4°C for 15–20 minutes); specimens decanted and resuspended in 1 to 2 ml PBS (pH 6.8); suspension mixed thoroughly and directly inoculated to solid media (Lowenstein-Jensen [LJ] × 2 (150–200 μl each) and broth-based media (Mycobacterial Growth Indicator Tube [MGIT] × 1 [500 μl]) (BD, Franklin Lakes, NJ).

Nonsputum Specimen Processing

Nonsputum specimens were handled similarly to sputum specimens with the following exceptions. Blood from patients was inoculated directly into Myco-F-Lytic bottles (BD, Franklin Lakes, NJ) and placed into an automated blood culture instrument (BACTEC 9050/9120 system; BD). Lymph node aspirates were directly inoculated into an MGIT tube; if sufficient specimen volume remained after inoculation into the MGIT tube, an LJ culture was inoculated, and a smear was prepared. Stool was prepared for decontamination by emulsifying 1 g in 10 ml of sterile water and sterile glass beads followed by filtering. Urine was concentrated by centrifugation, decanted, and adjusted to 10 ml using sterile water.

Smear Microscopy

After specimen processing and culture inoculation, smears were prepared, stained (Ziehl-Neelsen), and graded according to WHO recommendations. The actual number of AFB observed on smear was documented for all smears having 1 to 9 AFB per 100 fields. Smears were then documented and reported using the following scale: negative (no AFB seen per 100 fields), 1+ (10–99 AFB per 100 fields), 2+ (1–10 AFB per field in at least 50 fields), and 3+ (>10 AFB in at least 20 fields). Smears documented as having 1 to 3 AFB per 100 fields were recoded as negative for the purpose of analysis based on the known low correlation of 1 to 3 AFB and positive cultures (2527). Smears with 4 to 9 AFB per 100 fields and those reported as 1+, 2+, or 3+ were classified as positive.

Mycobacterial Culture

LJ cultures were evaluated twice within the first week of inoculation and then once per week for 42 days. Cultures with growth were confirmed as AFB positive by visual inspection or smear. MGIT tubes and MycoF-Lytic bottles were incubated for 42 days in the BACTEC MGIT 960 and BACTEC 9050/9120, respectively. Cultures flagged as positive by the instruments were removed. AFB smears were performed, and the specimens were subcultured on a blood agar plate (BAP) to check contamination. AFB-positive cultures were subcultured onto two LJ; AFB-positive but contaminated cultures were redecontaminated and subcultured onto two LJ; and cultures less than 42 days old with no organisms present on smear and BAP were returned to the instrument. AFB-negative and contaminated MGIT cultures were documented and discarded. AFB-negative MycoF-Lytic cultures that had growth on BAP underwent bacterial or fungal identification. All MGIT tubes and MycoF-Lytic cultures were removed after 42 days, visually inspected for growth, and discarded. Blood cultures were subcultured onto two LJ and incubated for an additional 3 weeks.

MTB Identification

Positive LJ or MGIT cultures were identified as MTB using the niacin production and nitrate reduction tests. In patients for whom no culture was positive for MTB and one or more MGIT cultures grew NTM, the MGIT cultures that grew NTM additionally were tested with a DNA hybridization system (AccuProbe; Gen-Probe, San Diego, CA) to confirm the presence or absence of MTB.

Quality Control

Extensive quality control procedures were implemented in each collection site and laboratory to reduce the possibility of false-positive results, including from cross-contamination. Microscopists were not blinded to the results of previous tests, but smears were judged to be positive only if they were confirmed by two different readers. Smears with 1 to 9 AFB per 100 fields were reread on site by an independent microbiologist from USCDC. Genotyping (spoligotyping and mycobacterial interspersed repetitive unit typing) was performed for any patient with only one culture positive for MTB; that isolated positive culture was genotyped along with positive cultures from any other patient who had specimens processed on the same day as the isolated positive culture. Genotyping was performed at USCDC-supported laboratories in the United States using established methods (28, 29).

Sputum specimens were obtained from all enrolled, analyzed patients. Laboratories evaluated the specimen quality (e.g., physical appearance, volume) of all specimens submitted. When specimens of poor quality were submitted, laboratories requested new specimens; if patients were unable to provide specimens of higher quality, the originally submitted sputum was processed.

Data Analysis

A pulmonary TB case was defined as a patient with at least one MTB culture–positive sputum specimen on LJ or MGIT and no MTB culture–positive nonsputum specimens. An extrapulmonary TB patient was defined as a patient with at least one MTB culture–positive nonsputum specimen on any culture media and no MTB culture–positive sputum specimens. A patient with both pulmonary and extrapulmonary TB was defined as a patient with at least one sputum specimen MTB culture positive on either LJ or MGIT and at least one MTB-positive nonsputum specimen.

The incremental yield (IY) of an additional specimen or different diagnostic method is equal to the number of subjects who were diagnosed using that additional specimen or using a different method who were not diagnosed with the baseline approach divided by the total number of patients diagnosed with disease. To calculate the IY for sputum smear, we defined the denominator as all patients with pulmonary TB or both pulmonary and extrapulmonary TB. For the yield of the first smear, the numerator was all patients with their first positive smear. For the yield of the second smear, the numerator was all patients whose first smear was not positive but whose second smear was positive. For the yield of the third smear, the numerator was all patients whose first and second smears were not positive but whose third smear was positive.

To calculate the IY for sputum culture, we defined the denominator as all patients with a diagnosis of pulmonary TB or both pulmonary and extrapulmonary TB. For the yield of the first culture, the numerator was all patients with their first sputum culture positive for MTB. For the yield of the second culture, the numerator was all patients with their first sputum culture not positive for MTB and their second sputum culture positive for MTB. For the yield of the third culture, the numerator was all patients with their first and second sputum cultures not positive for MTB and their third sputum culture positive for MTB.

To calculate the IY for additional MTB laboratory testing beyond two sputum smear examinations, we defined the denominator as all patients with a diagnosis of TB (including only pulmonary TB, only extrapulmonary TB, or both). We calculated the diagnostic yield of performing two sputum smears plus additional test(s) and subtracted this number from the diagnostic yield of only performing two sputum smears.

Exact, 95% confidence intervals (CIs) were calculated for incremental yields using established methods (30).

Ethical Review

The study was approved by human subjects review committees at USCDC and collaborating institutions in each country.


Of 1,111 patients screened for eligibility, 44 were ineligible: 35 because of current TB treatment or TB treatment within past year, two because they were screened for TB in the past 3 months, four because they took medication with anti-TB activity in the past month, and three because of unspecified reasons. Of the 1,067 eligible patients, three refused enrollment, and four did not supply sufficient specimens for analysis. Of the 1,060 patients enrolled and analyzed, 147 (14%) patients were diagnosed with TB: 61 (42%) had only pulmonary TB diagnosed, 21 (14%) had only extrapulmonary TB diagnosed, and 65 (44%) had both pulmonary and extrapulmonary TB diagnosed. The characteristics of enrolled patients and those diagnosed with TB are summarized in Table 1.



All Patients Enrolled (n = 1060)

Patients with TB (n = 147)
Age, median (IQR)30 (26–36)29 (26–34)
Male sex630 (59)115 (78)
Country of origin
 Thailand630 (59)34 (23)
 Vietnam430 (41)113 (77)
Receiving ART at time of enrollment132 (12)5 (3)
Previously treated for TB19 (2)2 (1)
Lymphadenopathy128 (12)47 (32)
CD4+ T-lymphocyte, cells/μl, median (IQR)
281 (123–429)
125 (41–321)

Definition of abbreviations: ART = anti-retroviral therapy; IQR = interquartile range; TB = tuberculosis.

*Values are n (%) unless otherwise noted.

Defined as having a peripheral, noninguinal lymph node more than 1 cm in diameter or a peripheral, inguinal lymph node more than 2 cm in diameter.

Number of Sputum Smears to Diagnose TB

Of 126 patients with pulmonary TB, the first sputum smear diagnosed 36 (29%; 95% CI, 21–37). The second sputum smear added nine patients (IY, 7%; 95% CI, 4–13) who would not have been diagnosed with the first sputum, and the third sputum smear added two patients (IY, 2%; 95% CI, 0–6) who would not have been diagnosed with the first two sputum smears. Three sputum smears failed to diagnose pulmonary TB in 79 patients (IY, 63%; 95% CI, 54–71).

Number of LJ Cultures to Diagnose TB

Of 126 pulmonary TB cases, LJ culture of the first sputum specimen diagnosed 60 (48%; 95% CI, 39–56). LJ culture of the second specimen added 18 patients (IY, 14%; 95% CI, 9–22) who would not have been diagnosed with the first sputum LJ culture, and the LJ culture of the third sputum specimen added eight patients (IY, 6%; 95% CI, 3–12) who would not have been diagnosed with the first two sputum LJ cultures. LJ culture of three sputum specimens failed to diagnose pulmonary TB in 40 patients (32%; 95% CI, 24–40).

Number of MGIT Cultures to Diagnose TB

Of 126 pulmonary TB cases, MGIT culture of the first sputum specimen diagnosed 90 cases (71%). MGIT culture of the second sputum specimen added 21 patients (IY, 17%; 95% CI, 11–24) who would not have been diagnosed with the first sputum MGIT culture, and MGIT culture of the third sputum specimen added 12 patients (IY, 10%; 95% CI, 5–16) who would not have been diagnosed with the first two sputum MGIT cultures. MGIT culture of three sputum specimens failed to diagnose pulmonary TB in three patients (2%; 95% CI, 1–7). The contamination rate for sputum specimens on MGIT was 13%.

Which Laboratory Method to Use when Testing Sputum for TB

To evaluate the yield of individual laboratory methods, we first analyzed the yield on the first sputum specimen collected. Of 96 culture-positive patients diagnosed from the first sputum specimen collected, sputum smear diagnosed 35 patients (36%; CI, 27–47). LJ culture added 30 patients (IY, 31%; 95% CI, 23–41) who would not have been diagnosed by sputum smear. MGIT culture added 31 patients (IY, 32%; 95% CI, 23–42) who would not have been diagnosed with smear or LJ. We obtained similar results when we analyzed the second and third sputum specimens cultured.

Culture of Nonsputum Specimens

Twenty-one patients had negative sputum cultures and at least one MTB positive nonsputum culture. These 21 patients were diagnosed by culture of blood (n = 1), stool (n = 7), urine (n = 5), lymph node aspirate (n = 7), and multiple sources (n = 1 patient with positive urine, stool, and lymph node cultures). The total yield of culture was 2% for blood (16/1,051), 6% for stool (61/1,052), and 3% for urine (27/1,057). Of 1,060 patients enrolled, 128 (12%) had enlarged peripheral lymph nodes; of these, 82 (64%) had fluid successfully obtained from the lymph node. Of the 82 aspirates, 34 (42%) had an MTB-positive culture. Of the 52 aspirates for which smear microscopy was performed, 16 (31%) had a positive AFB smear.

Evaluation of a Patient with Two Negative Sputum Smears

The TB diagnostic approach recommended by WHO is to conduct smear microscopy on two sputum specimens, but we found that this strategy would fail to diagnose TB in 102/147 (69%) patients with culture-confirmed TB. Therefore, we evaluated the incremental yield of different combinations of sputum and nonsputum tests over this strategy of doing two sputum smears. A strategy of performing LJ culture on a single sputum specimen left 52% of TB patients undiagnosed (Table 2). In contrast, we found that the testing strategies with the highest incremental yield were MGIT culture of the two initial sputum specimens (23% of patients undiagnosed), MGIT culture of three sputum specimens (16% of patients undiagnosed), or MGIT culture of two sputum specimens and of a lymph node aspirate (18% of patients undiagnosed) (Table 2).


Incremental Yield

TB Cases Not Diagnosed after all Tests Performed
Additional Test Performed
n (%)
95% CI
n (%)
95% CI
Examine third sputum with ZN§2 (1)0–5100 (68)60–75
Perform LJ culture on first sputum specimen25 (17)11–2477 (52);44–61
Perform LJ culture on first and second sputum specimens40 (27)20–3562 (42)34–50
Perform LJ culture on first, second, and third sputum specimens46 (31)24–3956 (38)30–46
Perform MGIT culture on first sputum specimen52 (35)28–4350 (34);27–42
Perform MGIT culture on first and second sputum specimen68 (46)38–5434 (23)17–31
Perform MGIT culture on first, second, and third sputum specimens79 (54)46–6223 (16)10–22
Perform ZN smear of lymph node aspirate6 (4)2–996 (65)57–73
Perform ZN smear and LJ culture of lymph node aspirate15 (10)6–1687 (59);51–67
Perform ZN smear and MGIT culture of lymph node aspirate16 (11)7–1786 (59)50–66
Perform MGIT culture on first and second sputum and ZN smear and MGIT culture of lymph node aspirate76 (52)43–6026 (18)12–25
Perform ZN smear of stool, urine, and lymph node aspirate7 (5)2–995 (65)57–72
Perform ZN smear and LJ culture of stool, urine, and lymph node aspirate31 (21)15–2871 (48)40–57
Perform ZN smear and MGIT culture of stool, urine, and lymph node aspirate46 (31)24–3956 (38)30–46
Perform mycobacterial blood culture
8 (5)
94 (64)

Definition of abbreviations: AFB, acid-fast bacilli; LJ, Lowenstein-Jensen; MGIT, mycobacterial growth indicator tube; TB, tuberculosis; ZN, Ziehl-Neelsen.

*A total of 1,060 people with HIV were enrolled in this study.

A total of 45 (31%) patients with TB had a positive sputum smear result on the first or second sputum specimen.

For incremental yield calculation, the denominator was all patients who were culture positive for TB from any site, using any media.

§The incremental yield of a third sputum smear is 2% when using a denominator of 126 pulmonary TB patients, as reported in the text.

Quality Assurance

Quality assurance evaluations, including genotyping, found only one possible instance of cross-contamination. In this instance, isolates from two patients who had specimens processed on the same day were matched by mycobacterial interspersed repetitive unit typing and spoligotyping. Moreover, this genotype pattern was not found in other isolates from the study. Repeating our analysis with both or either of these cases excluded did not affect our findings.

In HIV-infected patients evaluated for TB, we found that a third sputum smear adds little to the diagnosis of TB, that broth-based culture of at least two sputum specimens diagnoses most TB cases, and that lymph node aspiration provides the highest incremental yield of any nonpulmonary specimen test for TB. Our findings have important implications for global public health policy. First, we believe that our study provides convincing evidence to support the WHO recommendation of using two sputum smears, rather than three, in HIV-infected patients suspected of having TB. Second, because we found that over two thirds of all TB cases in people with HIV are not detected when sputum-smear microscopy alone is used, the WHO's recommendation for the evaluation of sputum smear–negative HIV-infected patients should be revised to recommend the routine use of broth-based culture. Ideally, broth-based culture would be done on three sputum specimens and possibly on a lymph node aspirate (for patients with an enlarged peripheral lymph node). However, when the availability of broth-based culture is limited, even broth-based culture of one or two sputum specimens provides substantial benefit compared with sputum smears alone.

Consistent with previous analyses, we found that sputum microscopy is highly insensitive in HIV-infected patients, diagnosing only a fraction of culture-positive TB patients (3). The incremental yield of microscopy diminished rapidly after the second smear, with an estimated yield of only 2% for the third sputum smear, consistent with a recent metaanalysis (7). Our study evaluated only one method of microscopy: bright-field microscopy performed on digested, decontaminated, concentrated, and carbol-fuchsin–stained specimens. It is likely that the overall yield of microscopy would have been greater had we used fluorescence microscopy (31). Nevertheless, assuming a 10% increase in cumulative sensitivity with this technique, almost 60% of TB cases would remain undetected (31). Broth-based culture of sputum identified substantially more cases than microscopy or solid media culture. In fact, a single sputum cultured on broth yielded a similar number of cases as three sputums cultured on solid media. Our study, therefore, provides further evidence to support the WHO's recent recommendation that countries with a high burden of TB develop laboratories that can perform broth-based culture (6). We found that the incremental yield for broth-based culture of a third specimen, compared with two specimens, was 10%, suggesting that settings that have financial and technical capacity to culture three specimens should do so.

In multiple studies from high-burden TB countries, mycobacterial blood culture has been found to aid TB diagnosis in HIV-infected patients with advanced immune suppression. In one series, 15% of HIV-infected persons with cough were diagnosed with TB by blood culture alone (23). We found that the yield of blood culture is relatively low, likely because our study focused on outpatients and previous studies evaluated patients presenting to emergency departments with fever and other symptoms. Mycobacterial blood culture should probably be reserved for use in severely ill HIV-infected patients. The yield of urine culture—which also has been reported to be high in many case series—was similarly low (3235). A third of smear-negative, sputum culture-negative TB patients were diagnosed by stool culture. Bacteriologic evidence of MTB in stool may be attributable to gastrointestinal TB or to sputum swallowed by persons with pulmonary TB; in patients with chronic diarrhea, stool culture may be particularly high yield (3639). We found that stool specimens were extremely difficult to decontaminate, making it problematic to recommend their use as a routine adjunct to TB diagnosis. We found a high yield for lymph-node aspiration, a technique that is widely recommended but infrequently used (18). The major limitations of aspiration are that it can only be performed in selected patients and that it requires training. Nevertheless, the high percentage of positive smears and cultures from successfully obtained aspirates provides convincing evidence of its value.

In the past 5 years, the WHO has made attempts to strengthen the evidence base for TB diagnosis (40). Unfortunately, most evidence cited for existing policies involves studies conducted in the pre-HIV era or case series in selected populations. The strength of our study is that it prospectively enrolled a large, unbiased cross-section of HIV-infected patients from resource-limited, high-TB-burden settings and applied uniform, high-quality diagnostic testing on all patients using pulmonary and nonpulmonary specimens. Studies with larger sample sizes have been done to study isolated questions, such as the number of smears performed or the yield of specific extrapulmonary tests, but our study is the most comprehensive evaluation of TB diagnostics in HIV-infected patients ever conducted and was uniquely designed to determine the diagnostic performance of multiple laboratory testing procedures and to evaluate current international policy. Microbiologists in this study were not formally blinded to the results of previously performed tests. We believe that the lack of is unlikely to have introduced substantial bias because multiple quality assurance steps were taken to reduce false-positive results. Although we have clearly demonstrated the increased diagnostic yield of different tests, we cannot quantify the impact this would have on morbidity and mortality because our study did not formally evaluate patient outcomes. It is possible that the outcomes of patients diagnosed through aggressive case finding and those diagnosed using routine methods would be similar, but we believe that the extremely high mortality of TB in HIV-infected patients—which spans all bacteriologic forms, occurs rapidly after clinical TB develops, and increases with diagnostic delay (4143)—strongly supports aggressive case finding rather than waiting for MTB bacilli to multiply until they cross the sensitivity threshold of smear microscopy.

Similar studies of new diagnostic methods in HIV-infected populations are urgently needed to reduce mortality from undiagnosed TB. While we await the development of new diagnostic tools, national programs should adapt current evidence by changing their TB diagnostic algorithms and by building capacity for broth-based culture and lymph node aspiration.

The authors thank the U.S. Agency for International Development for supporting this study. The authors thank the following members of the study team for their contributions to patient care, data collection, and laboratory testing: Thailand—Dr. Praphan Phanuphak, Dr. Tippawan Pankam, Dr. Nittaya Phanuphak, Dr. Channawong Burapat, Ms. Apiratee Kanphukiew. Vietnam—Dr. Thai Le, Dr. Trinh Thanh Thuy, Dr. Hoang Thi Quy, Dr. Nguyen Thi Ngoc Lan, Dr. Nguyen Ngoc Lan, Dr. Nguyen Huu Minh, Dr. Nguyen Hong Duc, Dr. Nguyen Tuan Tai, Mrs. Le Thi Ngoc Bich and staff of the Pham Ngoc Thach HIV outpatient Clinic, Mrs. Dai Viet Hoa and staff at the Pham Ngoc Thach microbiology laboratory, Dr. Le Truong Giang and staff at the People's AIDS Committee of Ho Chi Minh City, and the staff of the District 1 and District 2 HIV outpatient clinics.

1. World Health Organization. Global tuberculosis control: surveillance, planning, financing: WHO report 2008. Geneva, Switzerland: WHO.
2. Perkins MD, Cunningham J. Facing the crisis: improving the diagnosis of tuberculosis in the HIV era. J Infect Dis 2007;196(Suppl 1):S15–S27.
3. Getahun H, Harrington M, O'Brien R, Nunn P. Diagnosis of smear-negative pulmonary tuberculosis in people with HIV infection or AIDS in resource-constrained settings: informing urgent policy changes. Lancet 2007;369:2042–2049.
4. Kingkaew N, Sangtong B, Amnuaiphon W, Jongpaibulpatana J, Mankatittham W, Akksilp S, Sirinak C, Nateniyom S, Burapat C, Kittikraisak W, et al. HIV-associated extrapulmonary tuberculosis in Thailand: epidemiology and risk factors for death. Int J Infect Dis (In press)
5. World Health Organization. Improving the diagnosis and treatment of smear-negative pulmonary and extrapulmonary tuberculosis among adults and adolescents: recommendations for HIV-prevalent and resource-constrained settings. 2006;Geneva: WHO.
6. WHO Strategic and Technical Advisory Group for Tuberculosis. Report on conclusions and recommendations, 11–13 June, 2007, WHO Headquarters, Geneva, Switzerland. Available from:
7. Mase S, Ramsay A, Ng N, Henry M, Hopewell PC, Cunningham J, Urbanczik R, Perkins M, Aziz MA, Pai M. Yield of serial sputum specimen examinations in the diagnosis of pulmonary tuberculosis: a systematic review. Int J Tuberc Lung Dis 2007;11:485–495.
8. Crampin AC, Floyd S, Mwaungulu F, Black G, Ndhlovu R, Mwaiyeghele E, Glynn JR, Warndorff DK, Fine PE. Comparison of two versus three smears in identifying culture-positive tuberculosis patients in a rural African setting with high HIV prevalence. Int J Tuberc Lung Dis 2001;5:994–999.
9. Harvell JD, Hadley WK, Ng VL. Increased sensitivity of the BACTEC 460 mycobacterial radiometric broth culture system does not decrease the number of respiratory specimens required for a definitive diagnosis of pulmonary tuberculosis. J Clin Microbiol 2000;38:3608–3611.
10. Sprinson JE, Porco TC, Lawton E S, Mase SR, Flood JM. Yield of sputum smear examinations in evaluation of M. tuberculosis culture-positive pulmonary TB patients. Richmond, CA: State of California.
11. Cruciani M, Scarparo C, Malena M, Bosco O, Serpelloni G, Mengoli C. Meta-analysis of BACTEC MGIT 960 and BACTEC 460 TB, with or without solid media, for detection of mycobacteria. J Clin Microbiol 2004;42:2321–2325.
12. Pithie AD, Chicksen B. Fine-needle extrathoracic lymph-node aspiration in HIV-associated sputum-negative TB. Lancet 1992;340:1504–1505.
13. Bem C, Patil PS, Elliot AM, Namaambo KM, Bharucha H, Porter JD. The value of wide-needle aspiration in the diagnosis of tuberculous lymphadenitis in Africa. AIDS 1993;7:1221–1225.
14. Wilson D, Nachega JB, Chaisson RE, Maartens G. Diagnostic yield of peripheral lymph node needle core biopsies in HIV-infected adults with suspected smear-negative tuberculosis. Int J Tuberc Lung Dis 2005;9:220–222.
15. Jeena PM, Coovadia HM, Hadley LG, Wiersma R, Grant H, Chrystal V. Lymph node biopsies in HIV-infected and noninfected children with persistent lung disease. Int J Tuberc Lung Dis 2000;4:139–146.
16. Voetberg A, Lucas SB. Tuberculosis or persistent generalized lymphadenopathy in HIV disease. Lancet 1991;337:56–57.
17. Perenboom RM, Richter C, Swai ABM, Kitinya J, Mtoni I, Chande H, Kazema RR, Mwakyusa DH, Maselle SY. Diagnosis of tuberculous lymphadenitis in an area of HIV infection and limited diagnostic facilities. Trop Geogr Med 1994;46:288–292.
18. Hudson CP, Wood R, Maartens G. Diagnosing HIV-associated tuberculosis: reducing costs and diagnostic delay. Int J Tuberc Lung Dis 2000;4:240–245.
19. Oplustil CP, Leite OHM, Oliveira MS, Sinto SI, Uip DE, Boulos M, Mendes CF. Detection of mycobacteria in the bloodstream of patients with acquired immunodeficiency syndrome in a university hospital in Brazil. Braz J Infect Dis 2001;5:252–259.
20. Ramachandran R, Swaminathan S, Somasundaram S, Asgar VN, Paramesh P, Paramasivan CN. Mycobacteremia in tuberculosis patients with HIV infection. Ind J Tub 2002;50:29–31.
21. McDonald LC, Archibald LK, Rheanpumikankit S, Tansuphaswadikul S, Eampokalap B, Nwanyanawu O, Kazembe P, Dobbie H, Reller LB, Jarvis WR. Unrecognized Mycobacterium tuberculosis bacteremia among hospital inpatients in less developed countries. Lancet 1999;354:1159–1163.
22. Archibald LK, McDonald LC, Rheanpumikankit S, Tansuphaswadikul S, Chaovanich A, Eampokalap B, Banerjee SN, Reller LB, Jarvis WR. Fever and human immunodeficiency virus infection as sentinels for emerging mycobacterial and fungal bloodstream infections in hospitalized patients ≥ 15 years old, Bangkok. J Infect Dis 1999;180:87–92.
23. Talbot EA, Hay Burgess DC, Hone NM, Iademarco MF, Mwasekaga MJ, Moffat HJ, Moeti TL, Mwansa RA, Letsatsi P, Gokhale NT, et al. Tuberculosis serodiagnosis in a predominantly HIV-infected population of hospitalized patients with cough, Botswana, 2002. Clin Infect Dis 2004;39:e1–e7.
24. Von Reyn CF. The significance of bacteremic tuberculosis among persons with HIV infection in developing countries. AIDS 1999;13:2193–2195.
25. Rieder HL, Van Deun A, Kam KM, Kim SJ, Chonde TM, Trébucq A, Urbanczik R. Priorities for tuberculosis bacteriology services in low income countries. Paris: International Union Against Tuberculosis and Lung Disease; 2007.
26. World Health Organization. Laboratory services in tuberculosis control: microscopy, part II. Geneva, Switzerland: WHO; 1998.
27. Kent PT, Kubica GP. Public health mycobacteriology: a guide for the level III laboratory. Atlanta, GA: US Department of Health and Human Services, Centers for Disease Control and Prevention; 1985.
28. Cowan LS, Diem L, Brake MC, Crawford JT. Transfer of a Mycobacterium tuberculosis genotyping method, Spoligotyping, from a reverse line-blot hybridization, membrane-based assay to the Luminex multianalyte profiling system. J Clin Microbiol 2004;42:474–477.
29. Cowan LS, Diem L, Monson T, Wand P, Temporado D, Oemig TV, Crawford JT. Evaluation of a two-step approach for large-scale, prospective genotyping of Mycobacterium tuberculosis isolates in the United States. J Clin Microbiol 2005;43:688–695.
30. Blaker H. Confidence curves and improved exact confidence intervals for discrete distributions. Can J Stat 2000;28:783–798.
31. Steingart KR, Henry M, Ng V, Hopewell PC, Ramsay A, Cunningham J, Urbanczik R, Perkins M, Aziz MA, Pai M. Fluorescence versus conventional sputum smear microscopy for tuberculosis: a systematic review. Lancet Infect Dis 2006;6:570–581.
32. Shafer R, Kim DS, Weiss JP, Quale JM. Extrapulmonary tuberculosis in patients with human immunodeficiency virus infection. Medicine 1991;70:384–397.
33. Klotz SA, Penn RL. Acid-fast staining of urine and gastric contents is an excellent indicator of mycobacterial disease. Am Rev Respir Dis 1987;136:1197–1198.
34. Christensen WI. Genitourinary tuberculosis: review of 102 cases. Medicine 1974;53:377–390.
35. Simon HB, Weinstein AJ, Pasternak MS, Swartz MN, Lunz LJ. Genitourinary tuberculosis: clinical features in a general hospital. Am J Med 1977;63:410–420.
36. Allen BW. Comparison of three methods for decontamination of faeces for isolation of Mycobacterium tuberculousis. Tubercle 1991;72:214–217.
37. Allen BW. Isolation of Mycobacterium tuberculosis from faeces. Med Lab Sci 1989;46:101–106.
38. Manatsathit S, Tansupasawasdikul S, Wanachiwanawin D, Setawarin S, Suwanagool P, Prakasvejakit S, Leelakusolwong S, Eampokalap B, Kachintorn U. Causes of chronic diarrhea in patients with AIDS in Thailand: a prospective clinical and microbiological study. J Gastroenterol 1996;31:533–537.
39. Kramer F, Modilevsky T, Waliany AR, Leedom JM, Barnes PF. Delayed diagnosis of tuberculosis in patients with human immunodeficiency virus infection. Am J Med 1990;89:451–456.
40. Pai M, Ramsay A, O'Brien R. Evidence-based tuberculosis diagnosis. PLoS Med 2008;5:e156.
41. Sanguanwongse N, Cain KP, Suriya P, Nateniyom S, Yamada N, Wattanaamornkiat W, et al. Antiretroviral therapy for HIV-infected tuberculosis patients saves lives but needs to be used more frequently in Thailand. J Acquir Immune Defic Syndr 2008;48:181–189.
42. Cain KP, Kanara N, Laserson KF, Vannarith C, Sameourn K, Samnang K, et al. The epidemiology of HIV-associated tuberculosis in rural Cambodia. Int J Tuberc Lung Dis 2007;11:1008–1013.
43. Cain KP, Anekthananon T, Burapat C, Akksilp S, Mankhatitham W, Srinak C, et al. Causes of death in HIV-infected persons who have tuberculosis, Thailand. Emerg Infect Dis 2009;15:258–264.
Correspondence and requests for reprints should be addressed to Dr. Jay K. Varma, M.D., U.S. Embassy Beijing, No. 55 An Jia Lou Rd, Beijing 100600, China. E-mail:


No related items
American Journal of Respiratory and Critical Care Medicine

Click to see any corrections or updates and to confirm this is the authentic version of record