Subjects being investigated for pulmonary TB who were unable to produce sputum spontaneously or had one sample of spontaneously produced sputum that was smear-negative for acid fast bacilli (AFB) were prospectively enrolled at Pusan National University Yangsan Hospital between March 2010 and February 2011. No subject had started antituberculosis medication at the time of sputum induction.

Subjects were assigned into the TB group when 1) MTB was cultured in induced or spontaneously produced sputum or 2) there was clear evidence for a clinical diagnosis based on radiology with an appropriate response to treatment. Subjects with negative MTB cultures and an alternative diagnosis were assigned to the non-TB group. All subjects were followed up for at least 6 months after the final diagnosis

The protocol for this study was approved by the Institutional Review Board of the Pusan National University Yangsan Hospital (approval number: 02-2010-21) and each patient provided written informed consent prior to enrollment in the study.

Sputum induction was performed using 3% hypertonic saline solution generated by an ultrasonic nebulizer (Devilbiss Ultraneb 99; Sunrise Medical, Somerset, PA, USA) for 20 minutes in a room with negative-pressure ventilation. All subjects were pre-medicated with 200 µg of salbutamol via a metered-dose inhaler 30 minutes prior to sputum induction. Peak expiratory flow rate (PEFR) was measured every 5 minutes and induction was terminated if PEFR declined by 20%, or if major adverse events were reported. At 5-minute intervals, subjects were asked to clear saliva from their mouth and then expectorate sputum. A trained nurse supervised the entire process. An inadequate induced sputum sample was defined by the following criteria: sputum induction tolerated for <5 minutes, total cells isolated from sputum <1×10⁶, or squamous cell percentage >80%.

Collected induced sputum was processed within 3 hours of sampling. Approximate 3 mL of sputum was removed for microbiological examination including AFB smear, TB culture using Lowenstein-Jensen media, and TB polymerase chain reaction (PCR) (Advansure TB/NTM, LG Life Sciences, Seoul, Korea). Then, remaining sputum was processed for cell isolation. Briefly, after adding equal volume of working sputasol solution (sputasol:PBS, 1:9) (Sputasol, Oxoid SR 0233A), the sample was maintained at room temperature for 20 minutes, during which the sticky materials were frequently dissolved with a sterile empty disposable pipette and a vortex mixer. The sample was filtered through a 40-µm nylon mesh into another conical tube. After centrifugation (1500×g, 10 minutes), the supernatants were collected and Ficoll-gradient centrifugation of the pellet was performed to isolate the leukocytes. It was then washed twice with RPMI 1640 medium (Gibco; Gaithersburg, MD, USA), and the isolated cells were counted with a hemocytometer and cell viability was assessed using trypan blue. Counted cells were used for the preparation of cytospin slides and for the ELISPOT assay.

Cytospin slides were stained with Hemacolor (Merck; Darmstadt, Germany). Differential cell counts were performed under light microscope using a set of standardized morphometric appearance criteria (magnification, ×400), and the percentage of each cell population, including neutrophils, macrophages, lymphocytes, and bronchial epithelial cells, was individually recorded after counting more than 200 cells. And absolute cell numbers were calculated with total cell counts.

The ELISPOT assay (T SPOT-TB; Oxford Immunotec, Oxford, UK) was performed according to the manufacturer's instructions. This assay kit included nil control, early secretory antigen target (ESAT)-6-derived peptide (Panel A), CFP-10-derived peptide (Panel B), and phytohaemagglutinin in positive control wells. Briefly, isolated sputum cells were plated at 2.5×10⁵ cells/well in 96-well T-spot plates. After incubating the plate overnight in a humidified incubator at 37℃ and 5% CO₂, the plate was washed 3 times with fresh PBS solution and Conjugate Reagent solution was added. After incubating at 4℃ for 1 hour, the plate was washed 3 times with fresh PBS solution and substrate solution was added. The plate was incubated at room temperature for 7 minutes and wash with distilled water to stop the reaction. Then, the plate was allowed to dry and the spots were counted with Immuno.spot (Cellular Technology Ltd., Cleveland, OH, USA).

As per the manufacturer's instructions for ESAT-6 and CFP-10, a test was scored as qualitatively reactive if either (a) for nil control of 0-5 spots, the antigen well spot count minus the nil control was ≥6 or (b) for nil control of >5 spots, the antigen count was ≥2 times the nil control. Overall, a T SPOT-TB test was considered reactive if either ESAT-6 or CFP-10 was reactive. The results were considered as indeterminate when <20 spots were found in positive control well (positive control failure) or when >5 spots in the negative control well and indicated non-reactive after calculating with the spots in antigen-wells (negative control failure).

The chi-square test or Fisher's exact test was performed to compare categorical variables, and Student's t-test or Mann-Whitney test was performed to compare continuous variables. Simple logistic regression analysis was used to assess univariate associations between tested variables and IS ELISPOT result (determinate vs. indeterminate). Correlation analysis between variables was carried out by Spearman's rho test. All analysis were performed using SPSS version 15.0 (SPSS Inc., Chicago, IL, USA), and the results with p<0.05 were considered statistically significant.

A total of 43 subjects were enrolled in the study. Thirty one of them (72%) had AFB smear negative sputum and 12 (28%) did not produce sputum spontaneously. Among 43 subjects, 25 (58%) were assigned to the TB group and 18 (42%) to the non-TB group. In the TB group, 20 of 25 subjects had MTB-positive cultures from their induced sputum samples, while 13 of those 20 subjects were AFB smear positive. TB-PCR was positive in 12 of 13 smear positive and 3 of 12 smear negative samples. Final diagnosis of the 18 subjects in non-TB group included: former TB without reactivation (n=7), pneumonia (n=4), bronchiectasis (n=2), lung cancer (n=2), nontuberculous mycobacterial pulmonary disease (n=2), and idiopathic pulmonary fibrosis (n=1).

Of the 43 subjects, 12 (28%) had at least 1 comorbidity; malignancy was the most common (n=7: lung, n=2; stomach, n=2; colon, n=2; and kidney, n=1), followed by psychiatric disease (n=2), cardiovascular disease (n=1), gastrointestinal disease (n=1), and diabetes mellitus (n=1). HIV-ELISA tests were performed on 2 patients and both were negative.

There were no significant adverse events or PEFR reduction to terminated premature sputum induction. The median volume of induced sputum processed for cell isolation was 7.5 mL (range, 2-18) in the TB group and 7.5 mL (range, 3-18) in the non-TB group (p=0.992). The median value of total cell numbers obtained from induced sputum was 3.1×10⁶ cells (range, 0.2-35.0) in the TB group and 2.2×10⁶ cells in the non-TB group (range, 0.6-19.6) (p=0.808).

Of 43 sputum samples, 11 (26%) were inadequate to perform the ELISPOT assay. Eight samples had insufficient total cell numbers and 3 samples had >80% of squamous epithelial cells. The proportion of inadequate sputum was not significantly different between the TB (20%, 5/25) and non-TB group (33%, 6/18) (p=0.480) (Table 1).

The median number of total cells from adequate sputum was 3.3×10⁶ cells (range, 1.0-3.5), which was significantly higher than the value of 0.9×10⁶ cells (range, 0.2-1.4) from inadequate sputum (p=0.034). There was no correlation between total cell numbers and volumes of sputum (Spearman's rho=0.108, n=43, p=0.497). In addition, there were no significant differences in the tested variables such as sputum volume, gender, age, diagnosis of TB, smoking history, and underlying diseases between the adequate and inadequate samples (data not shown).

Cytospin slides were obtained from induced sputum cells, and differential cell counts were performed. Fig. 1 shows the percentage of each cell type. Alveolar macrophages were predominant in the TB group while neutrophil numbers were higher than any other cell types in the non-TB group. The percentage of lymphocytes was significantly higher in the TB group than that in the non-TB group, whereas the percentage of neutrophils was inverse between the groups.

IS ELISPOT was performed on 32 adequate sputum samples after excluding 11 inadequate samples. The median numbers of ESAT-6- and CFP-10-specific cells were higher in the TB group (24; range, 0-409, and 37; 0-215, respectively) compared to the non-TB group (11; 0-86, and 13; 0-112, respectively), but the differences were not statistically significant (p=0.095, and 0.184, respectively). However, after subtraction of spots from the negative controls, the median number of ESAT-6- and CFP-10-specific cells were significantly higher in the TB group (15; 0-210, and 20; 0-200, respectively) than in the non-TB group (0; 0-5 and 0; 0-3, p<0.001 and p<0.001, respectively) (Fig. 2).

Of 32 adequate sputum samples, 17 showed determinate results whereas 15 were indeterminate. The proportion of determinate results was higher in the TB (75%, 15/20) than in the non-TB group (17%, 2/12) (p=0.002), and all of these determinate results were consistent with the final diagnosis. The characteristics of these 17 subjects are shown in Table 2, and the reasons for inconclusive results are explained in Table 1.

Sensitivities of TB-PCR (60%, 15/25) and IS ELISPOT (60%, 15/25) in the TB group were same. However, in 12 AFB smear-negative samples, the sensitivity of IS ELISPOT (75%, 9/12) was much higher than that of TB-PCR (25%, 3/12).

Among the 32 subjects with adequate sputum, univariate analysis indicated the status of active TB as a unique predictor for determinate results (odds ratio, 15.0; 95% confidence interval, 2.42-93.01; p=0.003) (Table 3). In contrast, as for the TB group of 20 subjects with adequate sputum, no significant predictor for determinate results was observed. Unexpectedly, even smear positivity and extent of chest radiography (CXR) were not significant factors (Table 4).