INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The tuberculin skin test is one of the few diagnostic tests in current clinical use that was first introduced at the beginning of the twentieth century (1). It is therefore somewhat surprising that interpretation of the tuberculin skin test (TST) remains a subject of some difficulty and controversy (2). One common belief is that larger tuberculin reactions, always more impressive to patients and providers, indicate higher likelihood of active disease, or higher risk of disease in the future, as is associated with close contact with an active case, or radiographic abnormalities of inactive tuberculosis (TB).

We have examined the relationship between final diagnosis, type and extent of radiographic abnormalities, and size of tuberculin reaction, in a cohort of close contacts, and among participants in a study evaluating different tests for the diagnosis of minimal pulmonary tuberculosis.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patient Population, Clinical Information

Two patient populations were studied. The first included all close/ household contacts of patients with active TB identified between June 1996 and June 1997. The second was composed of patients at the Montreal Chest Institute who participated in a study evaluating tests for the diagnosis of minimal pulmonary TB between July 1995 and June 1998.

Contacts were tuberculin tested, and referred for further evaluation if they had a TST of 5 mm or greater. All these individuals had a chest radiograph and saw a chest specialist, who determined if further evaluation, including sputum induction, bronchoscopy, or further radiographs, as well as treatment were needed. Participants in the diagnostic study underwent at least one sputum induction, performed for up to 15 min with hypertonic (3%) saline administered by an ultrasonic nebulizer (De Vilbiss Ultra-Neb 90), which produced 5-6 ml/ min (maximum 90 ml). Sputum specimens were digested, decontaminated, and centrifuged, then a portion of the processed sample was examined with fluorescent microscopy (confirmation with Ziehl-Neelson staining), and another portion was used to inoculate 0.1 ml onto Loewenstein Jensen media, and the liquid media of the BACTEC 460 system. Positive cultures were sent to the provincial reference laboratory for identification and susceptibility testing.

Tuberculin skin testing was performed by one of three highly experienced nurses of the TB clinic, who administered 5-TU of PPD-T (Tubersol; Connaught Laboratories, Toronto, ON, Canada) using the Mantoux technique (6). The transverse diameter of induration was measured after 48-72 h using the ball-point method (7) and recorded in millimeters. Hospital records were reviewed to obtain demographic data and clinical information.

All chest X-rays were reviewed by one of the investigators (D.M.), who rated the overall likelihood of active TB (unlikely, possible, or probable), estimated percentage of total lung parenchyma affected, and classified the abnormalities using a categorization scheme developed for Immigration Canada. Where pairs of chest X-rays taken 2-3 mo apart were available, they were interpreted as better, worse, or unchanged. All such readings were made without knowledge of the order of the radiographs (when there were a pair taken months apart), nor final diagnosis or other clinical information (i.e., this was a blinded independent evaluation).

Diagnoses and Definitions

Active TB was considered confirmed if cultures from sputum, bronchoalveolar lavage, or other specimens were positive for Mycobacterium tuberculosis. It was considered a clinical case (culture negative) if compatible abnormal chest X-ray, which improved after treatment for 2-3 mo with three or four antituberculous drugs, as judged by the independent, blinded review of the chest X-ray.

Inactive TB was considered confirmed if all cultures were negative for M. tuberculosis, there was a compatible and stable abnormal chest X-ray, as judged by the independent review, there was a final clinical diagnosis of inactive TB, and no other pulmonary diagnosis was made.

Nontuberculous Mycobacterium (NTM) was confirmed if potentially pathogenic NTM was isolated from respiratory secretions (sputum or bronchoalveolar lavage), there was a compatible abnormal chest X-ray, and no other diagnosis was made.

Normal/unrelated to TB was considered confirmed if all mycobacterial cultures were negative and the chest X-ray was normal or abnormal but not consistent with a TB-related abnormality and stable. Tuberculin skin test was not considered in making this diagnosis. No clinical abnormality was detected (other than latent TB infection) or there was a diagnosis unrelated to TB such as asthma, chronic obstructive pulmonary disease (COPD), or heart failure.

Other diseases potentially associated with anergy included lung cancer diagnosed by histology, sarcoidosis, diagnosed by histology and/or CT of the chest, human immunodeficiency virus (HIV) infection, diagnosed by serology, and other pulmonary infections such as pneumonia or lung abscess, diagnosed by appropriate cultures and/or response to antibiotic therapy.

Radiographic findings were classified as normal/minimal if there were normal, apical pleural capping, granuloma(s), or minor costophrenic angle blunting; as apical fibronodular if upper lobe scarring was present, with or without granuloma(s), and with or without apical pleural thickening; and other major if there were lung masses, hilar and/or mediastinal lymphadenopathy, pleural effusion, or cavitary or noncavitary parenchymal infiltrates.

Statistical Analysis

All analyses were conducted using SAS (SAS Institute Inc., Cary, NC). The association of tuberculin reactions categorized as 0-4 mm (negative), 5-9 mm (doubtful), and 10+ mm (positive) with diagnostic group, radiographic category, or bacillus Calmette-Guérin (BCG) vaccination history was tested with chi squared tests. t tests or analysis of variance were used to test whether the mean size of tuberculin reactions was different between these groups, and those with and without a history of BCG vaccination (8).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In 1996-1997, 182 close contacts were identified and tuberculin tested, of whom 103 had a positive TST; six of these proved to have active TB. Between 1995 and 1998, 502 patients investigated for possible active TB participated in the diagnostic study. Of these 62 were diagnosed to have active TB, of whom 46 were culture confirmed, including two patients with extrapulmonary TB (genitourinary, and lymphadenitis). Inactive TB was diagnosed in 274, nontuberculous mycobacterial (NTM) disease in 14, bronchiectasis in 25, airways disease (COPD, bronchitis, asthma, and allergic bronchopulmonary aspergillosis) in 12, interstitial lung disease in 4, and bronchogenic cyst and congestive heart failure in one each. Seventy-three patients had no apparent pulmonary disease, and 36 had conditions associated with anergy including 13 with lung cancer, 12 with pneumonia or lung abscess, 5 with sarcoidosis, and 6 with HIV infection.

As shown in Table 1, the prevalence of tuberculin reactions of 5 mm or greater was significantly lower among contacts, subjects with anergy-associated conditions, conditions unrelated to TB, or normal (p < 0.001). However, once the threshold of 5 mm was exceeded, then the mean size of TST reactions was very similar (p = 0.6). Similarly the frequency distribution of tuberculin reactions was remarkably similar in all diagnostic groups as shown in Figure 1 (p = 0.8). Because of small numbers the curves for patients with NTM and anergy-associated conditions are not shown.

View larger version (18K): [in this window] [in a new window]   Figure 1.   Pattern of tuberculin reactions in patients with different diagnoses. Fourteen patients with nontuberculous mycobacteria and 36 with conditions associated with anergy were excluded. Solid line with small squares: active TB; solid line with triangles: close contacts; dashed line with diamonds: inactive TB; dotted line with stars: normal/not TB related. Curves smoothed by moving three-point averages. Tuberculin reactions of zero (0) mm not shown (see Table 1).

The 68 patients with active tuberculosis, 14 with NTM, and 36 with conditions associated with anergy were excluded from the analyses of association of tuberculin reaction size with BCG vaccination or radiographic findings. As seen in Table 2A, among the remaining 395 subjects who were tuberculin tested, positive reactions were less frequent with apical fibronodular or other more significant radiographic findings (p < 0.001), but mean size of reaction (Table 2A) and pattern of reactions (Figure 2) were not associated with type of radiographic abnormality. Positive reactions were somewhat more frequent with less extensive parenchymal abnormalities, yet mean reaction size was slightly smaller (Table 2B). As seen in Figure 3, among the 121 subjects who gave a history of BCG vaccination, size and pattern of tuberculin reactions were not different from the 176 who reported no prior BCG vaccination (Table 3). Vaccinal status was unknown in 98 (25%).

View larger version (17K): [in this window] [in a new window]   Figure 2.   Pattern of tuberculin reactions in patients with different radiographic abnormalities. Sixty-eight patients with active TB, 14 patients with NTM, and 36 with conditions associated with anergy were excluded. Solid line with large squares: lung masses, hilar or mediastinal lymph nodes, pleural effusion, parenchymal infiltrates, or cavities; Dotted line with stars: apical fibronodular changes; dashed line with stars: normal, minimal abnormalities, or granulomas on X-ray. Curves smoothed by moving three-point averages. Reactions of zero (0) mm not shown (see Table 2).

View larger version (33K): [in this window] [in a new window]   Figure 3.   Pattern of tuberculin reactions and history of BCG vaccination. Sixty-eight patients with active TB, 14 patients with NTM, and 36 with conditions associated with anergy were excluded. Filled bars: 121 subjects with history of BCG vaccination; cross-hatched bars: 176 subjects without BCG; Striped bars: 95 subjects whose BCG status was unknown.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This study emphasizes that the tuberculin reaction is an all-or-nothing phenomenon. Among patients with increased likelihood of true TB infection because of clinical and/or radiographic abnormalities, the size of the tuberculin test is of no diagnostic utility once the threshold of 5 mm is passed. In this series, patients with TST less than 5 mm were less likely to have active TB, but above that threshold, larger TST were not more likely to indicate active disease, and the size and pattern of TST were virtually identical in patients with different diagnoses, and different radiographic findings.

In our series among the patients with active TB, 6% had TST less than 5 mm, 12% had TST less than 10 mm, and 43% had TST less than 15 mm. Others have reported false-negative TST in 21% (9) or 30% (10) of patients with newly diagnosed active TB. Given that the pattern of TST was very similar among all diagnostic groups, use of larger reaction sizes to define a positive TST in this patient population would not have improved specificity, but would have substantially reduced sensitivity. For example, use of criteria of 5 mm, 10 mm, or 15 mm would have resulted in sensitivity of 94%, 88%, or 57%, respectively, for the diagnosis of active TB. Similarly, applying criteria of 5 mm, 10 mm, or 15 mm would have potentially misclassified 1%, 8%, or 35% of patients with clinically diagnosed inactive TB, and 2%, 16%, and 40% of contacts, respectively, as not having TB infection.

A potential criticism of this analysis is that the TST result may have influenced the final clinical diagnosis, so the analysis of association of TST with diagnosis (Table 1 and Figure 1) was inappropriate. However, the contacts were defined before tuberculin testing, and in the diagnostic study TST results were used only to support a diagnosis of inactive TB, which was also based on negative cultures, treating physician's impression, consistent clinical and radiographic findings including follow-up chest X-rays, and absence of other pulmonary diagnosis. The diagnosis of active TB and NTM was primarily based on microbiological results, while the diagnosis of other illnesses was based on clinical, laboratory, histological, and radiographic data, but not the TST. The chest radiographs were read independently, and without knowledge of clinical diagnosis or test results including TST. Therefore the analysis of the association of TST size and pattern with the extent and type of radiographic abnormality (Table 2 and Figure 2) should not have been biased.

The ATS has recommended that when interpreting tuberculin reactions, criteria of 5 mm, 10 mm, or 15 mm should be used to consider a TST positive, depending upon the clinical situation (11). This recommendation is based largely upon estimates of the positive predictive value, that is, the likelihood that tuberculin reactions of different sizes represent true positive reactions. The positive predictive value is dependent upon the prevalence of true infection in the population being tested, the clinical situation (contact [12] versus general population screening, abnormal chest X-ray, U.S. or Canadian born versus foreign born), and the likelihood of false-positive reactions due to BCG or sensitivity to NTM antigens (13). However, the ATS recommendation is often misinterpreted to mean that larger reactions are more serious, that is, more likely to indicate disease, or that a radiographic abnormality is related to TB infection. In this series, size of tuberculin reactions had no relationship to presence of active disease, nor risk factors for future diseasehistory of contact, or radiographic abnormalities. In this population BCG vaccination did not appear to affect tuberculin reactions, similar to one other study of contacts (14), but in contrast to studies conducted among general population samples (15). This difference is primarily because the expected prevalence of true infection in our study population, with risk factors for TB infection, was much higher than in general population studies (15). In the latter the lower prevalence of true TB infection resulted in a relatively greater impact of false-positive reactions, which were therefore easier to identify (15). As well, in our study BCG vaccination was based on patient recallthe resultant potential misclassification would have reduced our ability to identify any effects of BCG on tuberculin reactions.

These findings support the recommendation that in the presence of clinical or radiographic abnormalities, a tuberculin reaction of 5 mm or larger should be considered significant (11). However, above this threshold, size does not matter.