Outcome of untreated lung nodules with histological but no microbiological evidence of tuberculosis

This was a retrospective cohort study conducted in National Taiwan University Hospital (NTUH). The Research Ethics Committee of NTUH approved this study (REC No.: 201205025RIC). Under Taiwan’s National TB control policy, all cases of culture-confirmed and suspected TB should be reported to the Taiwan Centers for Disease Control (CDC), and all TB contacts should receive contact investigation and be reported to Taiwan CDC, as well. Permission is required to access the online reporting database of Taiwan CDC (https://​tb.​cdc.​gov.​tw/​slow/​ca/​loginbycard.​asp). Standard regimen used in Taiwan for treating new TB cases follows the guidelines of the World Health Organization (WHO), consisting isoniazid (5 mg/kg), rifampicin (10 mg/kg), pyrazinamide (25 mg/kg), and ethambutol (15 mg/kg) for 2 months, followed by isoniazid (5 mg/kg), rifampicin (10 mg/kg), plus ethambutol (15 mg/kg) if results of susceptibility test is not available, for 4 months [6, 7].

We searched the histopathology database of NTUH for findings of caseous necrosis or granulomatous inflammation from January 2008 to June 2013. Patients were excluded if (1) the M. tuberculosis complex or NTM were isolated within 60 days before or after biopsy, (2) biopsy samples were extrapulmonary, (3) biopsy was performed during anti-TB treatment, (4) computed tomography (CT) images of the chest were unavailable, (5) histological or microbiological evidence of fungal or parasitic infection was noted, (6) patients had pure or mixed with non-nodular radiographic patterns suggestive for active disease, such as tree-in-buds pattern, consolidation, and miliary lesions [8], and (7) a positive tissue AFS, since these cases were usually treated as TB until proven otherwise.

All included patients were followed up for 4 years after biopsy unless they died or received a diagnosis of active TB. Active TB was defined if either of the following two criteria was fulfilled [9, 10]: (1) mycobacterial cultures of sputum or other respiratory samples yielded the M. tuberculosis complex; and (2) chest radiography revealed new lesions without other proven etiology, which improved after standard anti-TB treatment, determined by serial chest radiography or by follow-up CT, if the pre-treat pulmonary lesion was not detectable by chest radiography.

Medical records were reviewed to obtain demographic data, including age, sex, symptoms, comorbidity, history of TB, biopsy method, histology, AFS results, mycobacterial culture of sputum and tissue specimens, and adverse events of anti-TB treatment. At the beginning of tissue diagnosis, we obtained laboratory test results, including leukocyte and differential leukocyte counts and hemoglobin, albumin, and C-reactive protein levels. CT images were further reviewed by a pulmonologist. Furthermore, radiographic patterns; presence of solitary or multiple nodules, background fibrocalcified lesions, and bronchiectasis; maximum diameter and density of lesions; cavitation; and mediastinal lymphadenopathy were recorded.

Categorical variables were compared using the chi-square test or Fisher’s exact test, as appropriate, whereas continuous variables were compared using independent sample t tests. Multivariate logistic regression analysis was performed in a backward manner to identify the predictors for prescribing anti-TB treatment. The analyzed variables included age, sex, history of TB, presence of respiratory or constitutional symptoms suggestive of TB, histological findings (caseous necrosis, granulomatous inflammation on biopsy histology, concomitant malignancy on biopsy histology), biopsy method, mycobacterial culture for respiratory specimens or biopsy specimens, comorbidity (diabetes mellitus, end-stage renal disease, malignancy, hepatitis B virus infection, human immunodeficiency virus infection, organ transplant recipient, alcoholism, and autoimmune disease), findings on chest computed tomography (multiple nodules, lesion size > 3 cm, cavitary lesions, ground-glass opacity, fibrocalcified lesions, mediastinal lymphadenopathy), and laboratory test results (leukocyte count and hypoalbuminemia [defined as serum albumin level < 3.5 g/dL]). Kaplan–Meier curves were generated to represent the time to subsequent development of active TB in different groups and were compared using the log-rank test. All analyses were performed using IBM SPSS statistics (version 23; IBM Corp., Armonk, NY, USA). Statistical significance was set at p < 0.05.

From January 2008 to June 2013, a total of 470 patients with histological findings suggestive of TB were identified; among them, 439 had granulomatous inflammation, 233 had caseous necrosis, and 143 had positive tissue AFS. In total, 363 patients were excluded because of various reasons shown in Fig. 1. The remaining 107 (22.8%) patients were analyzed further.

Among all 107 patients, 54 (51%) were clinically asymptomatic and were referred for further evaluation because of abnormal chest images noted during health examination or routine checkup for other medical problems. Among all 107 patients, 40 (37%) received anti-TB treatment (treated group); of them, 17 were clinically asymptomatic. The remaining 67 patients (63%) received follow-up without immediate treatment (untreated group). Among all 107 cases, 78 had multiple lung nodules or masses, while the remaining 29 patients had solitary nodule. Typical CT images of solitary and multiple lung nodules or masses are illustrated in Fig. 2.

The clinical characteristics of all 107 patients are presented in Table 1. The mean age was 56.3 ± 13.4 years, with a male predominance (60%). The most common underlying disease was malignancy (35%), followed by diabetes mellitus (10%). Tissue samples were collected through surgical resection in 65 patients (61%). In total, 91 patients (85%) were tested through sputum or tissue mycobacterial culture. The most common histological finding was granulomatous inflammation (99%), followed by caseous necrosis (51%); 18 patients (17%) had concomitant histological findings of malignancy.

The mean follow-up duration was similar between the treated and untreated groups. During the follow-up period, 3 treated and 8 untreated patients died. Of the 8 untreated patients, 3 received surgical resection of the lesion. Five died of malignancy, and the remaining 3 died of heart disease, upper gastrointestinal bleeding, and pneumonia, respectively. Half of the 8 untreated patients had the past history of treated pulmonary TB; seven of them (88%) had been tested through either sputum or tissue mycobacterium culture, or both. No evidence of active TB was noted during the entire clinical course of the 8 cases.

No significant difference was observed in age, sex, clinical symptoms, and history of TB between the treated and untreated groups. More patients in the untreated group had malignancy than did those in the treated group (46% vs. 15%, p = 0.001). More patients in the treated group had the histological features of caseous necrosis (78% vs. 36%, p < 0.001), but fewer patients had the histological features of concomitant malignancy (5% vs. 24%, p = 0.012). Anti-TB treatment was not administered to 24 of the 55 patients (44%) with biopsy histology showing caseous necrosis; none of these patients developed active TB in the subsequent 4 years.

The radiographic patterns and laboratory data of the patients are summarized in Table 1. In general, most patients had normal hemogram and serum albumin levels. The laboratory data were similar in the treated and untreated groups, except that the treated group had higher albumin levels. In both groups, the most common radiographic findings were multiple lung nodules and mediastinal lymphadenopathy. The associated findings, such as cavitary lesions and ground-glass opacity, did not differ significantly between the 2 groups. Fibrocalcified lesions, bronchiectasis, and mediastinal lymphadenopathy, implying inactive or chronic disease status, did not differ significantly between the 2 groups.

The multivariate logistic regression analysis identified 2 independent factors for predicting whether physicians initiate anti-TB treatment: caseous necrosis on biopsy histology (OR = 9.60, 95% CI = 3.38–27.23), and surgical resection (OR = 0.25, 95% CI = 0.09–0.69; Table 2).

In the untreated group, during a total of 251.2 follow-up patient-years, only 1 patient developed pulmonary TB approximately 19 months after biopsy (Day 574). None of the 16 cases having co-existing histologic finding of malignancy became incident TB case within a follow-up of 56.7 patient-years.

The patient, aged between 50 and 60, had a history of left maxillary acinic cell carcinoma, which had been surgically resected 13 years ago and had recurred 2 months before the index biopsy. Chest CT for staging work-up revealed a 1.8-cm nodule in the right upper lobe (Fig. 3a). The case reported weight loss and a tendency to become easily fatigued, but denied any airway symptoms. CT-guided biopsy revealed granulomatous inflammation with lymphocyte aggregation and negative tissue AFS. Sputum AFS and mycobacterial cultures were also negative. Anticancer chemotherapy and target therapy were administered consequently. Although follow-up chest CT at 10 months after the index biopsy showed progression in bilateral alveolar infiltrates, sputum studies for a total of 6 specimens collected at 9 and 11 months after the index biopsy were all negative for the M. tuberculosis complex. Follow-up chest CT at 17 months after the index biopsy demonstrated multiple bilateral cavitary consolidations (Fig. 3b). Sputum culture at that time yielded the M. tuberculosis complex. No records of recent contact to active TB case were found in the online reporting database of Taiwan CDC.

Among the 40 patients receiving anti-TB treatment after the index biopsy, adverse events occurred in 21 (53%). The most common adverse event was rash (20%), followed by hepatotoxicity (18%) and constitutional symptoms including fever, malaise, dizziness, and chest tightness (18%; Fig. 4). The development of adverse events led to treatment interruption in 8 patients (20%), including 6 patients who developed hepatotoxicity. Among the 17 clinically asymptomatic patients who received anti-TB treatment, 13 (76%) had adverse events, leading to therapeutic regimen modification in 3 patients (18%) and treatment interruption in 6 patients (35%).