Elsevier

Tuberculosis

Volume 87, Issue 1, January 2007, Pages 44-52
Tuberculosis

Volatile biomarkers of pulmonary tuberculosis in the breath

https://doi.org/10.1016/j.tube.2006.03.004Get rights and content

Summary

Pulmonary tuberculosis may alter volatile organic compounds (VOCs) in breath because Mycobacteria and oxidative stress resulting from Mycobacterial infection both generate distinctive VOCs. The objective of this study was to determine if breath VOCs contain biomarkers of active pulmonary tuberculosis. Head space VOCs from cultured Mycobacterium tuberculosis were captured on sorbent traps and assayed by gas chromatography/mass spectroscopy (GC/MS). One hundred and thirty different VOCs were consistently detected. The most abundant were naphthalene, 1-methyl-, 3-heptanone, methylcyclododecane, heptane, 2,2,4,6,6-pentamethyl-, benzene, 1-methyl-4-(1-methylethyl)-, and cyclohexane, 1,4-dimethyl-.

Breath VOCs were assayed by GC/MS in 42 patients hospitalized for suspicion of pulmonary tuberculosis and in 59 healthy controls. Sputum cultures were positive for Mycobacteria in 23/42 and negative in19/42 patients. Breath markers of oxidative stress were increased in all hospitalized patients (p<0.04). Pattern recognition analysis and fuzzy logic analysis of breath VOCs independently distinguished healthy controls from hospitalized patients with 100% sensitivity and 100% specificity. Fuzzy logic analysis identified patients with positive sputum cultures with 100% sensitivity and 100% specificity (95.7% sensitivity and 78.9% specificity on leave-one-out cross-validation); breath VOC markers were similar to those observed in vitro, including naphthalene, 1-methyl- and cyclohexane, 1,4-dimethyl-. Pattern recognition analysis identified patients with positive sputum cultures with 82.6% sensitivity (19/23) and 100% specificity (18/18), employing 12 principal components from 134 breath VOCs.

We conclude that volatile biomarkers in breath were sensitive and specific for pulmonary tuberculosis: the breath test distinguished between “sick versus well” i.e. between normal controls and patients hospitalized for suspicion of pulmonary tuberculosis, and between infected versus non-infected patients i.e. between those whose sputum cultures were positive or negative for Mycobacteria.

Introduction

The current global epidemic of pulmonary tuberculosis has highlighted the need for new screening tests that are rapid and accurate. The social burden of pulmonary tuberculosis has increased because many patients are also infected with human immunodeficiency virus (HIV), and the rates of multidrug-resistant tuberculosis are increasing.1 However, screening technology has not changed greatly during the past several decades. Many high-burden countries depend upon sputum smears and chest radiographs, supplemented by cultures when resources permit. This approach is highly specific for active pulmonary tuberculosis, but its value in primary screening is limited by low sensitivity and high cost.

We tested the hypothesis that volatile organic compounds (VOCs) in the breath might provide new biomarkers of active pulmonary tuberculosis. The rationale of this hypothesis is based on two observations: first, Mycobacteria produce distinctive patterns of VOCs in vitro, and second, patients with active pulmonary tuberculosis suffer from increased oxidative stress which also generates distinctive patterns of VOCs. Several species of Mycobacteria produce VOC metabolites that act as chemical “fingerprints”: M. avium, M tuberculosis, M. gordonae, M. gastri, M. kansasii, M. szulgai, and M. flavescens can be identified by their distinctive patterns of volatile metabolites, including C14–C26 fatty acids and their methylated and hydroxylated derivatives.2, 3, 4 Also, patients with active pulmonary tuberculosis suffer from increased oxidative stress: serum markers of oxidative stress, including lipid peroxidation products, conjugated dienes, malondialdehyde, and allantoin are generally increased in patients with active pulmonary tuberculosis, and decrease following a course of antituberculous therapy.5, 6, 7 Oxidative stress also liberates distinctive VOCs into the breath, particularly C4–C20 alkanes and methylated alkanes comprising the breath methylated alkane contour (BMAC).8 We have previously reported altered patterns of breath markers of oxidative stress in different diseases, including heart transplant rejection,9, 10 lung cancer,11, 12 breast cancer,13 ischemic heart disease,14 preeclampsia of pregnancy,15 and diabetes mellitus.16

We analyzed VOCs derived from Mycobacteria cultures, as well as VOC markers of oxidative stress in the breath of patients undergoing evaluation for Mycobacterial infection. Two different mathematical techniques, pattern recognition analysis and fuzzy logic, were employed to address two questions: First, could breath VOCs distinguish between healthy controls and all hospitalized patients undergoing evaluation for Mycobacterial infection (culture positive as well as culture negative)? Second, could breath VOCs distinguish between hospitalized patients whose sputum cultures for Mycobacteria were positive or negative?

Section snippets

Breath collection and assay

The method has been described.17 Subjects breathed in and out through the disposable mouthpiece of a portable breath collection apparatus for 2.0 min, and the VOCs in 1.0 l alveolar breath and 1.0 l room air were captured onto separate sorbent traps. VOCs captured on the sorbent traps were analyzed in the laboratory by automated thermal desorption, gas chromatography and mass spectroscopy (ATD/GC/MS).

Identification of VOCs produced by M. tuberculosis in vitro

Reference samples of M. tuberculosis were cultured in vitro (by VLB) utilizing VersaTREK Myco

In vitro studies

One hundred and thirty different VOCs were consistently detected in M. tuberculosis cultures in vitro, predominantly derivatives of benzene, naphthalene, and alkanes. The 10 most abundant VOCs are shown in Table 1.

Human studies

Patient characteristics are shown in Table 2.

Clinical course of hospitalized patients

All patients had three induced sputum for AFB. A total of 23/42 patients had sputum that was culture positive for M. tuberculosis. These patients were referred to the Bellevue Chest clinic/DOT clinic and followed. Of these patients, 16 had

Discussion

There were two main conclusions from this study: First, a set of breath VOCs accurately distinguished between normal controls and hospitalized patients undergoing screening for Mycobacterial infection. Second, another set of breath VOCs distinguished between hospitalized patients whose sputum cultures were positive or negative for Mycobacterial infection. Two different mathematical techniques—fuzzy logic analysis and pattern recognition analysis—independently generated similar conclusions.

These

Acknowledgements

This research was supported by SBIR award 1R43 AI52504-01 from the National Institute of Allergy and Infectious Diseases of the National Institutes of Health. Michael Phillips is President and CEO of Menssana Research, Inc.

Cited by (254)

  • Electronic nose based on metal oxide semiconductor sensors for medical diagnosis

    2024, Progress in Natural Science: Materials International
  • Detection of tuberculosis-associated compounds from human skin by GCxGC-TOFMS

    2023, Journal of Chromatography B: Analytical Technologies in the Biomedical and Life Sciences
  • A Clinical Breathomics Dataset

    2024, Scientific Data
View all citing articles on Scopus

Deceased.

View full text