original article volatile organic compounds arising from...

33

VOCs arising from stent biofi lm 69

Original Article J. St. Marianna Univ. Vol. 5, pp. 69–75, 2014

Volatile Organic Compounds Arising from Tracheobronchial

Stent-related Biofi lm Formation Detected in Patients’ Breath by

Ion Mobility Spectrometry

Teppei Inoue, Hiroshi Handa, Takeo Inoue, Masamichi Mineshita, and Teruomi Miyazawa

(Received for Publication: September 3, 2014)

AbstractBackground: Tracheobronchial stent-related biofi lm formation encountered in interventional pulmonology that can result in pneumonia and in tissue granulation. Breath analysis by means of ion mobility spectrometry (IMS) has been reported for detection of volatile organic compounds (VOCs) indicative of bacteria biofi lm. We hy-pothesized that IMS can detect specifi c VOCs in patients with bacterial biofi lm formation resulting from tra-cheobronchial stent placement, and we tested our hypothesis in a prospective study conducted between October 2011 and October 2012.Methods: During the study period, stent placement, removal, or replacement was performed in 21 patients with tracheobronchial stenosis. A silicone stent was used in 11 of these patients. In 11 cases, 5 cases were placement procedures and 6 cases were replacement or removal procedures. Breath samples were obtained from these 6 patients before and after stent removal or replacement, and the samples were analyzed by IMS. Peaks were characterized with the use of Visual Now 2.2 software (B&S Analytik, Dortmund, Germany).Results: Bacteriologic culture from the removed silicone stents showed Pseudomonas aeruginosa biofi lm for-mation in 4 of the 6 patients. Wilcoxon-Rank tests was applied to 36 peaks. Box-and-whisker plots were drawn, and 12 peaks were identifi ed. Five of the 12 peaks differed signifi cantly in signal intensity after removal of the initial stent. Comparison against an IMS database identifi ed the following: peak no. 22 (P22), unknown (p < 0.05); P23, limonene (p < 0.05); P24: 2,2,4,6,6-pentaheptylmethane (p < 0.05); P31, 1-octanol (p < 0.05); and P35: phenylacetaldehyde (p < 0.05).Conclusions: Using IMS, we were able to assess the degree of Pseudomonas aeruginosa biofi lm formation resulting from stent placement. IMS breath analysis in conjunction with clinical symptoms will provide for non-invasive assessment of the need for stent replacement or removal.

Key wordsbreath analysis, silicone stent, Pseudomonas aeruginosa

Division of Respiratory and Infectious Disease, Department of Internal Medicine, St. Marianna University School of Medi-cine, 2–16–1 Sugao Miyamae, Kawasaki, Kanagawa, 216–8511 Japan

Introduction

In patients with a tracheobronchial stent, the biofi lm that forms around the stent often becomes problematic. The formation of such biofi lm can result in pneumonia, mucus obstruction inside the stent, and granulation tissue formation. Previously, we re-ported stent-related complications in 35 patients

treated for central airway stenosis. The specifi c com-plications listed in the order of prevalence were as follows: mucus obstruction inside the stent (31%), granulation tissue formation at the edge of the stent (49%), tumor overgrowth (9%), and stent migration (11%)1). In previous patient series that we have re-ported2–4), we found that it was extremely important to control the development of Pseudomonas aerugi-

34

Inoue T Handa H et al70

nosa (P. aeruginosa) biofi lm. Moreover, it has been reported that stent-associated respiratory tract infec-tion (SARTI) can develop in as many as 20% of pa-tients who have undergone airway stenting, with P. aeruginosa as the main causative bacteria5). For pa-tients who have received an airway stent, it is impor-tant that any biofi lm be evaluated thoroughly and noninvasively.

Ion mobility spectrometry (IMS) can detect traces of volatile organic compounds (VOCs) in pa-tients’ exhaled breath and is used to diagnose lung cancer, interstitial pneumonia, asthma, and various other lung diseases, including infectious diseases6–14). To our knowledge, there have been no studies of VOCs emitted from bacteria on previously placed stents. We hypothesized that IMS can be used to detect specifi c VOCs in the exhaled breath of patients in whom P. aeruginosa biofi lm has developed as a result of silicone stent placement, and we conducted a prospective single-center study to investigate whether IMS detection of these specifi c VOCs can be used to prevent the development of bacteria-induced complications in patients with a silicone stent.

Materials and Methods

Ethics committee approvalThis study was approved by the Research Ethics

Committee of St. Marianna University (approval number 1820) and registered with the Medical Infor-mation Network (UMIN 000006696). Written in-formed consent was obtained from all study patients for their participation.

PatientsThe study was conducted between October 2011

and October 2012, and we enrolled patients with various lung diseases who were over 20 years of age and judged to be good candidates for necessary stent placement, replacement, or removal. The study pro-tocol involved the use of IMS to analyze breath samples obtained before and after silicone stent re-moval or replacement.

Breath sampling and IMSExhaled breath was obtained from each patient

with a spirometer that has a CO2-controlled sample inlet (Ganhorn Medizin Electronic, Niederlauer, Ger-many), and 10 mL samples were analyzed by means of a multi-capillary column coupled with an ion mo-bility spectrometer (MCC-IMS, B&S Analytik, Dort-mund, Germany). IMS for analysis of exhaled breath

has been described previously23). Briefl y, within the spectrometer, a 95 MBq 63Ni ß-radiation source was applied for ionization of the carrier gas (Nippon Megacare, Tokyo, Japan). The spectrometer was con-nected to a polar multi-capillary column (MCC, type OV-5, Multichrom Ltd, Novosibirsk, Russia), which was used as the pre-separation unit. The exhaled breath sample was sent through the MCC, which consists of 1000 parallel capillaries, each with an in-ner diameter of 40 µm and a fi lm thickness of 200 nm. The total diameter of the separation column is 3 mm.

Exhaled breath was taken directly into the spi-rometer, and the 500 mL dead volume was excluded from analysis. The 10 mL content of the sample loop was transferred to the inlet of the MCC and after pre-separation was transferred directly into the ionization region of the ion mobility spectrometer. The MCC and the drift tube of the spectrometer were held iso-thermally at 40°C.

IMS breath analysis was performed immediately before stenting or removal of a stent. In addition, we cultured the cells present on the stents that were re-moved to determine the types of bacteria that were prevalent. Breath analysis was performed again 1 week after the stent procedure for comparison of the pre- and post-operative results.

Statistical analysisIMS analysis was performed before and 1 week

after the stent procedure. Peaks on the IMS-chro-matograms were characterized with the use of Visu-alNow 2.2 software (B&S Analytik)19) 24–27). All peaks were characterized in relation to drift time (corre-sponding 1/K0 value) and retention time and their concentration related to the peak height. For the dif-ferent groups (pre-operative IMS and post-operative IMS) and each of the peaks, a box-and-whisker plot was drawn. The rank sum provided by Wilcoxon-Mann-Whitney test was obtained by VisualNow 2.2, and the peaks showing the greatest differences be-tween groups were ranked.

The observed peaks were compared against those recorded in the IMS 110801 databased (B&S Analytik), which contains reference measurements, as described elsewhere15) 28–30). The peaks observed in the present study were compared to those in the ref-erence database.

35


Results

Study patients and stent proceduresStenting, with either a metallic or silicone stent

was performed during the study period in 21 patients with tracheobronchial stenosis resulting from lung cancer, tuberculosis, relapsing polychondritis, medi-astinal seminoma, or esophageal cancer. Eleven in the 21 cases involved a silicone stent that required initial placement (n = 5) or removal and replacement (n = 6). P. aeruginosa was detected in 4 cases of the 6 cases: in 1 case of stent removal after successful chemotherapy and in 3 cases of stent replacement (Fig 1). Of the 4 patients (3 men, 1 woman), 2 were being treated for tracheobronchial malacia due to tuberculosis, 1 was being treated for tracheobron-chial malacia due to relapsing polychondritis, and 1 was being treated for tracheal obstruction due to me-diastinal seminoma.

Bacteriologic inspectionPre- and post-extraction bronchoscopic images

and an electron micrograph of an extracted stent are shown in Fig 2. In 4 patients, P. aeruginosa was isolated by bacteriologic culture of cells attached to the removed silicone stents. The pre-bronchoscopic image in Fig 2 shows P. aeruginosa biofi lm forma-tion on the inside surface of the silicone stent lumen.

Results of IMSAn example IMS-chromatogram is shown in Fig

3. Wilcoxon’s rank-sum test was applied to 36 peaks in total. Box-and-whisker plots were drawn, and 12 peaks were shown to have signal differences. Five of the 12 peaks differed signifi cantly in signal intensity from those observed before the stent procedures (Fig. 4). From the database provided, a signifi cant differ-ence was observed for peak no. 22 (P22) (1/K0 (drift time) = 0.7865, RT (retention time) = 26.2): un-known (p < 0.05); P23 (1/K0 = 0.6351, RT = 25.4):

Figure 1. Study fl ow chart.

Figure 2. Bronchoscopic and electron micrograph images obtained in a case of relapsing polychondritis.

A, Bronchoscopic image obtained before silicone stent replacement; B, Electron micrograph of

biofi lm extracted before stent replacement; C, Bronchoscopic image obtained after silicone

stent replacement.

36


limonene (p < 0.05), P24 (1/K0 = 0.6663, RT = 25.4): 2,2,4,6,6-Pentaheptylmethane (p < 0.05); P31 (1/K0 = 0.6985, RT = 45.3): 1-octanole (p < 0.05); and P35 (1/K0 = 0.7643, RT = 26.9): phenylacetaldehyde (p < 0.05). Furthermore, although not signifi cant, a de-crease was seen in P2 (1/K0 = 0.5839, RT = 39.2): acetophenone; P13 (1/K0 = 0.7551, RT = 10.4): cy-clohexanole; P20 (1/K0 = 0.7338, RT = 57.5): non-anal; and P30 (1/K0 = 0.5746, RT = 27.3): 4-Isopro-pyltoluene. In contrast, an increase was seen in P6 (1/K0 = 0.4922, RT = 5.3): propanol; P15 (1/K0 = 0.5487, RT = 7.3): 3-pentanone; and P32 (1/K0 = 0.5487, RT = 7.5): phenylacetylene (Table 1). Peaks were compared by their actual position to the nearest peak in the reference database.

Discussion

To our knowledge, this is the fi rst prospective clinical study to investigate the usefulness of exhaled

breath analysis before and after stenting. We found that IMS can detect specifi c VOCs in patients with P. aeruginosa biofi lm resulting from silicone stent placement. Furthermore, using IMS in conjunction with patients’ clinical symptoms, we were able to consider whether to replace or remove the silicone stents.

As noted above, bacteria-laden biofi lms can cause SARTI or formation of friable granulation tis-sue. Several bacterial species have been found in stent biofi lms, and these are all thought to produce various VOCs. P. aeruginosa, Staphylococcus spe-cies, and fungi are frequently identifi ed in cases of SARTI5). The main causative agent of SARTI is P. aeruginosa. Bacterial analysis of the stents in our study patients confi rmed the presence of P. aerugi-nosa in several biofi lms.

Recently, Bos et al. found that propanol and 2-pentanone (VOCs) are related to pathogens includ-

Figure 3. Best separation of 36 peaks on an example IMS-chromatogram obtained for a patient with relapsing

polychondritis. A, IMS-chromatogram obtained before replacement of the silicone stent; B, IMS-chro-

matogram obtained after replacement of the silicone stent.

37


Figure 4. Box-and-whisker plots of peaks on IMS chromatograms obtained before and after removal of a silicon

stent. After the infectious stent was removed from the respiratory tract, the signal intensity of the 5

peaks decreased signifi cantly. The peak (P) numbers are shown.

Pre: before removal of the silicon stent (n = 4).

Post: after removal of the silicon stent (n = 4).

Peak no. IMS Database P2 Acetophenone P6 Propanol

P13 Cyclohexanol P15 3-Pentanone P20 Nonanal P22 Unknown P23 Limonene P24

eneulotlyporposI-403PlonatcO-113P

enelytecalynehP23PedyhedlatecalynehP53P

Peaks ed according to the IMS Visual Now 110801 database.

2,2,4,6,6-Pentaheptylmethane

Table 1. Peaks and the Volatile Organic Compounds They Represent

ing P. aeruginosa31). Furthermore, Thomas et al. found 2-aminoacetophenon, which is excreted by P. aeruginosa, to be a potential breath biomarker for cystic fi brosis32). We found that a single bacterium may not necessarily produce only a single type of compound; even P. aeruginosa was identifi ed on the basis of a combination of VOC peaks. We also found that particular bacteria produced particular VOCs. However, considering that bacteria in biofi lms pro-

duce several compounds, the combined VOC profi le constructed from our total biofi lm data may explain the signal differences we observed between some peaks. In addition, the number of cases in which stents were removed and microbiologically analyzed was small. Future investigations should include the assessment of VOCs in patients with newly placed stents in addition to the 5 VOC peaks obtained in this study that were considered biofi lm markers useful for

38


evaluating the presence and degree of biofi lm forma-tion. The increased strength of combined multiple peak signals rather than the increased strength of a single peak signal could aid in biofi lm assessment. Other studies on assessment of bacterial infection by means of IMS have shown the importance of correla-tion when combining peaks13) 14). Thus, even though assessment can be performed on the basis of a single peak, we believe that the presence and degree of a biofi lm can be accurately inferred through the use of a combined peak analysis.

In addition to the breath analysis, we tested the phlegm of all four patients 1 week after the stent procedure; however, no new data were obtained. This is because the absolute numbers of bacteria after stent removal had decreased markedly in comparison to the numbers before stent removal, making our study of phlegm for bacterial composition and num-ber uninformative. We believe that highly accurate results can be obtained when the type of bacteria on the stent is examined microbiologically and the num-bers of bacteria are determined on the basis of IMS peak signal strength. Furthermore, our study showed that the presence and degree of biofi lm formation can be assessed accurately by means of IMS. By tracking the peaks identifi ed in this study, the removal or ex-change of silicone stents can be considered before the harmful effects of bacteria-laden stent biofi lms (i.e. SARTI) set in.

The results of our study may be especially im-portant to the future of stent-insertion. Until now, highly invasive methods such as bronchoscopy or highly inaccurate methods such as sputum assess-ment have been relied upon for assessing the risk associated with stent placement. However, IMS can be used to assess the presence of infectious biofi lm with minimal invasion and high accuracy. This will in turn allow for preventive measures to be taken before the onset of SARTI. Our long-term objective is to prevent the onset of SARTI by using the present study results, to perform periodic breath analysis of patients with stents, and to continuously monitor changes in the 5 biofi lm markers we identifi ed. The ultimate goal of our research is to use IMS to evalu-ate biofi lm development and control its progress. Studies in larger patient groups are needed to confi rm the fi ndings reported herein.

Acknowledgements

The authors would like to thank Mr. Jason Tonge of St. Marianna University School of Medicine for

reviewing this English manuscript and Prof. J. I. Ba-umbach of the Faculty of Applied Chemistry at Re-utlingen University (Reutlingen, Germany) for re-viewing our statistical analysis.

References

1) Miyazawa T, Arita K. Airway stenting in Japan. Respirology 1998; 3: 229–234.

2) Miyazawa T, Yamakido M, Ikeda S, Furukawa K, Takiguchi Y, Tada H, Shirakusa T. Implanta-tion of ultrafl ex nitinol stents in malignant tra-cheobronchial stenoses. Chest 2000; 118: 959–965.

3) Iwamoto Y, Miyazawa T, Kurimoto N, Miyazu Y, Ishida A, Matsuo K, Watanabe Y, Interven-tional bronchoscopy in the management of air-way stenosis due to tracheobronchial tuberculo-sis. Chest. 2004; 126: 1344–1352.

4) Nishine H, Hiramoto T, Kida H, Matsuoka S, Mineshita M, Kurimoto N, Miyazawa T. Assess-ing the site of maximal obstruction in the trachea using lateral pressure measurement during bron-choscopy. Am J Respir Crit Care Med 2012; 185: 24–33.

5) Agrafi otis M, Siempos II, Falagas ME. Infec-tions related to airway stenting. Respiration 2009; 78: 69–74.

6) McCulloch M, Jezierski T, Broffman M, Hub-bard A, Turner K, Janecki T. Diagnostic accu-racy of canine scent detection in early- and late-stage lung and breast cancers. Integr Cancer Ther 2006; 5: 30–39.

7) Ehmann R, Boedeker E, Friedrich U, Sagert J, Dippon J, Friedel G, Walles T. Canine scent detection in the diagnosis of lung cancer: revisit-ing a puzzling phenomenon. Eur Respir J 2012; 39: 669–676.

8) Montuschi P, Mores N, Trové A, Mondino C, Barnes PJ. The electronic nose in respiratory medicine. Respiration 2013; 85: 72–84.

9) D’Amico A, Di Natale C, Falconi C, Martinelli E, Paolesse R, Pennazza G, Santonico M, Sterk PJ. U-BIOPRED study. Detection and identifi ca-tion of cancers by the electronic nose. Expert Opin Med Diagn: 2012; 6: 175–185.

10) Darwiche K, Baumbach J. I, Sommerwerck U, Teschler H, Freitag L. Bronchoscopically ob-tained volatile biomarkers in lung cancer. Lung 2011; 189: 445–452.

11) Kikowatz A, Becher G, Dietze S, Steinhäusser W, Beck E. Differential ion mobility spectros-

39


copy: non-invasive real-time diagnostics and therapy control in metabolic diseases. Eur J Med Res 2009; 14: 121–125.

12) Westhoff M, Litterst P, Freitag L, Baumbach J. I. Ion mobility spectrometry in the diagnosis of sarcoidosis: results of a feasibility study. J Physiol Pharmacol 2007; 58: 739–751.

13) Rabis T, Sommerwerck U, Anhenn O, Darwiche K, Freitag L, Teschler H, Bödeker B, Maddula S, Baumbach J. I. Detection of infectious agents in the airways by ion mobility spectrometry of exhaled breath. Int. J. Ion Mobility Spectrom 2011; 14: 187–195.

14) Maddula S, Rabis T, Sommerwerck U, Anhenn O, Darwiche K, Freitag L, Teschler H, Baum-bach J. I. Correlation analysis on data sets to detect infectious agents in the airways by ion mobility spectrometry of exhaled breath. Int. J. Ion Mobility Spectrom 2011; 14: 197–206.

15) Jünger M, Bödeker B, Baumbach J. I. Peak as-signment in multi-capillary column - ion mobil-ity spectrometry using comparative studies with gas chromatography - mass spectrometry for exhalred breath analysis. Anal. Bioanal Chem 2010; 396: 471–482.

16) Westhoff M, Litterst P, Freitag L, Urfer W, Bader S, Baumbach J. I. Ion mobility spectrom-etry for the detection of volatile organic com-pounds in exhaled breath of patients with lung cancer: results of a pilot study. Thorax 2009; 64: 744–748.

17) Bunkowski A, Boedeker B, Bader S, Westhoff M, Litterst P, Baumbach J. I. MCC/IMS signals in human breath related to sarcoidosis-results of a feasibility study using an automated peak fi nd-ing procedure. J. Breath Res 2009; 3: 046001.

18) Westhoff M, Litterst P, Freitag L, Baumbach J. I. Ion mobility spectrometry in the diagnosis of sarcoidosis: results of a feasibility study. J Physiol Pharmacol 2007; 58: 739–751.

19) Westhoff M, Litterst P, Maddula S, Bödeker B, Baumbach J. I. Statistical and bioinformatical methods to differentiate chronic obstructive pul-monary disease (COPD) including lung cancer from healthy control by breath analysis using ion mobility spectrometry. Int. J. Ion Mobility Spectrom 2011; 14: 139–149.

20) Baumbach J. I. Process analysis using ion mo-bility spectrometry. Anal Bioanal Chem 2006; 384: 1059–1070.

21) Bunkowski A, Maddula S, Davies A. N,

Westhoff M, Litterst P, Bödecker B, Baumbach J. I. One-year time series of investigations of analytes within human breath using ion mobility spectrometry. Int. J. Ion Mobility Spectrom 2010; 13: 141–148.

22) Bödeker B, Davies A. N, Maddula S, Baumbach J. I. Biomarker validation—room air variation during human breath investigations. Int. J. Ion Mobility Spectrom 2010; 13: 177–184.

23) Maddula S, Blank L, Schmid A, Baumbach J. I. Detection of volatile metabolites of Escherichia coli by multi capillary column coupled ion mo-bility spectrometry. Anal Bioanal Chem 2009; 394: 791–800.

24) Bödeker B, Baumbach J. I. Analytical Descrip-tion of IMS-Signals. Int. J. Ion Mobility Spec-trom 2009; 12: 103–108.

25) Bödeker B, Vautz W, Baumbach J. I. Peak Find-ing and Referencing in MCC/IMS – Data. Int. J. Ion Mobility Spectrom 2008; 11: 83–87.

26) Bödeker B, Vautz W, Baumbach J. I. Peak Com-parison in MCC/IMS – Data – Searching for potential biomarkers in human breath data. Int. J. Ion Mobility Spectrom 2008; 11: 89–93.

27) Bödeker B, Vautz W, Baumbach J. I. Visualisa-tion of MCC/IMS – Data. Int. J. Ion Mobility Spectrom 2008; 11: 77–81.

28) Maddula S, Rupp K, Baumbach J. I. Recom-mendation for an upgrade to the standard format in order to cross-link the GC/MSD and the MCC/IMS data. Int. J. Ion Mobility Spectrom 2012; 15: 79–81.

29) Jünger M, Vautz W, Kuhns M, Hofman L, Ul-bricht S, Baumbach J. I, Quintel M, Perl T. Ion mobility spectrometry for microbial volatile or-ganic compounds: A new pathogen identifi cation tool by smelling human pathogenic bacteria. Appl. Microbiol. Biotechnol 2012; 93: 2603–2614.

30) Bunkowski A. Software tool for coupling chro-matographic total ion current dependencies of GC/MSD and MCC/IMS. Int. J. Ion Mobility Spectrom 2010; 13: 169–175.

31) Bos L. D. J, Sterk P. J, Schultz M. J. Volatile Metabolites of Pathogens: A Systematic Review. PLOS Pathogens 2013; 9:e1003311.

32) Scott-Thomas A, Syhre M, Pattemore P, Epton M, Laing R, Pearson J, Chambers S. 2-Amino-acetophenone as a potential breath biomarker for Pseudomonas aeruginosa in the cystic fi brosis lung. BMC Pulmonary Medicine 2013; 7; 10:56.

original article volatile organic compounds arising from...

Documents