| Sign In to gain access to subscriptions and/or personal tools. |
Clinical Assessment of Oral Malodor by the Electronic Nose System
1 Department of Preventive Dentistry, Graduate School of Dentistry, Osaka University, 1-8, Yamadaoka, Suita, Osaka 565-0871, Japan; and Correspondence: * corresponding author, tanakam{at}dent.osaka-u.ac.jp
A recently developed electronic nose has not yet been clinically applied to evaluations of oral malodor. This investigation sought to determine whether an electronic nose could clinically assess oral malodor. Twenty-nine healthy adults and 49 patients were assessed by results of an actual organoleptic test, a score representing malodor strength with an electronic nose in "top-note" mode (top-note score), and measurements of volatile sulfur compound (VSC) concentrations. The correlation coefficient between top-note and actual organoleptic scores (r = 0.71) was comparable with the log VSC and actual organoleptic scores (r = 0.63). However, the area under the receiver-operating characteristic plots for top-note score was significantly larger than that for log VSC. In logistic regression analyses with top-note score as a dependent variable, probing depth, tongue coating, and plaque control record each had independent associations. Our findings suggest that the top-note score from an electronic nose examination may be useful for the clinical evaluation of oral malodor.
Key Words: oral malodor • clinical assessment • electronic nose • oral health parameters
The two primary methods used for the clinical analysis of oral malodor are organoleptic measurement and instrumental analysis, which involves gas chromatography and a portable sulfide monitor (Rosenberg, 1996). The former is simple, since the examiner directly sniffs expelled mouth air and rates the malodor intensity on a scale from 0 to 5; however, training is required to minimize inter- and intra-examiner variability. In contrast, a gas chromatography unit has highly selective and sensitive flame photometric detection systems specific for volatile sulfur compounds (VSC), and can readily identify hydrogen sulfide (H2S), methylmercaptan (CH3SH), and dimethyl sulfide (CH3SCH3), as well as quantitatively determine VSC levels in breath (Oho et al., 2001; Furne et al., 2002). However, since this instrument is unable to detect other volatile compounds, there remains no ideal objective method by which the degree of oral malodor can be clinically assessed. In recent years, the advent of chemical sensors and chemical sensor systems (so-called "electronic noses") has made possible fast and simple odor analyses for use in many fields (Mantini et al., 2000). This powerful technology has only recently been introduced in the field of medicine; however, its potential to assist in diagnosis is promising (Thaler et al., 2001). Although several researchers have proposed the application of electronic noses for medical purposes, very few studies have been conducted to investigate their possible use for the clinical assessment of oral malodor. The electronic nose system used in the present study was developed for quantitative measurement of malodor present in food and beverages (Kita et al., 2000), and we applied it clinically to assess oral malodor and to examine the association between oral malodor strength and oral health status.
VSC Preparation as Standard Gases Using a PD-1B permeation device (Gastec Co., Ltd., Kanagawa, Japan), we prepared H2S, CH3SH, and CH3SCH3 and added them separately to polypropylene bags containing 3000 mL of air (GL Science Inc., Tokyo, Japan). Final H2S, CH3SH, and CH3SCH3 concentrations were adjusted to approximately 0.05, 0.1, 0.5, and 1 ppm, respectively, while an additional mixture of 0.01 ppm of CH3SCH3 was also prepared. These gases were analyzed by means of gas chromatography and the electronic nose in top-note mode in triplicate examinations.
Subjects
Measurement with the Electronic Nose
Statistical Analysis The following methods can be found in the Web-only APPENDIX: Oral Gas Sampling, Measurement with Gas Chromatography, Organoleptic Assessment, and Periodontal and Dental Examinations.
The numbers 1 and 2 sensors of the electronic nose yielded an exponential function-like response when exposed to a constant concentration of H2S, CH3SH, and CH3SCH3, whereas sensors 3 and 4 showed a weak response to each VSC (Fig. 1A  ). The concentrations of H2S, CH3SH, and CH3SCH3 estimated by the electronic nose in top-note mode were virtually identical to those found with the gas chromatograph (Fig. 1B).
Correlations between oral malodor levels determined from organoleptic and electronic nose top-note scores, and log total VSC by gas chromatography, were examined. The correlation coefficient between top-note scores and actual organoleptic scores was 0.71 (Fig. 2A ), which was nearly the same as that between log VSC and organoleptic scores (r = 0.63) (Fig. 2B). No significant difference was found between these two correlation coefficients (p = 0.366). ROC sensitivity-specificity plots provided an appropriate means for comparison of the top-note scores and log VSC data for classifying subjects with or without a  2.0 actual organoleptic score. The area under the ROC plots that represented the top-note scores (0.879) was significantly larger than that for log VSC (0.698) (p = 0.0052).
In bivariate analyses using age, gender, smoking, and 3 levels of oral health status as variables (Table 1 ), we found the top-note score to be significantly greater in subjects who showed higher values for age, smoking, probing depth, tongue-coating score, and plaque control record. Further, organoleptic score was significantly associated with age, pocket depth, tongue-coating score, and plaque control record. In contrast, total VSC level had a significant association with only probing depth and tongue-coating score.
The independent effects of those variables associated with top-note score levels were analyzed by multiple linear and logistic regression analyses (Table 2 ). In the linear regression analyses, the percentage of teeth with a 4 mm or greater probing depth, tongue-coating score, and plaque control record had a significantly independent effect on top-note score, with an explanatory power of 41%. Further, in logistic regression analyses of the possibility of subjects being in the upper 25th percentile of top-note score distribution as a dependent variable, gender, percentage of teeth with a 4 mm or greater probing depth, tongue-coating score, and plaque control record were each significantly associated with top-note score, with odds ratios of 22.7, 22.0, 9.8, and 12.5, respectively.
The independent effects of top-note scores and log VSC, as well as other clinical parameters for actual organoleptic score as dependent variables, were tested by linear regression analysis. As a result, top-note score, plaque control record, log VSC, and age each had a significantly positive independent effect, with an explanatory power of 51%, 7%, 5%, and 3%, respectively (data not shown).
We found that at least 2 of the 6 sensors in the electronic nose were sensitive for the detection of VSC, which enabled the device to detect malodorous substances in a manner that was highly correlated with the concentration of each. Thus, with the present device, malodor may be able to be precisely measured as VSC levels in mouth air samples. Further, since the present electronic nose system can also detect such volatile compounds as organic compounds, aromatic compounds, amine-containing compounds, and ammonia derivatives in food and beverages (Kita et al., 2000), those sensors insensitive to VSC may be able to determine other volatile compounds in mouth air samples.
In the present study, we measured oral malodor strength using an electronic nose system in top-note measurement mode. In a comparison with all-note and deep-note measurement modes, the top-note score showed the highest correlation with organoleptic score, which suggested that the main compounds related to oral malodor are volatile substances with a low boiling point. There is the possibility that volatile components of mouthrinses or toothpastes could influence top-note vales, but this was not observed with the toothpaste in our preliminary experiment. A traditional problem with the use of electronic noses is the influence of water vapor. In this study, a water vapor trap was used that eliminated this problem. The current instrumentation has been used on a nearly daily basis for 4 yrs, and no evidence of any effects of water vapor on the sensors has been observed. Further, the correlation coefficient between the top-note and organoleptic scores (r = 0.71) was nearly the same as that between log VSC by gas chromatography and organoleptic score (r = 0.63), suggesting that the electronic nose system may be useful for the measurement of non-sulfur gaseous compounds as well as VSC. The area under the ROC curve representing top-note score was significantly larger than that for log VSC. This result suggests that the top-note score determined with the electronic nose had a higher level of accuracy than log VSC for classifying subjects with and without an actual organoleptic score of In the present subjects, bivariate analyses showed that age, smoking, probing depth, tongue-coating score, and plaque control record each had a significant association with top-note score. However, only probing depth and tongue-coating score were identified as having significant relationships to VSC level measured with gas chromatography, while age, probing depth, tongue coating, and plaque control record were each associated with actual organoleptic score. In linear regression analyses, the explanatory power for top-note score was 41%, while that for total VSC level was 15%, and for organoleptic score, 45% (data not shown). In addition, in a linear regression analysis with actual organoleptic score as a dependent variable, the explanatory power of top-note score was much higher (51%) than that of log VSC (7%), suggesting that top-note score may be more strongly associated with oral malodor than VSC by gas chromatography. Associations between oral health parameters and malodor level have also been shown in previous studies, though with conflicting results. Using both sulfide measurements and organoleptic tests, Bosy et al.(1994) reported that oral malodor was not associated with periodontal probing depth, gingival index, or plaque index, whereas a significant correlation was observed between periodontal condition and oral malodor with both methods in other studies (Rosenberg et al., 1991; Kozlovsky et al., 1994; Morita and Wang, 2001). Further, tongue coating was found to have an association with oral malodor in organoleptic tests and portable sulfide monitor results (Miyazaki et al., 1995; Morita and Wang, 2001; Oho et al., 2001). Yaegaki and Sanada (1992) found a greater association of tongue coating with VSC formation in periodontitis subjects than in healthy individuals. Plaque accumulation has also been weakly associated with sulfide measurement and organoleptic test results in one report (Rosenberg et al., 1991), but with only organoleptic test results in another (Kozlovsky et al., 1994). Our present findings suggest that the effectiveness of the electronic nose can be best attributed to the cumulative effect of several factors, including the amount of supragingival plaque, the depth of periodontal pockets, and tongue coating. We concluded that top-note score as determined with the use of an electronic nose can provide objective halitosis-related measurements, though those obtained by the multiple linear regression method showed only relative and not absolute values. Further investigations regarding the absolute expression of malodor intensity with an electronic nose will be conducted.
This work was supported by a Grant-in-Aid for Scientific Research from the Ministry of Education, Science, Sports, and Culture of Japan (B-15390650).
A supplemental appendix to this article is published electronically only at http://www.dentalresearch.org. Received for publication April 28, 2003. Revision received January 23, 2004. Accepted for publication January 26, 2004.
Journal of Dental Research, Vol. 83, No. 4,
317-321 (2004) This article has been cited by other articles:
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|
|
 |
|
Sign In to gain access to subscriptions and/or personal tools.
TOP
ABSTRACT
INTRODUCTION
sensor 2, 
sensor 3, sensor 4, 
sensor 5, 
sensor 6) (A), as well as mean VSC concentrations estimated with the electronic nose in top-note mode (B).
