Assessment of the accuracy of portable monitors for halitosis evaluation in subjects without malodor complaint. Are they reliable for clinical practice?

Abstract Halitosis is defined as a foul odor emanated from the oral cavity, with great impact in quality of life and social restraints. Recently, the use of Breath Alert™ in research increased significantly. Halimeter™, another portable device, is often used in clinical practice. Nevertheless, not many studies have verified the accuracy and compared the results of both devices simultaneously. Objective: To verify the accuracy of Breath Alert™ and Halimeter™ in patients without chief complaint of halitosis, using the organoleptic test (OT) as "gold standard." The second aim was to verify whether their concomitant use could enhance the diagnostic accuracy of halitosis. Material and Methods: A cross-sectional analytical study was performed. The quality of expired air of 34 subjects without chief complaint of halitosis was assessed. Two experienced examiners carried out the OT. Afterward, a third blinded examiner performed Halimeter™ (HT) and Breath Alert™ (BA) tests. Results: The OT identified halitosis in 21 subjects (62%). The area under the ROC curve (95% confidence interval) was 0.67 (0.48-0.85) and 0.54 (0.34-0.75) for HT and BA, respectively. The accuracy for HT and BA was 59% and 47%, respectively. The combined usage of HT and BA provided 11 positive results, being 9 subjects (43%) out of the total of 21 positive cases. Conclusions: Halimeter™ and Breath Alert™ were not able to diagnose halitosis in non-complainer subjects at the same level as the organoleptic examination, since their accuracy were low. Our results suggest that such portable devices are not reliable tools to assess halitosis and may neglect or misdiagnose a considerable number of patients in clinical practice.


Introduction
Halitosis is a universally experienced condition that has a variety of physiologic, pathologic and adaptive etiologic factors and affects nearly 15 to 30% of the population 5, 7,11,21 . The gold standard method to evaluate halitosis is the organoleptic or "sniff" test (OT), which access all oral odorants collectively. In this test, the examiner uses the sense of smell to detect malodor and subjectively score patients' halitosis 21 .
It has some important drawbacks, such as being dependent on someone's interpretation about the quality of the odor and the offensiveness score of the smell. Subjective measures are an issue in research, since objective data are more likely to be standardized.
In addition, the organoleptic test requires calibration and can be an embarrassing procedure 1,15 . Therefore, of other methods in order to decrease the need of performing OT.
Many efforts have been made to create reliable and objective methods to evaluate halitosis. Halimeter™ (Interscan Corporation, Chatsworth, CA, USA) is a portable monitor that measure the amount of sulphur compounds responsible for bad breath. However, it still does not meet all the requirements to be considered an "ideal" device in the assessment of halitosis 25 . Breath Alert™ is another device that (Tanita Corporation, Tokyo, Kantõ, Japan) has been gaining special attention in clinical practice 3,6,12,13,17 . It is a small handheld breath-checking equipment that measures and calculates the volatile sulphur compounds and hydrocarbon gases in expired air. Nevertheless, not in clinical practice.
We believe not enough studies have accessed the breath odor quality in patients without chief complaint of halitosis. Since some patients with halitosis might be unaware of their condition, we aimed to study a population more representative of daily clinical practice. Thus, the main purpose of this study was to access simultaneously the accuracy of Breath Alert™ and Halimeter™ in patients without chief complaint of halitosis, using the OT as "gold standard." We also hypothesized that the concomitant use of both methods could enhance the diagnostic accuracy of halitosis, and thus could be helpful in decreasing the use of OT in clinical practice.

Material and methods
The study design has followed the guidelines for quality of evidence for studies of diagnostic accuracy 19 .
A cross-sectional analytical study was performed using a convenience and consecutive sample. The human subject protocol was in accordance with the principles laid down in the Declaration of Helsinki and was approved by the correspondent Institutional Board (033/04). Informed consent was obtained from those who accepted the invitation. Participants were examined in two moments. A periodontist performed intraoral examination and instructed about the following visit for halitosis assessment.
In the second visit, two experienced examiners who

Subjects
The sample was composed of men and women the following criteria were eligible: not diagnosed with gingival and or chronic periodontal diseases based on 2 and without complaint of halitosis. In addition, oralthe previous 3 weeks) represented the exclusion criteria.
The subjects were informed about the breath malodor examination and received written instructions for the next appointment. They were instructed to not consume garlic, onion, spicy foods, and alcohol.
Gargling oral rinse and breath fresheners during 24 hours before the breath examination were not allowed as well. They were also instructed to not smoke during the 12 hours preceding the evaluation. Within the previous 2 hours, the subjects had to abstain the use of aromatic beverage, such as tea or coffee. They were Assessment of the accuracy of portable monitors for halitosis evaluation in subjects without malodor complaint. Are they reliable for clinical practice? told to have a meal and to perform their usual oral hygiene practices 2 hours before the assessment. At the day of halitosis assessment, the use of any scented cosmetic, such as perfume and aftershave, was not allowed once it could interfere in the examiners' olfactory sense and with the equipment sensors.

Organoleptic test
Examiners were instructed to postpone the assessment in the presence of any unpredictable situations that could lead to olfactory disturbance, such as postnasal drip, rhinitis, sinusitis or cold. Volunteers were required to close their mouth and refrain from talking for 3 minutes before the evaluation. Aiming  to 167).
The OT was positive for 21 participants (62%) ( Figure 1). The area under the ROC curve was 0.67 BA, respectively. The sensitivity was 33% and 24% for HT and BA, respectively ( Table 1).
The combined usage of HT and BA provided 11 positive results, being 9 (43%) out of the total of 21 positive cases. Using simultaneously all the methods and eleven as negative in the same way. (Figure 2).
The ppb mean value of VSC of the three positive cases was 167, 156 and 110 ppb (data not shown).