104 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS was determined by a pre-study screening evaluation to identify the qualifying panelists. The early morning screening evaluations were conducted using the same method as described below in the test phase. The exclusion characteristics consisted of being a smoker, having hard or soft tissue tumors of the oral cavity, having extensive caries or periodontal problems, and using full or partial dentures and/or orthodontic bands. Also excluded from the study were pan- elists who used medication such as antihistamines, antibiotics, or other types of med- ication for two weeks prior to or during the study. The study consisted of three phases: control, test, and recovery. During the control phase, subjects used their normal hygiene practices for three days, with evaluations on the second and third days. During the test phase, subjects were asked to use a placebo dentifrice in the morning and prior to retiring at night for two days (days 1 and 2), with evaluations on the second and third day. Subjects were asked to brush with the placebo dentifrice for 60 seconds and rinse with water. During the recovery phase subjects resumed their normal hygiene practices for two days as in the control phase, and were evaluated on the second and third days of the recovery phase. The purpose of this phase was to ascertain any carryover effect from the test phase. Spearman's rank correlation was computed to determine the relationship between the instrumental reading of VSC and the sensory ratings. In all phases, subjects were asked to use their treatment twice a day, in the morning and at night before retiring (days 1 and 2). Following the final nighttime treatment on the morning of day 3, organoleptic evaluations and sample storage collections were con- ducted. For each morning evaluation subjects reported to the test site without brushing, rinsing, eating, or drinking in order not to influence or wash away the VSC formed in the mouth air. At this time, two baseline samples were evaluated organoleptically and two storage samples were collected from each subject. Subjects were then given breakfast and the appropriate treatment: normal hygiene in the control phase, the placebo den- tifrice in the test phase, or normal hygiene in the recovery phase. Following the morning treatments, the subjects returned in three hours for the post-treatment organoleptic evaluations and sample storage collections. During this 3-hr interim between evalua- tions, subjects are asked not to eat or drink. RESULTS Table II shows both the mean instrumental and sensory mouth odor scores of subjects grouped according to treatment period of the study. For example, the mean and stan- dard deviation for the sensory ratings for day 3 of the control period is 7.15 +- 0.42, while the mean for the instrumental evaluations is 21.86 ng/ml +- 4.18. The results show that during the control period the correlation was poor (r = 0.22), while, during the test and recovery period, the values had improved with an r value of 0.77 and 0.78, respectively. These latter values were significant at a p-value of less than 0.01. Table III shows the instrumental mean values that correspond to each mean integer of each sensory rating (i.e., 5, 6, 7, 8). All subjects with mean sensory ratings equal to 5, 6, 7, or 8 were identified, and their corresponding instrumental mean values were

MOUTH ODOR EVALUATION 105 Table II Mean Instrumental Values vs Mean Sensory Ratings Determination (mean -+ S.D.) Control period Test period Recovery period N Day 2 Day 3 Day 2 Day 3 Day 2 Day 3 Sensory rating Volatile sulfur ng/ml Correlation (r) coefficient 20 7.15 6.98 6.43 6.45 6.38 6.55 + 0.42 + 0.76 + 0.73 -+ 0.82 + 0.70 + 0.69 20 21.86 22.77 20.14 20.16 23.12 22.22 -+ 4.18 + 6.00 + 6.12 + 4.97 -+ 6.13 -+ 4•06 0.22 (N.S.) 0.77 (p (0.01) 0.78 (p (0.01) Table III Calculation of Prediction Factor Mean sensory Mean instrumental readings ratings (total sulfur volatiles, ng/ml) A N* B Factor = B/A 8 4O 25.85 3.23 7.0 31 22.70 3.24 6.0 21 19.78 3.29 5.0 28 16.59 3.32 Average = 3.27 * Total of 120 determinations from all three phases of the study. averages for each sensory rating (5, 6, 7, or 8). This allows us to calculate a prediction factor (B/A). Note that this prediction factor is quite consistent over the sensory and instrumentation values obtained in this study, providing another indication of correla- tion between sensory and instrumental scores. A sensory rating of 5 was the lowest rating reported for any subject by the judges during this study. From Table I, a rating of 5 is the lowest rating for "definite" offensive mouth odor. From Table III, a rating of 5 corresponds to an instrumental score of 16.59. This instrumental value could be considered a threshold for odor offensiveness. In other words, if the total sulfide measurement in nanograms/milliliter for a given subject is below 16.59 ng/ml, it could be ascertained that the subject's mouth odor would not be of "definite" offensiveness (5 rating) but of "faint" offensiveness. DISCUSSION The organoleptic (sensory) rating is currently an acceptable in vivo procedure for deter- mining the efficacy of mouth rinses and dentifrices in mouth odor reduction. On the other hand, previous chemical, mass spectrometric, and gas chromatographic methods have established VSC as being the primary source of offensive local mouth odor (4). Therefore, a number of investigations have been conducted to establish a correlation between instrumental techniques and a subjective sensory method of evaluating mouth