154 JOURNAL OF COSMETIC SCIENCE in the perception of sensitive skin by women from different geographical areas (4) and ethnic origins (3 ). Incidence of self-perceived sensitive skin is lower in the male popu- lation (30%) (2,4). It tends to decrease with age (4) and in summer (5 ). Although these unpleasant cutaneous signs tend to disappear quickly, this condition can render the use of certain kinds of cosmetic product very problematic (2,3). The uncomfortable sensa- tions with no visible signs represent 50% of the reported adverse reactions to cosmetics and toiletries (6) and could be the very first symptoms of an irritant contact dermatitis (7). Consequently, because of the awareness of that problem, cosmetic manufacturers have tried to develop formulations without ingredients likely to induce uncomfortable sensations or containing soothing ingredients intended to decrease skin over-reactivity (8). Products labeled for sensitive skins have met with a growing success: they are purchased by 80% of self-assessed sensitive skin subjects and by 25% of those with non-sensitive skin (9). Sensitive skin is not a pathological disorder (1). This condition is not easy to assess because it lacks both a consensual definition (9) and visible, physical, or histologically measurable signs. Consequently, there is neither a simple nor pertinent method available to assess sensitive skin (10), unlike the dry or greasy character of skin. The absence of any measurable physical features of self-declared sensitive skin by classic methods has even led some authors to question the reality of this skin condition (11). Owing to the difficulty in exploring this subjective disorder, psychophysical tests based on the report of sensations induced by topically applied chemical probes have been proposed to identify individuals with sensitive skin and subsequently to test products for "sensitive skin" in this population. The approach was recently validated by functional magnetic resonance imaging (£MRI). The neurophysiological reality of sensitive skin was illustrated by a specific pattern of brain activation in self-assessed sensitive-skin subjects when a chemical probe was applied to their face (12). Moreover, self-reported unpleasant sensations have been reported as a useful tool for irritancy assessment of detergents and soaps (7). In this study, panelists differentiated products in terms of sensation of dryness several washings before observable differences were detected. Nevertheless, much greater attention was given to the visually observable or instrumentally measurable signs of irritation such as redness and inflammation (13). The lactic acid stinging test, proposed in 1977 ( 14), relies on the intensity of stinging sensations induced by a lactic acid solution applied on the nasolabial folds. The subjects who report stinging sensations are called "stingers." Using a slightly modified procedure (1), this test is currently performed by cosmetic manufacturers for the selection of "stingers." This selected population prone to experience neurosensory problems with topical products is asked to test new products to substantiate claims indicating that they are appropriate for sensitive skin. Although very useful for product safety, the lactic acid stinging test does not fully render the complexity of self-assessed sensitive skin, as illustrated by the discrepancy between acid lactic response and self-perception of sen- sitive skin (5,9,10). In 2000, this difference was taken into account for the recommen- dation to include "stingers" with a concomitant self-declared sensitive skin as panelists for safety testing (9). Owing to the great similarity of symptoms induced by topically applied capsaicin to those associated with sensitive skin (13), a new elicitation test using a 0.075% emulsion of a pungent compound extracted from chili peppers was proposed in the 1990s to

DETECTION THRESHOLDS OF CAPSAICIN 155 identify people with sensitive skin (15 ). Cutaneous application of capsaicin leads to a short release of neurotransmitters (substance P) from peripheral nerve endings and causes the appearance of uncomfortable sensations such as itching, burning, or stinging asso- ciated or not with redness at the application site (13, 16,17). These unpleasant reactions are more frequent and more intense in self-declared sensitive-skin subjects (15 ). Both the lactic acid and capsaicin tests presented above are based on the reported sensations by the subject in terms of nature and intensity and have thus raised contro- versies owing to the use of a subjective individual pain scale. However, they have been proved to be linked with self-declared sensitive skin because reported sensations were globally stronger and more frequent in sensitive-skin subjects. This link with sensitive skin is stronger with capsaicin than with lactic acid (18). The application of a stimu- lating agent at a single dose (e.g., 10% lactic acid solution or 0.075% capsaicin emul- sion) was shown to induce painful sensations, particularly in subjects with very sensitive skin or in a Chinese population as recently reported for lactic acid (19). In 1998, another psychophysical test based on the assessment of peripheral sensitivity to thermal stimuli was suggested as a possible diagnosis of sensitive skin (20). This test involved the use of a thermal testing instrument-for example, the thermal sensory analyzer (TSA 2001) manufactured by Medoc (Ramat Yishai, Israel)-to assess the thermal functional components of cutaneous nerve endings. The device, called a ther- mode, delivered thermal stimuli capable of heating or cooling the skin. However, two recent studies showed that this promising test was of limited value in the diagnosis of sensitive skin. In a first study, the link between the detection threshold to thermal stimuli assessed by TSA and self-perceived sensitive skin was not as strong as that observed with skin reactivity to capsaicin and, to a lesser extent, with the lactic acid stinging test (18). In another study, a significant difference in the mean of cold-pain thresholds was reported between sensitive and non-sensitive self-assessed-skin subjects (21 ). This difference was too weak to consider this thermal parameter as a predictive indicator of sensitive skin. The interest in using capsaicin in the assessment of sensory skin irritation has recently been confirmed by other authors (22). In addition, the formulation of capsaicin in hydroalcoholic solution has the advantage of accelerating the action of capsaicin on the face in comparison with the previously used 0.075% capsaicin emulsion (13,18). In the present study, we used the hydroalcoholic solution of capsaicin to develop a test that combines the specific reactivity of sensitive skin to capsaicin, the simplicity of the lactic acid stinging test, and a method of detection threshold. The major goal of this large random population study was to determine if the new method had subsequent advantages in terms of painlessness and accuracy, i.e., if it would be able to rank the subjects on a large range of capsaicin concentration and provide a robust link with self-assessed sensitive skin without being associated with painful sensation. MATERIALS AND METHODS SUBJECTS A total of 150 healthy women participated in this monocenter study carried out in November 2002. Subjects ranged from 18 to 61 years of age (mean = 35). Most of them