DETECTION THRESHOLDS OF CAPSAICIN 163 sensations (e.g., burning and stinging sensations). All the sensations were considered as slightly (intensity 2 = 93 .6%) or moderately (intensity 3 = 4.6%) perceptible. They were never reported as strongly perceptible (intensity 4) or painful (intensity 5 ). In the course of the experiment, the time of onset of sensation at the detection threshold was recorded in 7 4 of the 109 subjects in order to check that the three-minute interval between applications was long enough. The mean time was 80 seconds +/-17 s. The shortest time was 60 s and the longest 120 s. On the other hand, a total of 61 subjects reported some sensations before the detection threshold. These sensations appeared on the vehicle side or on the capsaicin side, but they lasted less than 30 seconds in the latter case. These sensations were qualitatively similar to those at the detection threshold described above and concerned the vehicle side as often as the capsaicin side. DISCUSSION This study describes a new test for the assessment of skin neurosensitivity on a large cohort. This test consists of applying capsaicin solutions with geometrically increasing concentrations to determine the individual capsaicin detection threshold. In comparison with the other psychophysical tests using chemical probes (18, 19), the test is painless. As expected from topically applied capsaicin (23 ), a large variety of sensations at the detection threshold were reported. However, all of these sensations were considered as only slightly or moderately perceptible, even in volunteers with the most sensitive skin. The absence of pain is highly advantageous and confirms the right choice of the starting concentration and of the concentration gradient. This psychophysical test implies a simultaneous, split-face, single-blind application of capsaicin and vehicle solutions. This bilateral procedure, with the simultaneous presen- tation of the vehicle on the controlateral side as control, helps volunteers to differentiate background cutaneous sensations from those specifically caused by capsaicin (13 ), and thus avoids false-positive results (24). The single-blinded approach also helps to avoid false-positive results. The test based on the existence of a sensation or not on the capsaicin side is easy to explain to the volunteers, and thus far more reliable than those commonly used involving the highly variable individual pain scale. We observed a large distribution of this random adult female population according to capsaicin detection thresholds (Figure 2a). The lowest capsaicin concentration was de- tected by 23.3% of the panelists, whereas almost the same proportion (27.3%) did not perceive even the highest concentration, which is one hundred times more concentrated. This range of at least two log units, which looked sizeable, is in agreement with previous results obtained on the arm (23). A more striking feature about the distribution of the population according to the capsaicin detection threshold was its non-unimodal shape. It suggests the existence of (at least) two population groups. Individuals with an inter- mediate detection threshold account for a low proportion compared to those with either a low or a high detection threshold. Whatever this splitting may mean in terms of "sensitive skin," it appears that, in the tested population, two clearly separated pheno- types are seen with regard to the reactivity to capsaicin through neuron-mediated response. Considering the pattern of Figures 2a and 26, a segmentation of the population between two sub-populations could be proposed (see vertical line in Figure 2): women

164 JOURNAL OF COSMETIC SCIENCE having a detection threshold above or equal to 1 x 10- 3 % (44% of the population) could be identified as having a low skin neurosensitivity, and women having a detection threshold below 1 x 10- 3 % (56% of the population) could be identified as having a high skin neurosensitivity. As illustrated in Table II, capsaicin detection thresholds are more strongly linked to self-declared skin reactivity to environment and to emotional stress than to cosmetics. Thus, the capsaicin detection threshold test tends to explore a larger aspect of self- declared sensitive skin than the lactic acid stinging test. To address this question in a more detailed way, further development of the capsaicin detection threshold test shall involve a comparison with the lactic acid stinging test in the same population. This test should be of particular interest for epidemiological studies designed to assess possible sexual and ethnic variations in skin neurosensitivity. A gender difference ob- served with this test could be an explanation of the well-documented discrepancy between women and men regarding self-assessed sensitive skin (2,4). It would also enable us to address the question of a possible higher skin neurosensitivity in women, as recently suggested by a study indicating that females show a trend towards being more sensitive to the lactic acid stinging test than males (2 5 ). It would also be inter- esting to explore the reality of the widespread idea that Asian (8) and especially Japanese women (26) would be more prone to subjective symptoms. Significant subjective sensory differences between Japanese and German women have recently been documented using the lactic acid stinging test. Japanese women complained about stronger stinging sen- sations. In the absence of any measurable differences concerning barrier function and skin reactivity to sodium lauryl sulphate, the authors theorized that this result may reflect a different cultural behavior rather than measurable characteristics in skin physiology (27). Anyway, this possible heightened neurosensitivity in Asian populations does not seem to rely on structural or biochemical particularities of the epidermal fiber network (28). On the contrary, a decrease in epidermal nerve density with age (29) can be evoked to explain the decreased neurosensitivity with age observed in our experiment. CONCLUSION This new test of skin neurosens1t1v1ty, which is simple, inexpensive, and painless, appears to be a promising prototype for the diagnosis of sensitive skin. Its sensitivity could also provide a basis for the assessment of modulators of skin neurosensitivity. The large distribution of the test population according to capsaicin detection thresholds illustrates the very important interindividual differences concerning the activity of cutaneous nerve endings, which are certainly one of the major physiological features explaining the phenomenon of sensitive skin. ACKNOWLEDGMENTS The authors thank Mrs Carole Guiramand and Mrs Veronique Chevalier for providing the test solutions and for their involvement in the preliminary studies concerning the formulation of capsaicin. We also thank Mrs Daniela Giacchetti for her constant support, Mr Jean-Louis Gueret for providing the customized packaging, and Mrs Odette Jam-