- Email: [email protected]
Comparative study of normal and rough human skin hydration in vivo: Evaluation with four different instruments
Journal of Dermatological
Science, 2 (1991) 119-124
Elsevier
119
DESC 00072
Comparative study of normal and rough human skin hydration in vivo: Evaluation with four different instruments D. Van Neste Skin Study Center, Skinteface
SPRL,
Toumai, Belgium
(Received 31 March 1990; accepted 6 December
Key words: Electrical properties
of the skin; Cosmetics;
Skin hydration;
1990)
Skin capacitance;
Non-invasive
methods
Abstract Appropriate monitoring of skin hydration during clinical and/or experimental trials needs devices with acceptable reproducibility and sensitivity under conditions ranging from increased, and normal to low hydration. The aim of this study was to compare the variation of electrometric data generated by 4 different instruments (Skicon Hygrometer, 2 CM420 and a CM820 corneometer) in normal and experimentally damaged skin displaying surface roughness. Rough skin sites were observed during the healing process after repeated tape stripping of stratum corneum in humans (e.g. lo-14 days after insult). They displayed lower conductance and/or capacitance levels as compared to normal skin sites of the same subjects. The Skicon hygrometer showed higher variability as compared to the comeometers and was less sensitive, in relative terms, in the rough skin sites. This device also showed a moderate zero drift and re-zeroing was repeatedly utilized during the experiment. When the comeometer data were plotted against the hygrometer data, the slope of the regression line generated by the CM420a was different from CM420b and from CM820; the two latter were not significantly different from each other. Hence, comparison of absolute data obtained under comparable conditions (in this case CM420a and CM420b) in a single laboratory should not be made without prior calibration. Standards for evaluating interinstrumental variation are currently unavailable. This aspect of the measurement of electrical properties of the skin has not been investigated in great detail and has often been neglected in the past. Our findings also indicate that a constant control over the performances of a particular device should further improve the reliability of the data.
Introduction Clinical evaluation of the severity of skin lesions is subject to two main criticisms: 1) variation between observers and 2) non parametric description or scores for evaluating disease severity. The former can be standardized when observers’ scores are compared at regular Correspondence
du Sondart,
to: D. Van Neste; Skinterface B 7500 Tournai, ,Belgium.
0920-5497/91/$03.50
SPRL, 9 rue
0 1991 Elsevier Science Publishers
B.V.
intervals during a trial. It is clear that scores may change with the experience of the observers. Even in the absence of this observer related variation, and under appropriate experimental conditions (surfactant induced skin irritation on flexor forearm and back in 10 volunteers, evaluation of scores in comparison with an objective color ruler), we were able to demonstrate that severity scores of erythema showed significant regional variation (more sever scores in the back) while instrumental evaluation, i.e. quantification of
120
blood flow with laser Doppler flowmetry (LDF) did not. We knew from previous studies that LDF showed excellent correlation with erythema when both sites were taken separately [ 1, personal unpublished data]. In our experiment, reading of clinical scores of erythema appeared to be influenced by interfering signals from the epidermis which were in fact responsible for apparently more severe redness in the back as compared with flexor forearm. Indeed, skin roughness scores were more severe in the back as compared with the flexor forearm. The significant regional differences of skin feeling were corroborated by regional variations of skin capacitance: a more drastic reduction of skin capacitance was recorded in the back. This indicates that in complex skin lesions optical properties of the skin may be modified by the epidermal compartment making a clinical scoring of a dermal lesion more or less unreliable as suggested by other studies in allergic contact dermatitis and atopic dermatitis [2]. As yet, there is no means to demonstrate that a skin lesion with a severity score of 4 is twice or four times as severely damaged than a severity score 2 or 1, respectively. The development of commercially available measurement devices allows for quantitative evaluation of skin function in a noninvasive way. Contrary to the clinical scoring systems, continuous data are generated. Probes and measurement methods vary according to the manufacturer but it is not clear, in the absence of appropriate standards, that there is no change of readings over time with any single instrument. Few studies have been conducted to compare in vivo various devices that were built by the same or by different companies. In the case of skin surface hydration for example, a wide range of instrumental methods have been made available to the dermatological scientist [review in 31. The most commonly used methods for the evaluation of moisturization are measuring electric impedance of the skin [reviewed by 41 and there is indirect evidence that under physiological conditions electrical signals might correlate with skin surface hydration [ 51 and some authors express
their readings in terms of ‘water reading % ’ [ 61. Whether this is still applicable in dermatological conditions with surface roughness has not been clearly demonstrated. Hence, appropriate monitoring of skin hydration during clinical or experimental trials needs devices with acceptable reproducibility and sensitivity in both the lower and higher impedance load offered by the skin under conditions of normal and low hydration respectively. The aim of this study was to compare the variation of electrometric data generated by 4 different instruments in normal and experimentally damaged skin, namely roughness during the healing phase after stripping of the stratum corneum. Material and Methods Subjects
Five females (age average 40.2 years) volunteered for this experiment and gave written consent after being informed about the experimental procedure. In each individual 4 normal sites (N) and 4 stripped skin sites (S) were examined. S showed ‘rough dermatitic skin’ which appeared between 10 to 14 days after a well defined insult (25 adhesive tape strippings). Electrometric devices 4 different instruments
were utilized in the following order: Skicon (IBS, Skin Hygrometer, Hamamtsu, Japan), and 3 corneometers (Courage and Khazaka, Koln, FRG): 2 from the first (CM420a and CM420b) and 1 from the second generation (CM820). For each instrument, the measurements were made according to manufactures’ instructions: probes were applied vertically onto the skin surface; the skin was left unoccluded immediately after the data were displayed on the machine face (see also experimental protocol). The other preset instrumental characteristics could not be modified by the investigator: duration of contact between the probe and the skin was preset at 3 s after which data were displayed on the screen; pressure on the skin was controlled with a spring
121
mechanism built into the probe-holders (the spring of the IBS instrument was set a 0 graduation; pressure of 3 N with the CM420 or 2 N with the CM820 were necessary to switch on the current in the probe). The external diameters of the probeholders in contact with the skin surface were as follows: 25 mm, 28 mm and 15 mm for the Skicon, DM420 and CM820, respectively. The external diameters of the mobile part of the central probe inserted into the probe-holder were 6, 20 and 11 mm, respectively. The circular probes of the Skicon and the CM420 had an external diameter of 6 and 15 mm. The edges of the CM820 square probe were 7 mm long. The distance between the central (1 mm 0) and peripheral (internal 0: 4 mm) electrodes of the Skicon was 1 mm. The distance between the interdigitating electrodes of the CM probes was 100 pm and 70 ,um for the CM420 and CM820, respectively. The operating frequency of the electric current was 3.5 MHz with the Skicon and the central processing unit was equipped with a zeroing button; the frequency in the corneometers was in the range of 40-100 kHz. There was direct contact between the skin surface and the electrodes in the Skicon; this was not the case neither with the CM420 where the circuits were separated from the skin by a polythene foil, nor with the CM820 which was equipped with a glass covered measuring probe. Experimental protocol All measurements were made in a climatic room where temperature and relative humidity were stable throughout the experiment (average and standard deviation: 20.60 rtr 0.81 “C and 51.40 + 2.45%RH). The back of the panelists had not been washed for at least 12 h and no topicals had been applied during the last 6 days. Adhesive tape stripping (2 cm wide and 5 cm long) had been performed between 10 and 14 days before the first measurements, and the S sites felt rough and moderately dermatitic (occasional faint erythema). The back was explored on resting subjects, after a 10 min adaptation period to the ambient
conditions. 8 sites were investigated in sequence: upper back left (No. 1S and N), upper back right (No. 2s and N), lower back left (No. 3s and N), lower back right (No. 4s and N). From each area (from No. 1s to No. 4N) 5 measures were recorded for descriptive statistics and averaged for comparative studies. Immediately after the first data were displayed (i.e. 3 s after application), the probe was removed from the skin. The second measurement started 5 s later. The probe was applied to the skin immediately adjacent to the previously measured skin area in order to prevent any overlap with the previous measurement site and changes of surface hydration due to repeated periods of occlusion. When the 4th and 5th measurements were made, the skin area had been air exposed for at least 13 s in the case of the largest probe-holders (Skicon and CM420); in the case of CM820 no such overlap occurred. This sequence was adopted because prolonged contact results in accumulation of unbound water in the stratum corneum especially in skin lesions with high water permeability levels [ 31. Unfortunately, the CM420a went out of order midway during the trial which was nevertheless continued with all other instruments. Even though fragmentary, the data were included for the specific purpose of comparison between 2 CM420. When the electrometric measures were completed, the skin sites were also evaluated with the EPl evaporimeter (Servomed, Sweden) a PF3 laser Doppler flowmeter (Perimed, Sweden) and the tristimulus Chroma MeterTM (Minolta, Osaka, Japan). The details of these measures will not be included in this study but have been published in part elsewhere [2]. The evaporimeter showed that there was no impairment of the barrier function of the skin; indeed equal levels of transepidermal water loss were found on S and N sites. The laser Doppler flowmeter detected significantly higher levels of microvascular activity in S as compared to N sites ; the average relative variation did however not exceed 15%. The L*a*b* parameters obtained with the tristimulus chroma meter were not different at S as compared with N sites. These measures were included to ascertain that changes
122
reported with electrometric devices were not related to other changes than those present in the most superficial layers of the stratum corneum.
G
IAU
140. f
120. 100.
Statistical analysis
The individual data were processed using Excel1(Microsoft@ software). The descriptive statistics of each set of 5 measures comprised the average (av), standard deviation (SD) and coefficient of variation (CV% = [sd/av]* 100). The database of our five subjects comprises (5 subjects* 5 measures* 8 sites* 4 devices = 800 data boxes). Descriptive statistics of averaged measures (40 per instrument) of stripped (S) and normal (N) skin sites are shown as averages with SD. 3-factor analysis of variance for repeated measures (ANOVA) was used to evaluate variation according to site (S and N), location on the back (site 1-2: upper back or site 3-4: lower back) and subjects. Linear regression analysis between data generated by the various devices was used setting the 3 CM data (arbitrary units, AU) against the conductance (G or pmho obtained with the Skicon). Significance of linear correlation is analyzed with ANOVA (P < 0.05). Slopes were compared with Student’s ‘t’-test and if P < 0.05 differences were considered significant (two tail test).
/
I
60. 60.
i
I
_I
I
SNSNSNSN CM420+a
Skicon
CM420#b
CM820
Fig. 1. Electrometric data on stripped and normal skin sites. Average ( _+SD) of Skicon data are shown in G (pmho) while those of CM data are expressed in arbitrary units (AU). Comparison of averaged measures on stripped (S) and normal (N) adjacent skin sites reveals that all devices detected lower levels of skin surface hydration at stripped skin sites (P < 0.0001).
on stripped and normal adjacent skin sites reveals that all devices detected lower levels of skin surface hydration at stripped skin sites (P < 0.0001; Fig. 1). Comparison between 2 CM420
devices
For technical reasons, it was impossible to collect all data with the CM420a. Comparison of data obtained with CM420a and b, showed significant linearity (P < 0.0001) but the X-Y intercept was clearly different from 0 (Fig. 2). In practice, exchange of absolute data generated by dif-
Results
ANOVA for repeated measures showed that there was no significant variation according to individuals or site in the back but that the stripping was a significant factor influencing the readings of all skin electrometric devices (P < 0.000 1). Descriptive statistics of the original data sets (5 measures per S or N skin zone) shows remarkable differences of electrometric devices: Skicon data showed highest coefficients of variation especially in rough stripped skin areas (ranging from 21 to 101%) as compared to normal skin sites (range: 5.9-44.5%). The CM devices generated more homogeneous sets of data (CV% below 22.5% and 17.5% in S and N, respectively). Comparison of averaged measures
t 120
0.
0
100
a*
60' r=.763,p<0.0001 /@-I
0
20
40
60
60
e
.:
loo
120
140
NJ
Fig. 2. Comparison between 2 CM420 corneometers. Comparison of data obtained with CM420a and b, showed signiflcant linearity (y = 24.59 + 0.58x; r = 0.783; P < 0.0001) and the X-Y intercept was different from 0. This illustrates that the 2 instruments do not generate equivalent absolute data. In terms ofrelative changes further description of calibration is given in Fig. 3.
123
ferent CM420 devices would not appear valid but relative variation between data generated by the 2 devices should be acceptable. Comparison of Skicon with CM420 (a and b) and CM820 The linearity between averaged Skicon and each set of CM data is significant (P < 0.0001; Fig. 3). The X-Y intercepts with CM420a, and b (96.62 and 77.37 respectively) were higher than the one obtained with CM820 (51.71). This indicates that there is much more room for measurements in the high impedance zone offered by rough skin with the CM devices than with the Skicon hygrometer. The slope of the regression line was different from 0 in all cases (P < 0.0001). The slope was significantly different for the 2CM420(a: 1.19* 0.26andb:0.80 f O.ll;avs b: P < 0.05) and the CM820 (0.72 & 0.8; vs a: P < 0.05; vs b: not significant). Hence calibration of corneometers with the skin hygrometer appears feasible on the basis of in vivo generated data. ---The old CM420a was significantly different from the recently repaired CM420b and from the CM820; the 2 latter showed however similar slopes when calibrated against the Skicon.
AU
o CM420ta 0 CM420#b A CM820
01 0
IO
20
30
40
50
60 70 pmho
Fig. 3. Comparison of 3 corneometers with the Skicon, skin hygrometer. The arbitrary units (AU) displayed by the various CM devices are computed against the conductance recorded with the Skicon (G = pmho) on all skin test sites (S and N). The slope of the regression line is different from 0 in all cases (P < 0.0001). Even though linearity appears satisfactory, the X-Y intercepts with CM420a, b and CM820 is clearly different. The slope of the CM420a data differs significantly from that of the CM420b (P c 0.05); the latter is not different from the CM820.
Discussion The present study, which was performed during winter time and under strict control of the ambient conditions, reveals that the IBS hygrometer and all 3 conreometers gave coherent information in terms of relative changes occurring in ‘rough’ as compared to normal human skin. In a previous study on the electrical properties of the skin surface with the sodium lauryl sulphate induced rough dermatitic skin model, the CM420 data appeared indeed to be correlated with the roughness of the skin surface [8]. Several papers detailed technical assessment of these instruments individually [9-l I] or in combination [ 6,121 and some comparative assays with other methods such as infrared spectrography have also been published [ 131. All published experiments seem to indicate that electrometric data are influenced by the actual water content of the stratum corneum which increases the electrolyte mobility [4]. During our experiments, a ‘zero drift’ was observed with the skin hygrometer as reported earlier [ 111, re-zeroing during the experiment makes a direct comparison between data possible. The Skicon has the advantage of controlled quantification of the signal expressed in G (or pmho). The major disadvantage of this device is lack of resolution in the low conductance range which is unfortunately, that region where the experimental dermatologist needs precise data to control effect of moisturizers. The lower limit of reliable detection was estimated at 10 pmho [ 111. We also noted a high variability of individual data as well on normal as rough skin sites (coefficient of variation range up to 100 y0 in rough and 45 % in normal). Increasing the number of measures and taking the average (3 measurements in the papers 11 and 12; 5 in this case) can resolve this problem in terms of reproducibility during the clinical investigation. Such high variability had also been noted by other observers especially on so-called ‘dry skin’ [ 121. Absolute comparison of the hygrometer data and ours reveals slight differences : the average + SD in the back of the Danish patch test clinic was 22 f 12 pumho while ours
124
was 30 f 13 pumho. Similarly, with their CM420, the Danish data were slightly different: 97 + 9 AU vs our CM420a: 110 f 13.8 AU and CM420b: 83.9 f 10.5 AU indicating a relative variation of + 15y0 between various devices built by the same manufacturer. Coefficients of variation observed with the corneometers were below 25%. Even though the surface of the CM820 is substantially smaller than the probe of the CM420, it is still larger than the probe of the Skicon and less subject to variation. The high variability of the Skicon hygrometer data has been ascribed to the smaller size of the probe [ 121. We would like to propose that the distance between the electrodes rather than their absolute size is mandatory. Indeed, inserting a 2 mm instead of the standard 1 mm central electrode, reducing by the way the distance between the central and peripheral probe by 50%) the hygrometer shows a three times greater sensitivity [3] but even then, there is still a lot of data outside the optimum region of operation. The 1 mm distance occupied by the dielectric is still at least 10 times the ldistance between the interdigitating electrodes built into the CM probes. As a conclusion of our study, we obtained satisfactory linear correlations between the data generated by the various devices. In terms of extrapolation, it is possible to compare relative data obtained in different laboratories using the CM420, CM820 and Skicon, inasmuch as the devices are not running out of order (see data with CM420a). Some improvements are still needed to increase the sensitivity in the dielectric rough skin zones, and to standardize, control and ascertain the units of the ‘numbers’ appearing on the screens. Acknowledgements The technical
assistance of Sylvie Lyonnet, Bernadette de Brouwer and Marianne Dumortier was highly appreciated during collection and preparation of the database.
References 1 Van Neste D, Mahmoud G, Masmoudi M: Experimental induction of rough dermatitic skin in humans. Contact Derm 16: 27-33, 1987. 2 Van Neste D, Lyonnet S, de Brouwer B, Thivolet J: Noninvasive in vivo evaluation ofbiophysical parameters of atopic dermatitis. Comparison of inflamed lesions and clinically normal skin with changes recorded on allergic patch test reactions and after tape stripping on control subjects, in ‘Pharmacology of the Skin. vol. 4: Immunological aspects of atopic and contact eczema’. Edited by J Czemielewski. Karger, Basel, (in press). 3 Tagami H: Impedance measurement for evaluation of the hydration state of the skin. Cutaneous investigation in health and disease. Noninvasive methods and instrumentation. J.L. Leveque, Marcel Dekker, New York, 1989, pp 79-111. 4 Leveque JL, De Rigal J: Impedance methods for studying skin moisturization. J Sot Cosmet Chem 34: 419-428, 1983. 5 Murahata RI, Hing SAO, Maibach HI, Roheim JR: The use of a microwave probe to evaluate the hydration of human stratum corneum in vivo. Bioeng Skin 2: 235-247, 1986. 6 Dikstein S, Katz M, Zlotogorski A, Broun Y, Wilson D, Maibach H: Comparison of different instruments for measuring stratum corneum moisture content. Int J Cosmetic Sci 8: 289-292, 1986. I Van Neste D: in vivo evaluation of unbound water accumulation in stratum corneum. The influence of acute skin irritation induced by sodium laurylsulfate. Dermatologica 181: 197-201, 1990. 8 Van Neste D, Antoine JL: A vehicle controlled study of the effects of hydrating agents in a human model of rough dermatitic skin. Bioeng Skin 4: 243-262, 1988. 9 Tagami H, Ohi M, Iwatsuki K, Kanamaru Y, Yamada M, Ichijo B: Evaluation ofthe skin surface hydration in vivo by electrical measurement. J Invest Dermatol 75: 500-507, 1980. 10 Werner Y: The water content of the stratum corneum in patients with atopic dermatitis. Measurement with the Corneometer CM 420. Acta Derm Venereol (Stockh) 66: 28 l-284, 1986. 11 Moseley H, English JSC, Coghill GM, Mackie RM: Assessment and use of a new skin hygrometer. Bioeng Skin 1: 177-192, 1985. 12 Agner T, Serup J: Comparison of two electrical methods for measurement of skin hydration. An experimental study on irritant patch test reactions. Bioeng Skin 4: 263-269, 1988. 13 Triebskorn A, Gloor M, Greiner F: Comparative investigations on the water content of the stratum comeum using different methods of measurement. Dermatologica 167: 64-69, 1983.
Science, 2 (1991) 119-124
Elsevier
119
DESC 00072
Comparative study of normal and rough human skin hydration in vivo: Evaluation with four different instruments D. Van Neste Skin Study Center, Skinteface
SPRL,
Toumai, Belgium
(Received 31 March 1990; accepted 6 December
Key words: Electrical properties
of the skin; Cosmetics;
Skin hydration;
1990)
Skin capacitance;
Non-invasive
methods
Abstract Appropriate monitoring of skin hydration during clinical and/or experimental trials needs devices with acceptable reproducibility and sensitivity under conditions ranging from increased, and normal to low hydration. The aim of this study was to compare the variation of electrometric data generated by 4 different instruments (Skicon Hygrometer, 2 CM420 and a CM820 corneometer) in normal and experimentally damaged skin displaying surface roughness. Rough skin sites were observed during the healing process after repeated tape stripping of stratum corneum in humans (e.g. lo-14 days after insult). They displayed lower conductance and/or capacitance levels as compared to normal skin sites of the same subjects. The Skicon hygrometer showed higher variability as compared to the comeometers and was less sensitive, in relative terms, in the rough skin sites. This device also showed a moderate zero drift and re-zeroing was repeatedly utilized during the experiment. When the comeometer data were plotted against the hygrometer data, the slope of the regression line generated by the CM420a was different from CM420b and from CM820; the two latter were not significantly different from each other. Hence, comparison of absolute data obtained under comparable conditions (in this case CM420a and CM420b) in a single laboratory should not be made without prior calibration. Standards for evaluating interinstrumental variation are currently unavailable. This aspect of the measurement of electrical properties of the skin has not been investigated in great detail and has often been neglected in the past. Our findings also indicate that a constant control over the performances of a particular device should further improve the reliability of the data.
Introduction Clinical evaluation of the severity of skin lesions is subject to two main criticisms: 1) variation between observers and 2) non parametric description or scores for evaluating disease severity. The former can be standardized when observers’ scores are compared at regular Correspondence
du Sondart,
to: D. Van Neste; Skinterface B 7500 Tournai, ,Belgium.
0920-5497/91/$03.50
SPRL, 9 rue
0 1991 Elsevier Science Publishers
B.V.
intervals during a trial. It is clear that scores may change with the experience of the observers. Even in the absence of this observer related variation, and under appropriate experimental conditions (surfactant induced skin irritation on flexor forearm and back in 10 volunteers, evaluation of scores in comparison with an objective color ruler), we were able to demonstrate that severity scores of erythema showed significant regional variation (more sever scores in the back) while instrumental evaluation, i.e. quantification of
120
blood flow with laser Doppler flowmetry (LDF) did not. We knew from previous studies that LDF showed excellent correlation with erythema when both sites were taken separately [ 1, personal unpublished data]. In our experiment, reading of clinical scores of erythema appeared to be influenced by interfering signals from the epidermis which were in fact responsible for apparently more severe redness in the back as compared with flexor forearm. Indeed, skin roughness scores were more severe in the back as compared with the flexor forearm. The significant regional differences of skin feeling were corroborated by regional variations of skin capacitance: a more drastic reduction of skin capacitance was recorded in the back. This indicates that in complex skin lesions optical properties of the skin may be modified by the epidermal compartment making a clinical scoring of a dermal lesion more or less unreliable as suggested by other studies in allergic contact dermatitis and atopic dermatitis [2]. As yet, there is no means to demonstrate that a skin lesion with a severity score of 4 is twice or four times as severely damaged than a severity score 2 or 1, respectively. The development of commercially available measurement devices allows for quantitative evaluation of skin function in a noninvasive way. Contrary to the clinical scoring systems, continuous data are generated. Probes and measurement methods vary according to the manufacturer but it is not clear, in the absence of appropriate standards, that there is no change of readings over time with any single instrument. Few studies have been conducted to compare in vivo various devices that were built by the same or by different companies. In the case of skin surface hydration for example, a wide range of instrumental methods have been made available to the dermatological scientist [review in 31. The most commonly used methods for the evaluation of moisturization are measuring electric impedance of the skin [reviewed by 41 and there is indirect evidence that under physiological conditions electrical signals might correlate with skin surface hydration [ 51 and some authors express
their readings in terms of ‘water reading % ’ [ 61. Whether this is still applicable in dermatological conditions with surface roughness has not been clearly demonstrated. Hence, appropriate monitoring of skin hydration during clinical or experimental trials needs devices with acceptable reproducibility and sensitivity in both the lower and higher impedance load offered by the skin under conditions of normal and low hydration respectively. The aim of this study was to compare the variation of electrometric data generated by 4 different instruments in normal and experimentally damaged skin, namely roughness during the healing phase after stripping of the stratum corneum. Material and Methods Subjects
Five females (age average 40.2 years) volunteered for this experiment and gave written consent after being informed about the experimental procedure. In each individual 4 normal sites (N) and 4 stripped skin sites (S) were examined. S showed ‘rough dermatitic skin’ which appeared between 10 to 14 days after a well defined insult (25 adhesive tape strippings). Electrometric devices 4 different instruments
were utilized in the following order: Skicon (IBS, Skin Hygrometer, Hamamtsu, Japan), and 3 corneometers (Courage and Khazaka, Koln, FRG): 2 from the first (CM420a and CM420b) and 1 from the second generation (CM820). For each instrument, the measurements were made according to manufactures’ instructions: probes were applied vertically onto the skin surface; the skin was left unoccluded immediately after the data were displayed on the machine face (see also experimental protocol). The other preset instrumental characteristics could not be modified by the investigator: duration of contact between the probe and the skin was preset at 3 s after which data were displayed on the screen; pressure on the skin was controlled with a spring
121
mechanism built into the probe-holders (the spring of the IBS instrument was set a 0 graduation; pressure of 3 N with the CM420 or 2 N with the CM820 were necessary to switch on the current in the probe). The external diameters of the probeholders in contact with the skin surface were as follows: 25 mm, 28 mm and 15 mm for the Skicon, DM420 and CM820, respectively. The external diameters of the mobile part of the central probe inserted into the probe-holder were 6, 20 and 11 mm, respectively. The circular probes of the Skicon and the CM420 had an external diameter of 6 and 15 mm. The edges of the CM820 square probe were 7 mm long. The distance between the central (1 mm 0) and peripheral (internal 0: 4 mm) electrodes of the Skicon was 1 mm. The distance between the interdigitating electrodes of the CM probes was 100 pm and 70 ,um for the CM420 and CM820, respectively. The operating frequency of the electric current was 3.5 MHz with the Skicon and the central processing unit was equipped with a zeroing button; the frequency in the corneometers was in the range of 40-100 kHz. There was direct contact between the skin surface and the electrodes in the Skicon; this was not the case neither with the CM420 where the circuits were separated from the skin by a polythene foil, nor with the CM820 which was equipped with a glass covered measuring probe. Experimental protocol All measurements were made in a climatic room where temperature and relative humidity were stable throughout the experiment (average and standard deviation: 20.60 rtr 0.81 “C and 51.40 + 2.45%RH). The back of the panelists had not been washed for at least 12 h and no topicals had been applied during the last 6 days. Adhesive tape stripping (2 cm wide and 5 cm long) had been performed between 10 and 14 days before the first measurements, and the S sites felt rough and moderately dermatitic (occasional faint erythema). The back was explored on resting subjects, after a 10 min adaptation period to the ambient
conditions. 8 sites were investigated in sequence: upper back left (No. 1S and N), upper back right (No. 2s and N), lower back left (No. 3s and N), lower back right (No. 4s and N). From each area (from No. 1s to No. 4N) 5 measures were recorded for descriptive statistics and averaged for comparative studies. Immediately after the first data were displayed (i.e. 3 s after application), the probe was removed from the skin. The second measurement started 5 s later. The probe was applied to the skin immediately adjacent to the previously measured skin area in order to prevent any overlap with the previous measurement site and changes of surface hydration due to repeated periods of occlusion. When the 4th and 5th measurements were made, the skin area had been air exposed for at least 13 s in the case of the largest probe-holders (Skicon and CM420); in the case of CM820 no such overlap occurred. This sequence was adopted because prolonged contact results in accumulation of unbound water in the stratum corneum especially in skin lesions with high water permeability levels [ 31. Unfortunately, the CM420a went out of order midway during the trial which was nevertheless continued with all other instruments. Even though fragmentary, the data were included for the specific purpose of comparison between 2 CM420. When the electrometric measures were completed, the skin sites were also evaluated with the EPl evaporimeter (Servomed, Sweden) a PF3 laser Doppler flowmeter (Perimed, Sweden) and the tristimulus Chroma MeterTM (Minolta, Osaka, Japan). The details of these measures will not be included in this study but have been published in part elsewhere [2]. The evaporimeter showed that there was no impairment of the barrier function of the skin; indeed equal levels of transepidermal water loss were found on S and N sites. The laser Doppler flowmeter detected significantly higher levels of microvascular activity in S as compared to N sites ; the average relative variation did however not exceed 15%. The L*a*b* parameters obtained with the tristimulus chroma meter were not different at S as compared with N sites. These measures were included to ascertain that changes
122
reported with electrometric devices were not related to other changes than those present in the most superficial layers of the stratum corneum.
G
IAU
140. f
120. 100.
Statistical analysis
The individual data were processed using Excel1(Microsoft@ software). The descriptive statistics of each set of 5 measures comprised the average (av), standard deviation (SD) and coefficient of variation (CV% = [sd/av]* 100). The database of our five subjects comprises (5 subjects* 5 measures* 8 sites* 4 devices = 800 data boxes). Descriptive statistics of averaged measures (40 per instrument) of stripped (S) and normal (N) skin sites are shown as averages with SD. 3-factor analysis of variance for repeated measures (ANOVA) was used to evaluate variation according to site (S and N), location on the back (site 1-2: upper back or site 3-4: lower back) and subjects. Linear regression analysis between data generated by the various devices was used setting the 3 CM data (arbitrary units, AU) against the conductance (G or pmho obtained with the Skicon). Significance of linear correlation is analyzed with ANOVA (P < 0.05). Slopes were compared with Student’s ‘t’-test and if P < 0.05 differences were considered significant (two tail test).
/
I
60. 60.
i
I
_I
I
SNSNSNSN CM420+a
Skicon
CM420#b
CM820
Fig. 1. Electrometric data on stripped and normal skin sites. Average ( _+SD) of Skicon data are shown in G (pmho) while those of CM data are expressed in arbitrary units (AU). Comparison of averaged measures on stripped (S) and normal (N) adjacent skin sites reveals that all devices detected lower levels of skin surface hydration at stripped skin sites (P < 0.0001).
on stripped and normal adjacent skin sites reveals that all devices detected lower levels of skin surface hydration at stripped skin sites (P < 0.0001; Fig. 1). Comparison between 2 CM420
devices
For technical reasons, it was impossible to collect all data with the CM420a. Comparison of data obtained with CM420a and b, showed significant linearity (P < 0.0001) but the X-Y intercept was clearly different from 0 (Fig. 2). In practice, exchange of absolute data generated by dif-
Results
ANOVA for repeated measures showed that there was no significant variation according to individuals or site in the back but that the stripping was a significant factor influencing the readings of all skin electrometric devices (P < 0.000 1). Descriptive statistics of the original data sets (5 measures per S or N skin zone) shows remarkable differences of electrometric devices: Skicon data showed highest coefficients of variation especially in rough stripped skin areas (ranging from 21 to 101%) as compared to normal skin sites (range: 5.9-44.5%). The CM devices generated more homogeneous sets of data (CV% below 22.5% and 17.5% in S and N, respectively). Comparison of averaged measures
t 120
0.
0
100
a*
60' r=.763,p<0.0001 /@-I
0
20
40
60
60
e
.:
loo
120
140
NJ
Fig. 2. Comparison between 2 CM420 corneometers. Comparison of data obtained with CM420a and b, showed signiflcant linearity (y = 24.59 + 0.58x; r = 0.783; P < 0.0001) and the X-Y intercept was different from 0. This illustrates that the 2 instruments do not generate equivalent absolute data. In terms ofrelative changes further description of calibration is given in Fig. 3.
123
ferent CM420 devices would not appear valid but relative variation between data generated by the 2 devices should be acceptable. Comparison of Skicon with CM420 (a and b) and CM820 The linearity between averaged Skicon and each set of CM data is significant (P < 0.0001; Fig. 3). The X-Y intercepts with CM420a, and b (96.62 and 77.37 respectively) were higher than the one obtained with CM820 (51.71). This indicates that there is much more room for measurements in the high impedance zone offered by rough skin with the CM devices than with the Skicon hygrometer. The slope of the regression line was different from 0 in all cases (P < 0.0001). The slope was significantly different for the 2CM420(a: 1.19* 0.26andb:0.80 f O.ll;avs b: P < 0.05) and the CM820 (0.72 & 0.8; vs a: P < 0.05; vs b: not significant). Hence calibration of corneometers with the skin hygrometer appears feasible on the basis of in vivo generated data. ---The old CM420a was significantly different from the recently repaired CM420b and from the CM820; the 2 latter showed however similar slopes when calibrated against the Skicon.
AU
o CM420ta 0 CM420#b A CM820
01 0
IO
20
30
40
50
60 70 pmho
Fig. 3. Comparison of 3 corneometers with the Skicon, skin hygrometer. The arbitrary units (AU) displayed by the various CM devices are computed against the conductance recorded with the Skicon (G = pmho) on all skin test sites (S and N). The slope of the regression line is different from 0 in all cases (P < 0.0001). Even though linearity appears satisfactory, the X-Y intercepts with CM420a, b and CM820 is clearly different. The slope of the CM420a data differs significantly from that of the CM420b (P c 0.05); the latter is not different from the CM820.
Discussion The present study, which was performed during winter time and under strict control of the ambient conditions, reveals that the IBS hygrometer and all 3 conreometers gave coherent information in terms of relative changes occurring in ‘rough’ as compared to normal human skin. In a previous study on the electrical properties of the skin surface with the sodium lauryl sulphate induced rough dermatitic skin model, the CM420 data appeared indeed to be correlated with the roughness of the skin surface [8]. Several papers detailed technical assessment of these instruments individually [9-l I] or in combination [ 6,121 and some comparative assays with other methods such as infrared spectrography have also been published [ 131. All published experiments seem to indicate that electrometric data are influenced by the actual water content of the stratum corneum which increases the electrolyte mobility [4]. During our experiments, a ‘zero drift’ was observed with the skin hygrometer as reported earlier [ 111, re-zeroing during the experiment makes a direct comparison between data possible. The Skicon has the advantage of controlled quantification of the signal expressed in G (or pmho). The major disadvantage of this device is lack of resolution in the low conductance range which is unfortunately, that region where the experimental dermatologist needs precise data to control effect of moisturizers. The lower limit of reliable detection was estimated at 10 pmho [ 111. We also noted a high variability of individual data as well on normal as rough skin sites (coefficient of variation range up to 100 y0 in rough and 45 % in normal). Increasing the number of measures and taking the average (3 measurements in the papers 11 and 12; 5 in this case) can resolve this problem in terms of reproducibility during the clinical investigation. Such high variability had also been noted by other observers especially on so-called ‘dry skin’ [ 121. Absolute comparison of the hygrometer data and ours reveals slight differences : the average + SD in the back of the Danish patch test clinic was 22 f 12 pumho while ours
124
was 30 f 13 pumho. Similarly, with their CM420, the Danish data were slightly different: 97 + 9 AU vs our CM420a: 110 f 13.8 AU and CM420b: 83.9 f 10.5 AU indicating a relative variation of + 15y0 between various devices built by the same manufacturer. Coefficients of variation observed with the corneometers were below 25%. Even though the surface of the CM820 is substantially smaller than the probe of the CM420, it is still larger than the probe of the Skicon and less subject to variation. The high variability of the Skicon hygrometer data has been ascribed to the smaller size of the probe [ 121. We would like to propose that the distance between the electrodes rather than their absolute size is mandatory. Indeed, inserting a 2 mm instead of the standard 1 mm central electrode, reducing by the way the distance between the central and peripheral probe by 50%) the hygrometer shows a three times greater sensitivity [3] but even then, there is still a lot of data outside the optimum region of operation. The 1 mm distance occupied by the dielectric is still at least 10 times the ldistance between the interdigitating electrodes built into the CM probes. As a conclusion of our study, we obtained satisfactory linear correlations between the data generated by the various devices. In terms of extrapolation, it is possible to compare relative data obtained in different laboratories using the CM420, CM820 and Skicon, inasmuch as the devices are not running out of order (see data with CM420a). Some improvements are still needed to increase the sensitivity in the dielectric rough skin zones, and to standardize, control and ascertain the units of the ‘numbers’ appearing on the screens. Acknowledgements The technical
assistance of Sylvie Lyonnet, Bernadette de Brouwer and Marianne Dumortier was highly appreciated during collection and preparation of the database.
References 1 Van Neste D, Mahmoud G, Masmoudi M: Experimental induction of rough dermatitic skin in humans. Contact Derm 16: 27-33, 1987. 2 Van Neste D, Lyonnet S, de Brouwer B, Thivolet J: Noninvasive in vivo evaluation ofbiophysical parameters of atopic dermatitis. Comparison of inflamed lesions and clinically normal skin with changes recorded on allergic patch test reactions and after tape stripping on control subjects, in ‘Pharmacology of the Skin. vol. 4: Immunological aspects of atopic and contact eczema’. Edited by J Czemielewski. Karger, Basel, (in press). 3 Tagami H: Impedance measurement for evaluation of the hydration state of the skin. Cutaneous investigation in health and disease. Noninvasive methods and instrumentation. J.L. Leveque, Marcel Dekker, New York, 1989, pp 79-111. 4 Leveque JL, De Rigal J: Impedance methods for studying skin moisturization. J Sot Cosmet Chem 34: 419-428, 1983. 5 Murahata RI, Hing SAO, Maibach HI, Roheim JR: The use of a microwave probe to evaluate the hydration of human stratum corneum in vivo. Bioeng Skin 2: 235-247, 1986. 6 Dikstein S, Katz M, Zlotogorski A, Broun Y, Wilson D, Maibach H: Comparison of different instruments for measuring stratum corneum moisture content. Int J Cosmetic Sci 8: 289-292, 1986. I Van Neste D: in vivo evaluation of unbound water accumulation in stratum corneum. The influence of acute skin irritation induced by sodium laurylsulfate. Dermatologica 181: 197-201, 1990. 8 Van Neste D, Antoine JL: A vehicle controlled study of the effects of hydrating agents in a human model of rough dermatitic skin. Bioeng Skin 4: 243-262, 1988. 9 Tagami H, Ohi M, Iwatsuki K, Kanamaru Y, Yamada M, Ichijo B: Evaluation ofthe skin surface hydration in vivo by electrical measurement. J Invest Dermatol 75: 500-507, 1980. 10 Werner Y: The water content of the stratum corneum in patients with atopic dermatitis. Measurement with the Corneometer CM 420. Acta Derm Venereol (Stockh) 66: 28 l-284, 1986. 11 Moseley H, English JSC, Coghill GM, Mackie RM: Assessment and use of a new skin hygrometer. Bioeng Skin 1: 177-192, 1985. 12 Agner T, Serup J: Comparison of two electrical methods for measurement of skin hydration. An experimental study on irritant patch test reactions. Bioeng Skin 4: 263-269, 1988. 13 Triebskorn A, Gloor M, Greiner F: Comparative investigations on the water content of the stratum comeum using different methods of measurement. Dermatologica 167: 64-69, 1983.