Other studies also support a role for surfactant-protein interaction in the development of skin irritation. Imokawa et al. measured the specific rotation of bovine serum
albumin (BSA) in the presence of surfactant to assess surfactant-protein interaction (69). Changes in the specific rotation were the result of conformational changes in BSA due to interactions with surfactant. Studies conducted with a range of surfactants suggested that both ionic and hydrophobic interactions between the surfactant molecule and BSA determine the extent of denaturation. For example, the authors proposed a stepwise interaction between ionic surfactants and BSA that would ultimately lead to complete denaturation of the protein molecule. They reported an excellent correlation between surfactant-protein interaction, as determined by the BSA specific rotation method, and skin roughness measurements made with a circulation apparatus (69).
Imokawa also used a technique based on indigo carmine dye displacement to examine binding of surfactant to stratum corneum and reported that the skin roughening effects of surfactants are related to their ability to adsorb onto skin (11,51,70). Keratin denaturation was believed to follow surfactant adsorption, as in the BSA model, ultimately leading to skin roughness. Kawai and Imokawa later extended this model to explain the sensation of tightness (71). Their work showed that lipid removal from skin was related to tightness induction; however, delipidization of the skin with ether did not result in marked tightness, and surfactants’ ability to remove lipids did not always parallel their potential to induce tightness. There was, however, a strong correlation between surfactant adsorption and tightness, and removal of skin surface lipids enhanced tightness upon subsequent surfactant exposure. The authors proposed a model in which stratum corneum lipid extraction by surfactant is a necessary step to induce skin tightness, but is itself not sufficient to cause tightness.
Prottey et al. analyzed tape strip or cup scrub samples collected from the backs of hands following immersion in surfactant solutions or water for acid phosphatase activity (72). They found a decrease in enzyme activity following surfactant exposure, which was attributed to acid phosphatase denaturation and subsequent enzymatic inactivation resulting from surfactant interaction with the protein. The authors reported an inverse relationship between remaining acid phosphatase activity and hand dryness, and proposed this enzyme as a marker for monitoring interactions between surfactants and stratum corneum proteins. Ananthapadmanabhan et al. examined the binding behavior of several surfactants to isolated guinea pig or human stratum corneum and reported that the extent of surfactant binding to stratum corneum correlated well with the irritation potential predicted by in vitro and in vivo methods (73). Rhein et al. noted a time – dependent effect on stratum corneum swelling for SLS, which was attributed to the interaction of surfactant with keratin and disruption of the keratin’s secondary and tertiary structure (23). As noted earlier, swelling induced by this and other surfactants studied exhibited a maximum for Ci2-Ci4 chain lengths. The swelling response was for the most part reversible except following exposure to soap concentrations > 1% or prolonged (> 24 hours) soap exposure. In a later review Rhein proposed a model by which surfactants interact with stratum corneum proteins that explains the observed swelling behavior (74). This model incorporates ionic and hydrophobic binding interactions and accounts for the effect of pH on both stratum corneum proteins and on anionic and cationic surfactants.
Mukherjee et al. examined the interaction of pure anionic surfactants and cleansing bars based on anionic surfactants with isolated stratum corneum in vitro by measuring displacement of 1-anilinonaphthalene-8-sulfonic acid (ANS), a fluorescent probe known to bind to stratum corneum proteins (75). Their results showed agreement between surfactants’ ability to displace ANS from stratum corneum samples and their potential to irritate skin as predicted by in vitro and in vivo methods, suggesting that surfactants’ potential for binding to stratum corneum proteins determines their in-use skin
compatibility. Lopez et al. exposed porcine stratum corneum to solvent (chloroform – methanol) and nonionic surfactant (octyl glucoside) solutions (26). Solvent exposure removed stratum corneum lipids but did not affect stratum corneum adhesion. In contrast, surfactant exposure preserved epidermal lipids; however, the lipid domain structure was disrupted. The surfactant also damaged corneocyte envelopes and caused corneocyte dishesion, suggesting that surfactant-protein interaction plays a role in irritation development. Shukuwa et al. studied the impact of pure surfactants and 1% solutions prepared from full formula bars on corneocyte disaggregation and swelling, and on morphologic deterioration using stratum corneum disks isolated from forearm suction blisters (76). The test materials’ tendency to induce corneocyte disaggregation did not correspond well with induced swelling behavior, e. g., SLS caused significant corneocyte disaggregation but only slightly greater swelling than water. The ranking of the test soaps based on corneocyte swelling was consistent with irritation potential predicted by the soap chamber test (77), and the authors propose corneocyte swelling as an in vitro model for predicting cleansers’ skin effects. One caution with the extrapolations made in several of these studies, however, is that the results generated under controlled exposure protocols that are used to “validate” the in vitro test data are themselves not always predictive of consumer experience with personal care products (9,78,79).
Factors related to the personal cleanser use environment will also influence surfactant-skin interactions. For example, Berardesca et al. examined irritation resulting from 5% SLS applied to the forearm at temperatures of 4°C, 20°C, and 40°C (80). Measurements made after four days of once-daily treatment showed that barrier compromise and erythema production increased with temperature. Desquamation and reflectance (L – value) also exhibited temperature-dependent behavior. Clarys et al. demonstrated a temperature-dependent increase in the irritant response to two dish washing liquids over a much narrower temperature range: 37°C and 40°C (81). In both of these studies the increase in irritation with temperature was attributed to greater fluidity of the epidermal lipids and enhanced irritant penetration.
Water hardness is another variable that varies in different use situations. We showed that water hardness impacts the absolute and relative skin compatibility of commercial personal cleansing products; soap-containing bars being more affected than syndet-based cleansers (82). Fujiwara and coworkers conducted arm immersion experiments using solutions of sodium laurate to examine the relationship between water hardness (calcium ion) and calcium soap-deposition onto skin (83). They found that hardness in water increased soap deposition, driven especially by the presence of calcium in the rinse water. We extended this work using marketed cleansing bars tested under a consumer-relevant arm wash protocol (84,85). A syndet bar, a triethanolamine (TEA) soap bar, and a sodium soap bar were tested. Washing was divided into two phases: a wash phase and a rinse phase conducted with various combinations of deionized water and hardened water (11 grains/gallon calcium). The syndet and TEA soap bars produced significantly less dryness and erythema than the sodium soap bar in the presence of calcium, but the difference between the products was negligible in deionized water (PR0.48 for inter-product comparisons). Greater deposition of calcium soap onto skin occurred under the hard water conditions. As reported by Fujiwara, the rinse step was particularly important in determining the compatibility of these cleansing bars with the skin. Although the specific interaction between the calcium soap and skin was not examined in either of the above studies, both provide an example of the role surfactant-skin interaction (i. e., calcium soap deposition) plays in determining personal cleansers’ skin compatibility.
The pH is thermodynamically defined as the negative logarithm of the hydrogen ion activity in aqueous solution. The pH is often defined in more practical terms as the negative logarithm of hydrogen ion concentration. Strictly speaking the hydrogen ion activity and concentration are not identical but in dilute solution this is a reasonable assumption. Many publications refer to the pH of the skin, but since the skin is not an aqueous solution it clearly does not have a pH. When a wet pH electrode is placed onto the skin, water-soluble materials on the skin surface dissolve; the pH of this solution is what is actually measured. Also, personal cleansing products, and even the preparations made from them, are usually not dilute solutions. In what follows, “pH” is used to remain consistent with the original references, even though in many instances what is measured is more correctly called an apparent pH.
Soap dissolves in water to form free fatty acid and strong base, e. g., sodium soap will react with water to produce small quantities of free fatty acid and sodium hydroxide. As a result soap-based cleansing bars usually produce lather with a higher pH than do products based on synthetic detergents. The inherent tendency for soap-based cleansers to produce lather/solutions with pH values in the range of about 9-10, coupled with their generally poor skin compatibility, frequently forms the basis for a hypothesized cause – and-effect relationship between a cleanser’s pH and its potential to irritate the skin.
At a fundamental level, Ananthapadmanabhan et al. reported a pH dependence for sodium lauroyl isethionate adsorption to skin, showing a minimum from pH 7 to pH 9, suggesting that pH might play a role in determining surfactant-skin interactions (73). However, van Scott and Lyon examined the potential for tap water with its pH adjusted from 4.5-10.5 or 1% solutions of various soap and detergent products to denature keratin (86). Water had no effect on the denaturation of defatted keratin or keratin plus 1% sebum over the pH range studied. Similarly, there was no significant relationship between product pH values, which ranged from pH 6.7 to pH 10.1, and denaturation of any of the keratins studied. Robbins and Fernee reported no significant in vitro swelling change when stratum corneum was exposed to water with pH values adjusted to between 3 and 9 (49). They also examined the effect of pH on stratum corneum swelling response to three different surfactants: SLS, linear alkylbenzene sulfonate (LAS), or dodecyl trimethyl ammonium bromide (DTAB). SLS and LAS are anionic; DTAB is cationic. Decreasing the pH value from 9 to 3 reduced the swelling responses for SLS and LAS. However, the swelling response was unchanged or increased when the pH was lowered from pH 9 to 6, a range that is relevant to many personal cleansers. The swelling response for DTAB increased when the pH was lowered from pH 9 to pH 3. Dugard and Schueplein observed that buffer in the pH range 3.0-9.5 produced no increase in stratum corneum permeability in the absence of surfactant (48). These authors found no change in the rate of permeability increase as a function of pH for the three surfactants studied: sodium dodecanoate (pH range 7.5-9.5), sodium dodecyl sulphate (pH range 5.0-9.0), and sodium dodecylamine hydrochloride (pH range 3.0-7.5). Bettley and Donoghue also performed water permeability experiments using isolated human stratum corneum (87). Their work showed that water, pH 10 buffer, and “Teepol” (2° alkyl sulphate detergent) at its natural pH or buffered to pH 10 had a minimal effect on membrane permeability. However, membrane permeability was markedly increased by exposure to 1% or 5% solutions of sodium palmitate. Membrane permeability gradually recovered upon removal of the soap, which as mentioned earlier argues against epidermal lipid extraction as a mechanism of irritation. The authors instead suggested that irritancy is related to a surfactant’s ability to penetrate the stratum corneum.
In vivo studies show a similar trend. Bettley and Donoghue also conducted patch testing with toilet soap and TEA soap (88). The TEA soap was less irritating than the toilet soap even though the solutions prepared from each product had a comparable pH value. This may reflect a counterion effect; Rhein et al. also reported reduction in swelling response, i. e., a reduced potential for skin irritation, from TEA salts of surfactants (23). Frosch reported the relative skin irritation potential of 23 cleansing bars marketed in the United States and Germany determined using a soap chamber test (9). These products represented a range of surfactant compositions and covered a pH range from 5.4 to 10.7. The published results do not support a cause-and-effect relationship between a cleanser’s potential to irritate skin and its pH value. Van der Valk et al. conducted a similar experiment and assessed the skin compatibility of 13 marketed personal cleansers (89). Irritation from 2% aqueous solutions of the products applied to subjects’ volar forearms on stratum corneum barrier function was assessed by evaporimetry. All of the cleansers significantly increased transepidermal water loss (TEWL) compared to control, but the results showed no relationship between cleanser pH and irritation TEWL. In a similar study, van der Valk conducted patch testing with pure surfactant solutions on unaffected forearm skin of healthy subjects and subjects with irritant contact dermatitis or atopic dermatitis (90). The results again did not support a relationship between surfactant pH and irritation. Van Ketel et al. examined the irritation potential of several liquid hand cleansers spanning a pH range from 3.5 to 10.0 by applying 8% aqueous solutions of each product under patch (91). These authors concluded that the pH value of a cleanser is not a useful parameter for predicting its irritancy. Murahata et al. used a modified soap chamber test to study the skin irritation from a series of buffer solutions covering a pH range from 4.0 to 10.5, 8% (w/w) detergent solutions prepared from marketed syndet and soap bars, and 8% solutions prepared from altered soap base in which low molecular weight free fatty acids were added to the bars during processing (92). The buffers altered the skin surface pH but did not produce irritation. Likewise, the cleanser preparations changed the skin surface pH, but there was no correlation between pH and irritancy. One seeming exception is a patch test study by Baranda et al. conducted with 27 cleansing bars (tested as 8% emulsions), two undiluted liquid cleansers, and water (93). These authors reported a significant correlation between irritation and cleanser pH. However, the coefficient of determination calculated from the reported results is r2 = 0.244. Thus, only about 25% of the variability in irritation that was observed in the study is explained by differences in cleanser pH.
Taken together, these in vitro and in vivo results suggest that the skin irritation potential of a personal cleansing product is primarily driven by differences in the chemical and physical properties of its component surfactants rather than by the pH value. However, personal cleansing products could affect skin condition in other ways. For example, Ananthapadmanabhan et al. conducted experiments to study the effect of pH on the physical properties of the stratum corneum (94). A series of in vitro experiments was conducted using Yucatan piglet skin as a model substrate. Sections of isolated stratum corneum were placed into the wells of microtiter plates and buffer or buffered surfactant solutions were added. Samples intended for swelling analysis (optical coherence tomography) were soaked for five or 21 hours at 37°C. Samples intended for lipid fluidity analysis were soaked for about 16 hours at room temperature. These experiments showed an increase in stratum corneum swelling at pH 10 compared to the other pH values; this effect was increased by the addition of surfactant. Lipid fluidity decreased at pH 10 relative to the other pH values; surfactant again increased this effect. The authors conclude that there is a direct effect of pH on stratum corneum protein swelling and lipid rigidity; both are greater at pH 10 than at pH 6.5 or pH 4.
Sznitowska et al. studied the effect of pH on the lipoidal route of stratum corneum penetration (95). Suspensions of two model compounds, hydrocortisone and testosterone, were prepared at pH values ranging from 2.0 to 10.0 (hydrocortisone) and from 1.0 to 12.0 (testosterone). Penetration was studied using full thickness cadaver skin mounted in flowthrough diffusion cells. The studies were conducted with untreated skin and with skin pretreated with methanol-chloroform (11) or Azone, a material that alters stratum corneum lipid organization. The results from the experiments conducted with intact skin showed no significant effect of pH on the penetration of the model compounds in the range from 1.0 to 11.0. Removal of skin lipids with methanol-chloroform increased penetration, as did pretreatment with Azone. However, no significant pH effect on penetration was found for either pre-treatment method. A follow-up study was conducted to examine the effect of pH on lipid and free fatty acid extraction, lipid packing, and keratin conformation (96). Human stratum corneum samples were shaken for 24 hours with buffers ranging from pH 1 to 12. Buffer pH had no large impact on the amount of sterols or ceramides extracted, but free fatty acid extraction was pH-dependent, being maximal at pH 11 and 12. Differential scanning calorimetry showed some disordering of lipid packing in alkaline-treated samples. The changes were not instantaneous and required > 1 hour exposure, becoming maximal after about eight hours. Fourier transform infrared spectroscopic analysis showed that the stratum corneum was largely unaffected by exposure to the buffer solutions, with no major changes to lipid packing motifs. Keratin conformation also appeared to be largely unaffected by buffer exposure, though there was some evidence that intracellular keratin took on a more ordered conformation at alkaline pH values. These authors concluded that the stratum corneum is remarkably resilient to extended exposure to both highly acidic and highly alkaline environments.
In adults the skin surface is normally slightly acidic, giving rise to the concept of the so-called “acid mantle.” Healthy adult skin exhibits a very good ability to recover from pH changes even when challenged with alkaline solutions having a pH value around 13 (97). Literature indicates that personal cleansing products can transiently affect the skin surface pH in both adults and infants. As was mentioned previously, Gfatter et al. examined the effect of washing infants’ skin with synthetic detergent and soap-based cleansing products (63). Washes were conducted with water (pH 7.9-8.2), a synthetic detergent bar (pH 5.5), a liquid synthetic detergent cleanser (pH 5.5), or a soap bar (pH 9.5). Skin surface pH measurements were made 10 minutes after washing. All washes increased the skin surface pH, with the water control producing the smallest increase ( + 0.20 units). Both synthetic detergent cleansers increased the skin surface pH by + 0.29 units, significantly greater than the control. The soap produced the greatest skin surface pH increase, +0.45 units. This increase was significantly greater than that produced by the control or the synthetic detergent cleansers.
Changes in the skin surface pH resulting from washing with personal cleansing products can persist for longer periods. Bechor et al. examined the time course of changes in skin surface pH following controlled washing (62). Adult volunteers washed their faces for 30 seconds with one of 41 cleansing products covering the surfactant composition range from soap to synthetic. The skin surface pH was measured at defined times for up to 200 minutes after washing. The results from this study show that cleanser-induced elevation of skin surface pH persisted for as long as 94 minutes after washing.
Korting et al. conducted eight-week crossover studies to demonstrate the potential for personal cleansers to alter skin surface pH. Liquid syndet cleansing preparations adjusted to pH 5.5 or 8.5 were used as test products. Subjects washed sites on their forehead and the ventral forearm twice daily for one minute. One cleanser was used for the first four weeks, the other for the remaining four weeks. Skin parameters were assessed at
various times during each period, at least six hours after the previous wash. Both studies showed that washing with the pH 8.5 product resulted in a higher skin surface pH than washing with the pH 5.5 product. The cleansers produced no consistent difference in TEWL or skin surface roughness (98) but did influence the skin’s microflora (99). No cleanser effect was observed on coagulase-negative Staphylococci populations, but Propionibacteria counts were increased when the cleanser at pH 8.5 was used. A similar effect on bacterial populations was demonstrated in a crossover study in which subjects used a full syndet bar or a soap bar for cleansing (100). The authors report that overall the skin surface pH was higher by 0.3 units and that Propionibacteria counts were elevated during the period of soap washing. These products were later compared in a three-month use study conducted among adolescents and young adults with acne (101). Fewer inflammatory lesions were observed in the group using the full syndet bar product. The authors extrapolate results from the earlier study conducted with liquid cleansers to rule out an effect due to differences in cleanser composition.
Alteration of skin surface pH might also effect more fundamental changes in the stratum corneum. For example, Fluhr et al. examined the impact of pH on stratum corneum acidification and integrity in a murine model (102). The backs and flanks of hairless mice were treated twice daily for three days with secretory phospholipase inhibitor (bromphenacylbromide or 1-hexadecyl-3-trifluoroethylglycero-sn-2-phospho – methanol) or vehicle control. Free fatty acid (palmitic, stearic, or linoleic acid) was coapplied to some animals. The effect of pH was examined by immersing flanks of anesthetized mice in bffer solution (pH 5.5 or pH 7.4) for three hours. The authors found that treatment with secretory phospholipase inhibitor increased skin surface pH and decreased barrier function (TEWL) and integrity (tape stripping), demonstrating a role for phospholipid metabolism in both these processes. Co-application of free fatty acid or exposure to pH 5.5 buffer normalized these effects. However, exposure to pH 7.4 buffer alone produced barrier alterations similar to the inhibitors, and exacerbated barrier effects in inhibitor-treated mice.
Barel et al. compared the skin effects resulting from use of a syndet bar (pH of 2% solution = 6.9) or a soap bar (pH of 2% solution = 9.6) in a blinded home-use test (103). Subjects washed their entire body with the assigned product at least once daily for a period of 10 weeks. Skin surface pH, TEWL, redness (chromameter a – value) and stratum corneum hydration were measured at baseline and endpoint on the hand, forearm, upper arm, neck, and leg. The skin surface pH after using with soap was significantly higher than after using the syndet bar on the upper arm, neck, and leg. The difference between the mean pH values measured at study end was < 0.4 unit, and the mean skin surface pH was in all cases < 6.0. None of the other instrumental measurements showed a difference between the two treatment groups, and expert evaluation of dryness and erythema showed that daily use of the products did not induce visible skin changes. Subjective ratings of overall irritation/ mildness showed a trend favoring the syndet bar at the end of the 10-week use period, but it seems likely based on the earlier discussion (e. g., the work of Imokawa) that this was due to a factor other than the product pH. The results of this study again highlight the difficulty of predicting in-use experience with controlled exposure models.