• endogenous factors;
  • ethnicity;
  • irritant contact dermatitis;
  • irritation;
  • man;
  • race


  1. Top of page
  2. Abstract
  3. Black versus Caucasian irritation response
  4. Asian versus Caucasian irritation response
  5. Conclusion
  6. References

Irritant contact dermatitis (ICD) is a common dermatological diagnosis in which endogenous and exogenous factors have been implicated. This review explores ethnicity as a possible endogenous factor. While there is a clinical consensus that blacks are less reactive and Asians are more reactive than Caucasians, the data supporting this hypothesis rarely reaches statistical significance. The studies reviewed do not provide sufficient evidence to determine that race is a predisposing factor to the degree of irritation in ICD. We conclude that race could be a factor in ICD, which has practical consequences regarding topical product testing requirements, an ever-expanding global market, occupational risk assessment, and the clinical thinking about ICD. Pitfalls in defining differences are discussed.

Irritant contact dermatitis (ICD) is a common and potentially serious dermatological disorder (1–3). It is also the second most common occupational illness (4). Because contact dermatitis can develop into chronic skin disease, understanding the underlying factors of its aetiology is clinically important.

This condition is divided into several forms depending on the nature of exposure and the resulting clinical presentation. Two common entities are acute and cumulative contact dermatitis. Acute contact dermatitis presents the classic symptoms of irritation such as localized and superficial erythema, oedema, and chemosis. It occurs as a result of single exposure to an acute irritant (5). Cumulative irritant contact dermatitis presents similar symptoms, but occurs when exposure to a less potent irritant is persistent or repeated until signs and symptoms develop over weeks, years, or decades.

The ability of the offending irritant to cause contact dermatitis depends on both the nature of the irritant agent and the initial skin condition. The severity of symptoms depends on exogenous and endogenous factors (6–8). Exogenous factors include the irritant's chemical and physical properties, and the vehicle and frequency of application. Endogenous factors have been speculated to be age, sex, pre-existing skin diseases, skin sensitivity, genetic background, and – the subject of this review − race (6), or, in today's parlance, ethnicity.

Ethnic differences in skin physiology and pathophysiology exist (9–11), and so whether ethnicity is, in fact, an endogenous factor affecting ICD is an important question in dermatotoxicology. Ethnic predisposition to ICD has been studied by comparing the irritant responses of blacks and Asians to those of Caucasians as a benchmark. We review these studies to evaluate if ethnic differences in susceptibility to ICD do exist.

The answer to the question of ethnicity as a factor in ICD has clinical and practical research consequences. Pre-market testing of topical products (soaps, detergents, perfumes, and cosmetics), risk assessment for occupational hazards, and subject-inclusion requirements for product safety studies require knowledge about ethnic differences in irritation (12).

Black versus Caucasian irritation response

  1. Top of page
  2. Abstract
  3. Black versus Caucasian irritation response
  4. Asian versus Caucasian irritation response
  5. Conclusion
  6. References

Using erythema as the parameter to quantify irritation, early studies note that blacks display less redness than Caucasians. In a hallmark paper, Marshall et al. (13) showed that while 59% of Caucasians exhibit acute irritant contact dermatitis as defined by erythema from 1% dichlorethylsulfide (DCES), only 15% of blacks do. Later, Weigand and Mershon (14) performed a 24-h patch test using ortho-chlorobenzylidene malononitrile as an irritant, which confirmed that blacks are less susceptible than Caucasians to ICD as defined by erythema. Further studies, also using erythema as a measure of irritation, showed that blacks are less reactive than Caucasians to irritants(160 mm/L and 1280 mm/L methacoline) (15, 16).

Weigand and Gaylor (17) showed that if the stratum corneum of black and Caucasian subjects is removed, there is no significant difference in irritation as measured by erythema between the two groups. They conclude that there might be structural differences in the stratum corneum that provide more protection from chemical irritation to black skin than Caucasian skin. Indeed, while the stratum corneum thickness is the same in both races (18), the stratum corneum of black skin has more cellular layers and stronger cells (12), more casual lipids (19), increased desquamation (20), decreased ceramides (21), and higher electrical resistance (22) than Caucasian skin. Some of these anatomical and physiological differences of the stratum corneum could be used to explain the observed reduced irritation in black skin as measured by erythema (3).

It is difficult, however, to conclude that blacks are less susceptible to cutaneous irritation based only on studies using visual scoring. Erythema is notoriously difficult to measure in darker skin. Perhaps the difference in skin irritation between the two test groups is simply a result of the difficulty of assessing erythema in black subjects.

To understand this issue better it is necessary to analyse studies that use alternative accurate detection methods (23) to assess the level of induced cutaneous irritation. Berardesca et al. (24) performed such a study to determine the difference in irritation between young Caucasian and young black skin. They applied 0.5% and 2.0% sodium lauryl sulfate (SLS) to untreated, pre-occluded, and predelipidized skin. Then they quantified the resulting level of irritation using objective techniques: laser Doppler velocimetry (LDV), transepidermal water loss (TEWL), and water content of the stratum corneum (WC). They found no statistical difference in irritation between the two groups as measured by LDV and WC, but they did find a statistical difference in the TEWL results of the pre-occluded test with 0.5% SLS. In that test, blacks had higher TEWL levels than Caucasians, suggesting that in the pre-occluded state blacks are more susceptible to irritation than Caucasians. The finding of this study contradicts the hypothesis that blacks are less reactive than Caucasians.

Similarly Gean et al. (25) found no statistically significant difference in the maximum LDV response between black and Caucasian subject groups when they challenged skin with topical methyl nicotinate (0.1 M, 0.3 M, and 1.0 M). Further, unlike the earlier studies, they found no difference in the blood flow and erythema responses between the two groups.

Guy et al. (26) supports the results finding that LDV measurements of induced blood flow after application of 100-mm methyl nicotinate reveal no significant differences between black and Caucasian subject groups; however, a significant difference was found using photoplethysmography (PPG). Caucasians had a greater PPG value than blacks, suggesting that Caucasians may be more susceptible to irritation. The authors did not explain why blood flow measurements using PPG showed a statistically significant difference between the groups when LDV did not.

Berardesca et al. (27) also found decreased reactivity in blood vessels in the black test group than the Caucasian test group. They measured the postocclusive cutaneous reactive hyperaemia – temporary increase in blood flow after vascular occlusion – after an application of a potent corticosteriod, and measured vasoconstriction using LDV; the black subject group had several significantly different parameters of the hyperaemic reaction. They found a decreased area under the LDV curve response, a decreased LDV peak response, and a decreased decay slope after peak blood flow, showing that blacks have a decreased level of irritation-induced reactivity of blood vessels. These results are consistent with their previous work.

In conclusion, older studies using erythema as the only indicator for irritation show that blacks have less-irritable skin than Caucasians, but more recent studies using objective bioengineering techniques suggest that the eye may have misled us to an incorrect interpretation. Findings that do and do not show statistically significant differences in the irritation response between blacks and Caucasians are summarized in Tables 1 and 2.

Table 1.  Findings that show a statistically significant difference in the irritation response between blacks and Caucasians
1% DichlorethylsulfideErythemaUntreatedMarshall et al. (13)
Ortho-chorobenzylideneErythemaUntreatedWeigand et al. (14)
100 mm methyl nicotinatePPGUntreatedGuy et al. (26)
0.05% clobetasolLDVPre-occludedBerardesca et al. (27)
0.5–2.0% SLSTEWLPre-occludedBerardesca et al. (24)
Table 2.  Findings that do not show a statistically significant difference in the irritation response between blacks and Caucasians
0.5–2.0% SLSLDV and WCUntreated, pre-occluded, and pre-delipidizedBerardesca et al. (24)
100 mm methyl nicotinateLDVUntreatedGuy et al. (26)
0.1 m, 0.3 m, and 1.0 m methyl nicotninateLDV and ErythemaUntreatedGean et al. (25)

Asian versus Caucasian irritation response

  1. Top of page
  2. Abstract
  3. Black versus Caucasian irritation response
  4. Asian versus Caucasian irritation response
  5. Conclusion
  6. References

An early study comparing Caucasian and Japanese susceptibility to cutaneous irritation was done by Rapaport (28). He performed a standard 21-day patch test protocol on Caucasian and Japanese females in the Los Angeles area in which 15 irritants (different types or concentrations of cleansers, sunscreen, and SLS) were tested. The results were reported according to the cumulative readings of all subjects in an ethnicity group for each irritant. Japanese women had higher cumulative irritation scores for 13 of the 15 irritants tested; he interpreted these findings to confirm the common impression that Japanese are more sensitive to irritants than Caucasians. Also, this sensitivity was independent of the concentration or exact chemical formulation of the substance tested, suggesting that Japanese are in general more sensitive than Caucasians.

While these findings are important, it is difficult to interpret this data. First, as also noted by Robinson (12), Rapaport provides little experimental detail and data. For example, while the study required 21 separate days of irritation readings, only the end cumulative irritation scores are reported. If he had reported daily irritation readings, we would have been able to note the time pattern of response. Further, no statistical tests were conducted to ascertain if the differences between the Japanese and Caucasian subjects were statistically significant. Note, too, that the cumulative irritation test score does not distinguish between the intensity of a subject's response and the number of subjects responding. Thus it is possible, for example, for a few extremely sensitive Japanese subjects to inflate the overall irritation score. Therefore, at the minimum, it would be helpful to provide standard deviations to rule out such problems.

At first seemingly surprising, Basketter et al. (29) found that Germans are more sensitive than Chinese subjects. Subjects in Germany, China, and the United Kingdom were exposed to varying concentrations (0.1%-20%) of sodium dodecyl sulfate (SDS) for 4 h on the upper outer arm, and the resulting dose–response irritation was measured based on erythema. They concluded that Germans tend to be more sensitive than Chinese subjects, and the Chinese subjects slightly more sensitive than the British subjects. This conclusion runs contrary popular belief and to the Rapaport study, which indicated that Asians are more likely to develop irritant contact dermatitis than Caucasians.

There are, however, inherent flaws in this study, some of which the authors acknowledged. First and foremost, this study does not control the variables of time and location. The German and Chinese studies were performed over 3–6 weeks in the winter, while the UK study was spread over 15 months. Also, in particular, German winters are colder and drier than Chinese winters, and Chinese winters tend to be colder than English winters. These variables will distort the results in a predictable way if we assume that an individual becomes more sensitive to irritant contact dermatitis in colder and drier climates (2). We would then expect, based on climatic conditions, that the German subjects would be more reactive than the Chinese subjects, and the Chinese subjects more reactive than those from the UK. As these are the actual results, we cannot necessarily contribute the differences in irritant response to ethnicity, as it is possible that the differences are due to weather conditions. Also, they mention that 15% of the UK volunteers were black. While they account for this by showing that the black irritant response was similar to the overall UK group response, it is scientifically problematic to mix racial groups in a study testing for racial differences. Furthermore, they supplied no statistical tests for their conclusion that Germans are slightly more sensitive than the other ethnic groups. To shed more light on the results, we conducted simple binomial tests of the differences in the percentage response of the subject groups. Using the resulting statistics, we found a larger statistically significant difference between the two predominantly Caucasian groups than between each of the Caucasian and the Chinese groups (Table 3). These results indicate that race may not be the predominant factor affecting susceptibility to ICD in this study; other uncontrolled variables may dominate the results.

Table 3.  Statistical analysis of the Basketter et al. (29) study
 0.1% SDS0.25% SDS0.5% SDS1.0% SDS2.5% SDS5.0% SDS10% SDS20% SDS
  1. The numbers in the first three rows are the decimal value of the percentage of the group that developed a positive irritant reaction at a specific SDS concentration. The numbers in the last three rows are the Z-values. We applied the binomial test to ascertain the differences in the percentage response of the subject groups: Z = (r1r2)/[2r(1 − r)/100]50, where r1 and r2 are the ratios for the two ethnic groups and r is the weighted average. As the sample sizes for different groups are equal, r becomes the simple average. An asterisk indicates that the ratios are significant at the 5% level. Note that all the UK–Germany differences, except one, are statistically significant; however, more than half of the UK–China and almost half of Germany–China differences are not statistically significant. This indicates a larger statistically significant difference between the two Caucasian groups that between the Caucasian and Asian groups.

Z (Germany-China)1.753.07*4.79*4.29*2.84*1.65− 0.51NA
Z (UK-China)− 1.10− 1.74− 2.83*− 4.42*− 2.64*
Z (UK- Germany)− 1.01− 2.60*− 3.41*− 5.28*− 4.53*− 4.42*− 3.94*NA

Variables such as time and location were eliminated by the Goh and Chia (30) study that tested the susceptibility to acute irritant contact dermatitis in Chinese, Malaysian, and Indian subjects. These subjects were exposed to 2% SLS in the right scapular region, and resulting irritation measured using transepidermal water loss (TEWL). This technique is an objective way to indirectly quantify irritation – the higher the TEWL value, the greater the implicit irritation. There was no significant difference in the TEWL level of irritant skin in a 3-way statistical test of the three racial groups. There was a significant difference, however, between the TEWL values of Chinese and Malaysian subjects such that Chinese subjects were more susceptible to contact dermatitis. While this test does not contribute to the discussion of the difference in predisposition of irritation in Caucasian versus Asian skin, it does add to the overall question of whether race can be a predisposition to irritant contact dermatitis.

Foy et al. (31) clearly added to our knowledge of the difference in the acute and cumulative irritation response in Japanese and Caucasian female skin. They reduced some variables that compromised other studies; location, time, season, and scorer were the same for both study populations. 11 different materials were tested in the acute test; they were applied to the upper arms for 24 h, and irritation was measured based on erythema. The cumulative test consisted of testing 5 irritants using a 4-exposure cumulative patch protocol.

In the acute test, while there is a slight tendency to greater susceptibility to irritation among Japanese subjects, only 4 out of the 11 irritants caused a significant difference in reactivity between the two groups—these were the most-concentrated irritants used. This shows that perhaps for more-concentrated irritants there is indeed a statistical difference in the acute contact dermatitis response; of course, this study needs to be interpreted in context with others to follow. For the cumulative study the skin irritation scores between the two test groups are close – but the Japanese tended to have slightly higher numbers. The differences, however, only reached statistical significance in two instances. and as the authors noted, it is difficult to interpret the importance of those two instances, since the statistically significant differences are not maintained at later points in the timeline. It is safe to conclude therefore that while the acute irritant response to highly concentrated irritants was significantly different between the Japanese and Caucasian subjects, the cumulative irritant response rarely reaches a statistical difference.

Studies that include both acute and cumulative irritant tests, like the one above, are more informative than single tests since they give a more complete view of differences in skin irritation between groups. Robinson (32) conducted a series of studies that tested racial differences in acute and cumulative skin irritation responses between Caucasian and Asian populations. In the first acute tests, Caucasian and Japanese groups were exposed on the upper outer arm to 5 irritants under occlusion for up to 4 h. The resulting erythema was scored on an arbitrary visual scale. The results were represented as the cumulative percent incidence of positive test reactions to the different irritants.

It is curious to note that while Japanese subjects tend to be more susceptible to acute irritation than Caucasians, no one irritant nor one test time caused a significant response difference between the two groups; rather, the significant differences were scattered across 5 different test materials and time points. The acute irritation response data was then reanalysed in terms of possible differences in temporal response. It was shown that Japanese subjects generally react faster than their Caucasian counterparts, as indicated by their shorter TR50 values (the time it takes for the cumulative irritation score to reach 50%). While this result is interesting, and adds the new dimension of temporal differences in reactivity between the two groups, hard data was not given and statistical analysis was not conducted to see if this temporal pattern difference is indeed statistically significant.

The cumulative irritation test was conducted concurrently and on the same Japanese and Caucasian subjects. Four concentrations of SDS (0.025%, 0.05%. 0.1%, and 0.3%) were applied on the subjects' upper backs for 24 h for a total of 14 days. The resulting skin grades were summed for all subjects for all test days. For the two lower SDS concentrations the Japanese subjects reacted only slightly more than the Caucasian subjects − but only the difference in skin grades for 0.025% SDS reached statistical significance. When this data was analysed in terms of temporal response, the two lowest concentrations the Japanese reacted only slightly faster than their Caucasian counterparts. Whether the difference in reaction time is statistically significant in not known.

In the same study, Robinson then applied both the acute and cumulative irritation protocols to compare three new subject groups – Chinese, Japanese, and Caucasian – to each other. The cumulative irritation study found no statistically significant differences between the different groups. In the acute test, he found that, in most cases, the Chinese subjects were more reactive to irritants than Caucasians, but that in only one case was this difference significant, and he states that most likely this was an anomaly. There was no discernible difference between the Japanese and Chinese groups. and surprisingly, when the Japanese subjects were again compared to the Caucasian subjects as they were in the beginning of his study, the results showed no significant difference between the two groups.

While Robinson's first 2-way irritation response comparison test between Japanese and Caucasian subjects did show some statistical differences, the fact that they could not be confirmed in the second half of the study emphasizes the difficulty in obtaining repeatable results in this type of study. For one thing, in the statistical sense Robinson's sample sizes (approximately 20 people) were small, combined with the variability between human skin within an ethnic group, this makes it difficult to make concrete conclusions. His study showed, however, that there were essentially no significant differences between the Asian and Caucasian groups, at least none that could be repeated.

Robinson et al. (33) had similar results. Using the 4-h occlusion patch method, they compared the relative acute skin reactivity of Asian and Caucasian subjects using the irritation temporal response to measure the difference in reactivity between the test groups. They tested 5 chemicals, including 20% SDS and 100% decanol. Unlike the previously described study, they failed to find a statistical difference between the reactivity to multiple irritants between the two groups, even at the 4-h mark. Then they did something new: they separated the racial subpopulations into ‘sensitive’ and ‘normal’ groups to test any differences in percent cumulative scores and temporal responses within these new groups but across race (i.e. they compared sensitive Asians to sensitive Caucasians). There were no statistically significant differences between subjects of the same skin type in different racial groups. This further contradicts the hypothesis that Asians are more reactive to irritants than Caucasians.

Recently, Robinson (34) compiled 5 years of his previous data and compared the acute reactivity differences between Caucasian and Asian (combined Japanese and Chinese) subgroups using the 4-h human patch method. The data was represented in terms of the time it took subjects to have a positive response to the irritant chemical. Again, as in most experiments, Asians displayed a greater irritation response score than Caucasians. Note that while these results of this study are probably more representative of the population at large because of the relatively large sample size (200 plus), the data from this study was compiled from 3 different testing centres over 5 years. This could have potentially added uncontrolled and unaccounted-for variables.

In support of the long-held belief that Asians are more susceptible to irritant contact dermatitis, several studies do indeed demonstrate this tendency (31, 33, 34). Rarely, however, is this trend statistically significant, and even more rarely can the statistical significance be repeated in an another study. Therefore, it can be concluded from these studies that there is no fundamental difference between Asian and Caucasian cutaneous irritant reactivity, the overall irritant response and the time to reach that response is similar in both subgroups.

However, the lack of comparable studies, small sample sizes, external variability, and intravariability within the subgroups make it difficult to completely dismiss Rapaport's original findings that Asians are more reactive than Caucasians. For example, different studies apply the irritant test material on different parts of the body, which might have different reaction responses. This makes it difficult to compare the results of one study to another, and therefore raises the question of whether a more solid trend among studies would exist if the irritants were applied to the same anatomical site. Some potent factors that might influence refinement of interpretation in future investigations are listed in Table 4. For the time being, however, in terms of topical product safety, risk assessment for occupational hazards, and global product marketing it would be practical to assume that few statistically significant difference between Asian and Caucasian cutaneous reactivity exists.

Table 4.  Potent factors that might influence refinement of interpretation in future investigations
Experimental design
Baseline versus‘stress’ test differences
Anatomic site
Open versus occluded irritant stresses
Ethnic groups in the same versus varying geography
Comparable climatic conditions
Presentation of hard data and statistical analysis


  1. Top of page
  2. Abstract
  3. Black versus Caucasian irritation response
  4. Asian versus Caucasian irritation response
  5. Conclusion
  6. References

The studies reviewed demonstrate that there is little evidence of statistically significant differences in the irritant response between Caucasian, black and Asian groups. We can see no consensus on whether race is indeed an endogenous factor in ICD. Intuitively, we suspect that ethnic differences exist in skin function and may have evolved as have hair and other differences. Basically, the studies suggesting differences in skin (24, 26) are ‘stress’ in nature (pre-occluded). Presumably new insights into physiology, pharmacology, and toxicology may clarify this situation.

Also it is possible that the well-known divergent response to irritants is due to intra-individual variations in the skin irritation response (35–37). This is a relatively new idea, and therefore further studies need to be conducted in this area before a definitive statement can be made linking intra-individual variation to ethnic differences in the intensity of an irritation response.

We reviewed PubMed, MD Consult, the Melvyl catalogue in the CDL-Hosted Database, Yahoo, Google, and our library files from 1967 to June 2002.


  1. Top of page
  2. Abstract
  3. Black versus Caucasian irritation response
  4. Asian versus Caucasian irritation response
  5. Conclusion
  6. References

Address: Howard I. Maibach School of Medicine Department of Dermatology 90 Medical Center Way Surge 110 San Francisco California 941430989 USA Fax: + 415 7535304 e-mail: