Susceptibility to and incidence of hand dermatitis in a cohort of apprentice hairdressers and nurses.

OBJECTIVES
The role of atopic constitution, contact sensitization, transepidermal water loss, and dry skin in the development of hand dermatitis was investigated in a prospective study of 74 apprentice hairdressers and 111 apprentice nurses.


METHODS
Base-line measurements included a questionnaire on personal characteristics and anamnestic information, examination of hand skin, measurements of transepidermal water loss, patch tests, and prick tests. The condition of the hands, previous exposure, and transepidermal water loss were followed at intervals of four to six weeks. Cox proportional hazard models were used in the statistical analysis.


RESULTS
The average incidence rate of hand dermatitis was 32.8 cases per 100 person-years for the hairdressers and 14.5 cases per 100 person-years for the nurses. The rate ratio of having a dry versus normal skin type was 7.3 for the hairdressers [95% confidence interval (95% CI) 2.2-24.3] and 1.7 for the nurses (95% CI 0.5-6.4). Apprentice nurses with a history of (atopic) mucosal symptoms had a 3.4-fold increased incidence rate of hand dermatitis (95% CI 1.05-11.2). The rate ratio of mucosal atopy for the apprentice hairdressers was 2.2 (95% CI 0.7-6.7). Graphic display of the results suggested an increased risk of hand dermatitis among the apprentice hairdressers with transepidermal water loss on the hand greater than 15 g.m-2.h, but the relative risk of increased transepidermal water loss was not statistically significant.


CONCLUSION
The most important endogenous risk factors for hand dermatitis among the apprentice hairdressers and nurses were the presence of dry skin and a history of mucosal atopy. No relationship between increased transepidermal water loss and the risk of hand dermatitis was observed.

opme nt of contact dermatitis (10)(11)(12)(13). The bi ological mechanism underlying thi s hypothesi s is that a dimin ish ed barri e r functio n of the skin, re flec ted by increased transepidermal wa te r loss, allows ha rm ful age nts to pa ss through the stra tum co rne um mor e easi ly and cause dam age in th e und erl ying laye rs of th e skin . However, th e predictive value of an inc rea sed base-line transepiderrnal water loss for the risk of hand dermatitis has not ye t been evalua ted in epidem iologi c studies .
A pro spective st udy was performed amo ng apprentice hairdressers and nur ses with two obj ect ives. O ne was to obt ain qu antitative estimates of th e relationship between hand dermatitis and known risk fa ctor s suc h as ato pic back ground, co ntact allergy, a nd typ e of skin. The oth er wa s to inv es tigate th e hyp othesis that indiv id ua ls wi th increased base-line transepidermal water loss as a person al ch aracteristi c are mor e sus ceptible to developing hand dermatiti s during occupational ex pos ure than those wi th low transepid ermal water loss and similar expo su re. Hai rdressers and nur ses were chose n as study population s becau se they are know n to be at high ris k for the de vel opment of hand dermatiti s (2,(14)(15)(16)(17).

Study design and population under study
The study group consisted of III apprentice nurses and 74 apprentice hairdressers who were free of hand dermatitis at the start of the investigation. The apprentices were under observation during their training. They were examined at regular intervals to identify new cases of hand dermatitis and to record changes in occupational exposure. Figure I summarizes the design of the study.
Four groups of apprentice hairdressers, a total of 77 persons, were recruited at the start of their training in September 1990, December 1990, May 1991 or September 1991 (response rate 100%). Three apprentices left school before the first follow-up measurement was made. Therefore, 74 apprentice hairdressers were included in the study. The duration of the hairdressers' training averaged 10 months. Eight subjects, who failed their final examination, remained under observation until they were successful one to four months later. The study of hairdressers ended in May 1992, when the last determination was made in the fourth group.
Two groups of apprentice nurses, a total of 126 persons, were recruited in the third year of their training, which included the first period of practical training (response rate 86%). The first group was enrolled in October 1990. They were under observation for two years until the end of their training (July 1992). The second group was enrolled in October 1991 and was under observation until July 1992 (one school year). Four individuals who suffered from hand dermatitis at the start of the study were not included in the analysis. Eleven apprentice nurses dropped out for reasons unrelated to hand dermatitis before the first follow-up determination. Therefore, III apprentice nurses were included in the study. The first year of follow-up included two periods of practical work of 12 weeks each, with an interval of eight weeks of classes without occupational exposure. The second year of follow-up consisted of 12 weeks of classes followed by 20 weeks of practical work.

Data collection
Baseline measurements. Data collection took place in an office at the hairdressers' and nurses' schools. At the time of entry into the study, the following protocol was adhered to: Four hairdressers left school and were lost to follow-up after a minimum observation period of 22 weeks. Six apprentice nurses were lost during the first school year of observation for several reasons unlikely to be related to hand dermatitis (left school, lost interest in the study, lack of time, illness, or accident). Thirteen apprentice nurses of the first group were lost in the second school year, mostly because they had a different schedule for practical work.
Patch test readings were obtained for 97% of the apprentice hairdressers and for 75% of the apprentice nurses. Prick tests were performed for 99% of the hairdressers and 78% of the nurses. The major reasons for refusal to participate in the dermatological tests were fear or inconvenience. Prior dermatological testing was not given as a reason for refusal by any of the subjects.
2. An examination of the skin of the hands was performed by a trained physician, and positive skin findings (one "A" sign or two "B" signs) were recorded (18). "A" signs were grouped papules, grouped pustules, grouped vesicles, and exudation. "B" signs were erythema, scaling, edema, fissures, and lichenification. The type of skin on the hands was recorded and graded as dry, normal, or oily. I. A questionnaire was completed to determine relevant base-line characteristics such as age, gender, history of asthma or hay fever, history of childhood eczema, history of past skin disease, and prior exposure to hairdressing or nursing activities.
3. Base-line measurements of transepidermal water loss (expressed as grams per square meter per hour) were performed on the forearm and the back of the dominant hand as indicators of the susceptibility of the skin. Measurements on the forearm were made on the volar side, 8 cm from the wrist. Transepidermal water loss was measured on the dominant hand, since this hand was assumed to have a similar probability of occupational exposure among the subjects. The evaporimeter EP IC (ServoMed AB, Kinna, Sweden) was used for measuring transepidermal water loss. The operating principle of the instrument is based upon measuring the vapor pressure gradient through the skin (19). The measurements were performed according to the guidelines proposed by the Standardization Group of the European Society of Contact Dermatitis (20). The exception was that the temperature and relative humidity ranged more widely in this study than what was proposed in the rec- ommendations (20°C; 40% relative humidity). On the six occasions that base-line measu rement s were performed, the ambient air temperature varied between 16 and 24°C. The range in the relative humidit y of the ambi ent air was 27-62%. Fluctuations in the measurement condit ions were due mainly to the fact that the measurements were performed in the schools, whe re no air-conditioned room s were available. Ambient air temperature and humidity were recorded before each measurement. The measurements of transepidermal water loss were recorded continuously on a chart strip recorder. All of the measurements were made in duplicate. 4. A thin-l ayer rapid use ep icutan eous test (TRUE test, Pharm acia) was performed to determin e the existence of ce ll-mediated allergy to common cont act aller gen s at the start of the study (2 1). Patch testing was performed with the use of the Europe an Standard Series , which included 23 of the most common allergens and one negative control. The patches were applied to the subjects' back and remo ved after 48 h. The patch tests were read by a derm atologist after 48 and 72 h. They were graded acco rding to the recommendations of the International Contact Derm atiti s Research Group (22). 5. Prick tests (Pharmacia) were carried out on the forearm to determ ine the existence of immediate type allergy. Test substances were pollen of birch, alder , timothy grass, dander of cat, dog and guinea pig, and hou se dust mite. Histamine was used as a positi ve control. Prick tests were read after 20 min by an experienced dermatologist.
Follow-up measurements. The follow-up measurements of the apprentice hairdre ssers took place at intervals of about six weeks. Tho se of the apprentice nurse s were performed the week before and the week after each period of practical work. The following protocol was applied for each follow-up measurement. I. A que stionnaire was completed by all of the subjects to determine the incidence of symptoms of hand dermatitis over the preced ing peri od and the date of occurrence of first symptoms. For this purpose, a set of validated questi ons on symptoms of hand dermatitis (vesicles, scaling, itching, redness, swelling, and fissures) and their frequen cy and duration was used (23). The type and intensity of exposure during the preceding period was also assessed. For the hairdressers the questi ons included the frequen cy of hair washing, giving permanent waves and hair dyein g, the use of gloves, and the type of fluids used for the permanent waves. For the nurses the question s included the frequency of hand wash ing, washing of patients, use of disinfectant s, cont act with medica-Scand J Work Environ Health 1994. vol 20. no 2 rncnts, treatment of decub itus wounds, use of hand lotion or creams.

2.
A physician examined the hands and record ed the presence of signs of (recent) hand derma titis.
3. Transcpidermal water loss was measured on the forearm and the back of the dominant hand according to the same protocol used for the base-line measurements. The result s of measurements over time will be repo rted elsewhere.

Definition of relevant variables
One expe rienced dermatologist (PIC) identified the cases of hand dermatitis without knowled ge of the subjects' base-line characteristics. The diagnosis was based upon the occurrence of symptoms of hand dermatiti s during the previous follow-up period as reported in the validated questionnaire (23). The skin find ings , as recorded by the physician , were evaluated to exclude mild or doubtful cases or subjects with other skin diseases. Thus a subject was diagnosed as having hand dermatitis if one or more combinations of sympt oms, as reported in the questi onnaire, had occurred during the previous follow-up period , if the symptoms were recurr ent or had lasted for at least three weeks (definition used in validation study), and if this informati on was supported by the skin findings recorded durin g the examination of the hand s. The date of the onset of symptoms reported in the questionnaire was taken as the incidence date of hand derm atiti s.
Inform ation on the history of (atopic) muco sal symptoms was based upon the question " Have you ever suffered from asthma or hay fever?" A history of childhood eczema was based upon the quest ion "Did you suffer from eczema or dermatitis as a child?" The presence of dry skin was assessed by visual scoring, in which a positi ve score was give n when the skin of the hand s looked or felt dry or rough.
Patch test read ings with a grading of 2+ or more after 72 h were interpreted as positive patch tests. A positi ve prick test was defin ed as a reaction (in millimeters) to a test substance that was greater than or equal to the reaction to histamine. Any reaction less than the one to histamine was defined as negative . One or more positive reactions to any of thes e substances versus all negati ve tests served as an indicator of the presence of immed iate type allergy in the analysis.
The average value of the dupli cate measurements of transepidermal water loss was taken as a subject' s base-line value for the hand and fore arm. For a comparison of the relati ve risk of high versus low values, boundary points for "high," "intermediate," and "low" were established from each subject's average value according to the tertil es of the distribution of the value s in the total study population. Prelimi -nary analyses showed that the coefficient of variation of the duplicate measurements on a person was 11.7% for transepidermal water loss on the forearm, and the corresponding value was 13.3% for the hand.

Statistical analysis
The incidence rate, expressed as the number of incident cases per 100 person-years of observation, was used as a measure for disease occurrence. Observation time was accumulated from the date of entry into the study to the date of exit from the study. The date of exit was defined as the date of the onset of first symptoms for cases or the date of lost to follow-up or end of the study for noncases. For comparison with results from other studies, the cumulative incidence of hand dermatitis was derived from the incidence rates (24).
The relationship between the variables of interest and the risk of hand dermatitis was assessed by means of Cox proportional hazard models (25). This method makes use of the instantaneous incidence rate ("hazard rate") at each point in time that a subject developed hand dermatitis. The ratio of the hazard rates for subjects with and without the factor of interest (for example, atopic versus nonatopic constitution) is a generic measure of the relative risk of the factor of interest. The term "significant" has been used throughout this paper to refer to statistical significance at the 5% level. The following steps were taken in the analysis.
I. The validity of the proportional hazard assumption was explored for all potential risk factors of interest. The assumption was valid for all such factors. Given the substantial differences in exposure conditions between the hairdressers and nurses, however, separate analyses were initially performed for both groups. An overall relative risk for each factor among the hairdressers and nurses combined was estimated by stratification on the type of training in this analysis, the differences in the base-line hazard rate between the two groups thus being allowed for.
2. The crude and adjusted estimates of the relative risk for each risk factor and its 95% confidence interval were estimated from the regression coefficients of a Cox regression model, modeling each risk factor separately. Table I shows the base -line characteristics of the 74 hairdressers and III nurses who contributed persontime to the study population. Only small differences were observed in the distribution of characteristics between the subgroups of hairdressers and nurses who entered the study population at different times. The study population of apprentice hairdressers was four years younger on the average than apprentice nurses and contained a higher proportion of men. The proportion of self-reported (atopic) mucosal symptoms was higher among the hairdressers than among the nurses, whereas the proportion of sensitized subjects and the proportion of subjects with dry skin were smaller among the hairdressers. The differences in the distribution of the base-line characteristics between the men and women were difficult to interpret due to the small number of men. Nevertheless, gender was taken into account in all of the analyses as a potential confounder. The Pearson correlation coefficient between a history of (atopic) mucosal symptoms and the presence of one or more positive prick tests was 0.46 (P<O.OO I). Nevertheless, 10% (5 of 50) of the hairdressers and 10% (7 of 70) of the nurses without (atopic) mucosal symptoms showed a positive response to the prick tests, while 53% (8 of 15) of the hairdressers and 50% (5 of 10) of the nurses with a positive history of mucosal symptoms showed a negative response to all of the prick tests. No significant relationship was observed between a history of childhood eczema and the results of the prick or patch tests.

Characteristics of the study population
Before the start of the study, 35% of the apprentice hairdressers had been occasionally involved in hairdresser's activities, mainly hair washing and drying. None of the nurses had been wor king as a nurse before the start of the stud y. The mean transepiderm al water loss from the hand of the hairdressers was 1.3 g . m' . h less than that of the nurses. No difference was observed between the hairdre ssers and nurses in transepiderm al water loss of the forearm . In addition, no significant differe nces were observe d in the base -line charac teristics of the subjects with high and low transepidermal water loss. The measurement conditions were different on the six occasions that the base-line measurements were performed. However, the differences in the mean of the transep ider mal water loss were not assoc iated with the differences in the measurement conditions. depend ed mainly on the type of institution in which the period of practica l work was spent. Hands were washed more frequentl y in general hospitals and nursing homes (vary ing from 10 to 40 times per shift on the average) than in institutions for psych iatric patients and for ment ally defective patients (less than five times per shift on the average). Differences in the frequency of was hing and changi ng patients were less marked between these two types of institutio ns; nevertheless, they showe d some varia tio n betwee n persons. However, no asso ciation was observed between the exposure characteristics of eac h training peri od and the presence of endoge nous characteristics.  Incidence of hand dermatitis Seventeen hairdressers and 16 nurses deve loped hand dermatitis while they were und er observation. In most cases, the symptoms were mild to moderate and occ urred periodically. None of the apprentices discontinued their training due to hand dermatitis. However, one hairdresser who was diag nose d as having hand dermatitis during the study con sulted the dermatology clinic within six months after she had been professionally employed. The ave rage observation time was 36 weeks for the hairdressers (total of 269 1 person-weeks) and 52 week s for the nurses (total 5747 person-weeks ).
The type and intensity of exposure was fairly homogeneous within the study popul ation of hairdressers . Figure 2 shows that the expos ure freq uency during hair washing, permane nt waving. and hair dyeing steadil y increased dur ing the practical trainin g. Th e variation in reported freque ncies between persons was small. No associatio n was obse rved between the expos ure charac teris tics and the point of entry into the study as a hairdr esser. The inciden ce rate was highest for the period betwee n three to six months after the start of training, and it greatly declined thereafter. The average incidence rate was 32.8 cas es per 100 per son-years of obser vation for the hairdr essers. The one-yea r cumulative incidence was 27.9%.
For the apprentice nurses, the average incidence rate of hand der matitis ove r the total study period was 14.5 cases per 100 person-years of observati on (one-year cumulative inc idence of 13.5%). The incide nce rate was highest during the first two periods of practica l train ing and was lower durin g periods of classes and during the last peri od of practical trainin g ( figure 3). The differen ce in incidence rates between the per iods was not sign ifica nt. The average incidence rate in the first schoo l year was 19.8 cases per 100 person-yea rs of observa tion. In the second year, the average inciden ce rate was 5.2 cases per 100 person-years of observation. The type and intensity of expos ure were more heterogeneous for the apprenti ce nurses than for the haird ressers. They

Relation between the base-line characteristics and the incidence rate of hand dermatitis
The crude and adjusted relative risks of developing hand dermatitis are shown in table 2 for the risk factors of interest. Adjustment for mucosal atopy and skin type altered the point estimates of the relative risk to some extent. Other endogenous risk factors (including gender) made no significant contributions to the model, nor did they meaningfully influence the magnitude of the point estimates of these relative risks. When the fact that subgroups entered the study at different times (thus experiencing different sequential effects of the seasons) was taken into account in the analysis, the estimates of the relative risk did not change meaningfully.
The relative risk for hand dermatitis of dry versus normal skin was elevated both for the hairdressers and the nurses; however, the increased relative risk was significant at the 5% level only for the hairdressers. The relative risk of (atopic) mucosal symptoms was elevated for both groups as well, but the increase was significant at the 5% level only for the nurses. The relative risks of skin type and mucosal atopy differed in magnitude between the hairdressers and nurses, but these differences were not significant at the 5% level. The overall estimates, adjusted for type of training, showed a significant 2.5-fold increased relative risk for mucosal atopy and a significant 3.2-fold increase in the relative risk for dry skin. No significant association was observed between age, gender, childhood eczema, or presence of a positive skin test (prick or patch test) and the risk of hand dermatitis.
The lack of a relationship with transepidermal water loss as a continuous variable indicates that there was no linear increase in the risk of hand dermatitis with increasing trans epidermal water loss on the hand or forearm. Adjustment for the temperature and humidity of the ambient air did not alter the estimates of the relative risk. Figure 4 (hairdressers) suggests that, for transepidermal water loss below 15 g . rrr-. h, there was no clear relationship with the incidence rate, whereas the incidence rate was elevated for transepidermal water loss above ] 5 g. m? . h. Figure 5 (nurses) provides no evidence for a relationship between transepidermal water loss on   the hand and the incidence of hand dermatitis . It should be noted that the average incidence rate in some categories of transepidermal water loss was based on small numbers of subjects.

Discussion
The incidence rate of hand dermatitis was higher for the apprentice hairdressers than for the nurses (32.8 versu s 14.5 cases per 100 person-years). It is likel y that the differences in the type and inten sity of exposure between the hairdressers and the nurses are responsible for the differences in the incidence rates. A previous study among apprentice hairdressers (16) showed that about one-third of the group developed skin changes during the first year of the training. The one-year cumulative incidence of about 28 % that was found for the hairdressers in the present study was similar.
The cumulative incidence of hand dermatiti s among apprentice nurses over the entire study period of 21 months was 21.8 %. The cumulative incidence figure s for nurses, as observed in a Swedish (2) and a Dutch (26) study , were 41 % (20 months) and 13% (18 months), respectively. The differences are likely to be explained by variations in exposure history and disease status before employment (26).
The presence of dry skin and a positive history of (atopic) mucosal symptoms were associated with the risk of hand dermatitis. Since these characteristics were recorded at the start of apprenticeship, it is unlikely that the association can be explained by information bias . The increased relative risk is unlikely to be attributed to seasonal influences since the subgroups entering the study population at different time s of the year had a similar distribution of endogenous characteristics. In addition, accounting for subgroup in the analysis did not meaningfully alter the estimates of the relative risk . It cannot be excluded that some confounding by occupational exposure occurred; occupational exposure was not accounted for in the anal ysis since reliable characterization of exposu re and changes over time was not pos sible on the individual level in this study . However, there was no association between exposure characteristics and the presence of dry skin and (atopic) mucosal symptoms. Thi s lack of association suggested that it was unlikely that the increased relative risk was largely influenced by differences in occupational exposure.
The relationship between hand dermatitis and the presence of dry skin was observed in a previous study among hairdressers and bricklayers (8). An experimental stud y demonstrated that individuals with dry skin react more strongly to exposure to irritants (9). Other studies found that dry skin was a risk factor, particularly in combination with atopic dermatitis (27) or with atopic mucosal symptoms (28) . However, dry skin is known to be a diagnostic feature of atopic dermatitis (29). In the present study , skin dryness was not associated with atopic dermatitis or atopic mucosal symptoms, a finding suggesting that atopy cannot explain the association with dry skin. Dry skin may also be an early manifestation of hand dermatitis resulting from previous involvement in wet work, such as household chores, haird ressing, or nursing activities. Although the population s in this study were selected because previous occupational exposure was unlikely, it appeared that 35% of the hairdressers had occasionally been involved in hairdress ing acti vities before the start of the study . Inspection of the data showed that the presence of dry Scund J Work Environ Health 1994. vol 20. no 2 skin was not ass ociated with pre vious expos ure or with a history of childhood eczem a. Thi s finding sugges ts that dr y skin, at least in the present study , repres e nts a risk factor for hand derm atit is per se, rather than an earl y manifestati on of hand dermatitis.
Th e association bet ween (atopic) mucosal sy mptom s and the risk of hand derm atitis has also been obse rve d in pr eviou s studies (5 , 30) , whereas no assoc iation was obs er ved in other st udies (3,4) . The biological mechan ism ex plai ning the relati on ship between (atopic) mucosal sy mptoms and hand dermatitis is not fully understood . In additio n, the validity of se lf-reported atopic mucosal symptoms is unkn own. About 50 % of the subjects who reported (atopic) mucosal symptoms res po nded negatively to the prick tests. This fin din g suggests some overreport ing of (atopic) mucosal sympto ms , although loca l allergic reactions may occur in spite of negative sk in tests. The absence of an association between hand dermatitis and a positive prick test indicates that the mech ani sm responsibl e for an ele vated risk of hand derm atiti s is mo re closel y related to mucosal sy mp to ms than to an atop ic co nsti tutio n, as determined by prick tests.
Th e lack of an association with childhood eczema in the present study is inco nsis te nt with result s fro m othe r studies suggesti ng that childhoo d eczema is the mo st imp ort ant risk factor fo r hand dermatiti s (3,4) . It is possibl e that our study lacked the power to detect a pot enti al association, bec ause the study populatio n contained only a small nu mber of person s with childhoo d eczema . Furthermo re, fo ur indivi dua ls with atopi c derm atitis who showed hand derm atitis at the start of the study were excl uded from the ana lysis. They were exami ned at ea ch follow -up measurement , and hand de rmat itis co ntinued to ex ist thr ou ghout the study .
Th e risk of hand dermatitis in nick el-sen siti ve apprent ices was not signifi cantly increa sed . This finding is inconclusive, bec au se it is not clear whether e xpo sure to nickel or oth er se ns itizing agents had occ ur red in the study populati on . Th e absence of a re latio n between the risk of hand derm atitis and age or gender was in agreement with res ults from other studies (13, 3 1).
Fin ally, the pres ent study does not suppo rt the hypothes is that base-l ine transep idermal water loss is an indicator for the ris k of hand derm atiti s. The result s suggested that hai rdresser s may have an increas ed risk of hand derm atitis at tran sepidermal water loss abov e 15 g . m 2 • h in the hand , but the inc reased risk wa s not statistically significa nt. No ev ide nce of a rela tio nship was obse rve d for the nurses. At present , no other epidemio logic studies are ava ilable to refute or co nfirm the lack of an associatio n between tran sepid ermal wa ter loss and the risk of hand derm atiti s. It has been suggested that transe pidermal water loss merel y refl ect s the barrier functio n of the skin for substances with physical and che mica l prop erties si milar to those of wate r (32), whereas occ upa tio na l exposur e of nurses and ha irdressers may invo lve oth er typ es of che micals as well. Th e lack of a clear relationship in th is study wo uld see m to co nfirm that sugges tio n. On the other hand , a tru e association may have go ne und etected du e to the sma ll sample size and relatively large interindividual variation in tran sepid ermal water loss. A preliminary study indi cated that the intraindi vidu al variability of transepiderm al water loss over a period of three wee ks was 15.1 %, a fi ndi ng suggesting that transepi derrnal water loss is relatively stab le ove r that length of time (33). Nevertheless, it is likely that the fluctuatio ns are great er over longer peri od s (mon ths, seaso ns, years ) and grea t fluct uat io ns in lo ng-t erm transepidermal water loss wo uld limit the practical valu e of this variable as a predi ctor fo r the risk of hand der matitis.
In summary, the incidence of hand dermatitis was increased amo ng the apprentices studied, es pec ially amon g hairdresser s with dry skin. App ren tices with a hist ory of (atopic) mucosal symptoms, particul arly nurses, also had an incre ased risk of hand derm atitis. Th is rel ati on sh ip has been observed in ot her stud ies , but a biolog ica l mech ani sm that co uld explain this finding is not full y understood . Th e res ults of the st udy do not support the hyp oth esis that an eleva ted level of tran sepiderm al water loss increases the risk of hand derm atiti s.