Percutaneous absorption of 2-butoxyethanol in man.

B. Percutaneous absorptionof 2-butoxyethanol in man. Scand J Work Environ Health 14(1988) 101-109. The percutaneousabsorption of the commonlyused glycol ether 2-butoxyethanol (ethylene glycol monobutylether)wasinvestigated in 12exposureexperiments with fivemen.The subjectkepttwo or four fingers immersed in neatbutoxyethanolfor 2h. Arterializedcapil lary blood sampleswerecollectedfrom the other hand before, during, and up to 4 h after the exposure and analyzedfor butoxyethanolby gas chromatography. Urine wascollectedfor 24 h and analyzedfor the metabolitebutoxyacetic acid, also by gas chromatography.The presenceof butoxyethanolin blood and of butoxyacetic acid in urine confirmed that butoxyethanolenters the systemiccirculation in man in vivoduring dermalexposure. Percutaneous uptake rates werecalculatedfrom measured blood levels of butoxyethanolwith the use of kinetic parameters (clearance and volumeof distribution) obtained in earlier experiments with the same subjects. The uptake rates ranged from 7 to 96 nmol . min-I. cm- 2• Theresults indicate that persons exposing largeportionsof theirskinto butoxyethanol areat riskof absorbing acutely toxic doses.

The glycol ether 2-butoxyethanol (ethylene glycol rnonobutyl ether) is frequently used in cleansing products, paints, and lacquers (13). The compound causes hemolysis and reduced hemoglobin concentration and red blood cell count in laboratory animals (IS,21). In contrast to the lower homologues 2-methoxyethanol and 2-ethoxyethanol, butoxyethanol does not appear to be teratogenic (15,20). The low vapor pressure of butoxyethanol, combined with the high preference for water -a water/air partition coefficient of about 8 000 has been reported (9) -reduces the risk of attaining very high vapor concentrations in the atmosphere. On the other hand butoxyethanol may readily penetrate the skin, as shown with the guinea pig (10, I I) and the rat (I) in vivo and with human, pig, and rat skin in vitro (I, 5). This paper deals with the percutaneous uptake of butoxyethanol in man in vivo. The aims of the investigation were (i) to confirm qualitatively that percutaneous uptake occurs in man in vivo and (ii) to obtain a quantitative estimate of the percutaneous uptake rate in order that comparisons with other routes of exposure and exposure levels and other species could be made. The main tool was the analysis of butoxyethanol in blood. In addition, the urinary excretion of the acid metabolite butoxyacetic acid was studied. Butoxyacetic acid is a major metabolite in urine from men exposed to butoxyethanol (8,12).
Reprint requests to: Mr G Johanson, National Institute of Occupational Health, S-171 84 Solna, Sweden.

Subjects
The experiment was approved by the Ethical Committee at the Karolinska Institute, Stockholm. Five healthy men volunteered for the study. They had approximately one and a half years earlier participated in a study involving inhalation exposure to butoxyethanol (12). Thus toxicokinetic data, such as estimates of the respiratory uptake, apparent blood clearance, and apparent volume of distribution, were available for these individuals. None of them had experienced any occupational exposure to solvents. All were nonsmokers and stated low or no alcohol consumption. One of the volunteers (subject B) was allergic to chloropromazine and had atopic eczema on his chest, back, face, and the cubital fossa, and one (subject C) had neurodermatitis on his left leg. The subjects were told not to consume alcoholic beverages during the preceding 24 h and throughout the experiment and also to avoid activities that might cause exposure to organic solvents or induce irritation or damage to the skin.

Chemicals
Butoxyacetic acid and pentoxyacetic acid were synthesized in our laboratory (12). All other chemicals were of analytical grade and purchased from Merck (Darmstadt, Federal Republic of Germany), unless otherwise stated. Butoxyethanol was distilled prior to use.

Experimental procedure
Prior to the experiment the fingers were carefully examined for skin disorders (in one experiment, the ex- Table 1. Body weight , age, and estimates of the blood clearance (CL b ) and steady-state volume of distribution (Vss) for 2·butoxyethanol.

Blood analysis
Arterialized capillary blood samples were repeatedly collected from the fingers of the unexposed hand after the hand had been rinsed and warmed for 1 min in freshly prepared hot tap water . The blood samples (2 . 200 ul) were rapidly transferred to screw-cap vials and extracted into toluene, derivatized with pentafluorobenzoyl chloride (Pierce, Beijerland, Holland), and 'analyzed by gas chromatography with electron capture detection as described earlier (12). The method error was estimated to be II 010, expressed as the standard deviation of 43 duplicate sam ples of blood spiked to 1-10 umol/I with butoxyethanol. posure was switched to the right-hand fingers because of a small wound on one of the left-hand fingers). Blood and urine samples were collected, and finger diameter, skinfold thickness, and finger volume were measured as described below.
The subjects were exposed to neat butoxyethanol for 2 h by placing four fingers of the left hand through cut holes in a pol yethylene cap into a polyethylene jar filled with pure solvent. The temperature in the room and the solvent jar was approximately 21°C. The subject sat in a comfortable chair with a good support for the left arm and hand throughout the exposure. The solvent jar and the left arm were kept in a separately ventilated hood to eliminate the risk of additional dermal exposure or inhalation or contamination of blood samples. At the end of the exposure, each subject rinsed the exposed hand thoroughly under running tap water and then washed both hands with a mild soap . They were thereafter allowed to move freely in the building, except when finger measurements and inspections were performed and blood and urine were sampled.
At regular intervals, the skinfold thickness of the dorsal skin on the third phalanx of the exposed fingers was measured with a Harpenden caliper (British Indicators Ltd, St Albans, Great Britain). In parallel, the finger volume was assessed with a simple plethysmograph , a test tube filled to the edge with water. The finger volume was determined by weighing the tube before and after the dipping of each finger. In all experiments the corresponding unexposed fingers served as the control. a From Johanson et at (12). b Unpub lished observation .
Thi s approach has been discussed in mor e deta il in a previous paper (10) . Estimates of v' sand CL b were obtained in previously performed inhalation experiment s (12) and are listed in table I . C h was determined by gas chromatography, as has already been

Urine analysis
All urine was quantitatively collected for approximately 24 h, in the same manner as previousl y described (12), and stored at -20°C until analyzed . Butoxyacetic acid, the urinary metabolite of butoxyethanol, was determined as follows . Urine (200 ul) was added to a teflon-lined screw-cap vial containing 2 ml of tetrabutylammonium hydrogen sulfate (0.1 M) buffered to pH 6 with potassium phosphate (0.2 M), with 25 /.11 of an aqueous solution of pentoxyacetic acid (I mmol/I) as the internal standard, 2 ml of meth ylene chloride, and 10 /.ll of pentafluorobenzyl bromide (Pierce, Beijerland, Holland). The vial was capped, vigorously shaken, and then rotated for 1 h at room temperature. The methylene chloride phase was collected and evaporated under nitrogen at 40°C. The residues were dissolved in 2 ml of hexane, and aliquots of 1 /.11 were injected into the gas chromatograph (Varian 3400 equipped with a 63nickel electron capture detector, Varian 8000 autosampler and Vista 402 integrator) by the split/splitless technique. A fused silica capillary column [Oribond SE-30, 25 m . 0.32 mm (inner diameter), 0.25-/.lm phase layer] was used. The column was identical to the one used for the analysis of butoxyethanol. Nitrogen was used as the carrier (2.0 ml/min) and make up (30 ml/rnin) gas. The temperatures were: injector 220°C , column 130°C, detector 260°C. The concentration of butoxyacetic acid was calculated from the pentafluorobenzyl butoxyacetate: pentafluorobenzyl pentoxyacetate peak area ratio. The described method is a further development of previous ones (12,19) and comprises a reduced number of extractions, less use of reagent , close to neutral pH, and incr eased sensitivity. Urine samples spiked to 5-500 umol/I were used as the standard. Fresh standard curves were prepared for each experiment. The slope of the curve was obtained by linear regression on log-log values. The method error was estimated to be 14 %, expressed as the standard deviation of 60 dupli cate urine samples.

Calculations
The cumulated systemic uptake of a substance (aUPl ake) may be expressed as a function of time . Thu s, on the assumption of linear kinetics, the cumulated uptake at a given time may be calculated during steady state once the concentratio n in blood (C b ) , the steady state volume of distribution ( V,s), the blood clearance (CL h ) , and the "area under the concentration time curve" (A UC(o-l) for the sub stance are known. Accordingly, the uptake of butoxyethanol ma y be calculated as: enee appeared not to be an effect of low temperature only, as a similar but less prominent response was obtained when the subject was exposed to butoxyethanol heated to 30°C in a third experiment.
As seen in figure 3, the solvent could be detected in the blood of all the subjects when their fingers were exposed to liquid butoxyethanol. This is direct evidence of systemic uptake of butoxyethanol via the skin in man in vivo. The shape of the concentration profile varied considerably between individuals and also between experiments for some of the individuals. Esti- Figure 1. Relative finger volume (exposed/control) as measured by plethysmography after exposure to neat z-butoxvethanol. The squares represent the average of 11 experiments with five subjects. The vertical bars represent the 95 % confidence interval (. p <O.01 in Student's t-test versus the initial value).
described. Obviously, if steady state is not reached (such as when changes in C b are rapid) or if the value of v's is incorrect, the first term in the equation will contribute inadequately to the calculated uptake. Such contribution is iIlustrated by peaks in some of the "invasion curves" in figure 4 in the Results section . However, the contribution of the first term, and thus of this error, will gradually decrease with time and eventually become zero. The A UCto._ l ) was calculated by the trapezoidal method up to 360 min. To obtain the A UC(O_OO), a residual term was added on the assumption of monoexponential decay (17). The slope constant used in the residual term was an average of all 12 experiments, with values from 180 to 360 min. The systemic uptake of butoxyethanol was obtained as the product of A UC(f).-OO) and CL b • The uptake rate was calculated as the uptake divided by the exposure time and the exposed area.
The excretion rate and cumulated excretion of butoxyacetic acid in urine were calculated from the product of the measured concentration of acid and the urine volume of each sampling period.
The half-times of butoxyethanol and butoxyacetic acid were obtained by linear regression on the semilog values from the decay phases.
The exposed area was calculated with the use of a cylindrical shape for the fingers. The average of the measured diameters of the first and third phalanges was taken as the diameter of the cylinder.
Skinfold thickness and finger volume were expressed as a quotient of the average value of the exposed fingers to that of the control fingers. The effect of exposure to butoxyethanol was graphically visualized from the quotient, expressed as a percentage of the initial (preexposure) quotient, at various times after the end of the exposure. The statistical significance of deviations from 100 % were tested with the two-tailed Student's t-test .

Results
None of the subject s complained about or showed any signs of discomfort during or after the exposure to butoxyethanol. The skin of the exposed fingers was not irritated but appeared somewhat more rigid and less elastic after exposure and had a wrinkled appearance. This effect continued to develop, reached a maximum about 2-4 h after the end of the exposure, and then gradually disappeared. In parallel to the visible changes , the volume and skin fold thickness of the exposed fingers decreased and then returned to normal, as seen in figures 1 and 2. A dry, reticulate pattern with small fissures developed within a few hours after the exposure. In some cases the fissures became slightly erythematous. This reaction disappeared within 1 or 2 d. One subject (C) developed white fingers during the exposure in the second experiment. This occur-    .c 6 5 Exp no 1 BE uptake (umol)~~L  mates on the skin uptake of butoxyethanol for each experiment are given in figure 4 and table 2. The percutaneous uptake varied more than tenfold, from 127 to 1 891 umol, corresponding to 7-96 nmol . min-I .
cm", with a geometric mean of 20 nrnol . min-I.
cm-2 • The half-time of butoxyethanol during the decay phase ranged from 0.6 to 4.8 (geometric mean 1.3) h. The acid metabolite butoxyacetic acid was found in the urine after dermal exposure to butoxyethanol. As seen in figure 5, the excretion rate increased during the first hours after the exposure, reached a maximum at about 5 h (3 h postexposure), and then declined with an average half-time of 3.1 h (geometric mean). The calculated total uptake of butoxyethanol has been compared with the cumulated 24-h urinary excretion of a Two fingers exposed. b Solvent heated to 3D'C . c Fingers of right hand exposed .
Uptake BE upta ke rate BAA of BE (nrnol : min -" excreted (limol) cm --') (urnol) butoxyacetic acid in figure 6. Linear regression analysis for all the experiments, as well as the geometric mean, suggest that, on the average, 17 0J0 of the absorbed dose of butoxyethanol was excreted as butoxyacetic acid in 24 h. However, there was a large variation between experiments. The cumulated excretion of butoxyacetic acid ranged from 8.7 to 313 umol (table  2), corresponding to 2.5-39 070 of the uptake of butoxyethanol.

Discussion
A possible source of error in the presently used method to calculate the percutaneous uptake of butoxyethanol is the clearance (CL b ) values . Their correctness depends on the accuracy of the estimated respiratory uptake and the A UC in the inhalation experiments performed more than one year earlier. Metabolic clearance may change due to enzyme induction, liver changes, etc. However, there was nothing to indicate the occurrence of such a change, as the eating, smoking and drinking habits, body weight , health status, and occupation of the volunteers had remained unchanged.
The calculations presented in this study are valid only when linear kinetics may be applied. In experiments performed with perfused rat liver, the elimination kinetics of butoxyethanol were concentration dependent, the estimated values on the apparent Michaelis constant ranging from 0.32 to 0.70 mmol/I (14). The blood concentrations of butoxyethanol were one or two orders of magnitude lower in the present study, and linear kinetics were thus indicated under the present conditions, when the results were extrapolated from rat to man. ,. The time course of the concentration of butoxyethanol in blood during and after dermal exposure varied widely between the experiments, both regarding the shape of th e curve and the levels reached. The reasons for the large fluctuations during the exposure are not known to us , but the y appear to be related to changes in the uptake rate of the solvent. Diffusion through the stratum corneum is generally considered to be the rate-limiting factor in skin absorption (2), but other factors such as reduced peripheral blood flow may also be important. One subject developed "white finger s" when dipping his fingers in solvent at room temperature . This effect was less pronounced when the experiment was repeated with the solvent heated to 30°C. However, with the limited data available, no relationship between the uptake rate of butoxyethanol and the de velopment of white finger s could be determined .
Other explanations of the va riability , such as changes in distribution or metabolic rate, are not supported by the results of earlier inhalation experiments (12). Contamination of blood samples was unlikely becau se of the precautions undertaken in the blood sampling procedure, and exposure by inhalation was prevented . The strikingly high uptake of butoxyethanol in the second experiment with subject B is suppo rted by a similarly high 24-h excretion of butoxyacetic acid (table 2 and figure 6).
The slope of the "invasion cur ve" was used to calculate the rate of uptake into the systemic circulation in previous guinea pig experiments (10). Th is approach could not be used in the present study because of the variable uptake pattern. Instead, total uptake was divided by exposure time . In the guinea pig, uptake rates calculated from the total uptake were approximatel y twice as high as tho se calculated from the slope of the " invasio n curve" (unpublished ob servation). However, this difference is considerably smaller than the tenfold variation in skin uptake rates of butoxyeth anol obtain ed for both guinea pig and man.
The estimated uptake rate of butoxyethanol of 7-96 nmol . min-I. cm--2 in the pre sent study is comparable to that obtained in vitro with human skin by Dugard et al [28 (SD 99) nmol . min -I. cm-2 (5)] and by Bartnik et al [approxima tely 24 nmol · min -I . cm-2 (1), our estimate from their table 6]. The uptake rate of another polar solvent, namely, methyl n-butyl ketone (0.05 and 0.08 umol . min -I . cm-2 ) in two men (4) was of similar magnitude .
The skin uptake of butoxyethanol in man is also comparable to that obser ved in the guinea pig [52-462 nmol . min-I. cm-2 (10)] and the hairle ss rat [appro ximatel y 170 nmol· min-I. cm-2 (I), our estima te from the ir table 4], con sider ing the known species differences with regard to the percutaneous uptake of xenobiotics (24).
The skin of the fingers exposed to butoxyethanol developed visible changes, as indicated by a decreased finger volume and skinfold thickness and a wrinkled and less elastic appearance. Generally, if skin is ex-posed to an organic solvent, lipids are extracted , and permeability to water increases. The most effective in skin delipidization is chloroform or a 2: 1 mixture of chloroform and methanol (16). The properties of butoxyethanol resemble those of the chloroform: methanol mixture in its ability to extract both hydrophilic and hydrophobic substances. Skin delipidization and dehydration may account for the development of the rigid, wrinkled, dry, and fissured appearance of the skin, as well as for the observed decrease in finger volume and skinfold thickness.
In most experiments the concentration in blood continued to increase after the exposure to butoxyethanol had ceased. This increase may partly be due to a depot effect. An additional explanation may be that the barrier properties of the skin were impaired by rehydration of the stratum corneum when the hands were washed at the end of the exposure (23). A similar effect was seen in guinea pigs dermally exposed to neat butoxyethanol (II).
With the possible exceptions of four experiments (with subjects B, D, and E), no lag time in the absorption of butoxyethanol could be seen. The absence of a lag time contrasts the 21-60 min observed for the guinea pig (10) and the < 1 h reported for human skin in vitro (5).
The percutaneous uptake of butoxyethanol was comparable to the respiratory uptake observed in the same men . The percutaneous uptake rate ranged from approximately 1 to 16 umol/rnin when four fingers were exposed, while the respiratory uptake rate ranged from 8 to 14 (average 10) umol/rnin when they were exposed to butoxyethanol vapor at 20 ppm (the current Swedish occupational exposure limit) during light physical exercise at 50 W on a bicycle ergometer (12). Thus, expo sing four fingers to liquid butoxyethanol corresponds roughly to being exposed to butoxyethanol vapor at 20 ppm.
However , care should be exercised when dermal and respiratory uptake rates are compared, for several reasons. First, variability in the dermal uptake rate may exist both between and within individuals, as suggested by the limited number of experiments presented in this paper. Th e uptake rates obtained in this investigation may not be repres entative of a large population. Second, the skin uptake of butoxyethanol may be considerably higher in other regions of the human body (22). Third, the condition of the skin may heavily influence the uptake rate, as has been shown for several compounds (18,22). Fourth, butoxyethanol is mostly used together with water or other solvents (13). The presence of water ha s been shown to enhance the percutaneous uptake of butoxyethanol in guinea pigs (1, 11).
Feelings of discomfort and increased erythrocyte osmotic fragility have been reported after experimental exposure to 98 ppm , 113 ppm , and 195 ppm of butoxyethanol vapor (3). Exposure to 98-195 ppm cor responds to an absolute respiratory uptake rate of ap-proximately 23-46 umol/rnin, if the ventilation rate and the relative respiratory uptake are assumed to be 10 lImin and 60 010, respectively. It is very probable that such uptake rates may be obtained with skin exposure if the exposed area is larger than in the present study or if other regions of the body, more permeable to solvents, are exposed.
The pattern of excretion for the urinary excretion of butoxyacetic acid (time to reach maximum excretion rate and half-time of the decay) was comparable to that determined in previous inhalation experiments (12). However, the excretion rate, as well as the cumulated 24-h excretion of acid metabolite, varied considerably both within and between experiments and between individuals, as seen in figures 5 and 6. Consequently, analysis of the acid metabolite in urine appears to be an unreliable method for assessing the percutaneous uptake of butoxyethanol in man.
Glycine conjugates of the butoxyacetic acid homologues ethoxyacetic acid and isopropoxyacetic acid have been detected in urine from experimental animals exposed to ethoxyethanol and isopropoxyethanol, respectively (6,7). The presently used derivatization method was carried out at pH 6 to avoid the hydrolysis of conjugates, if present. The previous methods involved an extraction step at pH 2 followed by derivatization at pH II (12,19). Acid-catalyzed hydrolysis of butoxyacetic acid conjugates in the urine may therefore explain the higher percentage of acid metabolite found in the previous inhalation experiments [geometric mean 41 (range 15-55) %, as compared to 17 (range 2.5-39) %]. This possibility is presently being further investigated .
In conclusion, our study clearly shows that butoxyethanol is absorbed through human skin in vivo and enters the systemic circulation. From calculations of the dermal uptake rate, a comparison with inhalatory uptake could be made. The comparison suggested that both skin and respiratory uptake should be considered when chemical products containing butoxyethanol are handled. Exposure of large areas of the skin to butoxyethanol may cause acute toxic effects in man.