[Clinical relevance of N-acetyltransferase type 2 (NAT2) genetic polymorphism]. / Pertinence clinique du polymorphisme génétique de la N-acétyltransférase de type 2 (NAT2).

Polymorphic N-acetyltransferase (NAT2) is involved in the metabolism of several compounds relevant in pharmacology or toxicology, with diverse clinical consequences. Inter-ethnic variations in distribution of the acetylation phenotype are significant. The caffeine test is most often used to assess the acetylation phenotype and to identify rapid and slow acetylators. The NAT2 phenotype could account for the increased risk of certain side effects in slow acetylators treated with isoniazid (particularly peripheral neuropathies and lupus erythematosus), although therapeutic efficacy seems to be independent of the acetylation status. Hypersensibility reactions with sulfonamides (including Lyell and Stevens-Johnson syndromes) are more frequent in slow acetylators, who also show poor tolerance to sulfasalazine and dapsone. In contrast, myelotoxicity induced by amonafide is more frequent in rapid acetylators, probably because of increased production of a toxic metabolite of the drug. In carcinogenesis, NAT2 may play a protective role against bladder cancer, although studies have shown contradictory results. Slow acetylators may have a risk of developing primitive liver cancer. For lung cancer, data are not conclusive, but slow acetylation status may predispose to mesothelioma in subjects exposed to asbestos. No relation has been found between acetylation phenotype and breast cancer. Contradictory results were reported on its role in colorectal cancer. Non-smoking type 1 diabetics may be at increased risk of nephropathy if they are rapid acetylators. Parkinson's disease may be more frequent among slow acetylators, but again, data have shown contradictory results. Finally, a poor acetylator phenotype may predispose to atopic diseases.

[The acetylator polymorphism in a Khmer population: clinical consequences]. / Polymorphisme d'acétylation dans une population khmère: conséquences cliniques.

The great variability of slow acetylator (SA) and/or rapid acetylator (RA) frequency is mainly due to ethnic-racial origin. Using the urinary elimination ratio of three metabolites of caffeine--acetylamino formylamino methyluracil (AFMU) to AFMU + 1-methyl urate (1U) + 1-methyl xanthine (1X)--we settled the acetylation phenotype in 54 independent subjects of Khmer and 70 independent subjects of Caucasian origin. Using DNA from peripheral leucocytes, we determined by PCR, in 32 Khmer and 122 Caucasian subjects, the frequencies of wild-type alleles (NAT-2 *4) and of mutated alleles (NAT-2 *5A, *6A, *7A). The frequency of SA was respectively 28 per cent and 61 per cent in Khmer and Caucasian subjects. The antimode of the distribution of the ratio was different in the two populations: 0.07 in Khmers and 0.18 in Caucasians showing a reduced acetylation capacity in the Khmer population in spite of a higher frequency of RA. The frequencies of alleles were also different between the two populations. Between Khmers and Caucasians respectively: *4: 48.4-23.8 per cent *5A: 15.6-44.2 per cent. *6A: 29.7-32.0 per cent. *7A: 6.3-0 per cent. These differences might be taken into account to define a therapeutic strategy in the treatment of tuberculosis by isoniazide.

Caffeine metabolism before and after liver transplantation.

AIM: To study drug metabolism in patients before and after liver transplantation using caffeine as a probe drug. Forty-five patients undergoing liver transplantation for various liver diseases and who had well documented dossiers were selected for the study. Before the liver transplantation and 1 month, 1 year, and 6 years after liver transplantation, they were given 200 mg of caffeine by the oral route in the morning after voiding their bladder. Twenty-four-hour urine samples were collected and caffeine and metabolites were determined by HPLC: 1-methylurate (1U), 1-methylxanthine (1X), 1.7-dimethylurate (17U), 1.7-dimethylxanthine (17X), 7-methylxanthine (7X), 3-methylxanthine (3X), 1.3-dimethylurate (13U), 3.7-dimethylxanthine (37X), 1.3-dimethylxanthine (13X), 1.3.7-trimethylxanthine = caffeine (137X). Indices of enzyme activities were calculated from the following urinary elimination ratios: (AFMU+1U+1X)/17U for CYP1A2, 17U/17X for CYP2A6, 1U/1X for xanthine oxidase (XO), AFMU/(AFMU+1U+1X) for N-acetyltransferase (NAT-2). RESULTS: Compared with results obtained in a group of 70 healthy subjects, caffeine metabolism before liver transplantation was deeply depressed with a decreased elimination rate in the case of all metabolites and a decreased CYP1A2 activity. Caffeine metabolism began to return to the control values one month after transplantation. One year and 6 years after liver transplantation, quantitatively, the metabolism of caffeine was stable and not different from control, but with qualitative modifications. CYP1A2 activity was decreased with reduced urinary elimination rates of 1X and 17X. XO and CYP2A6 activities and 1U and 17U urinary elimination rates were increased. Immunosuppressive treatment was possibly responsible for the metabolic pathway changes. Almost the same modifications were observed in 9 patients after bone marrow transplantation who had been treated with the same immunosuppressive drugs, cyclosporine and azathioprine. During severe rejection phases in 6 of the liver transplant patients, caffeine metabolism was progressively decreased when the usual liver function tests showed moderate but uniform changes. CONCLUSION: Despite an apparent normal drug-metabolic function, immunosuppressive treatment induces stable variations in drugmetabolic pathways after liver transplantation which can be detected from the changes in caffeine metabolism.

Caffeine metabolism in a group of 67 patients with primary biliary cirrhosis.

OBJECTIVE: To evaluate the polygenic regulated caffeine metabolism in a group of 67 patients with a documented primary biliary cirrhosis (PBC) classified according to the histologic stage proposed by Scheuer. METHODS: Over a 14-year period, drug liver metabolism, using caffeine as a probe drug, has been systematically carried out in addition to the usual clinical, histological and biochemical investigations performed in patients with PBC. The "Caffeine test" consisted of a 200 mg caffeine oral intake. Urines were collected over 24 hours: caffeine (137X), 1-7-dimethylxanthine (17X), 1-3-dimethylxanthine (13X), 1-3-dimethylurate (13U), 3-7-dimethylxanthine (37X), 1-7-dimethylurate (17U), 1-methylxanthine (1X), 1-methylurate (1U), 7-methylxanthine (7X), 3-methylxanthine (3X), and 5-acetylamino-6-formylamino-3-methyluracyl (AFMU) were analyzed by high performance liquid chromatography (HPLC). Total and individual metabolite urinary elimination rates were expressed in micromol/24 hours. Enzyme activities were evaluated from the following urinary metabolite ratios: (AFMU+1U+1X)/17U for CYP1A2, 17U/17X for CYP2A6, AFMU/(AFMU+U+ 1X) for NAT-2, 1U/1X for XO. RESULTS: Compared to healthy subjects, patients with PBC presented a reduced metabolism of caffeine due to a decreased CYP1A2 activity, all the more important since the patients had an advanced histological stage. This picture was nearly identical to the observed picture in chronic liver diseases from various origins. PBC affected the various metabolic pathways of caffeine in a differential manner. CYP1A2 activity was decreased but XO and mainly CYP2A6 activities were increased as shown by the raised urinary ratio 17U/total metabolite elimination. In contrast to the described loss of bimodality of the NAT-2 index distribution in patients with alcoholic cirrhosis, we found a clear-cut, bimodal distribution in patients with PBC, without a high incidence of slow acetylator status. CONCLUSION: Metabolism of caffeine is strongly and differentially disturbed in patients with PBC and apparently not exactly in the same way as that in alcoholic cirrhosis which is more often taken as an index of chronic liver disease. This suggests the need for caution with medicines whose metabolism is under polygenic regulation. Because of the relationships between caffeine metabolism modifications and histological stages, the caffeine test might be used along with the usual tests to safely follow-up the evolution of the disease.

Relationship between the severity of alcoholic liver cirrhosis and the metabolism of caffeine in 226 patients.

OBJECTIVES: To evaluate the polygenic regulated caffeine metabolism in a group of 226 patients with liver alcoholic cirrhosis classified according to the Child score. METHODS: Over a 14-year period an hepatic function test, using caffeine as probe drug, has been systematically associated to the usual clinical and biochemical investigations performed in patients with liver alcoholic cirrhosis. "Caffeine test" consisted in a 200 mg caffeine oral intake. Urines were collected over 24 hours: caffeine (137X), 1-7 dimethylxanthine (17X), 1-3 dimethylxanthine (13X), 1-3 dimethylurate (13U), 3-7 dimethylxanthine (37X), 1-7 dimethylurate (17U), 1-methylxanthine (1X), 1-methylurate (1U), 7-methylxanthine (7X), 3-methylxanthine (3X), and 5-acetylamino-6-formylamino-3-methyluracyl (AFMU) were analyzed by high performance liquid chromatography (HPLC). Total and individual metabolite urinary elimination rates were expressed in micromol/24 hours. Enzyme activities were evaluated from the following urinary metabolites ratios: (AFMU+1U+1X)/17U for CYPIA2, 17U/17X for CYP2A6, AFMU/(AFMU+ 1U+1X) for NAT-2, 1U/1X for XO. RESULTS: Compared to healthy subjects, whatever the Child score, caffeine metabolism was reduced by half in patients with alcoholic cirrhosis. The main cause was the decreased CYP1A2 activity. On the other hand, XO and CYP2A6 activities were increased and NAT-2 activity remained unchanged in slow acetylators (SA) and decreased in rapid acetylators (RA) Child B and C. Bimodality of NAT-2 distribution was unclear, but a right assignment of RA and SA phenotype in cirrhotic patients, confirmed by comparison with genotype, was obtained, using the antimode value of NAT-2 distribution used in healthy subjects. At last, there was an interindividual variability in caffeine metabolism as great as in the usual laboratory parameters. CONCLUSION: Metabolism of caffeine is decreased in patients with alcoholic liver cirrhosis. This decrease paralleled the modifications of the usual laboratory tests and does not bring additional information on the severity of the disease. But the equilibrium between the various metabolic pathways of caffeine is impaired. Beyond the changes of a specific enzymatic activity, this must be taken into account particularly for drugs whose metabolism is of the polygenic regulation type.

Similarities and/or dissimilarities of CYP2D6 polymorphism in three Tunisian ethnic groups: Arabs, Berbers, Numides.

Using the validated probe drug debrisoquine and the 8 h urinary metabolic ratio debrisoquine/4 hydroxy-debrisoquine, we have determined the phenotype of the debrisoquine CYP2D6 dependent polymorphic metabolism in 464 Arabs, 227 Berbers and 215 Numides to elicit similarities or dissimilarities of poor metabolizer (PM) prevalence. We found 2.36 per cent of PM in Arabs, 3.08 per cent in Berbers and 2.33 per cent in Numides. These figures are similar to those observed in Middle East populations, and cannot be considered as different from those observed in Caucasians.

Analysis of N-acetyltransferase (NAT2) in three ethnic groups in Tunisia.

Acetylation status was compared, using caffeine as a probe drug, in the three main racial/ethnic groups living in Tunisia: Arabs, Berbers and Numides. The frequency of slow acetylators appears identical in these three groups and is different from that observed in Caucasians. However, the NAT-2 activity as a whole is lower in Tunisians than in Caucasians. These differences might be attributable to the various population mixings which occurred in the past, given the geographical position of Tunisia. It may be asked whether these differences are relevant in term of efficiency and/or frequency of adverse drug reactions when medicines whose metabolism is NAT-2 dependent are used. This hypothesis deserves to be tested.

On the assessment of drug metabolism by assays of codeine and its main metabolites.

Codeine and its main metabolites appear to have advantages for assessing drug metabolic phenotypes. The authors have further developed a high-performance liquid chromatography (HPLC) method for the quantification of codeine and six of its metabolites in urine. Quantification was performed by electrochemical detection for morphine, normorphine, morphine-6-glucuronide, and the internal standard 4-O-methyldopamine; and by ultraviolet detection for codeine, norcodeine, and morphine-3-glucuronide. The method had a detection limit of 2 nmol/L(-1) for morphine and normorphine, 4 nmol/L(-1) for morphine-6-glucuronide, 3 nmol/L for the internal standard, 20 nmol/L(-1) for morphine-3-glucuronide, and 60 nmol/L(-1) for codeine and norcodeine. The coefficients of variations were <9% for intraday and <10% for interday analyses. The recovery of codeine and its metabolites ranged from 55% (for morphine-3-glucuronide) to 90% (for codeine, norcodeine, morphine, and morphine-6-glucuronide). Eleven healthy volunteers were phenotyped for CYP2D6 using codeine as well as debrisoquine and dextromethorphan. Ten subjects were extensive metabolizers (EM) and one a poor metabolizer (PM) of codeine, debrisoquine, and dextromethorphan. Significant correlations between the metabolic ratios (MRs) of the different probe drugs were obtained (r2 > 0.95, p < 0.001). This HPLC method is simple, sensitive, accurate, and reproducible for assessing the CYP2D6 phenotype.

Caffeine metabolism differences in acute hepatitis of viral and drug origin.

The 24-h urinary excretion rate of caffeine metabolites following 200 mg caffeine intake has been proved to be a valuable safe quantitative test of liver function. The pathological mechanism of acute hepatitis of viral and drug origin is different. In both diseases, the patient's caffeine metabolic capacity during the acute and the recovery period was compared. In the acute period, in both diseases, the strongly reduced metabolism of caffeine paralleled the variations of the usual biochemical tests. During the recovery period, in viral hepatitis, caffeine metabolism and biochemical tests returned to the normal values. In drug-induced hepatitis during the recovery period, caffeine metabolism remained severely impaired at a time when biochemical tests were back to the control levels. This discrepancy might be due to the histological or molecular toxic effects of the drug(s), irrespective of cytolysis. After drug-induced hepatitis, a caffeine test might be used to check the total recovery or to choose an adapted dosage of medicines.

Acetylation polymorphism expression in patients before and after liver transplantation: influence of host/graft genotypes.

The consequences of liver transplantation on NAT2 activity were studied in 58 patients of Caucasian origin and compared with a group control of 119 unrelated healthy individuals of the same ethnic origin. Acetylation phenotypes were determined using caffeine as a probe drug before and repeatedly after liver transplantation. NAT2 genotypes were determined with three separate polymerase chain reactions to detect either the NAT2*4 wild-type allele or the NAT2*5A, NAT2*6A and NAT2*7A mutated alleles, associated with a decrease in NAT2 enzyme activity. In patients, the molar urinary elimination ratio AFMU/(AFMU+1X+1U) appeared more reliable than AFMU/1X for assessing the acetylation phenotype and fitted better with the various haplotypes. The variation of xanthine oxidase activity as measured by the 1U/1X urinary elimination ratio, appeared to be responsible for the poor phenotype prediction from the AFMU/1X ratio in post-transplanted patients. Regardless of the pathologic conditions of the treatment in progress, the genotype of the liver played an overwhelming role in the phenotypic expression of NAT2 compared with the genotype of other organs, where NAT2 was expressed in patients who presented a chimerism after liver transplantation.

Influence of donor and recipient genotypes on CYP2D6 phenotype after liver transplantation: a study of mutations CYP2D6*3 and CYP2D6*4.

OBJECTIVE: After liver transplantation (LT), genotypic differences between the recipient and the transplanted liver, medications and post-LT complications may all affect drug metabolism. We have studied the effect of two CYP2D6 mutations in the donor and the recipient on post-LT CYP2D6 phenotype. METHOD: The CYP2D6 phenotype was assessed in 48 patients before and after LT with debrisoquine or dextromethorphan. CYP2D6*3 (CYP2D6A) and CYP2D6*4 (CYP2D6B) mutations were detected in the donor and the recipient using polymerase chain reaction. RESULTS: Before LT, 40 subjects were classified as extensive metabolisers (EM) and 8 as poor metabolisers (PM); after transplantation, 41 were EMs and 7 were PMs. Genotype and phenotype were in agreement in 100% of EMs and 40% of PMs. The low percentage of agreement in PMs could not be explained by severely altered liver function. The phenotype of 13 subjects was apparently changed by LT: 6 EMs became PMs and 7 PMs became EMs. All four subjects in whom genotype changed following LT had a corresponding change in phenotype: two EM subjects became PMs and two PM subjects became EMs. CONCLUSION: The low percentage of agreement in PMs may be partly explained by mutations other than CYP2D6*3 and CYP2D6*4. Nevertheless, our study shows that the CYP2D6 genotype of the donor controls the phenotype of the recipient of a liver transplantation.

Effects of liver diseases on drug metabolism.

Although the liver is the main site of drug metabolism, conflicting results have been reported on drug elimination during liver diseases. Drug metabolism may depend on histological changes in the liver (acute or chronic hepatitis, cirrhosis) but may also depend on their origin (viral, toxic or immunological). Drug metabolism is also influenced by the severity of liver dysfunction. Cytochrome P450 isozymes and conjugation pathways may be differently affected by these conditions, and specific probe drugs have to be used in order to study the effect of diseases on each enzyme of drug metabolism. Probe-based assays must be validated during disease, since the pharmacokinetics of the parent drug and/or of its metabolites may be altered. Because of these limitations, therapeutic drug monitoring may be the most reliable way to adjust drug dosing at present.

Inhibition of caffeine metabolism by 5-methoxypsoralen in patients with psoriasis.

Eight patients with psoriasis were given 200 mg caffeine orally with or without 1.2 mg kg-1 of 5-methoxypsoralen. Blood and urine samples were collected over a 2-day period. During 5-methoxypsoralen coadministration, the apparent volume of distribution of caffeine remained unchanged, but oral clearance (CLp.o.) decreased from 9.5 +/- 3.8 (mean +/- s.d.) to 3.2 +/- 0.51 h-1 (P < 0.01). The area under the plasma concentration-time curve (AUC) increased from 24 +/- 9 to 73 +/- 29 mg 1(-1) h (P < 0.001). This decrease in CLp.o. with increased AUC was consistent with a CYP1A2-dependent inhibition of caffeine N-demethylation which was further supported by significant decreases in the (AFMU+1U+1X)/17U and (AFMU+1U+1X)/17X urinary metabolic ratios.

[Hematotoxicity caused by azathioprine genetically determined and aggravated by xanthine oxidase deficiency in a patient following renal transplantation]. / Hématotoxicité de l'azathioprine à déterminisme génétique aggravée par un déficit en xanthine oxydase chez une transplantée rénale.

Azathioprine is an immunosuppressor used with ciclosporin and corticosteroids after organ transplantation. Azathioprine is rapidly transformed into 6-mercaptopurine which in turn is metabolized by three competitive pathways: a) intracellular hypoxanthine guanine phosphoribosyl transferase leads to 6-thioguanine nucleotides which can damage chromosome DNA; b) thiopurine methyltransferase produces inactive methylated derivatives; c) xanthine oxidase produces thiouric acid. Due to inter-individual variations in the later two pathways, azathioprine dose must be adapted to each patient. A 48-year-old female patient underwent renal transplantation in 1994 and was given immunosuppressive therapy combining thymoglobulins, azathioprine and ciclosporin. Severe leukopenia (< 3000/mm3) occurred on day 5 requiring withdrawal of azathioprine. Known hypouricaemia (< 50 mumol/l) suggested xanthine oxidase deficiency. Laboratory results confirmed xanthine oxidase deficiency and also revealed reduced thiopurine methyltransferase activity (14.9 pmol/h/mg Hb). Azathioprine toxicity was confirmed by regression of the leukopenia after withdrawal and recurrence at rechallenge. Xanthine oxidase deficiency occurs in 2% of the general population. Reduced thiopurine methyltransferase activity affects 11% of the population. The combined presence of these two genetic anomalies led to early and sudden intolerance to azathioprine and emphasize the need to develop new immunosuppressor agents degraded by other metabolic pathways.

Stability of debrisoquine (CYP2D6) phenotype in liver transplant patients.

Liver metabolism may be modified after liver transplantation according to the phenotype of the donor and may be influenced by posttransplantation complications. The CYP2D6 phenotype was assessed in 13 patients (group I) before and after liver transplantation using debrisoquine. CYP2D6 activity was also assessed in vitro on microsomes from the liver of the recipients and the donors, using dextromethorphan. Twelve patients were extensive metabolizers both before and after transplantation. One apparently poor metabolizer was transplanted with the liver of another poor metabolizer. The intrinsic clearance of dextromethorphan (CL(int)) measured on recipient liver microsomes was significantly lower than that on donor liver microsomes (p < 0.05). In extensive metabolizers, the debrisoquine metabolic ratio was correlated with CL(int) before (r = 0.78, p < 0.05) and after (r = 0.89, p < 0.0005) transplantation. Debrisoquine phenotype was measured repeatedly in nine additional patients (group II) up to 3 years after liver transplantation. Their phenotype was stable during the follow-up observation, although the variations observed may be clinically relevant. Therefore, no change in CYP2D6 phenotype (extensive/poor metabolizer) was observed because of the liver transplantation, and the debrisoquine log metabolic ratio was largely unaffected by the liver complications observed during the posttransplantation follow-up observation.

A population and family study of CYP1A2 using caffeine urinary metabolites.

CYP1A2 is a cytochrome P450 which is inducible by polycyclic aromatic hydrocarbons. This induction could be mediated via the Ah locus, which encodes a cytosolic receptor responsible for the regulation of the CYP1A1 gene. Enzyme activity in vivo can be measured by the urinary caffeine metabolite ratio (AFMU + 1X + 1U)/17U. Our goal was to determine, using this ratio, the possible existence of a genetic polymorphism in CYP1A2 induction. For this purpose, a population and family study, including smokers, were undertaken. In a first step, we investigated factors influencing enzyme activity in a population of 245 unrelated individuals. The induction effect of smoking and inhibiting effect of oral contraceptive use were confirmed. None of the other factors examined (age, sex, level of cigarette consumption, nicotine or tar amounts, filter, inhalation) accounted for the interindividual variability in the metabolic ratio. Using the statistical SKUMIX method, a unimodal (one peak) distribution of the ratio was concluded in 164 unrelated smokers, since a second distribution did not significantly improve the fit to the data (chi 2(1) = 1.39, P > 0.2). Segregation analysis was performed on 68 nuclear families and no major gene effect could be shown. Furthermore, the polygenic model did not provide a higher likelihood than the sporadic one, which argues against the existence of any familial resemblance. Although we cannot rule out the possibility that some environmental factors could obscure the phenotypes and occult a genetic determinism, we conclude that genetic factors are probably negligible in the determination of CYP1A2 activity measured by this method.(ABSTRACT TRUNCATED AT 250 WORDS)

Surface recovery and stripping methods to quantify percutaneous absorption of caffeine in humans.

The percutaneous absorption of caffeine from two vehicles, an emulsion and an acetone solution, was quantified by in vivo techniques in humans. A surface recovery technique over a 6-h application and a stripping method after a 30-min application were performed on the volar aspect of the forearm on 12 volunteers. Caffeine was assessed by HPLC. Two phases were distinguished in the percutaneous absorption of caffeine: a higher filling up of the stratum corneum with the oil-in-water emulsion than with the acetone solution, which was then followed by a steady-state flux corresponding to the penetration in the living tissues. The permeability constants (Kp) with emulsion and acetone were 1.59 x 10(-4) and 9.53 x 10(-8) cm/h, respectively. The stripping method showed concentrations of caffeine in stratum corneum that were five times higher with emulsion (212 ng/cm2) than with acetone (37 ng/cm2). With acetone as a vehicle, approximately 40% of caffeine of the cornfield layer was found around the treated area. This sizeable lateral spread within the stratum corneum was not observed with the emulsion.

A population and family study of N-acetyltransferase using caffeine urinary metabolites.

Population and family studies were undertaken to validate caffeine as a probe drug to establish the genetic status of rapid acetylators and slow acetylators. The acetylator status was established from the urinary metabolic ratio of 5-acetylamino-6-formylamino-3-methyluracil to 1-methylxanthine (AFMU/1X) after oral administration of caffeine. We confirmed a bimodal distribution (chi 2(1) = 229.48; p << 10(-9)) of the AFMU/1X ratio in 245 unrelated subjects. A third distribution did not significantly improve the fit to the data (chi 2(1) = 0.04; p = 0.84). Complex segregation analysis of 76 nuclear families confirmed the monogenic inheritance of N-acetyltransferase, with incomplete dominance of the rapid allele over the slow one. We observed a slight shift between the mean activities of heterozygous and homozygous rapid acetylators (t = 2.89; p < 0.01). However, the 30 obligate heterozygotes belonging to the 76 families were evenly distributed among the rapid acetylators and never located in a hypothetic intermediary group between slow acetylators and rapid acetylators.

Effect of pressure on in vitro percutaneous absorption of caffeine.

The effect of increased pressure, which is a mechanical property of massage, was investigated on the percutaneous absorption of an amphiphilic compound (caffeine) in vitro on Franz diffusion cells, using excised human skin. 50 microliters of either a 320 micrograms/ml or a 15 mg/ml acetone solution of caffeine were pipetted onto the surface of each skin sample, which represented caffeine skin deposits of 5 micrograms/cm2 and 240 micrograms/cm2 respectively. During each experiment, a pressure device delivering 0.25 bar over the atmospheric pressure was applied for the first 30 min on half of the cells. At 2, 4, 6, 8, 12 and 24 h the aqueous dermal bathing solution, containing 14 g/l albumin, was removed and chromatographed. With the applied dose of 5 micrograms/cm2 no statistical difference was found between the cumulated absorbed amount under atmospheric pressure and increased pressure. On the other hand, with the applied dose of 240 micrograms/cm2, the permeation of caffeine was 1.8 times higher under increased pressure than the permeation under atmospheric pressure (p < 0.05). This enhancing effect of increased pressure was probably connected to either an improved transappendageal route during the percutaneous absorption process or a higher stratum corneum filling-up.