1Department of Pharmaceutics and 3Department of Pharmacology, The Royal Danish School of Pharmacy, Universitetsparken 2, DK-2100 Copenhagen, Denmark. 2Department of Anaesthesiology, Copenhagen University Hospital, Blegdamsvej 9, DK-2100 Copenhagen, Denmark
Accepted for publication:March 8, 2000
![]() |
Abstract |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Br J Anaesth 2000; 85: 5129
Keywords:analgesics non-opioid, paracetamol; children; pharmacokinetics,paracetamol
![]() |
Introduction |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Paracetamol has beenfound to pass readily from blood to saliva via the salivary glandsresulting in similar concentrations in adult volunteers15 and patients.16
The purpose of this study was to confirmthe correlation between serum and saliva concentrations and to investigatethe pharmacokinetics of paracetamol after repeated rectal administration of25 mg kg1 every 6 h for 5 days inchildren.
![]() |
Methods |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Anaesthesia wasinduced intravenously or with an inhalational agent and nitrousoxide/oxygen. Post- operatively, some children were given epiduralbupivacaine/fentanyl and others had i.v. morphine either as patientcontrolled analgesia (PCA) or on a as required basis. Afterinduction of anaesthesia, the first dose of paracetamol 25 mgkg1 was given. One or two of the five availablesuppository doses (50, 125, 250, 500, 1000 mg) were used to administer adose as close as possible to 25 mg kg1 for each child.Subsequent doses of paracetamol 25 mg kg1 were given atapproximately 6 hourly intervals for up to 5 dayspostoperatively.
Saliva and blood samples were obtained immediatelybefore insertion of suppositories and at 1, 2, 3 and 4 h after theadministration of the first dose. On the following days, blood and/orsaliva samples were taken before administration of the morning dose and at1, 2, 3 and 4 h after these morning doses. Paracetamol was administered forup to 5 days. Blood samples were drawn from a central venous catheter or aperipheral vein. Saliva samples were obtained using the Salivette citrusdevice (Sarstedt Laboratories, Nümbrecht, Germany) developedespecially for saliva sampling. The Salivette citrus consists of a cottonwool swab (containing 25 mg citric acid to stimulate saliva production),which is placed in the oral cavity for 23 min.17 A string was attached to the swab to avoidaccidental swallowing. Binding of paracetamol to the swab was investigatedprior to this study and was found to be a maximum of 2% whencompared with saliva samples collected by spitting. Citric acid did notinterfere with the analysis and stimulation of saliva production and had noinfluence on the concentration of paracetamol in saliva.18 The exact dose and time of administration wasnoted along with the time of collection of blood and salivasamples.
Analysis ofsamples
Saliva samples were kept refrigerated at 4°C for amaximum 1 h until the saliva was centrifuged out of the swab. Blood sampleswere allowed to clot for 1 h before serum was separated by centrifugation.Immediately after centrifugation, saliva and serum samples were frozen at20°C until assayed. Saliva and serum samples were stable at20°C for at least 6 months. After thawing, 60 µl of samplewas mixed with 120 µl of a protein precipitating solution (2%ZnSO4.7H2O in 1:1 methanol/water) containing theinternal standard (metacetamol, 6 mg litre1) andwhirl-mixed for 30 s before centrifugation at 10 000 g for 10 min.
Paracetamol concentration was determined byhigh performance liquid chromatography (HPLC). The HPLC system consisted ofa Merck Hitachi automatic integrated system with a L-7100 pump,L-7400 UV-detector, L-7200 autosampler and a D-7500integrator. A Spherisorb RP-18 (250x4.6 mm, 5 µm) columnand a LiChrospher 100 RP-18 (5 µm) pre-column were used.The chromatographic conditions were as follows: wavelength 248 nm, mobilephase flow 1 ml min1, the mobile phase consisted ofacetonitrile0.1 M sodium acetate pH=4 (20:80 v/v).Triethylamine 150 µl was added per litre of mobile phase. Theinjection volume was 20 µl. Standard curves were linear in the range040 mg litre1. The intra day coefficient ofvariation was 6.7% at 0.05 mg litre1 and0.4% at 40 mg litre1. The between-daycoefficient of variation was 0.82.2% when concentrationsranged from 140 mglitre1.
Statistics andpharmacokinetic modelling
Data are presented as mean values withSD, median and ranges as appropriate. Pearsonscorrelation was used for analysis of salivaserum correlations and toevaluate the influence of demographic data on pharmacokinetic parameters.Statistical significance was defined as P<0.05.
Pharmacokinetic modelling was performed usingthe computer program ADAPT II, release three (D.Z. Dargenio and A.Schumitzky, University of Southern California, USA, 1992).
Aone-compartment linear model with first-order elimination andabsorption and lag-time was fitted to the data. The goodness of fitwas evaluated by visual inspection of predicted vs observeddata and from plots of residuals.
The absorption rate constant (ka), lag-time (tlag),elimination rate constant (ke), eliminationhalf-life (t), apparent volume ofdistribution (Vd/F, F=absorption fraction), and apparent clearance (CL/F)were estimated by the ADAPT II program for each patient. The maximumconcentration Cmax(1) and the time to reach theCmax(1) Tmax(1) in thefirst dosing interval was calculated by Cmax(1)=[(D·F/Vd)(ka/ke)(ke/(keka))] and Tmax(1)=[(1/(kake))(ln(ka/ke))]respectively [19], where D represents thedose.
The average concentration at steady state (Css) was calculated for each patient by using the vanRossums equation20 Css=[(D/)/(CL/F)], where
is the dosinginterval. The median of
for each patient was used for thispurpose.
In each case the time to reach 90% of the steadystate concentration (TCss(90%)) was calculated using the equation [21] TCss(90%)=3.3·t.
Simulation of the recommended dosingschedule was done in Quattro Pro 6.0 using the equation,
where n represents the number of dosesgiven.19
![]() |
Results |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
|
Salivaserumcorrelation
Thirty-two paired samples (n=64) were available for correlation analysis. Thesamples were obtained from 12 different children aged between 11 weeks and11 yr. The data are depicted in Figure 1A. Regression analysis resulted in a Pearsons Correlationcoefficient (r) of 0.91 (P<0.001). Theconstant of the regression line was not significantly different from zero(P=0.6) and the data were then fitted to the simplermodel forcing through the origin. The relationship between serum and salivaconcentrations could be described by the following equation (mglitre1):
|
Theratio between saliva and serum was calculated to be 0.92±0.26 (Fig.1B).
Pharmacokineticmodelling
Data from 17 children were used in the followingmodelling as four children (8, 11, 12 and 15 in Table 1) were excluded because of insufficientdata.
Individual parameters of ka, tlag, ke, t, Vd/F, CL/F, Css and TCss(90%) are given in Table 2.
|
|
|
|
![]() |
Discussion |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
After administrationof paracetamol suppositories, the absorption rate was faster than or equalto the rate of elimination in all children, but large variations were seenin the absorption phase and the resulting steady state concentrations. Theabsorption rate constant (ka) observed in ourstudy was 1.4 (2.2) h1. After administration ofsuppositories to children, Birmingham and colleagues3 reported a typical kaof 0.3 h1 and Coulthard and colleagues8 found a mean ka of 1.4h1.
Various explanations have been offered for thisunpredictable absorption pattern from paracetamol suppositories, but nonehave been proven. Premature defaecation would reduce the amount of drugavailable for absorption, but no premature defaecation was observed in ourstudy. Rectal pH and rectal contents may also influence the absorptionpattern of rectally administered drugs [27]. In children, therectal pH varies from 7.811.4 (95%-confidenceinterval).28 For paracetamol(pKa=9.5) this may influence bioavailability, since onlyundissociated drug passes biological membranes and the degree ofdissociation for paracetamol will vary from 2 to 99% in this pHrange (% molar dissociation=[100/(1+antilogpKapH)]).29
Childrenseem to absorb paracetamol from suppositories faster and to a greaterextent than adults. Using the same type of suppositories in a single doseof 2636 mg kg1 in adults after minorgynaecological laparoscopic surgery,16the highest mean (SD) observed concentration was 8.4 (3.5)mg litre1 compared with the Cmax after the first dose in our study which was 10.7(3.1) mg litre1 after administration of 2127 mgkg1. Giving 23.9 (4.2) mg kg1 tochildren younger than 160 days, Hansen and colleagues30 observed a Cmax of10.9 mg litre1. In a study of adults,16 Tmax was not reachedwithin 4 h after administration. In contrast, Tmax in the first dosing interval for the children inour study was 2.4 (1.1) h. Values of Tmax foundby other investigators after single dose administration of paracetamolsuppositories to children are between approximately 1 and 3 h.3 4 8 10 3032 In pre-term neonates, the absorptionseems to be slower, van Lingen and colleagues found Tmax values of 3.95.1 h in children2836 weeks of gestational age.33
The limited space in the rectum ofinfants and small children compared with suppository size may favour a moreefficient contact with the rectal mucosa and, consequently, improvedabsorption. Also, the rectal mucosa and colonic blood flow may be alteredin children, but this has not been proven.
Differences in theconcentration reached in individual patients may also be explained by avariability of venous drainage from the rectum. Paracetamol undergoesvariable hepatic first-pass metabolism.31 34 If adrug is administered in the upper part of rectum it will be subject to thehepatic first pass effect whereas drugs in the lower part of the rectumwill bypass the liver via the inferior vena cava.35 Administration of multiple suppositories, aswas the case for some children in our study, could also contribute to thevariable absorption.36 The degree ofrectal anastotomic channels has been proposed to counteract the rectalvenous drainage pattern.7
Thevolume of distribution (Vd/F)in our study is in agreement with values found in other paediatricstudies.11 32 37 Thelower correlation coefficient (r=0.57) between ageand Vd/F implies thatcalculation of a loading dose should not be based on age but rather onweight, height or BSA, which show higher correlations to the apparentvolume of distribution (r=0.670.70).
Only 4% of paracetamolis excreted unchanged in the urine in both neonates, children and adults.The rest is metabolized to paracetamol-sulphate and glucoronide invarying amounts depending on age.3839 Even though neonates, infants andchildren under 12 yr have a limited ability to conjugate phenolic drugsthey are able to compensate for this by a well-developed capabilityfor sulphate conjugation, resulting in overall paracetamol eliminationhalf-lifes after single doses similar to adult values.39 40 In ourstudy t, after repeated dosing, wasbetween 1.1 and 11.5 h with a mean (SD) of 3.5 (2.6) h, andfor one patient (14) the t
was as long as20.8 h, but the predicted steady state concentration for this patient wasin the therapeutic range. The long t
mayreflect a very slow rate of absorption, i.e. a flip-flop model, andnot a reduced clearance, as the steady state concentration for this patientwas comparable with the steady state concentration of the other patients.Repeated doses to children aged between 6 months and 6 yr in a study byNahata and colleagues resulted in a t
of2.2 h.14
The relatively goodcorrelation (r=0.72, P<0.05)between CL/F and age suggests that the ability to eliminate paracetamoldoes increase slightly with age (Fig. 3A). However, body weight (Fig. 3B) and BSA (Fig. 3D) seemto correlate better with elimination capacity (r=0.83and 0.82, P<0.05). In six pharmacokinetic studies ofparacetamol including a total of 270 children between 2 months and 17 yrthe terminal half-life after single doses ranged from 1.7 to 4.7h.4 811 3032 37In pre-term neonates the half-life seems to be prolonged. Forexample, van Lingen33 observed a t of 4.811.0 h in this agegroup.
The average concentration reached at steady state (Css) in our study (15.2 (6.8) mglitre1) is well above the lower limit for the acceptedantipyretic concentration range, which has been suggested to be 1020mg litre1.4143 Only one patient did not approach thisconcentration range at steady state. Although no solid link has been provento exist between serum concentration of paracetamol and degree of analgesiaproduced, the analgesic effect of paracetamol is believed to be related,directly, to its serum concentration44because of its high lipid solubility and low protein binding in thetherapeutic concentration range. After tonsillectomy in children twostudies have reported relationships between paracetamol concentration anddegree of postoperative pain. Anderson and colleagues2 observed that 50% of children experiencedsatisfactory analgesia at a concentration of 17 mglitre1, and in another study,13 acceptable pain scores were associated with aparacetamol concentration above 10 mg litre1.
Thesafety of repeated dosing of paracetamol in children has beenquestioned,14 33 45 andcases of toxicity after repeated therapeutic doses46 47 andsupratherapeutic doses48 have beenreported. In our study the highest predicted concentration at steady statewas 36.9 mg litre1 which is well below the limit oftoxicity49 50 at 120 mg litre1. However,care should be taken when situations of malnutrition are present, becausedepletion of glutathione stores could make children susceptible whensulphation and glucoronidation pathways are saturated.5153
The maximum daily dose of paracetamol iscontroversial,54 but the current oralpaediatric dosing recommendation for paracetamol is a loading dose of 20 mgkg1 and a maintenance dose of 15 mg kg1every 4 h55 with an upper limit of 90 mgkg1 per day.56Rectally 1520 mg kg1 is given every 4 hcommonly,57 but this may produceinadequate concentrations.58 59
To reduce the time to reach steady stateconcentrations, which in our study were not reached until nearly 12 h afteradministration of the first dose, a loading dose should be given.
Weused mean parameter values found in this study (ka=1.4 h1, tlag=0.5 h, ke=0.3 h1, Vd/F=1.3 litrekg1, CL/F=0.3 litrekg1h1) to simulate a paracetamol dosingschedule with a loading dose of 35 mg kg1 and amaintenance dose of 25 mg kg1 administered assuppositories every 6 h (Fig. 4).
|
![]() |
Acknowledgements |
---|
![]() |
Footnotes |
---|
![]() |
References |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
2 AndersonBJ, Holford NHG. Rectal paracetamol dosingregimens: determination by computer simulation. Paediatr Anaesth1997; 7:4515[ISI][Medline]
3 BirminghamPK, Tobin MJ, Henthorn TK etal. Twenty-four-hour pharmacokinetics of rectalacetaminophen in children. Anesthesiology 1997;87:24452[ISI][Medline]
4 HopkinsCS, Underhill S, Booker PD. Pharmacokinetics ofparacetamol after cardiac surgery. Arch Dis Child1990; 65:9716[Abstract]
5 AndersonB, Kanagasundarum S, Woollard G. Analgesic efficacyof paracetamol in children using tonsillectomy as a pain model.Anaesth Intens Care 1996; 24:66973[ISI][Medline]
6 LinY-C, Sussman HH, Benitz WE. Plasmaconcentrations after rectal administration of acetaminophen in pretermneonates. Paediatr Anaesth 1997; 7:4579[ISI][Medline]
7 AndersonBJ, Holford NHG, Woollard GA, Chan PLS. Paracetamolplasma and cerebrospinal fluid pharmacokinetics in children. Br J ClinPharmacol 1998; 46:23743[ISI][Medline]
8 CoulthardKP, Nielsen HW, Schroeder M etal. Relative bioavailability and plasma paracetamol profiles ofPanadol® suppositories in children. J Paediatr ChildHealth 1998; 34:42531
9 Keinänen-KiukaanniemiS, Similä S, Luoma P, Kangas L, SaukkonenA-L. Antipyretic effect and plasma concentrations of rectalacetaminophen and diazepam in children. Epilepsia1979; 20:60712[ISI][Medline]
10 MontgomeryCJ, McCormack JP, Clayton CR, Marsland CP. Plasmaconcentrations after high-dose (45 mg/kg) rectal acetaminophen inchildren. Can J Anaesth 1995; 42:9826[Abstract]
11 AndersonBJ, Woollard GA, Holford NHG. Pharmacokinetics ofrectal paracetamol after major surgery in children. PaediatrAnaesth 1995; 5:23742[ISI][Medline]
12 KollöffelWJ, Driessen FGWHM, Goldhoorn PB. Plasmaconcentration profiles after pre-operative rectal administration of asolution of paracetamol in children. Pharm World Sci1996; 18:1058[ISI][Medline]
13 AndersonBJ, Holford NHG, Woollard GA, Kanagasundarum S,Mahadevan M. Perioperative pharmacodynamics of acetaminophen analgesia inchildren. Anesthesiology 1999; 90:41121[ISI][Medline]
14 NahataMC, Powell DA, Miller MA. Acetaminophenaccumulation in pediatric patients after repeated therapeutic doses.Eur J Clin Pharmacol 1984; 27:579[ISI][Medline]
15 KamaliF, Fry J R, Bell GD. Salivary secretion ofparacetamol in man. J Pharm Pharmacol 1987;39:1502[ISI][Medline]
16 HahnTW, Mogensen T, Lund C, Schouenborg L, Rasmussen M.High dose rectal and oral acetaminophen in postoperativepatientsserum and saliva concentrations. Acta AnaesthesiolScand 2000; 44:3026.[ISI][Medline]
17 BaileyB, Klein J, Koren G. Noninvasive methods for drugmeasurement in pediatrics. Pediatr Clin North Am1997; 44:1526[ISI][Medline]
18 DorbitchRK, Svensson CK. Therapeutic drug monitoring insaliva. Clin Pharmacokinet 1992; 23:36579[ISI][Medline]
19 RasmussenSN. Noter i Farmakokinetik, 2nd Edn.Copenhagen: DSR, 1985
20 JuulP. Almen Farmakologi. Copenhagen:Forlaget for Faglitteratur, 1986; 65, 192
21 RowlandM, Tozer T. Multiple dose regimens. In: Rowland M,Tozer T, eds. Clinical Pharmacokinetics, Concepts andApplications, 3rd Edn. London: Williams & Wilkins,1995;83105
22 MaloneM. Parenteral nutrition. In: Walker R, Edwards C,eds. Clinical Pharmacy and Therapeutics. London: ChurchillLivingstone, 1999;6584
23 DuBois D, Du Bois E. A formula to estimate theapproximate surface area if height and weight be known. Arc IntMed 1916; 17:86371
24 PrescottLF. Adverse reactions and interactions. In:Prescott LF, ed. Paracetamol (Acetaminophen). A CriticalBibliographic Review. London: Taylor & Francis, 1996;397400
25 Al-ObaidySS, Li Wan Po A, McKiernan PJ, Glasgow JFT,Millership J. Assay of paracetamol and its metabolites in urine, plasma andsaliva of children with chronic liver disease. J Pharm BiomedAnal 1995; 13:10339
26 GorodischerR, Burtin P, Hwang P, Levine M, Koren G. Salivaversus blood sampling for therapeutic drug monitoring in children: Patientand parental preferences and an economic aaanalysis. Ther DrugMonit 1994; 16:43743
27 MuranishiS. Characteristics of drug absorption via therectal route. Methods Find Exp Clin Pharmacol 1984;6:76372[ISI][Medline]
28 JantzenJ-PAH, Tzanova I, Witton PK, Klein AM. RectalpH in children. Can J Anaesth 1989;36(6):6657[Abstract]
29 FlorenceAT, Attwood D. Physiochemical properties of drug insolution. In: Florence AT, Attwood D eds, PhysiochemicalPrinciples of Pharmacy, 2nd Edn. London: MacMillan, 1988;4780
30 HansenT, OBrian K, Morton NS, Rasmussen SN. Plasmaparacetamol concentrations and pharmacokinetics following rectaladministration in neonates and young infants. Acta AnaesthesiolScand 1999; 43:8559[ISI][Medline]
31 DangeSV, Shah KU, Deshpande AS, Shrotri DS.Bioavailability of acetaminophen after rectal administration. IndianPediatr 1987; 24:3312
32 BrownD, Wilson JT, Kearns GL, Eichler VF, Johnson VA,Bertrand KM. Single-dose pharmacokinetics of ibuprofen andacetaminophen in febrile children. J Clin Pharmacol1992; 32:23141
33 van LingenRA, Deinum JT, Quak JME etal. Pharmacokinetics and metabolism of rectally administeredparacetamol in preterm neonates. Arch Dis Child Fetal NeonatalEd 1999; 80:F59F63
34 PrescottLF. The absorption of paracetamol. In: Prescott LF,ed. Paracetamol (Acetaminophen). A Critical BibliographicReview. London: Taylor & Francis, 1996;369
35 MorselliPL, Franco-Morselli R, Bossi L. Clinicalpharmacokinetics in newborns and infants. Age-related differences andtherapeutic implications. Clin Pharmacokinet 1980;5:485527[ISI][Medline]
36 NärvänenT, Haslas M, Smal J, Marvola M. Is one paracetamolsuppository of 1000 mg bioequivalent with two suppositories of 500 mg?Eur J Drug Metab Pharmacokinet 1998; 23:2036[ISI][Medline]
37 KellyMT, Walson PD, Edge JH, Cox S, Mortensen ME.Pharmacokinetics and pharmacodynamics of ibuprofen isomers andacetaminophen in febrile children. Clin Pharmacol Ther1992; 52:1819[ISI][Medline]
38 MillerRP, Roberts RJ, Fischer LJ. Acetaminophenelimination kinetics in neonates, children, and adults. Clin PharmacolTher 1976; 19:28494[ISI][Medline]
39 RumoreMM, Blaiklock RG. Influence of age-dependentpharmacokinetics and metabolism on acetaminophen hepatotoxicity. JPharm Sci 1992; 81:2037[ISI][Medline]
40 LevyG, Khanna NN, Soda DM, Tsuzuki O, Stern L.Pharmacokinetics of acetaminophen in the human neonate: Formation ofacetaminophen glucoronide and sulfation in relation to plasma bilirubinconcentration and D-glucaric acid excretion. Pediatrics1975; 55:81825[Abstract]
41 WalsonPD, Mortensen ME. Pharmacokinetics of commonanalgesics, anti-inflammatories and antipyretics in children.Clin Pharmacokinet 1989; 17:11637[ISI][Medline]
42 WilsonJT, Brown D, Bocchini JA, Kearns GL. Efficacy,disposition and pharmacodynamics of aspirin, acetaminophen and cholinesalicylate in young febrile children. Ther Drug Monit1982; 4:14780[ISI][Medline]
43 RumackBH. Aspirin versus acetaminophen: a comparativeview. Pediatrics 1978; 62:9436[ISI][Medline]
44 LevyG. Pharmacokinetic analysis of the analgesic effectof a second dose of acetaminophen in humans. J Pharm Sci1987; 76:889[ISI][Medline]
45 PennaA, Buchanan N. Paracetamol poisoning in childrenand hepatotoxicity. Br J Pharmacol 1991;32:1439
46 MortonNS, Arana A. Paracetamol-induced fulminanthepatic failure in a child after 5 days of therapeutic doses. PaediatrAnaesth 1999; 9:4635[ISI][Medline]
47 PershadJ, Nichols M, King W. The silentkiller: chronic acetaminophen toxicity in a toddler. PediatrEmerg Care 1999; 15:436[ISI][Medline]
48 Rivera-PeneraT, Gugig R, Davis J etal. Outcome of acetaminophen overdose in pediatric patients andfactors contributing to hepatotoxicity. J Pediatr1997; 130:3004[ISI][Medline]
49 RumackBH, Matthew H. Acetaminophen poisoning andtoxicity. Pediatrics 1975; 55:8716[ISI][Medline]
50 PrescottLF, Wright N, Roscoe P, Brown SS.Plasma-paracetamol half-life and hepatic necrosis in patientswith paracetamol overdosage. Lancet 1971:51922
51 PessayreD, Dolder A, Artigou J-Y et al. Effect of fasting on metabolite-mediatedhepatotoxicity in the rat. Gastroenterology 1979;77:26471[ISI][Medline]
52 PriceVF, Miller MG, Jollow DJ. Mechanism offasting-induced potentiation of acetaminophen hepatotoxicity in therat. Biochem Pharmacol 1987; 36:42733[ISI][Medline]
53 HeubiJE. Acetaminophen use in children: more is notbetter. J Pediatr 1997; 130:1757[ISI][Medline]
54 ZacharisM, Watts D. Pain relief in children. BMJ1998; 316: 1552
55 ShannF. Paracetamol: When, why and how much. JPaediatr Child Health 1993; 29:845
56 PennaAC, Dawson KP, Penna CM. Is prescribing paracetamolpro re nata acceptable? J Pediatr Child Health1993; 29:1046
57 VernonS. Rectal paracetamol in small children with fever.Arch Dis Child 1979; 54:46979[Abstract]
58 AndersonBJ. Variability of concentrations after rectalparacetamol. Paediatr Anaesth 1998; 8:274
59 AndersonBJ. What we don't know about paracetamol in children. Paediatr Anaesth 1998; 8: 45160[ISI][Medline]