- Email: [email protected]
Bioequivalence and Pharmacokinetic Profile of Promethazine Hydrochloride Suppositories in Humans
Bioequivalence and Pharmacokinetic Profile of Promethazine Hydrochloride Suppositories in Humans SALOMON STAVCHANSKY”, JACKE. WALIACE~, RICHARD GEARY’,GERALDHEM§, CHARLES A. ROBE§,AND PETER Wu’ Received Au us1 28,1986,from the ‘University of Texas at Austin, College of Pharma Division of Phsrmaceutics, Austin, TX 78772, the *University 07 Texas Health Science Center, Path01 y Department, San Antonio, TX 7%4, and 5Alcon Laboratories, Inc., Fort Worth, TX 76707. Accepted for publication April 16. 198.
the iv studies, averaged 52.2 -t 13.2 ml/min/kg. The volume of distribution, calculated by the area-under-the-curve method,was 22.9 2 6.8 Lkg, and the terminal t,,, was 5.3 1.7 h. Reddrop et al.” monitored the time course of unchanged different formulations.The formulations tested were a 50-mg promethapromethazine in a single rate following a single ip dose of 25 zine hydrochloride polyethylene glycol Suppository,a 50-mg promethamg/kg. Of the dose, 0.2% was found to be eliminated renally zine hydrochloride cocoa butter-white wax suppository, and a 50-mg as unchanged promethazine in 24 h. The t,, was estimated to oral dose of promethazine hydrochloride syrup. Each subject received be 3.5 h. single doses of each of the three formulationson each of three different Recently, Moolenaar et al.I2 investigated the absorption days on a crossover basis. From the measured serum levels, estimates rate and bioavailability of promethazine from rectal and oral of the bioavailability parameters (area under the serum concentration dosage forms in six subjects using a crossover study and versus time curve, time-to-peakserum concentration, and peak serum found that after oral administration of 25 mg of promethaconcentration) were obtained by least-squaresdigital computer fitting. Also, a one-compartmentpharmacokineticopen model with two conseczine-HC1 in solution the maximum concentration (C,,,), the utive first-order input steps is proposed. Statistical analysis of the results time to C,,, (tmax),and t,, were 11.2 ng/mL, 3 h, and 12.7 ? was performed by using a linear multiple regression approach for the 2.4 h, respectively. The authors suggested that rectal absorpanalysis of variance. No significant differences between the syrup and tion is rather fast and fairly complete. In addition, they the polyethylene glycol suppositories were obtained (p > 0.05) for the suggested that rectal dosing with promethazine dissolved in above three bioavailability parameters. However, the polyethylene glyan aqueous micro-enema does not guarantee a more procol suppositories provided statistically higher peak serum concentration, nounced bypass of first-pass metabolism, as compared with shorter time-to-peak serum concentration, and larger area under the oral dosing. serum concentration versus time curve than the cocoa butter-white wax Schwinghammer and Juhl13 reported a comparison of the suppositories. bioavailability of oral, rectal, and intramuscular (im) promethazine. They concluded that rectal suppositories of promethazine are more slowly absorbed than oral solutions or Promethazine (l0-[2-(dimethylarnino)propyl]-phenothi- im injections, and that rectal suppositories and oral solutions are less bioavailable than im injections. The authors suggestazine) is used as an antihistamine, antiemetic, and sedative. ed that diminished systemic bioavailability may result from As a phenothiazine derivative, this drug usually undergoes extensive first-pass hepatic metabolism that occurs after extensive metabolism14 and would be expected to produce both oral and rectal dosing. In addition, a high degree of very low serum levels after a normal therapeutic dose in intersubject variation in the bioavailability of promethazine humans. Therefore, an assay method capable of determining rectal suppositories and oral solutions was observed. promethazine in human serum a t the ng/mL level was Zaman et al.14 reported the bioequivalency and dose prorequired. Due to the difficulty of assay a t such a low level, portionality of three tableted promethazine products. Data only a few bioavailability studies in humans have been from a five-way crossover study in human subjects show conducted. linear dose proportionality. This result is in disagreement Quinn and Calverts investigated the disposition of prowith that reported by Moolenaar et al.12 Intersubject variamethazine in humans and found that promethazine is highly tion was high for all bioavailability parameters. bound to plasma proteins (93% 5 2.5 bound a t 200 mg/mL, The objective of this study is to conduct a three-way 92.5% k 2.3 bound at 400 ng/mL). Binding appeared to be crossover bioequivalence study involving two rectal dosage constant over the expected range of plasma concentrations. forms and an oral solution containing a single therapeutic The data obtained after a 12.5-mg iv dose did not follow a dose of promethazine HC1, with the purpose of establishing multiexponential decline, but showed a second peak a t 1-2 h the rate and extent of absorption of the rectal dosage forms. a h r the dose. The terminal phase of the curve indicated a The results of this investigation will help clarify some of the half-life (tin) of 4.4 h. The t,, from urine data was 8.4 h. The disagreements which exist in the pharmacokinetic literature volume of distribution was found to be 171 L, and 2% of the about promethazine and will add to the existing knowledge dose was excreted unchanged. After a 30-mg oral dose, a t,, of in this field. 7 h was determined from plasma levels measured over a period of 24 h, and a tln of 10 h was measured from urine Experlmental Section data. The disposition of [36Slpromethazine ( 5 ng/kg) in rabbits MaterialgPromethazine hydrochloride (10-[2-(dimethylamino) was investigated by Taylor and Houston10 following iv, oral, propyll-phenothiazine monohydrochloride; Napp Chemicals, Inc., and direct hepatic portal administration of an aqueous soluh d i , NJ),phenergan suppository (Wyeth Labs., Inc., Philadelphia, tion. Blood samples were collected over a 10-h period followP A lot no. 1782592), phenergan syrup fortis (Wyeth Labs., Inc., Philadelphia, PA; lot no. 17825931, trifluopromethazine (E.R. ing dosing. The clearance of promethazine, calculated from Abstract 0A bioequivalence study of promethazine hydrochloride (1 0-
[2-(dimethylamino)propyl]-phenothiazinemonohydrochloride) was conducted in 20 male human subjects with the purpose of comparing, under blind condition, the human serum levels of promethazine in three
m22-3!549/87/W-W 7$0l.OO/O Q 7987, American Pharmaceutical Association
*
Journal of Pharmaceutical Sciences / 441 Vol. 76, No. 6, June 7987
Squibb and Sons, Inc., Princeton, NJ), and 2,2,2-trichloroethylchloroformate (Aldrich Chemical Co., Milwaukee, WI) were obtained from commercial sources. Other solvents and chemicals were analytical reagent grade. Assay of Serum Promethazine Hydrochloride-A high-performance liquid chromatographic procedure was ueed to determine the serum levels of promethazine.16 The method was b a d on the extraction of promethazine from alkalified serum with hexane containing a known amount of trifluopromethazine. The hexane was evaporated to dryness with air at 40 ‘C. Fifty microliters of ethyl acetate and 25 FL of trichloroethylchloroformate were added to the reeidue. The reaction mixture was incubated for 20 min at 120 “c. The solvent was then evaporated to dryness with air at 40 “C, and the residue was dissolved in 100 p L of methanol. Fifty-microliter samples were injected into a high-performance liquid chromatograph. A model 5000 Varian high-performance liquid chromatograph (Varian Inc., Palo Alto, CA) equipped with a Variochrom ultraviolet (254 nm) detector was used. Columna (30 cm in length, 0.60.d. and 0.4-cm i.d.) were packed with MCH-10 (Varian Inc., Palo Alto, CA). Columns were equilibrated for 30 min with 84% methanol:l6% HzO. The flow rate was 2.0 ml/min. Column temperature was maintained at 40 “C. The samples were injected using a 50-pL loop injector. Subject Selectio-Normal healthy male volunteers, 21 through 33 years of age, were selected for the study. All of the subjecta conformed to the Metropolitan Life Insurance Company, November 1959, bulletin for height and weight. Within one week preceding the study, subjects were given a thorough physical examination that included extensive laboratory tests. Subjecta with test results outside normal limits were excluded, as were any individuals showing a medical history of any significant physical or organ abnormality or disease. Informed consent was obtained from every subject. Study Design-Tkenty-four subjects were selected to receive single doses of each of the three different treatments on each of three different days on a crossover basis. Treatment A represents one phenergan suppository containing 50 mg of promethazine hydrochloride (Wyeth Laboratories), treatment B represents 10 mL of phenergan syrup fortis containing 50 mg of promethazine hydrochloride (Wyeth Laboratories), and treatment C represents an experimental polyethylene glycol (PEG) suppository containing 50 mg of promethwine hydrochloride (Alcon Laboratories, lot ZE-1424). The study design is an exact balance of formulation over a period of three weeks. Each subject received each formulation. In addition, each formulation occurred eight times in each weekly administration. The design is eight replicates of a 3 x 3 Latin square. Treatment administrations were separated by a 1-week washout period. Study Period-During the evening preceding the study day, each subject was instructed to drink at least 500 mL offluids. No alcoholic beverages were allowed between supper and midnight. Between midnight and 12:00 to 1:00 p.m. of the test day, subjecta fasted. At approximately 7:00 to 8:OO a.m., subjects were dosed. The PEG suppositories were moistened in water before insertion. The phenergan suppositories were inserted without premoistening. After insertion of the suppositories, the subjects were maintained in a standing position for 30 min. After this 30-min period, the subjects were ambulatory. Ten milliliters of the syrups were given orally from a syringe. Subjects were ambulatory throughout the test day, not recumbent, and were available for blood sampling. Subjects refrained from strenuous work. All subjects were given 200 mL of water at the time of treatment regardless of the dosage administered. An additional 200 mL of water was given every 2 h after treatment while fasting was still being maintained. Following 4 h of postdrug fasting, the subjects were allowed to eat and drink. Sample Collection-Blood samples (15 mL), drawn by serial vein puncture, were collected at zero (0) time (just prior to drug administration) and a t 20,40, and 60 min, and 1.5,2.0,3.0,4.0,6.0,8.0,12.0, and 24.0 h after drug administration. All blood apecimene were collected in vacutainer tubes. The tubes were immereed in chipped ice and the specimen8 were allowed to coagulate. The specimen tubes were quickly centrifuged (30006000 rpm for 5-10 min) in a refrigerated centrifuge. The serum from each tube waa transferred to polypropylene tubes. ARer tramfemng the serum to the tubes, the contents of the tubes were immediately flash t h a n by immersion of the tubes in an alcoho1:dry ice bath, and were kept frozen until WYed. 442 /Journal of Pharmaceutical Sciences Vol. 76, No. 6, June 1987
Results and Dlscusslon The semilogarithmic plots of mean serum promethazine concentration versus time for the three treatments, A, B, and C, are illustrated in Figure 1. The observed values for the individual peak serum concentrations, the time-to-peak serum concentrations, and the areas under the serum concentration versus time c w e s (estimated by the trapezoidal rule) are presented in Tables I, II, and 111, respectively. Missing D a t a - O f the 24 subjects enrolled for this study, four subjects did not complete the trials (subjects 8, 13, 20, and 24) and are ignored in this evaluation. There were two instances where serum samples could not be analyzed because of insufficient quantities; these were for subject 3, treatment C at 4 h, and for subject 18, treatment C a t 12 h. The loss of these two data points did not preclude determining the area under the c w e , time-to-peak concentration, or peak concentration. The most serious consideration concerning use of data had to do with the expulsion of suppositories and/or defecation within -30-46 min after dosing. This occurred with six subjects and always involved treatment C. For three of the six subjects (16,21, and 231, a portion of the suppository was either seen or retrieved. For one subject (61, no part of the suppository was seen or recovered. In any case, it appears that data resulting from these trials, in which the suppository was expelled so early after insertion, cannot be expected to adequately or accurately reflect the availability profile over a 24-h period. All data resulting from these specific subjects were omitted from this evaluation. The resultant data set that was analyzed is comprised of 14 subjecta on all three treatments and an additional six sub-
i P
TIME, MIN
FIQW I-Promethazine mean serum levels obtained after single &minisfration of freafmenf A (+), axxxr-borref suppositories; treatment 6 (*), syrup; and treatment C (m), PEG suppositories. Llnes represent nonllnear compuler fils.
jecta on treatments A and B only. There are two minor exceDtions in that subiect 3 does not have a 4-h observation for tieatment C and sibject 18 does not have a 12-h observation for treatment C. Statistical Analysis of the Data-The missing and unusable data described above resulted in an unbalanced design, making the use of standard analysis of variance procedures impractical. Also, the usual average values obtained from the different treatments would not be comparable in that they would not be equally affected by days and subjects. Because of this, the analysis of variance was performed using a “general linear model” approach: Y observed = mean + subject + treatment + day. This approach consisted of using multiple linear regression to fit the model with the aid of the SPSS computer program.’6 The system of dummy coding was applied to each categorical variable, such as treatments, days, and subjects. In this system, a set of dummy variables is created by treating each category of a nominal variable a s a separate variable, and arbitrary scores were given for all cases depending on their presence or absence in each of the categories (such as, “1”for presence and “0” for absence in a given category). The number of dummy variables necessary to exhaust the information about a given category is equal to the number of cases of that category minus one. After assigning the dummy values, the data were analyzed by simple multiple linear regression analysis following the order of subjects, treatments, and days. A set of partial correlation coefficients was obtained. Because these partial correlation coefficients are not orthogonal, an adjustment for the intercorrelations among coded categories was made. Analysis of variance was then performed. First the variation due to subjects was isolated. Then the differences in treatment adjusted for days were tested. Finally the differences in days adjusted for treatments were tested. No signifi-
Tabk ICTlmbto-Peak of the Indlvldual Serum C o m t r a t l o r t Time curve8
Tabk Hndlvldual Peak of the Serum Concentmtlon-Tlme Curve’
Table Ill-Area Curve’
Subject
Treatment A
Treatment B
Treatment C
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
366 367 245 490 481 240 723
182 248 239 183 184 182 180
182 241 181 189 181 362
480 359 479 476
180 185 242 182
184 183 239 237
363
480
190 180 117 185 239 240
250 240 480 502 250 194
238 361 181
123 362 127
249 360 190
Mean, min
394
198
263 (267)’
Standard Deviation
122
54
98 (95)’
-
-
480 358 362 360
-
-
-
-
360
-
-
.Data for subjects 1, 6, 11, 16, 21, and 23 were omitted due to defecation.
Under the lndlvldual Serum Concentratlon-llnn
Subject
Treatment A
Treatment B
Treatment C
Subject
Treatment A
Treatment B
Treatment C
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
22.6 10.6 59.5 3.6 22.3 12.9 5.1
23.1 14.7 70.7 7.5 26.7 12.8 19.2
11.7 34.3 92.1 2.3 46.2 9.8 19.5
13 708 10 565 28 122 2 734 18 031 6 389 3 473
8 765 6 579 50 681 5 146 14 510 6 519 10 818
5 902 13 478 70 924 1 705 30 939 3 813 5 828
4.5 12.4 12.5 17.0
51.9 10.5 31.6 32.8
29.4 15.7 4.9 33.3
3 755 9 082 11 568 14 410
31 709 3 168 14 049 15 054
19 223 10 043 4 427 24 302
7.5 8.9 10.3 5.2 8.5 9.5
6.4 23.0 23.1 13.5 12.5 10.4
8.1 22.0 1.6 10.9 11.8 13.7
7300 5 309 7 699 3 645 5 914 7 629
10 675 8 399 6 596 7 558 8446 5 858
4 802 16 513 1526 8 575 7 578 7 476
3.9 37.2 6.8
34.0 55.2 99.0
9.7 49.5 21.8
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
1 458 34 861 6411
14 946 32 233 60 163
7 931 45 175 12 131
10 103
16 094
15 116 (19040)b
8 517
15589
17018(19044)b
-
-
-
-
-
-
-
-
-
23
-
24
Mean, nglmL
14.0
28.9
22.4 (27.8)
Mean
Standard Deviation
13.4
23.7
21.3 (23.2)b
Standard Deviation
Data expressed as nglmL. Data for subjects 1, 6, 11, 16, 21, and 23 were omitted due to defecation.
-
-
-
-
-
-
-
-
-
-
Data expressed as ngmL-’.rnin-’. Data for subjects 1, 6, 11, 16, 21, and 23 were omitted due to defecatlon. Journal of Pharmaceutical Sciences / 443 Vol. 76, No. 6, June 1987
cant difference among days (p > 0.05) for area under the curve, time-to-peak concentration, and peak concentration is found. Among subjects, a significant difference at p < 0.05 is found for area under the curve and peak concentration, and no significant difference a t p > 0.05 is found for time-to-peak concentration. Significant differences among treatments A, B, and C at p < 0.05 for area under the curve, time-to-peak concentration, and peak concentration are found. A null hypothesis of pA = pB = pC, where pA, pB, and $2 are the population means of treatments A, B, and C, respectively, was rejected a t p = 0.05. In addition, with a n F test, the squared multiple correlation coefficientswere significant a t p < 0.05 for area under the curve, time-to-peak concentration, and peak concentration. Therefore, post hoc comparisons were performed. Scheffe's" test for multiple comparisons was used to compare the differences between treatment A versus treatments B and C, and treatment B versus treatment C. Comparisons between treatment A versus treatment C indicate significant differences (p < 0.05) for area under the curve, time-to-peak concentration, and peak concentration. Comparisons between treatment B versus treatment C indicate no significant differences for area under the curve, time-to-peak concentration, and peak concentration at p > 0.05. As a result, the rates of absorption, which are related to the time-to-peak concentration, are significantly slower for treatment A than treatment C. The extent of absorption, which is related to peak concentration and area under the curve, are significantly smaller for treatment A than treatment C. Estimation of Pharmacokinetic Parameters-The semilogarithmic plots of mean serum promethazine concentrations versus time reveal a monoexponential decline in the promethazine serum levels after the peak, with no significant distribution phase, as depicted in Figure 1. Each set of mean serum promethazine concentration data followingtreatments A, B, and C was fitted to eq 1 by the aid of NONL1N.l8
C(t) = Ale-B1(" + A2e-B2(t) + A3e-B3("
model described in Scheme I provided a significantly better fit upon comparison with a reduced (two-exponential) model. The mean serum promethazine concentration data following oral administration of the promethazine syrup is described by
C(t) = A l e
-kat
+ A2e-ke1'
(2)
where A1 = -A2, C(t) is the serum concentration of promethazine at running time t, A1 and A2 are two pre-exponential coefficients, and k, and k,1 are first-order rate constants describing rate of absorption and elimination, respectively. Therefore, mean serum promethazine concentration data for the two suppository treatments can be expressed by the following equation:
(1)
where A2 = -(A1 + A3), C(t) is the serum concentration of promethazine at the running time t, A l , A2, and A3 are three pre-exponential coefficients, and B1, B2, and B3 are three hybrid rate constants. Based on these data, a one-compartment open model with two consecutive first-order input steps is proposed as shown in Scheme I, where kl and kz are two consecutive first-order input rate constants and k,l is the first-order elimination rate constant. The process from compartment A to compartment B may represent the dissolution of suppositories. The process from compartment B to compartment C represents a first-order absorption process; compartment C is designated as the central compartment. The
C(t) = (FD/Vd)klk2[e-k1(tl/(-k1+ k2)(-k1 + kel) + e-ka't'/(-k2 + kl)(-kz + kel) + e-kel(t'/(-kel + kl)(-kel + (3) where F is the fraction of the dose, D,absorbed, and Vd is the apparent volume of distribution for the central compartment including blood serum, and k l , k2, and k,l correspond to B2, B3, and B1, respectively. The ratio FDIVd was calculated from the average o f A l (-kel + kl)(-kel + k2)lklk2,A2 ( - k 1 + kz)(-kl + k,J/klkz, and A3 ( - k z + kl)(-kz + kel)/klkz. The final pharmacokinetic parameters calculated from A l , B1, A2, B2, A3, and B3 are presented in Table IV for treatments A, B, and C. The serum clearance, CL, was calculated from the following equation:
CLIF
=
(D/AUCo,(BW)
(4)
where AUCo-, is the area under the serum promethazine concentration-time curve from zero time to infinity and BW is the individual body weight in kg. Assume that the oral dose is the standard to which the two suppository forms are to be compared. The data from the suppositories are normalized with respect t o the oral dosage form. The apparent volume of distribution for the central compartment, Vd, can be calculated from the following equation:
Vd/F = C L / ( k e f l
(5) where k,1 is the overall first-order elimination rate constant. The serum promethazine concentration-time curves derived from NONLIN16 fits are depicted in Figure 1. The peak concentration, time-to-peak concentration, and area and relative bioavailability under the curve from zero time to infinity, calculated from the NONLIN fits, are presented in Table V. The area under the curve from zero time to infinity was estimated from
The relative bioavailability, Frel, of the promethazine suppositories was calculated using AUCO, values corrected
Scheme I
Table IV-Pharmacokinetic Parameter Estimates for Treatments A, 8, and C followlng NONLIN Fit to One-Compartment Open Model with Two Consecutive First-Order input Steps
Treatment
4, min-' x 1 0 - ~
AB Bb
6.256
Ca
12.566
k,, min-' x 1 0 - ~
-
-
7.001
k2, min-' x 1 0 - ~ 9.541
-
21.835
kl, min-' x 1 0 - ~ 1.514 2.532 1.635
FDNd, ng/mL
CUF, mUmin/kg
Vd/F, Ukg
19.46 28.74 35.51
52.66 59.48 30.95
34.78 23.51 18.93
aFollowing NONLIN fit to one-compartment open model with two consecutive first-order input steps (kl and k2). bFollowing NONLIN fit to onecompartment open model with first-order input (k.). 444 /Journal of Pharmaceutical Sciences Vol. 76, No. 6. June 1987
Table V-Summary
of Bloavallablllty Parameters for Both lndlvldual and Fltted Data
Parameter Treatment
From Individual Data AUCsi4a
A B C
10 103 (8517)b 16 094 (15589)b 19 041 (19044)*
b, 394 (122)b 198 (54) 267 (95)
,C 14 (13.4)b 28.9 (23.7)b 27.8 (23.2)b
From Nonlinear Regression AUCsi440
b,
C-
Frel'
62.8
10 327
363
12.1
65.3
100.0
15 814
181
26.2
100.0
118.3
19 113
241
25.6
120.9
Fre, a
'Relative bioavailability parameter is represented as a percentage. Numbers in parentheses represent standard deviation. for the intraindividual variation in the elimination rate Frel =
(AUCo2, suppositories)(kel, suppository) (7) (AUCo-, syrup)(k,l, syrup)
where treatment B (syrup) was chosen as the standard. Comparing the results from the model-independent approach and the model-dependent approach, which are presented in Table I, 11, 111, and V, it can be observed that the magnitude of the area under the curve, time-to-peak concentration, and peak concentration for the three treatments follows the same rank: for area under the curve (AUC),C > B > A; for time-to-peak concentration (t,,,), A > C > B; and for peak concentration (C,,,), B > C > A. However, with the analysis of variance, the difference between treatment A and treatment C is proved to be significant a t p < 0.05, but the difference between treatment B and treatment C is not significant at p > 0.05. Treatment A appeared to have slower dissolution in vivo than treatments B and C, as indicated from the mean ( 5 standard deviation) t,, values of 394 5 122, 198 2 54,and 267 ? 95 min for treatments A, B, and C, respectively (Table 11). This is also indicated from NONLIN-fitted t,,, values of 446,204,and 252 min for treatments A, B, and C, respectively (Table V). Furthermore, the rate constant kl(6.256 x min-') for treatment A is slower than the k l (12.566x min-'1 for treatment C. In addition, the kz rate constant represented in Scheme I should approximate the absorption rate from solution. Treatment C appears to have an increased kz compared with that of treatment A, which may suggest an absorption enhancement that is possibly due to the polyethylene glycol formulation. However, because of large intersubject variability, it is difficult to conclude unambiguously that the rates are significantly enhanced. The NONLIN-fitted C,,, after treatment A was 11.0 ng/ mL. This C,,, was smaller than the C,,, (26.6ng/mL) for treatment B and the C,, (24.6ng/mL) for treatment C. The same results were presented in Table I where the mean ( * standard deviation) C,,, for treatments A, B, and C were 14.0 ? 13.4,28.92 23.7,and 27.85 23.2ng/mL, respectively. As a consequence, the higher C,,, is consistent with the shorter tmax and faster k1 observed for these three treatments. After the adjustment for the intrasubject variation with the overall elimination rate constants, the relative bioavailability (47.2% of treatment B) for treatment A is smaller than the relative bioavailability (87.8% of treatment B) for treatment C obtained fron NONLIN fit of the area under the curve from time zero to time infinity. Promethazine HCl elimination kinetics after oral and rectal administrations were different. The k,l obtained for oral treatment B was 2.53 x min-' and the corresponding half life, t,,,, was 4.6h. The kel and t,,, for rectal treatment A were 1.514 x min-' and 7.6 h, respectively, and the
kel and t,,, for rectal treatment C were 1.635 x lo-' min-' and 7.1h, respectively. These elimination parameters are not exactly in agreement with those reported for promethazine by other i n v e s t i g a t o r ~ . ~ ~ ~ J ~ Promethazine polyethylene glycol suppositories a r e thought to have many practical advantages over promethazine solutions. For example, the suppositories evaluated in this study could be easily applied to pediatric use. In addition, promethazine polyethylene glycol suppositories don't melt at room temperature and prevent the problems associated with suppositories that have a cocoa butter base. Thus, refrigeration during handling and storage is no longer needed. The promethazine polyethylene glycol suppositories evaluated appeared to have rapid dissolution in vivo. They produce serum promethazine concentrations comparable to an oral promethazine solution, and the relative bioavailability is approximately the same as the oral solution. Thus, in practice, these suppositories would act as a reliable dosage form for rectal promethazine administration.
References and Notes 1. Hansson, E.; Schmiterlow, C. G. Arch. Intern. Pharmacodyn. Ther. 1961,131,309. 2. Rusiecki, W.; Wvsocka-Paruszewska. B. Diss. Pharm. Pharmacol. 1969;21 ,. 73: 3. Nadeau, G.; Sobolewski, G. Can. Med. Assoc. J . 1959,81,658; Chem. Abstr. 54,4904h. 4. Robinson, A. E.; Beaven, V. H. J . Pharm. Pharmacol. 1964,16, 342. 5. Robinson, A. E. J . Pharm. Phurmacol. 1966. 18,19. 6. Beckett, A. H.; Al-Sarrqj, S.; Essien, E. E. Xenobiotica 1975,5, 325. 7. Beckett, A. H. Xenobiotica 1975,5, 449. 8. Bornschien, I.; Pfeifer, S. Phurmazie 1979,34,750. 9. Quinn, J.; Calvert, R. J . Phurm. Pharmacol. 1976,28,Suppl. 59P. 10. Ta lor, G.; Houston, J. B. J . Pharm. Phurmmol. 1979,31,Suppl. 40% 11. Reddrop, C. J.; Riess, W.; Slater, T. F. J . Chromatogr. 1980,192, 375.
12. Mdenaar, F.;Ensing, J. G.; Bohuis, B. G.; Visaer, J. I d . J . Phurm. 1981,9,353. 13. Schwinghammer, T. L.;Juhl, R. P. Biophurm. Drug Dispos. 1984.5. 185. 14. Zaman,' R.; Honigber I. L: Francisco G. E.; Kotzan, J. A.; Steward, J. T.; Brown,%. J.; Shaf, V. P.;helsor, F. R. Biophurm. Dr Dkps.1986,7,281. 15. Wayace, J. E.; Shimek, E. L.; Stavchansky, S.; Harris, S. C. Anal. Chem. 1981,53,960. 16. Nie, N. H.; Hull, C. H.; Jenlins, J. G.; Steinbrenner, K.; Bent, D.H. Statistical Package for the Socual Science, 2nd ed.; McGraw-Hill: New York, 1975. 17. Kerlinger, F. M.; Pedhazur, E. J. Multiple Re ression in Behavioral Research; Holt, Rinehart and Winston: 8 e w York, 1973;p 128. 18. Metzler, C.M. NONLIN Computer Program; Technical Report 7292/60/7292/005, The Upjohn Co., Kalamazoo, MI,1969.
Acknowledgments This'work was supported by a grant from Alcon Laboratories, Inc. Journal of Pharmaceutical Sciences / 445 Vol. 76.No. 6, June 1987
the iv studies, averaged 52.2 -t 13.2 ml/min/kg. The volume of distribution, calculated by the area-under-the-curve method,was 22.9 2 6.8 Lkg, and the terminal t,,, was 5.3 1.7 h. Reddrop et al.” monitored the time course of unchanged different formulations.The formulations tested were a 50-mg promethapromethazine in a single rate following a single ip dose of 25 zine hydrochloride polyethylene glycol Suppository,a 50-mg promethamg/kg. Of the dose, 0.2% was found to be eliminated renally zine hydrochloride cocoa butter-white wax suppository, and a 50-mg as unchanged promethazine in 24 h. The t,, was estimated to oral dose of promethazine hydrochloride syrup. Each subject received be 3.5 h. single doses of each of the three formulationson each of three different Recently, Moolenaar et al.I2 investigated the absorption days on a crossover basis. From the measured serum levels, estimates rate and bioavailability of promethazine from rectal and oral of the bioavailability parameters (area under the serum concentration dosage forms in six subjects using a crossover study and versus time curve, time-to-peakserum concentration, and peak serum found that after oral administration of 25 mg of promethaconcentration) were obtained by least-squaresdigital computer fitting. Also, a one-compartmentpharmacokineticopen model with two conseczine-HC1 in solution the maximum concentration (C,,,), the utive first-order input steps is proposed. Statistical analysis of the results time to C,,, (tmax),and t,, were 11.2 ng/mL, 3 h, and 12.7 ? was performed by using a linear multiple regression approach for the 2.4 h, respectively. The authors suggested that rectal absorpanalysis of variance. No significant differences between the syrup and tion is rather fast and fairly complete. In addition, they the polyethylene glycol suppositories were obtained (p > 0.05) for the suggested that rectal dosing with promethazine dissolved in above three bioavailability parameters. However, the polyethylene glyan aqueous micro-enema does not guarantee a more procol suppositories provided statistically higher peak serum concentration, nounced bypass of first-pass metabolism, as compared with shorter time-to-peak serum concentration, and larger area under the oral dosing. serum concentration versus time curve than the cocoa butter-white wax Schwinghammer and Juhl13 reported a comparison of the suppositories. bioavailability of oral, rectal, and intramuscular (im) promethazine. They concluded that rectal suppositories of promethazine are more slowly absorbed than oral solutions or Promethazine (l0-[2-(dimethylarnino)propyl]-phenothi- im injections, and that rectal suppositories and oral solutions are less bioavailable than im injections. The authors suggestazine) is used as an antihistamine, antiemetic, and sedative. ed that diminished systemic bioavailability may result from As a phenothiazine derivative, this drug usually undergoes extensive first-pass hepatic metabolism that occurs after extensive metabolism14 and would be expected to produce both oral and rectal dosing. In addition, a high degree of very low serum levels after a normal therapeutic dose in intersubject variation in the bioavailability of promethazine humans. Therefore, an assay method capable of determining rectal suppositories and oral solutions was observed. promethazine in human serum a t the ng/mL level was Zaman et al.14 reported the bioequivalency and dose prorequired. Due to the difficulty of assay a t such a low level, portionality of three tableted promethazine products. Data only a few bioavailability studies in humans have been from a five-way crossover study in human subjects show conducted. linear dose proportionality. This result is in disagreement Quinn and Calverts investigated the disposition of prowith that reported by Moolenaar et al.12 Intersubject variamethazine in humans and found that promethazine is highly tion was high for all bioavailability parameters. bound to plasma proteins (93% 5 2.5 bound a t 200 mg/mL, The objective of this study is to conduct a three-way 92.5% k 2.3 bound at 400 ng/mL). Binding appeared to be crossover bioequivalence study involving two rectal dosage constant over the expected range of plasma concentrations. forms and an oral solution containing a single therapeutic The data obtained after a 12.5-mg iv dose did not follow a dose of promethazine HC1, with the purpose of establishing multiexponential decline, but showed a second peak a t 1-2 h the rate and extent of absorption of the rectal dosage forms. a h r the dose. The terminal phase of the curve indicated a The results of this investigation will help clarify some of the half-life (tin) of 4.4 h. The t,, from urine data was 8.4 h. The disagreements which exist in the pharmacokinetic literature volume of distribution was found to be 171 L, and 2% of the about promethazine and will add to the existing knowledge dose was excreted unchanged. After a 30-mg oral dose, a t,, of in this field. 7 h was determined from plasma levels measured over a period of 24 h, and a tln of 10 h was measured from urine Experlmental Section data. The disposition of [36Slpromethazine ( 5 ng/kg) in rabbits MaterialgPromethazine hydrochloride (10-[2-(dimethylamino) was investigated by Taylor and Houston10 following iv, oral, propyll-phenothiazine monohydrochloride; Napp Chemicals, Inc., and direct hepatic portal administration of an aqueous soluh d i , NJ),phenergan suppository (Wyeth Labs., Inc., Philadelphia, tion. Blood samples were collected over a 10-h period followP A lot no. 1782592), phenergan syrup fortis (Wyeth Labs., Inc., Philadelphia, PA; lot no. 17825931, trifluopromethazine (E.R. ing dosing. The clearance of promethazine, calculated from Abstract 0A bioequivalence study of promethazine hydrochloride (1 0-
[2-(dimethylamino)propyl]-phenothiazinemonohydrochloride) was conducted in 20 male human subjects with the purpose of comparing, under blind condition, the human serum levels of promethazine in three
m22-3!549/87/W-W 7$0l.OO/O Q 7987, American Pharmaceutical Association
*
Journal of Pharmaceutical Sciences / 441 Vol. 76, No. 6, June 7987
Squibb and Sons, Inc., Princeton, NJ), and 2,2,2-trichloroethylchloroformate (Aldrich Chemical Co., Milwaukee, WI) were obtained from commercial sources. Other solvents and chemicals were analytical reagent grade. Assay of Serum Promethazine Hydrochloride-A high-performance liquid chromatographic procedure was ueed to determine the serum levels of promethazine.16 The method was b a d on the extraction of promethazine from alkalified serum with hexane containing a known amount of trifluopromethazine. The hexane was evaporated to dryness with air at 40 ‘C. Fifty microliters of ethyl acetate and 25 FL of trichloroethylchloroformate were added to the reeidue. The reaction mixture was incubated for 20 min at 120 “c. The solvent was then evaporated to dryness with air at 40 “C, and the residue was dissolved in 100 p L of methanol. Fifty-microliter samples were injected into a high-performance liquid chromatograph. A model 5000 Varian high-performance liquid chromatograph (Varian Inc., Palo Alto, CA) equipped with a Variochrom ultraviolet (254 nm) detector was used. Columna (30 cm in length, 0.60.d. and 0.4-cm i.d.) were packed with MCH-10 (Varian Inc., Palo Alto, CA). Columns were equilibrated for 30 min with 84% methanol:l6% HzO. The flow rate was 2.0 ml/min. Column temperature was maintained at 40 “C. The samples were injected using a 50-pL loop injector. Subject Selectio-Normal healthy male volunteers, 21 through 33 years of age, were selected for the study. All of the subjecta conformed to the Metropolitan Life Insurance Company, November 1959, bulletin for height and weight. Within one week preceding the study, subjects were given a thorough physical examination that included extensive laboratory tests. Subjecta with test results outside normal limits were excluded, as were any individuals showing a medical history of any significant physical or organ abnormality or disease. Informed consent was obtained from every subject. Study Design-Tkenty-four subjects were selected to receive single doses of each of the three different treatments on each of three different days on a crossover basis. Treatment A represents one phenergan suppository containing 50 mg of promethazine hydrochloride (Wyeth Laboratories), treatment B represents 10 mL of phenergan syrup fortis containing 50 mg of promethazine hydrochloride (Wyeth Laboratories), and treatment C represents an experimental polyethylene glycol (PEG) suppository containing 50 mg of promethwine hydrochloride (Alcon Laboratories, lot ZE-1424). The study design is an exact balance of formulation over a period of three weeks. Each subject received each formulation. In addition, each formulation occurred eight times in each weekly administration. The design is eight replicates of a 3 x 3 Latin square. Treatment administrations were separated by a 1-week washout period. Study Period-During the evening preceding the study day, each subject was instructed to drink at least 500 mL offluids. No alcoholic beverages were allowed between supper and midnight. Between midnight and 12:00 to 1:00 p.m. of the test day, subjecta fasted. At approximately 7:00 to 8:OO a.m., subjects were dosed. The PEG suppositories were moistened in water before insertion. The phenergan suppositories were inserted without premoistening. After insertion of the suppositories, the subjects were maintained in a standing position for 30 min. After this 30-min period, the subjects were ambulatory. Ten milliliters of the syrups were given orally from a syringe. Subjects were ambulatory throughout the test day, not recumbent, and were available for blood sampling. Subjects refrained from strenuous work. All subjects were given 200 mL of water at the time of treatment regardless of the dosage administered. An additional 200 mL of water was given every 2 h after treatment while fasting was still being maintained. Following 4 h of postdrug fasting, the subjects were allowed to eat and drink. Sample Collection-Blood samples (15 mL), drawn by serial vein puncture, were collected at zero (0) time (just prior to drug administration) and a t 20,40, and 60 min, and 1.5,2.0,3.0,4.0,6.0,8.0,12.0, and 24.0 h after drug administration. All blood apecimene were collected in vacutainer tubes. The tubes were immereed in chipped ice and the specimen8 were allowed to coagulate. The specimen tubes were quickly centrifuged (30006000 rpm for 5-10 min) in a refrigerated centrifuge. The serum from each tube waa transferred to polypropylene tubes. ARer tramfemng the serum to the tubes, the contents of the tubes were immediately flash t h a n by immersion of the tubes in an alcoho1:dry ice bath, and were kept frozen until WYed. 442 /Journal of Pharmaceutical Sciences Vol. 76, No. 6, June 1987
Results and Dlscusslon The semilogarithmic plots of mean serum promethazine concentration versus time for the three treatments, A, B, and C, are illustrated in Figure 1. The observed values for the individual peak serum concentrations, the time-to-peak serum concentrations, and the areas under the serum concentration versus time c w e s (estimated by the trapezoidal rule) are presented in Tables I, II, and 111, respectively. Missing D a t a - O f the 24 subjects enrolled for this study, four subjects did not complete the trials (subjects 8, 13, 20, and 24) and are ignored in this evaluation. There were two instances where serum samples could not be analyzed because of insufficient quantities; these were for subject 3, treatment C at 4 h, and for subject 18, treatment C a t 12 h. The loss of these two data points did not preclude determining the area under the c w e , time-to-peak concentration, or peak concentration. The most serious consideration concerning use of data had to do with the expulsion of suppositories and/or defecation within -30-46 min after dosing. This occurred with six subjects and always involved treatment C. For three of the six subjects (16,21, and 231, a portion of the suppository was either seen or retrieved. For one subject (61, no part of the suppository was seen or recovered. In any case, it appears that data resulting from these trials, in which the suppository was expelled so early after insertion, cannot be expected to adequately or accurately reflect the availability profile over a 24-h period. All data resulting from these specific subjects were omitted from this evaluation. The resultant data set that was analyzed is comprised of 14 subjecta on all three treatments and an additional six sub-
i P
TIME, MIN
FIQW I-Promethazine mean serum levels obtained after single &minisfration of freafmenf A (+), axxxr-borref suppositories; treatment 6 (*), syrup; and treatment C (m), PEG suppositories. Llnes represent nonllnear compuler fils.
jecta on treatments A and B only. There are two minor exceDtions in that subiect 3 does not have a 4-h observation for tieatment C and sibject 18 does not have a 12-h observation for treatment C. Statistical Analysis of the Data-The missing and unusable data described above resulted in an unbalanced design, making the use of standard analysis of variance procedures impractical. Also, the usual average values obtained from the different treatments would not be comparable in that they would not be equally affected by days and subjects. Because of this, the analysis of variance was performed using a “general linear model” approach: Y observed = mean + subject + treatment + day. This approach consisted of using multiple linear regression to fit the model with the aid of the SPSS computer program.’6 The system of dummy coding was applied to each categorical variable, such as treatments, days, and subjects. In this system, a set of dummy variables is created by treating each category of a nominal variable a s a separate variable, and arbitrary scores were given for all cases depending on their presence or absence in each of the categories (such as, “1”for presence and “0” for absence in a given category). The number of dummy variables necessary to exhaust the information about a given category is equal to the number of cases of that category minus one. After assigning the dummy values, the data were analyzed by simple multiple linear regression analysis following the order of subjects, treatments, and days. A set of partial correlation coefficients was obtained. Because these partial correlation coefficients are not orthogonal, an adjustment for the intercorrelations among coded categories was made. Analysis of variance was then performed. First the variation due to subjects was isolated. Then the differences in treatment adjusted for days were tested. Finally the differences in days adjusted for treatments were tested. No signifi-
Tabk ICTlmbto-Peak of the Indlvldual Serum C o m t r a t l o r t Time curve8
Tabk Hndlvldual Peak of the Serum Concentmtlon-Tlme Curve’
Table Ill-Area Curve’
Subject
Treatment A
Treatment B
Treatment C
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
366 367 245 490 481 240 723
182 248 239 183 184 182 180
182 241 181 189 181 362
480 359 479 476
180 185 242 182
184 183 239 237
363
480
190 180 117 185 239 240
250 240 480 502 250 194
238 361 181
123 362 127
249 360 190
Mean, min
394
198
263 (267)’
Standard Deviation
122
54
98 (95)’
-
-
480 358 362 360
-
-
-
-
360
-
-
.Data for subjects 1, 6, 11, 16, 21, and 23 were omitted due to defecation.
Under the lndlvldual Serum Concentratlon-llnn
Subject
Treatment A
Treatment B
Treatment C
Subject
Treatment A
Treatment B
Treatment C
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
22.6 10.6 59.5 3.6 22.3 12.9 5.1
23.1 14.7 70.7 7.5 26.7 12.8 19.2
11.7 34.3 92.1 2.3 46.2 9.8 19.5
13 708 10 565 28 122 2 734 18 031 6 389 3 473
8 765 6 579 50 681 5 146 14 510 6 519 10 818
5 902 13 478 70 924 1 705 30 939 3 813 5 828
4.5 12.4 12.5 17.0
51.9 10.5 31.6 32.8
29.4 15.7 4.9 33.3
3 755 9 082 11 568 14 410
31 709 3 168 14 049 15 054
19 223 10 043 4 427 24 302
7.5 8.9 10.3 5.2 8.5 9.5
6.4 23.0 23.1 13.5 12.5 10.4
8.1 22.0 1.6 10.9 11.8 13.7
7300 5 309 7 699 3 645 5 914 7 629
10 675 8 399 6 596 7 558 8446 5 858
4 802 16 513 1526 8 575 7 578 7 476
3.9 37.2 6.8
34.0 55.2 99.0
9.7 49.5 21.8
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
1 458 34 861 6411
14 946 32 233 60 163
7 931 45 175 12 131
10 103
16 094
15 116 (19040)b
8 517
15589
17018(19044)b
-
-
-
-
-
-
-
-
-
23
-
24
Mean, nglmL
14.0
28.9
22.4 (27.8)
Mean
Standard Deviation
13.4
23.7
21.3 (23.2)b
Standard Deviation
Data expressed as nglmL. Data for subjects 1, 6, 11, 16, 21, and 23 were omitted due to defecation.
-
-
-
-
-
-
-
-
-
-
Data expressed as ngmL-’.rnin-’. Data for subjects 1, 6, 11, 16, 21, and 23 were omitted due to defecatlon. Journal of Pharmaceutical Sciences / 443 Vol. 76, No. 6, June 1987
cant difference among days (p > 0.05) for area under the curve, time-to-peak concentration, and peak concentration is found. Among subjects, a significant difference at p < 0.05 is found for area under the curve and peak concentration, and no significant difference a t p > 0.05 is found for time-to-peak concentration. Significant differences among treatments A, B, and C at p < 0.05 for area under the curve, time-to-peak concentration, and peak concentration are found. A null hypothesis of pA = pB = pC, where pA, pB, and $2 are the population means of treatments A, B, and C, respectively, was rejected a t p = 0.05. In addition, with a n F test, the squared multiple correlation coefficientswere significant a t p < 0.05 for area under the curve, time-to-peak concentration, and peak concentration. Therefore, post hoc comparisons were performed. Scheffe's" test for multiple comparisons was used to compare the differences between treatment A versus treatments B and C, and treatment B versus treatment C. Comparisons between treatment A versus treatment C indicate significant differences (p < 0.05) for area under the curve, time-to-peak concentration, and peak concentration. Comparisons between treatment B versus treatment C indicate no significant differences for area under the curve, time-to-peak concentration, and peak concentration at p > 0.05. As a result, the rates of absorption, which are related to the time-to-peak concentration, are significantly slower for treatment A than treatment C. The extent of absorption, which is related to peak concentration and area under the curve, are significantly smaller for treatment A than treatment C. Estimation of Pharmacokinetic Parameters-The semilogarithmic plots of mean serum promethazine concentrations versus time reveal a monoexponential decline in the promethazine serum levels after the peak, with no significant distribution phase, as depicted in Figure 1. Each set of mean serum promethazine concentration data followingtreatments A, B, and C was fitted to eq 1 by the aid of NONL1N.l8
C(t) = Ale-B1(" + A2e-B2(t) + A3e-B3("
model described in Scheme I provided a significantly better fit upon comparison with a reduced (two-exponential) model. The mean serum promethazine concentration data following oral administration of the promethazine syrup is described by
C(t) = A l e
-kat
+ A2e-ke1'
(2)
where A1 = -A2, C(t) is the serum concentration of promethazine at running time t, A1 and A2 are two pre-exponential coefficients, and k, and k,1 are first-order rate constants describing rate of absorption and elimination, respectively. Therefore, mean serum promethazine concentration data for the two suppository treatments can be expressed by the following equation:
(1)
where A2 = -(A1 + A3), C(t) is the serum concentration of promethazine at the running time t, A l , A2, and A3 are three pre-exponential coefficients, and B1, B2, and B3 are three hybrid rate constants. Based on these data, a one-compartment open model with two consecutive first-order input steps is proposed as shown in Scheme I, where kl and kz are two consecutive first-order input rate constants and k,l is the first-order elimination rate constant. The process from compartment A to compartment B may represent the dissolution of suppositories. The process from compartment B to compartment C represents a first-order absorption process; compartment C is designated as the central compartment. The
C(t) = (FD/Vd)klk2[e-k1(tl/(-k1+ k2)(-k1 + kel) + e-ka't'/(-k2 + kl)(-kz + kel) + e-kel(t'/(-kel + kl)(-kel + (3) where F is the fraction of the dose, D,absorbed, and Vd is the apparent volume of distribution for the central compartment including blood serum, and k l , k2, and k,l correspond to B2, B3, and B1, respectively. The ratio FDIVd was calculated from the average o f A l (-kel + kl)(-kel + k2)lklk2,A2 ( - k 1 + kz)(-kl + k,J/klkz, and A3 ( - k z + kl)(-kz + kel)/klkz. The final pharmacokinetic parameters calculated from A l , B1, A2, B2, A3, and B3 are presented in Table IV for treatments A, B, and C. The serum clearance, CL, was calculated from the following equation:
CLIF
=
(D/AUCo,(BW)
(4)
where AUCo-, is the area under the serum promethazine concentration-time curve from zero time to infinity and BW is the individual body weight in kg. Assume that the oral dose is the standard to which the two suppository forms are to be compared. The data from the suppositories are normalized with respect t o the oral dosage form. The apparent volume of distribution for the central compartment, Vd, can be calculated from the following equation:
Vd/F = C L / ( k e f l
(5) where k,1 is the overall first-order elimination rate constant. The serum promethazine concentration-time curves derived from NONLIN16 fits are depicted in Figure 1. The peak concentration, time-to-peak concentration, and area and relative bioavailability under the curve from zero time to infinity, calculated from the NONLIN fits, are presented in Table V. The area under the curve from zero time to infinity was estimated from
The relative bioavailability, Frel, of the promethazine suppositories was calculated using AUCO, values corrected
Scheme I
Table IV-Pharmacokinetic Parameter Estimates for Treatments A, 8, and C followlng NONLIN Fit to One-Compartment Open Model with Two Consecutive First-Order input Steps
Treatment
4, min-' x 1 0 - ~
AB Bb
6.256
Ca
12.566
k,, min-' x 1 0 - ~
-
-
7.001
k2, min-' x 1 0 - ~ 9.541
-
21.835
kl, min-' x 1 0 - ~ 1.514 2.532 1.635
FDNd, ng/mL
CUF, mUmin/kg
Vd/F, Ukg
19.46 28.74 35.51
52.66 59.48 30.95
34.78 23.51 18.93
aFollowing NONLIN fit to one-compartment open model with two consecutive first-order input steps (kl and k2). bFollowing NONLIN fit to onecompartment open model with first-order input (k.). 444 /Journal of Pharmaceutical Sciences Vol. 76, No. 6. June 1987
Table V-Summary
of Bloavallablllty Parameters for Both lndlvldual and Fltted Data
Parameter Treatment
From Individual Data AUCsi4a
A B C
10 103 (8517)b 16 094 (15589)b 19 041 (19044)*
b, 394 (122)b 198 (54) 267 (95)
,C 14 (13.4)b 28.9 (23.7)b 27.8 (23.2)b
From Nonlinear Regression AUCsi440
b,
C-
Frel'
62.8
10 327
363
12.1
65.3
100.0
15 814
181
26.2
100.0
118.3
19 113
241
25.6
120.9
Fre, a
'Relative bioavailability parameter is represented as a percentage. Numbers in parentheses represent standard deviation. for the intraindividual variation in the elimination rate Frel =
(AUCo2, suppositories)(kel, suppository) (7) (AUCo-, syrup)(k,l, syrup)
where treatment B (syrup) was chosen as the standard. Comparing the results from the model-independent approach and the model-dependent approach, which are presented in Table I, 11, 111, and V, it can be observed that the magnitude of the area under the curve, time-to-peak concentration, and peak concentration for the three treatments follows the same rank: for area under the curve (AUC),C > B > A; for time-to-peak concentration (t,,,), A > C > B; and for peak concentration (C,,,), B > C > A. However, with the analysis of variance, the difference between treatment A and treatment C is proved to be significant a t p < 0.05, but the difference between treatment B and treatment C is not significant at p > 0.05. Treatment A appeared to have slower dissolution in vivo than treatments B and C, as indicated from the mean ( 5 standard deviation) t,, values of 394 5 122, 198 2 54,and 267 ? 95 min for treatments A, B, and C, respectively (Table 11). This is also indicated from NONLIN-fitted t,,, values of 446,204,and 252 min for treatments A, B, and C, respectively (Table V). Furthermore, the rate constant kl(6.256 x min-') for treatment A is slower than the k l (12.566x min-'1 for treatment C. In addition, the kz rate constant represented in Scheme I should approximate the absorption rate from solution. Treatment C appears to have an increased kz compared with that of treatment A, which may suggest an absorption enhancement that is possibly due to the polyethylene glycol formulation. However, because of large intersubject variability, it is difficult to conclude unambiguously that the rates are significantly enhanced. The NONLIN-fitted C,,, after treatment A was 11.0 ng/ mL. This C,,, was smaller than the C,,, (26.6ng/mL) for treatment B and the C,, (24.6ng/mL) for treatment C. The same results were presented in Table I where the mean ( * standard deviation) C,,, for treatments A, B, and C were 14.0 ? 13.4,28.92 23.7,and 27.85 23.2ng/mL, respectively. As a consequence, the higher C,,, is consistent with the shorter tmax and faster k1 observed for these three treatments. After the adjustment for the intrasubject variation with the overall elimination rate constants, the relative bioavailability (47.2% of treatment B) for treatment A is smaller than the relative bioavailability (87.8% of treatment B) for treatment C obtained fron NONLIN fit of the area under the curve from time zero to time infinity. Promethazine HCl elimination kinetics after oral and rectal administrations were different. The k,l obtained for oral treatment B was 2.53 x min-' and the corresponding half life, t,,,, was 4.6h. The kel and t,,, for rectal treatment A were 1.514 x min-' and 7.6 h, respectively, and the
kel and t,,, for rectal treatment C were 1.635 x lo-' min-' and 7.1h, respectively. These elimination parameters are not exactly in agreement with those reported for promethazine by other i n v e s t i g a t o r ~ . ~ ~ ~ J ~ Promethazine polyethylene glycol suppositories a r e thought to have many practical advantages over promethazine solutions. For example, the suppositories evaluated in this study could be easily applied to pediatric use. In addition, promethazine polyethylene glycol suppositories don't melt at room temperature and prevent the problems associated with suppositories that have a cocoa butter base. Thus, refrigeration during handling and storage is no longer needed. The promethazine polyethylene glycol suppositories evaluated appeared to have rapid dissolution in vivo. They produce serum promethazine concentrations comparable to an oral promethazine solution, and the relative bioavailability is approximately the same as the oral solution. Thus, in practice, these suppositories would act as a reliable dosage form for rectal promethazine administration.
References and Notes 1. Hansson, E.; Schmiterlow, C. G. Arch. Intern. Pharmacodyn. Ther. 1961,131,309. 2. Rusiecki, W.; Wvsocka-Paruszewska. B. Diss. Pharm. Pharmacol. 1969;21 ,. 73: 3. Nadeau, G.; Sobolewski, G. Can. Med. Assoc. J . 1959,81,658; Chem. Abstr. 54,4904h. 4. Robinson, A. E.; Beaven, V. H. J . Pharm. Pharmacol. 1964,16, 342. 5. Robinson, A. E. J . Pharm. Phurmacol. 1966. 18,19. 6. Beckett, A. H.; Al-Sarrqj, S.; Essien, E. E. Xenobiotica 1975,5, 325. 7. Beckett, A. H. Xenobiotica 1975,5, 449. 8. Bornschien, I.; Pfeifer, S. Phurmazie 1979,34,750. 9. Quinn, J.; Calvert, R. J . Phurm. Pharmacol. 1976,28,Suppl. 59P. 10. Ta lor, G.; Houston, J. B. J . Pharm. Phurmmol. 1979,31,Suppl. 40% 11. Reddrop, C. J.; Riess, W.; Slater, T. F. J . Chromatogr. 1980,192, 375.
12. Mdenaar, F.;Ensing, J. G.; Bohuis, B. G.; Visaer, J. I d . J . Phurm. 1981,9,353. 13. Schwinghammer, T. L.;Juhl, R. P. Biophurm. Drug Dispos. 1984.5. 185. 14. Zaman,' R.; Honigber I. L: Francisco G. E.; Kotzan, J. A.; Steward, J. T.; Brown,%. J.; Shaf, V. P.;helsor, F. R. Biophurm. Dr Dkps.1986,7,281. 15. Wayace, J. E.; Shimek, E. L.; Stavchansky, S.; Harris, S. C. Anal. Chem. 1981,53,960. 16. Nie, N. H.; Hull, C. H.; Jenlins, J. G.; Steinbrenner, K.; Bent, D.H. Statistical Package for the Socual Science, 2nd ed.; McGraw-Hill: New York, 1975. 17. Kerlinger, F. M.; Pedhazur, E. J. Multiple Re ression in Behavioral Research; Holt, Rinehart and Winston: 8 e w York, 1973;p 128. 18. Metzler, C.M. NONLIN Computer Program; Technical Report 7292/60/7292/005, The Upjohn Co., Kalamazoo, MI,1969.
Acknowledgments This'work was supported by a grant from Alcon Laboratories, Inc. Journal of Pharmaceutical Sciences / 445 Vol. 76.No. 6, June 1987