Br. J. clin. Pharmac. (1977), 4, 185-191
PHENYTOIN CONCENTRATIONS IN MIXED, PAROTID AND SUBMANDIBULAR SALIVA AND SERUM MEASURED BY RADIOIMMUNOASSAY J.W. PAXTON & B. WHITING Department of Materia Medica, University of Glasgow, Stobhill General Hospital, Glasgow, G21 3UW
K.W. STEPHEN Department of Oral Medicine, University of Glasgow Dental School, 378 Sauchiehall Street, Glasgow, G23 JZ
I Concentrations of phenytoin in mixed, parotid and submandibular saliva and serum were determined in normal subjects after an oral dose, using a specific double antibody radioimmunoassay which requires only 20 ,l fluid 2 Semi-log concentration-time plots of phenytoin concentration in mixed saliva and serum gave good parallelism after the initial 14 h post-administration period. 3 The mean ratio of the mixed saliva: serum phenytoin concentration was 10.3% ± 1.5 (s.d.) in seven normal subjects. 4 Phenytoin concentrations found in separate parotid and submandibular salivary fractions did not differ but were significantly greater (P < 0.001) than those found in mixed saliva. 5 Phenytoin concentrations in all salivary fractions were independent of the volume of fluid produced and the degree of stimulation. 6 The rate of phenytoin secretion in the parotid and submandibular fluid was proportional to the salivary flow rate. 7 These data suggest that mixed saliva may be a suitable medium for the monitoring of phenytoin concentrations and may provide a non-invasive alternative to the direct determination of phenytoin in serum.
Introduction
Several recent studies have reported good linear correlations between phenytoin concentrations in mixed saliva and those in serum, cerebrospinal fluid, and plasma water (the ultrafiltrate or free fraction) (Bochner, Hooper, Sutherland, Eadie & Tyrer, 1974; Cook, Amerson, Poole, Lesser & O'Tuama, 1976; Paxton, Rowell, Ratcliffe, Lambie, Nanda, Melville & Johnson, 1976; Reynolds, Ziroyanis, Jones & Smith, 1976). These studies suggest that phenytoin in saliva may be in equilibrium with the free therapeutically active phenytoin in plasma water. Reynolds et aL (1976) also suggested that phenytoin therapy could be more appropriately monitored by measurement of salivary, rather than plasma, drug concentrations especially in circumstances such as renal failure when protein binding is disturbed. Obviously, if measurements of phenytoin salivary concentrations are to have any practical value in routine
monitoring of therapy, the method of collection should be as simple as possible. In the four reports mentioned, mixed saliva was collected by voiding into vials without exogenous stimulation apart from the subjects studied by Bochner et aL (1974) who chewed on a small piece of Teflon and those by Reynolds et al. (1976) who received 10 mg citric acid to stimulate salivary flow. With the exception of Reynolds et al. (1976) who reported a 6% decrease in salivary phenytoin concentration after stimulation, the effect of exogenous stimulation on salivary concentration was not investigated. All four studies were also conducted using mixed saliva, which consists of parotid, submandibular, sublingual and minor gland saliva together with oral bacteria, desquamated epithelial cells and gingival fluid. No information was provided regarding the origin of the phenytoin in mixed saliva; whether it was secreted by one
186
J.W. PAXTON, B. WHITING & K.W. STEPHEN
particular salivary gland or was exuding from the gingival tissue. If estimation of phenytoin concentration in saliva is to be accepted as a reliable index of the biologically active fraction of phenytoin, the dependency of its concentration on the type and volume of saliva produced and on degree of stimulation and salivary flow rate must be determined. It is the aim of this study to investigate these factors and also to investigate whether changes in serum concentrations are reflected by corresponding changes in salivary concentrations.
Methods The subjects in this study were freely consenting healthy male laboratory staff aged 20-40 years, who received oral doses of phenytoin sodium BP and samples were taken as follows: 1 Seven subjects received a 100 or 300 mg oral dose of phenytoin. Two subjects received both doses. Simultaneous blood and mixed saliva samples were obtained at hourly intervals for the first 12 h and then thrice daily where possible over the subsequent 3 to 5 days. 2 Three subjects received an oral dose of 300 mg phenytoin. Twenty-four hours later, blood, mixed, parotid and submandibular saliva fractions were collected at hourly intervals for 4 hours. The mixed saliva was collected after chewing washed rubber bands for several minutes (Shannon, 1962). For the parotid and submandibular fractions, five sequential collections of 1 ml each were obtained where possible at each hourly interval, after stimulation of salivary flow rate by application to the dorsum of the tongue of 1 ml of 10% citric acid solution every 30 s (Stephen & Speirs, 1976). 3 One subject received an oral dose of 500 mg phenytoin. Fourteen hours later a series of five mixed, five parotid and five submandibular collections were obtained after increasing degrees of stimulation of salivary flow rate by application, as described above of 1 ml of a solution containing either 0.1%, 0.5%, 1%, 5% or 10% citric acid. A second set of samples was obtained by repeating this procedure 2-3 h later. Blood samples were allowed to clot, serum separated by centrifugation and stored at -20o C. Parotid and submandibular saliva was collected by techniques described by Stephen & Speirs (1976). All saliva specimens were frozen and stored at -20 C until required for analysis. Before assaying, the mixed saliva samples were centrifuged at 2500g for 15 min to remove any debris and 20 gl of the clear supematant used for the assay.
Radioimmunoassay Phenytoin concentrations in serum and saliva samples were measured by a specific double antibody radioimmunoassay (RIA) using an antiserum raised against phenytoin-valerate-BSA (Paxton, Rowell & Ratcliffe, 1976) and a [12511-labelled derivative of phenytoin (Paxton, Rowell & Ratcliffe, unpublished observations). In outline the method involves incubation of 20 Ml of sample with the appropriate dilution of antiphenytoin antiserum, donkey anti-rabbit precipitating serum (Wellcome Reagents Ltd, Beckenham, England) and trace amounts of [(12SI1-phenytoin for 1 h at 370 C. After centrifugation of the incubate at 40 C for 30 min at 2500g the supernatants were aspirated and precipitates (containing the antibody bound [12s5]j-phenytoin counted for 10-20s on an automatic 7-counter. Duplicate determinations were made on each sample. Assay precision was determined using three quality control samples made up in normal human serum or saliva at concentrations within the clinically relevant ranges (serum 2-200,umol/I and saliva 0.02-20,umol/1). Inter-assay precision varied between 3 and 6% and 4 and 9% (C.V.) for the serum and saliva assays respectively. These were calculated from the quality control sample values determined in at least twelve consecutive assays. The detection limit of the assay was approximately 0.4 nmoVl and there was a good correlation with a conventional gas-chromatographic method (r = 0.98 with gradient = 0.95). The time taken to carry out the assay, count and calculate the results on up to 100 samples was less than 3 hours.
Results
After simultaneous collection of mixed saliva and blood, semi-log concentration-time plots of phenytoin in saliva and serum showed good parallelism after the initial 14 h in all seven subjects at various doses as exemplified in Figure 1. During the initial 14 h period variations in phenytoin concentrations were observed in both mixed saliva and serum which were not due to experimental error. These variations could not be attributed to periods of fluid or food intake and did not seem to follow a regular pattern. Parallelism between the concentrations of phenytoin in mixed saliva and serum was demonstrated by the similar half-lives in these fluids. Half-life values shown in Table 1 were calculated from the con-
PHENYTOIN CONCENTRATIONS IN SEPARATE SALIVARY FRACTIONS
Figure 1 Phenytoin concentrations in the serum (* a) and saliva (o, o) as a function of time after oral administration of phenytoin capsules (-, o 300 mg; a, o 100 mg) to one subject (JWP).
centrations of phenytoin in saliva and serum from 30-100 h after administration of the phenytoin dose. Elimination of phenytoin appeared monoexponential within this period. However, although the forms of the serum and mixed salivary concentration-time curves were very similar, significantly greater variations were observed in the mixed salivary levels of phenytoin than in the serum levels. This was not entirely due to the greater coefficient of variation for the salivary phenytoin assay compared with the serum assay. No significant trends in the mixed saliva: serum Table 1
concentration ratio were detected, either with time or with serum concentration. Concentrations of phenytoin were also determined in five sequential 1 ml samples from separate submandibular and parotid salivary secretions obtained from three subjects 24 h after administration of phenytoin (300 mg). These concentrations together with the corresponding concentrations observed in mixed saliva and serum are shown in Table 2. Submandibular samples were not obtained from one subject (JGR) due to a broad lingual frenum which precluded the use of the submandibular collection device. The concentrations of phenytoin in the parotid and submandibular fluids were not significantly different but were significantly higher (P < 0.001 ) than the concentrations in the mixed salivary secretions when compared by Student's t-test on paired observations. A similar comparison of paired values for the percentage ratio of mixed salivary phenytoin concentration: serum concentration (mnean ± s.d. 13.1% ± 2.7, n = 12) and the percentage ratio for parotid and submandibular phenytoin concentration: serum concentration (mean ± s.d. 14.5% ± 1.2, n = 20) were also significantly different (P < 0.001). No significant trends in the phenytoin concentrations in either the parotid or submandibular fractions were observed in the five sequential 1 ml samples obtained over a period of 5-10 min (Table 2). The mean ± s.d. value (6.0% ± 2.0, n = 12) for the coefficients of variation of repeated analysis (n = 4) of the single mixed saliva samples, and the mean ± s.d. value (6.3% ± 2.6, n = 20) of the
Phenytoin mixed salivary serum concentration ratios and half-lives in normal subjects Dose
Subject
(mg)
JWP JWP JGR JGR FJR
100 300 100 300 300 300 300 300 100
VM CM
DJG DS Mean n
187
Mixed salivary total serum concentration n s.d. Range % mean ratio
10.1 9.7 9.9 12.9 8.9 9.1 11.0 12.1 8.7 10.3
1.7
14
1.9 1.1
11 13
3.3 1.5 2.8 1.8 1.8 1.3 1.5
7 11 8 8 7 14 9
6.8-13.8 6.5-12.7 8.6-11.7 7.7-16.9 6.8-10.8 6.2-12.4 7.8-13.0 7.0-16.0 6.8-12.0 6.2-16.9
Saliva
Serum
Ty.2 (h)
T %2 (h)
15.4 15.0 9.7 10.0 17.5 10.5 *18.2 11.0 9.0 12.3 3.2 (n=8)
14.8 14.1 9.7
±
9.8 17.2 12.0
11.3 8.3 12.2 3.0 (n=8) ±
number of combined saliva and serum samples
Ty, half-life of disappearance of phenytoin ' This value has been omitted in the calculation of mean saliva half-life as the corresponding half-life for phenytoin disappearance in serum was not available. The mean half-lives for phenytoin disappearance in serum and saliva were not significantly different
.o
'CtLoL)D tJ CCv)CLO a._
cs qt 00
0) N LO r-+l +1 +1 +1
x
+l +l +l +l
E
N (0000 r0'- r% (C CD ~CY)) I- I" a
,^
0
Le)D oCD LOf Cs N
DOD o 00 LDo
l
'q dooo oo66 0) CD
0)00 C)N
_
.xt
r% C')IV r'(0(L)0L
0
00 oo66 -(D ( C.') C')N O N LCDVCVC)Cl0q r%~(D C Cooo oooo
6
*_
0
.00 -6
N
0 r CN C,,)0
E
t
0 a,
C; C; C;O C; C") LO tt o o C; 0 C; 00 0 0 C
C; C;C')
CD
.
.
CD)
.
0
._
c
000~0 C' to
00'- CD C')
(D 000'-LO
(6 4 c6 (6
0 0 C 0
C
1%r co c0
-H +4 +4
+l
_N to t
(D C') 00 00 +l +4
+ l-H
0:N 00 0 co
c
0t
LO CD C' ( CD N0C0') (D IN
a,
V L.
._3 a
C; C; LO
.0
LO Lc (1 )
C
E
C; (W
co
6666d
C
a, m
Or% C
C^
c
O-N
rIlr~r'CD
L
) C,,
,co I G
x
E
8 0.
W U)
r
00 CI) 0) 0
o
o t)0 C) aw CO % co co
0
r6 666e 6666ACC eQ0 rtODN 6 6 6 d Or'.(C;C C LO IT
0.
c ._
000btDE v N N>
0) N 0 N N4
100 N N
0 a
0fC OD00 0O% 00 '- 0 l% (N CY)
0
+1 +1
+l
.co
N _% CD CD 00 0CD
E0a,
E C V
>
00oo _ C'
c
E
co(m Lo CD N C) C) o0 o
N CD > co
cuC
-Ca' 0.' O n
E t
m
._
a,
CD Ca
.Wa,
c c
v co N co
qt CD CD a)
+,+, +4 +4
+
a
N
cotoo
.5s
a) r-
a
_- _. CV) Cl 000r +l +l C+l c-H O CD 10 0
.a>C ,A
C.-E a,
co ot CDo co
4) C(
+1 +4 -H
qt00
0 co g
00 CY) 00
CD 0) (D q
-H
-H +i +1
+, +4 +4 +H
+H
C(00 LC'C)N N
CDCN C')f% uC
C0)e N co
UZi li li
C
Eo0 -C ._. 4.' .4
o o' t LO O P, N N N N
t LoO%
N N N N
Lo D f%00 N C N N
CL
Co
I.-
3:
C) ci
2
le
cc
t
~0
PHENYTOIN CONCENTRATIONS IN SEPARATE SALIVARY FRACTIONS
189
-
O 25
E
a)
.C 20-
0~~~~~~~~~~~~~
c L-
C
±
gco 1.0
E
C:
C0. ~.0
0 ._
ax)
0
5.0 10 05 1.0 % concentration citric acid for stimulation Gi
Figure 2 Effect of increasing stimulation of salivary flow rate on the concentration of phenytoin in mixed saliva in one subject (BW). Each point represents the mean ± s.d. of four analyses. (e) first and (a) second experiment.
coefficients of variation obtained from the five sequential submandibular and parotid samples were not significantly different when compared by Student's t-test. Similarly, no significant trends and similar coefficients of variation were observed in nine sequential mixed salivary samples obtained over a short period of time without exogenous stimulation from two subjects during the concentration-time course experiments. These results are shown in Table 3. Further investigation of the dependency of salivary phenytoin concentration on flow rate was Table 3 Phenytoin concentrations in sequential mixed saliva samples. These samples were obtained over a short period of time without stimulation from two subjects during a concentration-time course experiment
Sample number 1
2 3 4 5 6 7 8 9 Mean s.d. C.V.
Phenytoin concentrations (umol/l) Subject 1 Subject 2 1.35 1.38 1.31 1.28 1.40 1.38 1.27 1.48 1.23 1.34 0.08 5.7%
C.V. Coefficient of variation 13
E 0.a
0.65 0.67 0.58 0.64 0.67 0.59 0.57 0.62 0.58 0.04 6.8%
co
0. V
0
0~
0.5
1.0
1.5
2.0
Parotid saliva flow rate (ml/min per gland) Figure 3 Correlation between the rate of phenytoin secretion in the parotid fluid and flow rate of parotid saliva in one subject (BW). The circles (o) represent the first experiment and have a linear regression equation, y = 1.956x - 0.044 with correlation coefficient (r) = 0.9999. The triangles (Lx) represent a repeat experiment 2 h later and have a linear regression equation, y = 1.688x + 0.024 with r = 0.9994. The gradients of the lines represent the mean phenytoin concentration (nmol/ml) in the parotid fluid during the time of the experiment.
carried out on one subject. As accurate estimation of mixed salivary flow rate is exceedingly difficult this was not attempted, but Figure 2 illustrates the phenytoin concentrations obtained after stimulation by increasing concentrations of citric acid. These results indicate that the concentration of phenytoin in mixed saliva is independent of the degree of stimulation. For the individual parotid and submandibular saliva fractions, flow rate measurement proved possible by collection of a measured volume in a measured time interval at the different degrees of stimulation. Comparison of the rate of parotid salivary phenytoin secretion and parotid fluid flow rate gave a good linear correlation (r = 0.999) with the gradient equal to the mean phenytoin concentration in the fluid at that time. Repetition of this experiment 2 h later gave a similar good linear correlation (r = 0.999) but with a lower gradient corresponding to a lower phenytoin concentration at that time (Figure 3). Similar, but less significant straight line correlations, were obtained between flow rate and phenytoin secretion by the submandibular gland (r = 0.94 and 0.74).
190
J.W. PAXTON, B. WHITING & K.W. STEPHEN
Discussion
If estimation of phenytoin concentration in mixed saliva is to be accepted as a reliable index of the therapeutically active fraction, the following points must be elucidated: (a) the origin of the phenytoin in'the mixed salivary secretions must be established, (b) whether changes in serum phenytoin concentrations are reflected by corresponding changes in mixed salivary concentrations at relatively constant protein-binding and (c) whether salivary phenytoin concentrations are independent of the type and volume of saliva produced and of salivary flow rate. Our results indicate that phenytoin is a true salivary secretion, being found in similar concentrations in both the parotid and submandibular fluid. The results of the semi-log concentrationtime plots of phenytoin concentrations in mixed saliva and serum also indicate that changes in serum phenytoin concentrations are rapidly reflected in the mixed saliva. Furthermore we have shown that phenytoin concentrations in mixed saliva are independent of the volume of fluid produced and are also independent of the degree of stimulation. The excellent linear correlations obtained on comparison of the rate of parotid phenytoin secretion and fluid flow rate indicate that the phenytoin secretion is directly proportional to flow rate and thus its concentration is independent of the degree of stimulation of the parotid gland. The lower correlation coefficients obtained on similar comparison of the values obtained from the submandibular gland are partly due to the difficulty experienced in obtaining higher rates of salivary flow. After stimulation with 0.5% citric acid solution, there was only a marginal increase in flow rate at higher citric acid concentrations. In the second experiment flow rate actually diminished with increasing stimulation by citric acid. These observations were possibly due to partial fatigue of the submandibular gland.
These results, together with reports of saliva phenytoin concentrations comparable to cerebrospinal fluid and free phenytoin concentrations (Bochner et al., 1974; Cook et al., 1976; Paxton et al., 1976), suggest that it is the free drug which is in equilibrium between serum and saliva. The ease of diffusion across the lipid membranes will depend on lipid solubility characteristics and the degree of ionization of the drug. Since phenytoin has an ionization constant (pKa) of 9.2 and the pH of plasma is 7.4 and that of saliva between 6-8, depending on gland type and flow rate, very little drug should be ionized in either fluid. Also phenytoin can be considered to possess good lipid solubility characteristics since it is soluble in alcohol, chloroform and ether. Thus the movement of phenytoin across lipid membranes would not be hindered. Distribution at equilibrium which is dependent on serum and salivary pHs will ensure that salivary concentrations will be the same as that of free phenytoin in plasma water, assuming that saliva is essentially an albumin-free fluid. The independence of salivary concentration on volume produced and flow rate of fluid suggests that active transport of phenytoin is not involved in the transfer of the drug molecules across lipid membranes in the salivary glands. It appears likely that a simple filtration or diffusion process is involved which would indicate that mixed saliva is indeed a suitable fluid for phenytoin estimations. In addition these salivary estimations might provide a more accurate index of the biologically active fraction of the drug especially in situations where the serum protein-binding of the drug is abnormal or where there is interaction with other drugs or endogenous substances. We thank Professor A. Goldberg and Dr J.G. Ratcliffe for helpful discussion and comment and the volunteers without whom these investigations would not have been possible. Correspondence to be addressed to Dr J.W. Paxton.
References BOCHNER, F., HOOPER, W.D., SUTHERLAND, J.M.,
EADIE, MJ. & TYRER, J.H. (1974). Diphenylhydantoin concentrations in saliva. Arch. NeuroL, 31, S7-59. COOK, C.E., AMERSON, E. POOLE, W.E., LESSER, P. & O'TUAMA, L. (1976). Phenytoin and phenobarbital concentrations in saliva and plasma measured by radioimmunoassay. ClAn. Pharmac. 7her., 18, 742-747. PAXTON, J.W. ROWELL, F.J. & RATCLIFFE, J.G. (1976). Production and characterization of antisera to diphenylhydantoin suitable for radioimmunoassay. J.
ImmunoL Methods, 10, 317 327. PAXTON, J.W. ROWELL, F.J.. RATCLIFFE, J.G., LAMBIE, D.G.. NANDA, R., MELVILLE, I.D. &
JOHNSON, R.H. (1977). Salivary phenytoin radioimmunoassay: a simple method for the assessment of non-protein bound drug concentrations. Eur. J. clin. Pharmac. 11,71-74. REYNOLDS, F.. ZIROYANIS, P.N., JONES, N.F.
SMITH, S.E. (1976). Salivary phenytoin concentrations in epilepsy and in chronic renal failure. Lancet, ii, 384-386.
PHENYTOIN CONCENTRATIONS IN SEPARATE SALIVARY FRACTIONS SHANNON, I.L. (1972). Parotid fluid flow rate as related to the whole saliva volume. Archs. Oral Biol., 7, 391-394. STEPHEN, K.W. & SPEIRS. C.F. (1976). Methods for collecting individual components of mixed saliva: the
191
relevance to clinical pharmacology. Br. J. clin. Pharmac., 3, 315-319.
(Received July IS, 1976)
PHENYTOIN CONCENTRATIONS IN MIXED, PAROTID AND SUBMANDIBULAR SALIVA AND SERUM MEASURED BY RADIOIMMUNOASSAY J.W. PAXTON & B. WHITING Department of Materia Medica, University of Glasgow, Stobhill General Hospital, Glasgow, G21 3UW
K.W. STEPHEN Department of Oral Medicine, University of Glasgow Dental School, 378 Sauchiehall Street, Glasgow, G23 JZ
I Concentrations of phenytoin in mixed, parotid and submandibular saliva and serum were determined in normal subjects after an oral dose, using a specific double antibody radioimmunoassay which requires only 20 ,l fluid 2 Semi-log concentration-time plots of phenytoin concentration in mixed saliva and serum gave good parallelism after the initial 14 h post-administration period. 3 The mean ratio of the mixed saliva: serum phenytoin concentration was 10.3% ± 1.5 (s.d.) in seven normal subjects. 4 Phenytoin concentrations found in separate parotid and submandibular salivary fractions did not differ but were significantly greater (P < 0.001) than those found in mixed saliva. 5 Phenytoin concentrations in all salivary fractions were independent of the volume of fluid produced and the degree of stimulation. 6 The rate of phenytoin secretion in the parotid and submandibular fluid was proportional to the salivary flow rate. 7 These data suggest that mixed saliva may be a suitable medium for the monitoring of phenytoin concentrations and may provide a non-invasive alternative to the direct determination of phenytoin in serum.
Introduction
Several recent studies have reported good linear correlations between phenytoin concentrations in mixed saliva and those in serum, cerebrospinal fluid, and plasma water (the ultrafiltrate or free fraction) (Bochner, Hooper, Sutherland, Eadie & Tyrer, 1974; Cook, Amerson, Poole, Lesser & O'Tuama, 1976; Paxton, Rowell, Ratcliffe, Lambie, Nanda, Melville & Johnson, 1976; Reynolds, Ziroyanis, Jones & Smith, 1976). These studies suggest that phenytoin in saliva may be in equilibrium with the free therapeutically active phenytoin in plasma water. Reynolds et aL (1976) also suggested that phenytoin therapy could be more appropriately monitored by measurement of salivary, rather than plasma, drug concentrations especially in circumstances such as renal failure when protein binding is disturbed. Obviously, if measurements of phenytoin salivary concentrations are to have any practical value in routine
monitoring of therapy, the method of collection should be as simple as possible. In the four reports mentioned, mixed saliva was collected by voiding into vials without exogenous stimulation apart from the subjects studied by Bochner et aL (1974) who chewed on a small piece of Teflon and those by Reynolds et al. (1976) who received 10 mg citric acid to stimulate salivary flow. With the exception of Reynolds et al. (1976) who reported a 6% decrease in salivary phenytoin concentration after stimulation, the effect of exogenous stimulation on salivary concentration was not investigated. All four studies were also conducted using mixed saliva, which consists of parotid, submandibular, sublingual and minor gland saliva together with oral bacteria, desquamated epithelial cells and gingival fluid. No information was provided regarding the origin of the phenytoin in mixed saliva; whether it was secreted by one
186
J.W. PAXTON, B. WHITING & K.W. STEPHEN
particular salivary gland or was exuding from the gingival tissue. If estimation of phenytoin concentration in saliva is to be accepted as a reliable index of the biologically active fraction of phenytoin, the dependency of its concentration on the type and volume of saliva produced and on degree of stimulation and salivary flow rate must be determined. It is the aim of this study to investigate these factors and also to investigate whether changes in serum concentrations are reflected by corresponding changes in salivary concentrations.
Methods The subjects in this study were freely consenting healthy male laboratory staff aged 20-40 years, who received oral doses of phenytoin sodium BP and samples were taken as follows: 1 Seven subjects received a 100 or 300 mg oral dose of phenytoin. Two subjects received both doses. Simultaneous blood and mixed saliva samples were obtained at hourly intervals for the first 12 h and then thrice daily where possible over the subsequent 3 to 5 days. 2 Three subjects received an oral dose of 300 mg phenytoin. Twenty-four hours later, blood, mixed, parotid and submandibular saliva fractions were collected at hourly intervals for 4 hours. The mixed saliva was collected after chewing washed rubber bands for several minutes (Shannon, 1962). For the parotid and submandibular fractions, five sequential collections of 1 ml each were obtained where possible at each hourly interval, after stimulation of salivary flow rate by application to the dorsum of the tongue of 1 ml of 10% citric acid solution every 30 s (Stephen & Speirs, 1976). 3 One subject received an oral dose of 500 mg phenytoin. Fourteen hours later a series of five mixed, five parotid and five submandibular collections were obtained after increasing degrees of stimulation of salivary flow rate by application, as described above of 1 ml of a solution containing either 0.1%, 0.5%, 1%, 5% or 10% citric acid. A second set of samples was obtained by repeating this procedure 2-3 h later. Blood samples were allowed to clot, serum separated by centrifugation and stored at -20o C. Parotid and submandibular saliva was collected by techniques described by Stephen & Speirs (1976). All saliva specimens were frozen and stored at -20 C until required for analysis. Before assaying, the mixed saliva samples were centrifuged at 2500g for 15 min to remove any debris and 20 gl of the clear supematant used for the assay.
Radioimmunoassay Phenytoin concentrations in serum and saliva samples were measured by a specific double antibody radioimmunoassay (RIA) using an antiserum raised against phenytoin-valerate-BSA (Paxton, Rowell & Ratcliffe, 1976) and a [12511-labelled derivative of phenytoin (Paxton, Rowell & Ratcliffe, unpublished observations). In outline the method involves incubation of 20 Ml of sample with the appropriate dilution of antiphenytoin antiserum, donkey anti-rabbit precipitating serum (Wellcome Reagents Ltd, Beckenham, England) and trace amounts of [(12SI1-phenytoin for 1 h at 370 C. After centrifugation of the incubate at 40 C for 30 min at 2500g the supernatants were aspirated and precipitates (containing the antibody bound [12s5]j-phenytoin counted for 10-20s on an automatic 7-counter. Duplicate determinations were made on each sample. Assay precision was determined using three quality control samples made up in normal human serum or saliva at concentrations within the clinically relevant ranges (serum 2-200,umol/I and saliva 0.02-20,umol/1). Inter-assay precision varied between 3 and 6% and 4 and 9% (C.V.) for the serum and saliva assays respectively. These were calculated from the quality control sample values determined in at least twelve consecutive assays. The detection limit of the assay was approximately 0.4 nmoVl and there was a good correlation with a conventional gas-chromatographic method (r = 0.98 with gradient = 0.95). The time taken to carry out the assay, count and calculate the results on up to 100 samples was less than 3 hours.
Results
After simultaneous collection of mixed saliva and blood, semi-log concentration-time plots of phenytoin in saliva and serum showed good parallelism after the initial 14 h in all seven subjects at various doses as exemplified in Figure 1. During the initial 14 h period variations in phenytoin concentrations were observed in both mixed saliva and serum which were not due to experimental error. These variations could not be attributed to periods of fluid or food intake and did not seem to follow a regular pattern. Parallelism between the concentrations of phenytoin in mixed saliva and serum was demonstrated by the similar half-lives in these fluids. Half-life values shown in Table 1 were calculated from the con-
PHENYTOIN CONCENTRATIONS IN SEPARATE SALIVARY FRACTIONS
Figure 1 Phenytoin concentrations in the serum (* a) and saliva (o, o) as a function of time after oral administration of phenytoin capsules (-, o 300 mg; a, o 100 mg) to one subject (JWP).
centrations of phenytoin in saliva and serum from 30-100 h after administration of the phenytoin dose. Elimination of phenytoin appeared monoexponential within this period. However, although the forms of the serum and mixed salivary concentration-time curves were very similar, significantly greater variations were observed in the mixed salivary levels of phenytoin than in the serum levels. This was not entirely due to the greater coefficient of variation for the salivary phenytoin assay compared with the serum assay. No significant trends in the mixed saliva: serum Table 1
concentration ratio were detected, either with time or with serum concentration. Concentrations of phenytoin were also determined in five sequential 1 ml samples from separate submandibular and parotid salivary secretions obtained from three subjects 24 h after administration of phenytoin (300 mg). These concentrations together with the corresponding concentrations observed in mixed saliva and serum are shown in Table 2. Submandibular samples were not obtained from one subject (JGR) due to a broad lingual frenum which precluded the use of the submandibular collection device. The concentrations of phenytoin in the parotid and submandibular fluids were not significantly different but were significantly higher (P < 0.001 ) than the concentrations in the mixed salivary secretions when compared by Student's t-test on paired observations. A similar comparison of paired values for the percentage ratio of mixed salivary phenytoin concentration: serum concentration (mnean ± s.d. 13.1% ± 2.7, n = 12) and the percentage ratio for parotid and submandibular phenytoin concentration: serum concentration (mean ± s.d. 14.5% ± 1.2, n = 20) were also significantly different (P < 0.001). No significant trends in the phenytoin concentrations in either the parotid or submandibular fractions were observed in the five sequential 1 ml samples obtained over a period of 5-10 min (Table 2). The mean ± s.d. value (6.0% ± 2.0, n = 12) for the coefficients of variation of repeated analysis (n = 4) of the single mixed saliva samples, and the mean ± s.d. value (6.3% ± 2.6, n = 20) of the
Phenytoin mixed salivary serum concentration ratios and half-lives in normal subjects Dose
Subject
(mg)
JWP JWP JGR JGR FJR
100 300 100 300 300 300 300 300 100
VM CM
DJG DS Mean n
187
Mixed salivary total serum concentration n s.d. Range % mean ratio
10.1 9.7 9.9 12.9 8.9 9.1 11.0 12.1 8.7 10.3
1.7
14
1.9 1.1
11 13
3.3 1.5 2.8 1.8 1.8 1.3 1.5
7 11 8 8 7 14 9
6.8-13.8 6.5-12.7 8.6-11.7 7.7-16.9 6.8-10.8 6.2-12.4 7.8-13.0 7.0-16.0 6.8-12.0 6.2-16.9
Saliva
Serum
Ty.2 (h)
T %2 (h)
15.4 15.0 9.7 10.0 17.5 10.5 *18.2 11.0 9.0 12.3 3.2 (n=8)
14.8 14.1 9.7
±
9.8 17.2 12.0
11.3 8.3 12.2 3.0 (n=8) ±
number of combined saliva and serum samples
Ty, half-life of disappearance of phenytoin ' This value has been omitted in the calculation of mean saliva half-life as the corresponding half-life for phenytoin disappearance in serum was not available. The mean half-lives for phenytoin disappearance in serum and saliva were not significantly different
.o
'CtLoL)D tJ CCv)CLO a._
cs qt 00
0) N LO r-+l +1 +1 +1
x
+l +l +l +l
E
N (0000 r0'- r% (C CD ~CY)) I- I" a
,^
0
Le)D oCD LOf Cs N
DOD o 00 LDo
l
'q dooo oo66 0) CD
0)00 C)N
_
.xt
r% C')IV r'(0(L)0L
0
00 oo66 -(D ( C.') C')N O N LCDVCVC)Cl0q r%~(D C Cooo oooo
6
*_
0
.00 -6
N
0 r CN C,,)0
E
t
0 a,
C; C; C;O C; C") LO tt o o C; 0 C; 00 0 0 C
C; C;C')
CD
.
.
CD)
.
0
._
c
000~0 C' to
00'- CD C')
(D 000'-LO
(6 4 c6 (6
0 0 C 0
C
1%r co c0
-H +4 +4
+l
_N to t
(D C') 00 00 +l +4
+ l-H
0:N 00 0 co
c
0t
LO CD C' ( CD N0C0') (D IN
a,
V L.
._3 a
C; C; LO
.0
LO Lc (1 )
C
E
C; (W
co
6666d
C
a, m
Or% C
C^
c
O-N
rIlr~r'CD
L
) C,,
,co I G
x
E
8 0.
W U)
r
00 CI) 0) 0
o
o t)0 C) aw CO % co co
0
r6 666e 6666ACC eQ0 rtODN 6 6 6 d Or'.(C;C C LO IT
0.
c ._
000btDE v N N>
0) N 0 N N4
100 N N
0 a
0fC OD00 0O% 00 '- 0 l% (N CY)
0
+1 +1
+l
.co
N _% CD CD 00 0CD
E0a,
E C V
>
00oo _ C'
c
E
co(m Lo CD N C) C) o0 o
N CD > co
cuC
-Ca' 0.' O n
E t
m
._
a,
CD Ca
.Wa,
c c
v co N co
qt CD CD a)
+,+, +4 +4
+
a
N
cotoo
.5s
a) r-
a
_- _. CV) Cl 000r +l +l C+l c-H O CD 10 0
.a>C ,A
C.-E a,
co ot CDo co
4) C(
+1 +4 -H
qt00
0 co g
00 CY) 00
CD 0) (D q
-H
-H +i +1
+, +4 +4 +H
+H
C(00 LC'C)N N
CDCN C')f% uC
C0)e N co
UZi li li
C
Eo0 -C ._. 4.' .4
o o' t LO O P, N N N N
t LoO%
N N N N
Lo D f%00 N C N N
CL
Co
I.-
3:
C) ci
2
le
cc
t
~0
PHENYTOIN CONCENTRATIONS IN SEPARATE SALIVARY FRACTIONS
189
-
O 25
E
a)
.C 20-
0~~~~~~~~~~~~~
c L-
C
±
gco 1.0
E
C:
C0. ~.0
0 ._
ax)
0
5.0 10 05 1.0 % concentration citric acid for stimulation Gi
Figure 2 Effect of increasing stimulation of salivary flow rate on the concentration of phenytoin in mixed saliva in one subject (BW). Each point represents the mean ± s.d. of four analyses. (e) first and (a) second experiment.
coefficients of variation obtained from the five sequential submandibular and parotid samples were not significantly different when compared by Student's t-test. Similarly, no significant trends and similar coefficients of variation were observed in nine sequential mixed salivary samples obtained over a short period of time without exogenous stimulation from two subjects during the concentration-time course experiments. These results are shown in Table 3. Further investigation of the dependency of salivary phenytoin concentration on flow rate was Table 3 Phenytoin concentrations in sequential mixed saliva samples. These samples were obtained over a short period of time without stimulation from two subjects during a concentration-time course experiment
Sample number 1
2 3 4 5 6 7 8 9 Mean s.d. C.V.
Phenytoin concentrations (umol/l) Subject 1 Subject 2 1.35 1.38 1.31 1.28 1.40 1.38 1.27 1.48 1.23 1.34 0.08 5.7%
C.V. Coefficient of variation 13
E 0.a
0.65 0.67 0.58 0.64 0.67 0.59 0.57 0.62 0.58 0.04 6.8%
co
0. V
0
0~
0.5
1.0
1.5
2.0
Parotid saliva flow rate (ml/min per gland) Figure 3 Correlation between the rate of phenytoin secretion in the parotid fluid and flow rate of parotid saliva in one subject (BW). The circles (o) represent the first experiment and have a linear regression equation, y = 1.956x - 0.044 with correlation coefficient (r) = 0.9999. The triangles (Lx) represent a repeat experiment 2 h later and have a linear regression equation, y = 1.688x + 0.024 with r = 0.9994. The gradients of the lines represent the mean phenytoin concentration (nmol/ml) in the parotid fluid during the time of the experiment.
carried out on one subject. As accurate estimation of mixed salivary flow rate is exceedingly difficult this was not attempted, but Figure 2 illustrates the phenytoin concentrations obtained after stimulation by increasing concentrations of citric acid. These results indicate that the concentration of phenytoin in mixed saliva is independent of the degree of stimulation. For the individual parotid and submandibular saliva fractions, flow rate measurement proved possible by collection of a measured volume in a measured time interval at the different degrees of stimulation. Comparison of the rate of parotid salivary phenytoin secretion and parotid fluid flow rate gave a good linear correlation (r = 0.999) with the gradient equal to the mean phenytoin concentration in the fluid at that time. Repetition of this experiment 2 h later gave a similar good linear correlation (r = 0.999) but with a lower gradient corresponding to a lower phenytoin concentration at that time (Figure 3). Similar, but less significant straight line correlations, were obtained between flow rate and phenytoin secretion by the submandibular gland (r = 0.94 and 0.74).
190
J.W. PAXTON, B. WHITING & K.W. STEPHEN
Discussion
If estimation of phenytoin concentration in mixed saliva is to be accepted as a reliable index of the therapeutically active fraction, the following points must be elucidated: (a) the origin of the phenytoin in'the mixed salivary secretions must be established, (b) whether changes in serum phenytoin concentrations are reflected by corresponding changes in mixed salivary concentrations at relatively constant protein-binding and (c) whether salivary phenytoin concentrations are independent of the type and volume of saliva produced and of salivary flow rate. Our results indicate that phenytoin is a true salivary secretion, being found in similar concentrations in both the parotid and submandibular fluid. The results of the semi-log concentrationtime plots of phenytoin concentrations in mixed saliva and serum also indicate that changes in serum phenytoin concentrations are rapidly reflected in the mixed saliva. Furthermore we have shown that phenytoin concentrations in mixed saliva are independent of the volume of fluid produced and are also independent of the degree of stimulation. The excellent linear correlations obtained on comparison of the rate of parotid phenytoin secretion and fluid flow rate indicate that the phenytoin secretion is directly proportional to flow rate and thus its concentration is independent of the degree of stimulation of the parotid gland. The lower correlation coefficients obtained on similar comparison of the values obtained from the submandibular gland are partly due to the difficulty experienced in obtaining higher rates of salivary flow. After stimulation with 0.5% citric acid solution, there was only a marginal increase in flow rate at higher citric acid concentrations. In the second experiment flow rate actually diminished with increasing stimulation by citric acid. These observations were possibly due to partial fatigue of the submandibular gland.
These results, together with reports of saliva phenytoin concentrations comparable to cerebrospinal fluid and free phenytoin concentrations (Bochner et al., 1974; Cook et al., 1976; Paxton et al., 1976), suggest that it is the free drug which is in equilibrium between serum and saliva. The ease of diffusion across the lipid membranes will depend on lipid solubility characteristics and the degree of ionization of the drug. Since phenytoin has an ionization constant (pKa) of 9.2 and the pH of plasma is 7.4 and that of saliva between 6-8, depending on gland type and flow rate, very little drug should be ionized in either fluid. Also phenytoin can be considered to possess good lipid solubility characteristics since it is soluble in alcohol, chloroform and ether. Thus the movement of phenytoin across lipid membranes would not be hindered. Distribution at equilibrium which is dependent on serum and salivary pHs will ensure that salivary concentrations will be the same as that of free phenytoin in plasma water, assuming that saliva is essentially an albumin-free fluid. The independence of salivary concentration on volume produced and flow rate of fluid suggests that active transport of phenytoin is not involved in the transfer of the drug molecules across lipid membranes in the salivary glands. It appears likely that a simple filtration or diffusion process is involved which would indicate that mixed saliva is indeed a suitable fluid for phenytoin estimations. In addition these salivary estimations might provide a more accurate index of the biologically active fraction of the drug especially in situations where the serum protein-binding of the drug is abnormal or where there is interaction with other drugs or endogenous substances. We thank Professor A. Goldberg and Dr J.G. Ratcliffe for helpful discussion and comment and the volunteers without whom these investigations would not have been possible. Correspondence to be addressed to Dr J.W. Paxton.
References BOCHNER, F., HOOPER, W.D., SUTHERLAND, J.M.,
EADIE, MJ. & TYRER, J.H. (1974). Diphenylhydantoin concentrations in saliva. Arch. NeuroL, 31, S7-59. COOK, C.E., AMERSON, E. POOLE, W.E., LESSER, P. & O'TUAMA, L. (1976). Phenytoin and phenobarbital concentrations in saliva and plasma measured by radioimmunoassay. ClAn. Pharmac. 7her., 18, 742-747. PAXTON, J.W. ROWELL, F.J. & RATCLIFFE, J.G. (1976). Production and characterization of antisera to diphenylhydantoin suitable for radioimmunoassay. J.
ImmunoL Methods, 10, 317 327. PAXTON, J.W. ROWELL, F.J.. RATCLIFFE, J.G., LAMBIE, D.G.. NANDA, R., MELVILLE, I.D. &
JOHNSON, R.H. (1977). Salivary phenytoin radioimmunoassay: a simple method for the assessment of non-protein bound drug concentrations. Eur. J. clin. Pharmac. 11,71-74. REYNOLDS, F.. ZIROYANIS, P.N., JONES, N.F.
SMITH, S.E. (1976). Salivary phenytoin concentrations in epilepsy and in chronic renal failure. Lancet, ii, 384-386.
PHENYTOIN CONCENTRATIONS IN SEPARATE SALIVARY FRACTIONS SHANNON, I.L. (1972). Parotid fluid flow rate as related to the whole saliva volume. Archs. Oral Biol., 7, 391-394. STEPHEN, K.W. & SPEIRS. C.F. (1976). Methods for collecting individual components of mixed saliva: the
191
relevance to clinical pharmacology. Br. J. clin. Pharmac., 3, 315-319.
(Received July IS, 1976)
