- Email: [email protected]
Enhanced Stability of Codeine Sulfate: Effect of pH, Buffer, and Temperature on the Degradation of Codeine in Aqueous Solution
Enhanced Stability of Codeine Sulfate: Effect of pH, Buffer, and Temperature on the Degradation of Codeine in Aqueous Solution MICHAELF. POWELL Received December 16, 1985, from the Institute of Pharmaceutical Sciences, Syntex Research, Palo Alto, CA 94303. June 26, 1986. Abstract G In the absence of strong buffer catalysts,the degradation of codeine sulfate (7,8-didehydro-4,5a-epoxy-3-methoxy-l7-methylmorphinan-6a-01sulfate) in aqueous solution is described by the expression bS= k,.[H+] c k, + kHo [HO-1, where kH+= (3.9 2 1.3) x M - ' . s - ' , k, = (2.7 ? 0.5) x 10-8s-', and h0 = (5.1 f 1.0) x M '.s-' at 80°C.The activation energies associated with these rate constants are 27.7, 21 .O. and 28.3 kcal.mol-', respectively. In the absence of buffer catalysis, codeine sulfate is predicted to have a room temperature shelf life of -44 years between pH 1 and 10, significantly longer than the 1.1 year shelf life of codeine phosphate reported earlier.
Codeine has enjoyed enormous popularity as an analgesic for over a century. Of the various salt forms, codeine phosphate is generally preferred because of its high aqueous solubility (31 wt% at 25 "C). This alkaloid is fairly stable in the solid state, but in aqueous solution, it degrades relatively rapidly in the presence of strong acids or bases, and has been reported to have a room temperature shelf life of less than two years in buffer solutions of neutral pH.1.2One of these studies' concluded that codeine was most stable at pH 3.5, whereas another reported a gradual, 26-fold increase in degradation rate from pH 1.6 to pH 10.5.2 In these studies, different rate constants were sometimes determined at the same pH, indicating that factor(s) other than pH contributed to the observed rates of codeine degradation. In addition, the degradation of codeine has been studied at various temperat u r e s i 3 u n d e r d i s s i m i l a r r e a c t i o n conditions i n such a way that the temperature dependence of codeine degradation in aqueous solution could not be established. In this report we describe pH-rate profiles at 60, 80, and 100 "C for codeine degradation using data reported previousl ~ ' and . ~ herein. From these profiles, it is possible to calculate, by the Arrhenius equation, the pH-rate profile for this reaction a t room temperature. In what follows, i t is shown that: (a)the shelf life of codeine in aqueous solution at 25 "C is much greater than previously reported' when buffer catalysis is accounted for; and ( b ) codeine sulfate solutions are intrinsically more stable than codeine phosphate solutions due to catalysis in the latter by phosphate ion.
Experimental Section Apparatus and Reagents-Reversed-phase chromatography of codeine was carried out using an HPLC system consisting of a Micromeritics model 725 autoinjector, model llOA Altex pump, model 770 Spectra Physics spectrophotometric detector, and an SP 4000 computing integrator. A 250 x 4.6 mm Partisil5 ODS 3 column (Whatman) was used for analysis. Codeine sulfate (USP) from Mallinckrodt (lot 1580Tll)containing -7% (by weight) water, was used for all experiments. Methanol (HPLC-grade)and tetrahydrofuran (Burdick and Jackson) were used for the preparation of the mobile phase. Potassium hydrogen phosphate, sodium hydroxide, hydrochloric acid, and sodium heptanesulfonate were of the highest 0022-3549/86/0900-0901$01.00/0 C 1986, American Pharmaceutical Association
Accepted for publication
grade commercially available (Aldrich or Mallinckrodt), and were used without further purification. High-Performance Liquid Chromatography Conditions-A linear response throughout the range of 0.1-6 pg of codeine injected, and a Beparation of codeine from its degradation products, were achieved using the following conditions. The mobile phase was: water:methanol:tetrahydrofuran(70:29.5:0.5,v/v) containing 0.02 M sodium heptanesulfonate and 0.02 M KH2P04;the flow rate was 1.2 mumin; detection was at 235 nm; the injection volume was 10-50 pL; and the typical codeine retention time was 22 min. Kinetic-For all experiments,buffer solutionscontainingcodeine sulfate (0.1mg/mL)were prepared shortly before use and the pH was determined at the various temperatures. For strongly acidic or alkaline solutions, the hydronium and hydroxide ion activities at 60-100 "C were calculated from published activity coefficients4.5or acidity function Ho values.6.7 In a typical experiment, 5-mL aliquots of solution containing codeine sulfate were transferred to pretreated amber ampules, flame-sealed, and stored (with HCI and (NH4)2S04) at 60, 80, or 100°C. Several of these samples were refrigerated immediately after flame-sealing,and were later used as controls for the initial time points. At known time intervals, ampules were removed and refrigerated until 6-10 samples for each kinetic run were taken. Upon removal of the last sample, the stored solutions were allowed to warm to room temperature, and then all samples were analyzed on the same day. Determination of the pH of these solutions later showed no change in acidity. When necessary, Samples were neutralized with acid or base to facilitate the HPLC analysis. Peak area integration values were used directly when fitting the data to first-order kinetics; typically, reactions were followed to <90% of the codeine remaining, and often for more than two half-lives.
Results and Discussion Effect of pH and Temperature-For all experiments, first-order drug loss was observed, and the pseudo-first-order rate constants for codeine degradation are summarized in Table I and Fig. 1. In some instances, different rate constants were observed at similar pH values, especially when varying amounts of buffer were used. Larger rate constants were often observed for solutions containing significant amounts of buffer (but at constant pH), indicating a rate contribution by buffer catalysis. To reduce the effect of buffer catalysis, only rate constants determined at low buffer concentration were used for the log rat+pH profile of Fig. 1;rate constants determined for solutions where significant amounts of buffer were used are omitted from Table I because the enhanced rates due to buffer catalysis tend to obscure the true nature of the log rate-pH profile. Inspection of Fig. 1 shows that the degradation of codeine adheres to eq. 1:
(1) where the rate constants k H + , ko, and kHO- are the coefficients for catalysis by hydronium ion, water (or a spontaneous reaction), and hydroxide ion, respectively. The introduction of a pK, term (codeine pK, = 8.2 at 25 "C8)into eq. 1 did not Journal of Pharmaceutical Sciences / 901 Vol. 75, No. 9, September 1986
Table &Summary of Rate Data with Negllglble Buffer Catalysis for the Degradatlon of Codelne In Aqueous Solution Temperature, "C
Rate Constant,
PH
Buffer
Reference
S-1
100
80
-0.29" 1.126 2.05 6.21 11.128 1 1 .47g
6.72 x 9.89 x 6.59 X 1.34 x 2.05 x 3.68 x
- 1.65'
2.37 x lo-' 6.37 x lo-' 1.33 X lo-'' 2.77 X lo-' 3.35 x 10-8' 4.19 X lo-' 4.76 X lo-' 3.19 x 10-7 8.50 x 1 0 - ~ 2.17 x lo-' 2.80 x lo-'
-0.23d 1.60 2.39 3.50 10.29 10.45 11.484 1 1 .7!j8 12.04O 12.34g -0.18h
60
7.89 x 2.17 x 3.70 x 3.30 x 3.47 x 9.43 x 7.24 x 1.38 x
2.32 2.54 3.50 3.50 10.70 10.88 12.86g
10-7 10-8
1 .O M HCI 0.1 M HCI 0.01 M HCI 0.004M phosphate 0.1 M KOH 0.2M KOH
lo-' 10-7 lo-' 10-6
5.0M HCI 1 .O M HCI citric acid:phosphate 0.004M phosphate citric acid:phosphate 0.004M phosphate 0.008M phosphate 0.1 M NaOH 0.2M NaOH 0.4M NaOH 0.8M NaOH
10-9 10-9 10-9
1.0 M HCI 0.004M phosphate 0.008M phosphate
10-9' 10-9 10-9 10-6
citric acid:phosphate citric acid:phosphate 0.004M phosphate 0.008M phosphate 1 .O M NaOH
lo-''
Ho of 1 .O m HCI at 100 "C; extrapolated from data in Ref. 6.bHoof 0.1 M HCI at 100 "C; extrapolatedfrom data in Ref. 6.'Ho of 5 M HCI at 80 "C; see Ref. 6."Ho of 1 M HCI at 80 "C; see Ref. 6." Determinedin this study. 'Calculated from an Arrhenius plot of data obtained at 25 and 70 "C; see Ref. 2. 8Calculated from temperature-dependentactivity coefficients and autoprotolysis K, values; see Ref. 4. hHoof 1 M HCI at 60 "C; see Ref. 6. 'With added antioxidants:ascorbic acid or sodium thiosulfate.
result in a significant better fit of the rate data, and so was not included. Moreover, the similar rates determined at pH 3 and 10 at 80 "C,and those at pH 2 and 7 at 100 "C,show that the log rate-pH profile from pH 1 to 10 is a broad plateau without any distinct features attributable to a change in rate caused by codeine ionization. The rate constants derived from nonlinear least squares analysis of the data of Table I and eq. 1 are given in Table 11, and the corresponding activation parameters are summarized in Table 111. Table ICEffect of Temperature on the Rate Constants for the Degradation of Codelne In Aqueous Solution Rate Constant a
Temperature, "C ~~
~
100 80 60 25
kH+,M-' s.-'
ko, S -l
h-, M-'
S-'
~
2.87 X 3.96 x lo-' 3.23 x 1 0 - ~ 2.46 x lo-"
8.95 X lo-' 2.14 x lo-' 3.00 x 1 0 - ~ 7.60 x 10. "
4.94 x 10-5 5.24 x lo-' 5.01 x 10-7 3.22 X lo-'
"Obtained from a nonlinear least-squaresfit (Ref. 9) of the rate data to eq. 1. Rates at 25 "C were calculated from E, and log A given in Table 111.
0
8
4
12
H, or pH
Figure 1-Log rate-pH profile for the degradation of codeine sulfate in aqueous solution at 60 (C), 80 (O), and 100°C (a).Rate constants were determined from dilute buffer (or buffer-free)solutions to minimize the effects of buffer.cata1ysi.s.The dashed line is the calculated log r a t e pH profile for codeine at 25°C.Ho is used instead of pH below pH 1 . 902 /Journal of Pharmaceutical Sciencos Val. 75, No. 9, September 1986
Table IlCSummary of Activation Parameters for the Degradation of Codeine in Aqueous Solution Rate Constant
hi+ kJ
1610-
" + SD.
En,
kcal.mol-
'
Log A
AM,
ASS,
kcabmol-'
ca l m o l - 'K '
27.7 2 1.0" 9.72 0.6 27.0 2 1.0 5.3 f 0.8 20.3 t 1.2 21.0 2 1.2 28.3 f 0.6 12.3 f 0.3 27.6 2 0.5
-16.3 2 1.5 -36 f 12 -4.6 ? 0.2
Buffer Catalysis and Shelf Life-In several of the previously reported1.2 rate determinations, phosphate buffers were used to maintain solution pH. In this study, phosphate buffers were found to accelerate the rate of codeine degradation: for example, the buffer plots of Fig. 2 show that the rate of codeine degradation in 0.05 M phosphate buffer (pH 7) is almost 20-fold faster than in unbuffered solution (i.e., the intercept of this plot). It was also observed, in the reversedphase HPLC chromatograms, that norcodeine was one of the primary degradation products in phosphate buffer, as identified by coinjection of an authentic sample. Other, more polar compounds were also found as minor degradation products but were not identified. In phosphate buffers of neutral pH, the major buffer species are H2PO; and HPOiZ,either of which may act as a catalyst for codeine degradation. Determining which of these was the stronger catalyst was done by studying the degradation reaction at two different pH values. These kinetic studies were carried out a t 100°C to study codeine degradation under conditions used for autoclaving. In neutral pH solutions, the observed rate constants obeyed the general expression of eq. 2:
0.03
0.09
0.06
0.12
BT
where kHm- and kHpOi2 are the rate constants denoting catalysis by kzPO; and HPO;', respectively. Rearranging and dividing by the total phosphate buffer concentration BT gives:
Thus a plot of the pseudo-first-order rate constants, kobs, versus BT a t a constant pH affords k o as the intercept and k' as the slope (Fig. 2). A secondary plot of the apparent rate constants, k', versus the fraction of the acid buffer component present, IH2PO;1/BT, gives k ~ m iand - kH2mi as the 0- and 1intercepts, respectively. Such a plot for the degradation of codeine in phosphate buffers a t pH 6-7 a t 100 "C is given in the inset of Fig. 2: the derived values of k H p 1- and kH2pOi 5.1 x M-'.s-'and 4.9 x M-'* s-' ,respectively. Although these rate constants were determined quite carefully (for example, by using constant ionic strength), the relative inherent error is larger for kHzpOi than for k ~ p o ,i and so the existence of H2PO; catalysis is less certain than that of HPO;-. Since most pharmaceutical codeine preparations are formulated in the neutral pH region, the catalysts H,PO, and PO!- (found only a t the pH extremes, pH < 3 or > 9) make little, if any, contribution to the overall rate of degradation. This finding has important ramifications regarding the stability of codeine phosphate solutions since phosphate ion generated from the dissolution of codeine phosphate will act as a catalyst for codeine degradation. Comparison of the calculated rate constant a t 25 "C from Fig. 1 for codeine sulfate (ko = 7.6 x lo-" s-') and the experimentally determined rates of codeine phosphate degradation at 25 "C reported earlier (for example, k = 3.0 x lO-'s-l at pH 3.51) shows that the experimentally determined rates are 40 times faster than rates predicted from the Arrhenius rate data herein. The predicted shelf life of codeine sulfate at pH 4 is 44 years, much longer than the 1.1 years observed for a 2.5% solution of codeine phosphate in citrate:phosphate buffer.' Thus, most codeine phosphate solutions of sufficient concen-
-
Figure 2-Buffer plots of LS for codeine degradation at 100°C (p = 0.5) versus total phosphate buffer concentration, B,: upper line, pH 7.04; lower line, pH 6.00. The linear least-squaresslope of these lines is K ; the positive intercept is k,. The inset figure shows the dependence of the apparent rate constant. K , on the buffer composition. The 0- and 1intercepts of the inset figure are kHpeand kHpoi, respectively.
tration for therapeutic efficacy (i.e., a few percent) will exhibit significantly shorter shelf lives than the 'intrinsic' shelf life of codeine in aqueous solution; i.e., the shelf life determined by the reaction rate of codeine with water and not with buffer species such as phosphate ion. To ensure long-term stability of codeine in aqueous solution, the concentration of phosphate ion or other buffering species should be minimized, e.g., by use of codeine sulfate instead of codeine phosphate. The preparation of weaklybuffered codeine solutions should not present stability problems associated with pH drift because codeine has a wide pH range of maximum stability (vide supra, Fig. 1).
References and Notes 1. Gundermann,P.; Pohloudek-Fabini, R. Pharmcuie 1983,38,92QA
2. &welczyk, E.; Wachowiak, R. Herba Pol. 1974,20, 253-263. 3. Goeneohea, V. S.; Kobbe, K.; Goebel, K . J . Arzneirn. Forsch. 1978,28, 1070-1071. 4. Harned, H. W. "Physical Chemistry of Electrol tic Solutions"; Reinhold Publishing: New York, 1958; p 716, 7rO. 5. Greelet, R. S. Anal. Chern. 1960,32, 1717. 6. Kresge, A. J.; Chen, H. J.; Capen, G . L.; Powell, M. F. Can. J . Chem. 1983,61,249-256. 7. Rodchester, C. H. "Acidity Functions"; Academic Press: New York, 1971. 8. Oberts,F. W.; Andrews, H. L. J . Pharmacol. Exp. Ther. 1941,71, 38. 9. Bevingtan, P. R. "Data Reduction and Error Analysis for the Physical Sciences"; McGraw Hill: New York, 1969; p 140.
Acknowledgments The author expresses thanks to Drs. D. Johnson and L. Gu for helpful comments during the preparation of this manuscript, and to the referees for pointing out the usefulness of using concentrated phosphate buffers. Journal of Pharmaceutical Sciences / 903 Vol. 75, No. 9, September 1986
Codeine has enjoyed enormous popularity as an analgesic for over a century. Of the various salt forms, codeine phosphate is generally preferred because of its high aqueous solubility (31 wt% at 25 "C). This alkaloid is fairly stable in the solid state, but in aqueous solution, it degrades relatively rapidly in the presence of strong acids or bases, and has been reported to have a room temperature shelf life of less than two years in buffer solutions of neutral pH.1.2One of these studies' concluded that codeine was most stable at pH 3.5, whereas another reported a gradual, 26-fold increase in degradation rate from pH 1.6 to pH 10.5.2 In these studies, different rate constants were sometimes determined at the same pH, indicating that factor(s) other than pH contributed to the observed rates of codeine degradation. In addition, the degradation of codeine has been studied at various temperat u r e s i 3 u n d e r d i s s i m i l a r r e a c t i o n conditions i n such a way that the temperature dependence of codeine degradation in aqueous solution could not be established. In this report we describe pH-rate profiles at 60, 80, and 100 "C for codeine degradation using data reported previousl ~ ' and . ~ herein. From these profiles, it is possible to calculate, by the Arrhenius equation, the pH-rate profile for this reaction a t room temperature. In what follows, i t is shown that: (a)the shelf life of codeine in aqueous solution at 25 "C is much greater than previously reported' when buffer catalysis is accounted for; and ( b ) codeine sulfate solutions are intrinsically more stable than codeine phosphate solutions due to catalysis in the latter by phosphate ion.
Experimental Section Apparatus and Reagents-Reversed-phase chromatography of codeine was carried out using an HPLC system consisting of a Micromeritics model 725 autoinjector, model llOA Altex pump, model 770 Spectra Physics spectrophotometric detector, and an SP 4000 computing integrator. A 250 x 4.6 mm Partisil5 ODS 3 column (Whatman) was used for analysis. Codeine sulfate (USP) from Mallinckrodt (lot 1580Tll)containing -7% (by weight) water, was used for all experiments. Methanol (HPLC-grade)and tetrahydrofuran (Burdick and Jackson) were used for the preparation of the mobile phase. Potassium hydrogen phosphate, sodium hydroxide, hydrochloric acid, and sodium heptanesulfonate were of the highest 0022-3549/86/0900-0901$01.00/0 C 1986, American Pharmaceutical Association
Accepted for publication
grade commercially available (Aldrich or Mallinckrodt), and were used without further purification. High-Performance Liquid Chromatography Conditions-A linear response throughout the range of 0.1-6 pg of codeine injected, and a Beparation of codeine from its degradation products, were achieved using the following conditions. The mobile phase was: water:methanol:tetrahydrofuran(70:29.5:0.5,v/v) containing 0.02 M sodium heptanesulfonate and 0.02 M KH2P04;the flow rate was 1.2 mumin; detection was at 235 nm; the injection volume was 10-50 pL; and the typical codeine retention time was 22 min. Kinetic-For all experiments,buffer solutionscontainingcodeine sulfate (0.1mg/mL)were prepared shortly before use and the pH was determined at the various temperatures. For strongly acidic or alkaline solutions, the hydronium and hydroxide ion activities at 60-100 "C were calculated from published activity coefficients4.5or acidity function Ho values.6.7 In a typical experiment, 5-mL aliquots of solution containing codeine sulfate were transferred to pretreated amber ampules, flame-sealed, and stored (with HCI and (NH4)2S04) at 60, 80, or 100°C. Several of these samples were refrigerated immediately after flame-sealing,and were later used as controls for the initial time points. At known time intervals, ampules were removed and refrigerated until 6-10 samples for each kinetic run were taken. Upon removal of the last sample, the stored solutions were allowed to warm to room temperature, and then all samples were analyzed on the same day. Determination of the pH of these solutions later showed no change in acidity. When necessary, Samples were neutralized with acid or base to facilitate the HPLC analysis. Peak area integration values were used directly when fitting the data to first-order kinetics; typically, reactions were followed to <90% of the codeine remaining, and often for more than two half-lives.
Results and Discussion Effect of pH and Temperature-For all experiments, first-order drug loss was observed, and the pseudo-first-order rate constants for codeine degradation are summarized in Table I and Fig. 1. In some instances, different rate constants were observed at similar pH values, especially when varying amounts of buffer were used. Larger rate constants were often observed for solutions containing significant amounts of buffer (but at constant pH), indicating a rate contribution by buffer catalysis. To reduce the effect of buffer catalysis, only rate constants determined at low buffer concentration were used for the log rat+pH profile of Fig. 1;rate constants determined for solutions where significant amounts of buffer were used are omitted from Table I because the enhanced rates due to buffer catalysis tend to obscure the true nature of the log rate-pH profile. Inspection of Fig. 1 shows that the degradation of codeine adheres to eq. 1:
(1) where the rate constants k H + , ko, and kHO- are the coefficients for catalysis by hydronium ion, water (or a spontaneous reaction), and hydroxide ion, respectively. The introduction of a pK, term (codeine pK, = 8.2 at 25 "C8)into eq. 1 did not Journal of Pharmaceutical Sciences / 901 Vol. 75, No. 9, September 1986
Table &Summary of Rate Data with Negllglble Buffer Catalysis for the Degradatlon of Codelne In Aqueous Solution Temperature, "C
Rate Constant,
PH
Buffer
Reference
S-1
100
80
-0.29" 1.126 2.05 6.21 11.128 1 1 .47g
6.72 x 9.89 x 6.59 X 1.34 x 2.05 x 3.68 x
- 1.65'
2.37 x lo-' 6.37 x lo-' 1.33 X lo-'' 2.77 X lo-' 3.35 x 10-8' 4.19 X lo-' 4.76 X lo-' 3.19 x 10-7 8.50 x 1 0 - ~ 2.17 x lo-' 2.80 x lo-'
-0.23d 1.60 2.39 3.50 10.29 10.45 11.484 1 1 .7!j8 12.04O 12.34g -0.18h
60
7.89 x 2.17 x 3.70 x 3.30 x 3.47 x 9.43 x 7.24 x 1.38 x
2.32 2.54 3.50 3.50 10.70 10.88 12.86g
10-7 10-8
1 .O M HCI 0.1 M HCI 0.01 M HCI 0.004M phosphate 0.1 M KOH 0.2M KOH
lo-' 10-7 lo-' 10-6
5.0M HCI 1 .O M HCI citric acid:phosphate 0.004M phosphate citric acid:phosphate 0.004M phosphate 0.008M phosphate 0.1 M NaOH 0.2M NaOH 0.4M NaOH 0.8M NaOH
10-9 10-9 10-9
1.0 M HCI 0.004M phosphate 0.008M phosphate
10-9' 10-9 10-9 10-6
citric acid:phosphate citric acid:phosphate 0.004M phosphate 0.008M phosphate 1 .O M NaOH
lo-''
Ho of 1 .O m HCI at 100 "C; extrapolated from data in Ref. 6.bHoof 0.1 M HCI at 100 "C; extrapolatedfrom data in Ref. 6.'Ho of 5 M HCI at 80 "C; see Ref. 6."Ho of 1 M HCI at 80 "C; see Ref. 6." Determinedin this study. 'Calculated from an Arrhenius plot of data obtained at 25 and 70 "C; see Ref. 2. 8Calculated from temperature-dependentactivity coefficients and autoprotolysis K, values; see Ref. 4. hHoof 1 M HCI at 60 "C; see Ref. 6. 'With added antioxidants:ascorbic acid or sodium thiosulfate.
result in a significant better fit of the rate data, and so was not included. Moreover, the similar rates determined at pH 3 and 10 at 80 "C,and those at pH 2 and 7 at 100 "C,show that the log rate-pH profile from pH 1 to 10 is a broad plateau without any distinct features attributable to a change in rate caused by codeine ionization. The rate constants derived from nonlinear least squares analysis of the data of Table I and eq. 1 are given in Table 11, and the corresponding activation parameters are summarized in Table 111. Table ICEffect of Temperature on the Rate Constants for the Degradation of Codelne In Aqueous Solution Rate Constant a
Temperature, "C ~~
~
100 80 60 25
kH+,M-' s.-'
ko, S -l
h-, M-'
S-'
~
2.87 X 3.96 x lo-' 3.23 x 1 0 - ~ 2.46 x lo-"
8.95 X lo-' 2.14 x lo-' 3.00 x 1 0 - ~ 7.60 x 10. "
4.94 x 10-5 5.24 x lo-' 5.01 x 10-7 3.22 X lo-'
"Obtained from a nonlinear least-squaresfit (Ref. 9) of the rate data to eq. 1. Rates at 25 "C were calculated from E, and log A given in Table 111.
0
8
4
12
H, or pH
Figure 1-Log rate-pH profile for the degradation of codeine sulfate in aqueous solution at 60 (C), 80 (O), and 100°C (a).Rate constants were determined from dilute buffer (or buffer-free)solutions to minimize the effects of buffer.cata1ysi.s.The dashed line is the calculated log r a t e pH profile for codeine at 25°C.Ho is used instead of pH below pH 1 . 902 /Journal of Pharmaceutical Sciencos Val. 75, No. 9, September 1986
Table IlCSummary of Activation Parameters for the Degradation of Codeine in Aqueous Solution Rate Constant
hi+ kJ
1610-
" + SD.
En,
kcal.mol-
'
Log A
AM,
ASS,
kcabmol-'
ca l m o l - 'K '
27.7 2 1.0" 9.72 0.6 27.0 2 1.0 5.3 f 0.8 20.3 t 1.2 21.0 2 1.2 28.3 f 0.6 12.3 f 0.3 27.6 2 0.5
-16.3 2 1.5 -36 f 12 -4.6 ? 0.2
Buffer Catalysis and Shelf Life-In several of the previously reported1.2 rate determinations, phosphate buffers were used to maintain solution pH. In this study, phosphate buffers were found to accelerate the rate of codeine degradation: for example, the buffer plots of Fig. 2 show that the rate of codeine degradation in 0.05 M phosphate buffer (pH 7) is almost 20-fold faster than in unbuffered solution (i.e., the intercept of this plot). It was also observed, in the reversedphase HPLC chromatograms, that norcodeine was one of the primary degradation products in phosphate buffer, as identified by coinjection of an authentic sample. Other, more polar compounds were also found as minor degradation products but were not identified. In phosphate buffers of neutral pH, the major buffer species are H2PO; and HPOiZ,either of which may act as a catalyst for codeine degradation. Determining which of these was the stronger catalyst was done by studying the degradation reaction at two different pH values. These kinetic studies were carried out a t 100°C to study codeine degradation under conditions used for autoclaving. In neutral pH solutions, the observed rate constants obeyed the general expression of eq. 2:
0.03
0.09
0.06
0.12
BT
where kHm- and kHpOi2 are the rate constants denoting catalysis by kzPO; and HPO;', respectively. Rearranging and dividing by the total phosphate buffer concentration BT gives:
Thus a plot of the pseudo-first-order rate constants, kobs, versus BT a t a constant pH affords k o as the intercept and k' as the slope (Fig. 2). A secondary plot of the apparent rate constants, k', versus the fraction of the acid buffer component present, IH2PO;1/BT, gives k ~ m iand - kH2mi as the 0- and 1intercepts, respectively. Such a plot for the degradation of codeine in phosphate buffers a t pH 6-7 a t 100 "C is given in the inset of Fig. 2: the derived values of k H p 1- and kH2pOi 5.1 x M-'.s-'and 4.9 x M-'* s-' ,respectively. Although these rate constants were determined quite carefully (for example, by using constant ionic strength), the relative inherent error is larger for kHzpOi than for k ~ p o ,i and so the existence of H2PO; catalysis is less certain than that of HPO;-. Since most pharmaceutical codeine preparations are formulated in the neutral pH region, the catalysts H,PO, and PO!- (found only a t the pH extremes, pH < 3 or > 9) make little, if any, contribution to the overall rate of degradation. This finding has important ramifications regarding the stability of codeine phosphate solutions since phosphate ion generated from the dissolution of codeine phosphate will act as a catalyst for codeine degradation. Comparison of the calculated rate constant a t 25 "C from Fig. 1 for codeine sulfate (ko = 7.6 x lo-" s-') and the experimentally determined rates of codeine phosphate degradation at 25 "C reported earlier (for example, k = 3.0 x lO-'s-l at pH 3.51) shows that the experimentally determined rates are 40 times faster than rates predicted from the Arrhenius rate data herein. The predicted shelf life of codeine sulfate at pH 4 is 44 years, much longer than the 1.1 years observed for a 2.5% solution of codeine phosphate in citrate:phosphate buffer.' Thus, most codeine phosphate solutions of sufficient concen-
-
Figure 2-Buffer plots of LS for codeine degradation at 100°C (p = 0.5) versus total phosphate buffer concentration, B,: upper line, pH 7.04; lower line, pH 6.00. The linear least-squaresslope of these lines is K ; the positive intercept is k,. The inset figure shows the dependence of the apparent rate constant. K , on the buffer composition. The 0- and 1intercepts of the inset figure are kHpeand kHpoi, respectively.
tration for therapeutic efficacy (i.e., a few percent) will exhibit significantly shorter shelf lives than the 'intrinsic' shelf life of codeine in aqueous solution; i.e., the shelf life determined by the reaction rate of codeine with water and not with buffer species such as phosphate ion. To ensure long-term stability of codeine in aqueous solution, the concentration of phosphate ion or other buffering species should be minimized, e.g., by use of codeine sulfate instead of codeine phosphate. The preparation of weaklybuffered codeine solutions should not present stability problems associated with pH drift because codeine has a wide pH range of maximum stability (vide supra, Fig. 1).
References and Notes 1. Gundermann,P.; Pohloudek-Fabini, R. Pharmcuie 1983,38,92QA
2. &welczyk, E.; Wachowiak, R. Herba Pol. 1974,20, 253-263. 3. Goeneohea, V. S.; Kobbe, K.; Goebel, K . J . Arzneirn. Forsch. 1978,28, 1070-1071. 4. Harned, H. W. "Physical Chemistry of Electrol tic Solutions"; Reinhold Publishing: New York, 1958; p 716, 7rO. 5. Greelet, R. S. Anal. Chern. 1960,32, 1717. 6. Kresge, A. J.; Chen, H. J.; Capen, G . L.; Powell, M. F. Can. J . Chem. 1983,61,249-256. 7. Rodchester, C. H. "Acidity Functions"; Academic Press: New York, 1971. 8. Oberts,F. W.; Andrews, H. L. J . Pharmacol. Exp. Ther. 1941,71, 38. 9. Bevingtan, P. R. "Data Reduction and Error Analysis for the Physical Sciences"; McGraw Hill: New York, 1969; p 140.
Acknowledgments The author expresses thanks to Drs. D. Johnson and L. Gu for helpful comments during the preparation of this manuscript, and to the referees for pointing out the usefulness of using concentrated phosphate buffers. Journal of Pharmaceutical Sciences / 903 Vol. 75, No. 9, September 1986