536

Determination of Codeine in Human Plasma and Drug Formulation Using a Chemically Modified Electrode Jyh-Myng Zen,*þ Ming-Ren Chang,þ Hsieh-Hsun Chung,þ and Ying Shihþþ þ þþ

Department of Chemistry, National Chung-Hsing University, Taichung 402, Taiwan Department of Applied Cosmetology, Hung-Kuang Institute of Technology, Taichung 433, Taiwan

Received: January 28, 1998 Final version: March 24, 1998 Abstract Both flow injection methodology and square-wave voltammetry were developed and evaluated for determining codeine in human plasma and pharmaceutical formulations using a Nafion/ruthenium oxide pyrochlore chemically modified electrode. Combining the electrocatalytic function of the ruthenium-oxide pyrochlore with charge-exclusion and the preconcentration features of Nafion perform well in codeine detection. Compared to a bare glassy carbon electrode, the chemically modified electrode exhibits a shift of the oxidation potential in cathodic direction and a marked enhancement of the current response. A linear calibration plot is obtained over the 0–32 mM range in 0.05 M HClO4 solution with a detection limit (3j) of 10 nM in the square-wave voltammetric method. While, in flow-injection analysis, a linear calibration plot is obtained over the 0.5–40 mM range with a detection limit of 0.86 ng. Quantitative analysis was performed by the standard addition method for codeine content in human plasma and a commercially available drug. Keywords: Codeine, Flow injection analysis, Human plasma, Square-wave voltammetry, Chemically modified electrode

1. Introduction Codeine has long been used as an effective analgesic and antitussive agent in pharmaceutical preparations. A sensitive and specific bioanalytical method is essential for studying the bioavailability of codeine from oral formulations. Previous approaches, including GC, RIA, and HPLC, were reported for codeine determination [1–6]. However, these methods either involve a tedious extraction process prior to the determination or are time consuming. A relatively simple and rapid electrochemical method was recently reported, which uses plastic membrane electrodes for the potentiometric determination of codeine in pharmaceutical preparation [7]. Nevertheless, till date, more sensitive and selective methods for the determination of codeine in human plasma or pharmaceutical formulations are still needed. We report here both a square-wave voltammetric (SWV) method and flow injection analysis (FIA) for the determination of codeine using a Nafion/ruthenium oxide pyrochlore chemically modified electrode (CME). This CME has been used for several determinations with excellent sensitivity and selectivity in our previous studies [8–12]. Significant advantages have been achieved by combining the electrocatalytic function of the ruthenium oxide pyrochlore with charge-exclusion and preconcentration features of Nafion. In this article, the CME is further applied to the determination of codeine in human plasma and pharmaceutical formulations. The optimal experimental conditions were thoroughly investigated. Practical analytical utility was illustrated by selective measurements of codeine in human plasma and a commercially available drug.

2. Experimental Nafion perfluorinated ion-exchange powder, 5 wt. % solution in a mixture of lower aliphatic alcohols and 10 % water, was obtained from Aldrich Chemical Co. (Milwaukee, WI). Codeine and all the other compounds (ACS-certified reagent grade) were used Electroanalysis 1998, 10, No. 8

without further purification. Aqueous solutions were prepared with double distilled deionized water. Electrochemistry was performed on a BAS-50W electrochemical analyzer. A BAS VC-2 electrochemical cell was employed in these experiments. The three-electrode system consisted of either a glassy carbon electrode (GCE) or a Nafion/ruthenium oxide pyrochlore CME working electrode, a Ag/AgCl reference electrode (Model RE-5, BAS), and a platinum wire auxiliary electrode. Since dissolved oxygen did not interfere with the anodic voltammetry, no deaeration was performed. The flow injection system consisted of a carrier reservoir, a BAS PM-80 solvent delivery system, a Rheodyne Model 7125 sample injection valve (20 mL loop), interconnecting Teflon tubing, and a BAS CC-5 thin-layer electrochemical detector with a BAS MR-3068 dual GCE. A CH-660 electrochemical workstation was connected in the FIA experiments. The preparation of the Nafion/ruthenium oxide pyrochlore CME was described elsewhere [8]. Electrodes were prepared with the optimum coating solution of 1.25 wt. % Nafion at 3000 rpm spincoating rate. The Nafion/ruthenium oxide pyrochlore CME was equilibrated in the test solution containing codeine before measurement. SW voltammograms were obtained by scanning the potential from þ0.6 to þ1.4 V at a SW frequency of 15 Hz and SW amplitude of 50 mV. At a step height of 4 mV, the effective scan rate is 60 mV/s. The codeine quantitation was achieved by measuring the current of the oxidation peak after background subtraction. For most of the experiments, a 0.05 M HClO4 solution was used as supporting electrolyte. In order to fit into the linear range, all samples used for detection were suitably diluted with the supporting electrolyte. A stock solution prepared by dissolving 194 mg of codeine phosphate in 100 mL 0.05 M HClO4 solution. An aliquot was diluted to the appropriate concentrations with 0.05 M HClO4 before actual analysis. The standard addition method was used to evaluate the content of codeine in real samples. Human plasma was obtained from healthy volunteers and was mixed with 5 M HClO4 (3:1) to cause deproteination. After gentle agitation, the mixture was centrifuged at 1000 g for 15 min at 4 8C. The upper layer of solution was collected and then spiked with exogenous codeine. In order to fit into the linear range, the drug sample used for detection was suitably diluted.

q WILEY-VCH Verlag GmbH, D-69469 Weinheim, 1998

1040-0397/98/0807-0536 $ 17.50þ.50/0

Determination of Codeine in Human Plasma and Drug Formulation

537

Fig. 2. Dependence of the anodic peak current on [HClO4 ] in SWV for 10 mM codeine at the CME. SWV parameters are as in Figure 1. Fig. 1. SW voltammograms for 10 mM codeine in 0.05 M HClO4 solution at a bare GCE (a), the Nafion/GCE (b), and the Nafion/ruthenium oxide pyrochlore CME (c). SWV amplitude: 50 mV; SW frequency: 15 Hz; step height: 4 mV.

3. Results and Discussion 3.1. Voltammetric Behavior Figure 1 demonstrates the catalytic function of the Nafion/ ruthenium oxide pyrochlore CME in the determination of codeine by SWV. On scanning from þ0.6 V toward a positive potential at a bare GCE, only a much smaller anodic peak around þ1.2 V was observed for 10 mM codeine (Fig. 1, curve a). A clear increase in anodic peak was observed when a Nafion-coated GCE was used (Fig. 1, curve b). The increase in codeine response for the Nafioncoated GCE is an indication of the ion-exchange process between Nafion and codeine. Whereas, a large increase in the peak current at þ1.07 V was observed when the CME was used (Fig. 1, curve c). The enhancement in current response and the shift in oxidation potential are clear evidences of the catalytic effect of the CME toward codeine oxidation. Further investigation was made to the transport characteristics of codeine at the CME. The linear scan voltammetric current response obtained at the CME was found to be linearly proportional to the square root of the scan rate, which illustrated that the process was diffusion-controlled. Further evidence for the nonadsorptive behavior of codeine was demonstrated by the following experiment. When the CME was switched to a medium containing only supporting electrolyte after being used in measuring a codeine solution, no voltammetric peak signal was observed at all. Note that, since the process is diffusion-controlled, the response time for the CME is very fast. All the measurements can be done immediately after the CME is immersed into the test samples.

the CME as indicated in our previous studies, so that the effect of the concentration of HClO4 on the voltammetric oxidation of codeine was first studied. The results obtained are shown in Figure 2. As can be seen, the current response starts to increase rapidly in more acidic environment and the optimal condition was found to be around 0.05 M HClO4. The above supporting electrolyte was thus used in the subsequent caffeine detection work. Note that the optimum concentration of HClO4 supporting electrolyte is the same as that used in the detection of caffeine at the CME [11]. Apparently, the same catalytic mechanism of the ruthenium oxide pyrochlore is operated in both cases. The effects of preconcentration potential and preconcentration time on the SW response for codeine were studied next and the results obtained are shown in Figures 3A and B, respectively. As

3.2. Analytical Characterization Both the electrode and the detection aspects should be considered to arrive at the optimum conditions for codeine determination. As to the electrode aspect, the optimum conditions generally follow those used in previous studies [7–11]. The pH range is critical for

Fig. 3. The effects of preconcentration potential (A) and preconcentration time (B) on the SWV response for codeine at the CME. SWV parameters are as in Figure 1. Electroanalysis 1998, 10, No. 8

538

Fig. 4. Hydrodynamic voltammograms at a bare GCE (a) and the CME (b) for codeine obtained under flow injection conditions with 0.05 M HClO4 solution as a carrier at 100 mV intervals between þ0:8 and þ1:6 V.

can be seen in Figure 3A, the peak current increases, as the potential of the electrode becomes more negative between þ0.6 and 0.0 V. This behavior is explained by the fact that codeine bears a positive charge in strong acidic environment; as a result, the accumulation of codeine is favored at more negative potentials. However, the peak current drops rapidly if the potential is more negative than ¹0.2 V. A preconcentration potential of 0.0 V was therefore chosen in all subsequent work. As to the effect of preconcentration time, for 10 mM of codeine, the peak current increases as the preconcentration time increases and starts to level off around 15 s as shown in Figure 3B. Note that it takes longer time for the peak current to level off for a lower concentration of codeine (not shown). This phenomenon is as expected and further confirms the ion-exchange process between Nafion and codeine. Therefore, to increase the sensitivity of detection, a longer preconcentration time is needed for the lower concentration of codeine. The peak current obtained in SWV is dependent on various instrumental parameters such as SW amplitude, SW frequency, and step height. These parameters are interrelated and have a combined effect on the response, but here only the general trends will be examined. It was found that these parameters had little effect on the peak potential. When the SW amplitude was varied in the range of 100–100 mV, the peak currents increased with increasing amplitude until 50 mV. However, the peak width increased at the same time, in particular when the amplitude was greater than 50 mV. Hence, 50 mV was chosen as the SW amplitude. The step height together with the frequency defines the effective scan rate. Hence an increase with either the frequency or the step height results in an increase in the effective scan rate. The response for codeine increases with SW frequency, however, above 15 Hz the peak current was unstable and obscured by a large residual current. By maintaining the frequency as 15 Hz, the effect of step height was studied. At step height of 4 mV the response was more accurately recorded. Overall, the optimized parameters can be summarized as follows: SW frequency, 15 Hz; SW amplitude, 50 mV; step height, 4 mV. The effective scan rate was 60 mV/s. Under optimum conditions, the SWV current response was linearly dependent on the concentration of codeine between 0 and 32 mM in 0.05 M HClO4 solution with slope ( mA/mM), intercept ( mA), and correlation Electroanalysis 1998, 10, No. 8

J.-M. Zen et al.

Fig. 5. FIA responses with increasing codeine concentrations of 5 mM to 25 mM at 5 mM intervals using the CME.

coefficient of 0.796, 0.589, and 0.998, respectively. The detection limit (3j) was 10 nM. To characterize the reproducibility of the CME, repetitive measurement-regeneration cycles were carried out. The results of 15 successive measurements showed a relative standard deviation of 2.7 % for 10 mM codeine. Thus, the electrode renewal gives a good reproducible surface. The stability of the CME was also evaluated and the results showed an excellent long-term stability. No decrease in codeine response with regard to either peak potential or peak current was observed after the electrode was repeatedly used and stored in 1 M NaOH solution for more than 2 months.

3.3. FIA Amperometric Detection Parameters affecting the amperometric response to codeine such as the applied potential, the concentration of carrier, and the flow rate were optimized. Figure 4 shows the hydrodynamic voltammograms for codeine obtained under flow injection conditions with 0.05 M HClO4 solution as a carrier, at 100 mV intervals between þ0.8 and þ1.6 V. As can be seen, only a much smaller anodic peak current was observed for 10 mM codeine at the bare GCE ( Fig. 4, curve a). Whereas, a large increase in the peak current was observed when the CME was used (Fig. 4, curve b). The enhancement in current response of codeine in FIA is once again clear evidence of the catalytic effect of the CME toward codeine oxidation. The potential to obtain the maximum response for codeine was at about þ1.5 V, which is an agreement with the SW voltammetric data. An applied potential of þ1.5 V was selected for subsequent work to maintain a low background current and to increase the electrode stability. The current increased with flow rate from 0.3 to 1.0 mL/min and then decreased rapidly. The increase of the response arises from the reaction being able to go to completion in a shorter time. Beyond 0.6 mL/min the kinetics are limiting leading to lower currents as the reaction fails to keep up with the passing analyte. A flow rate of 0.6 mL/min was chosen as the optimum flow rate in FIA. A typical FIA response for codeine using the CME is shown in Figure 5. A good reproducible flow injection response was obtained at þ1.5 V applied potential with 0.05 M HClO4 carrier at a flow rate of 0.6 mL/min. The calibration plot was linear between 0.5 and 40 mM with a detection limit (S/N ¼ 3) of 0.14 mM (i.e.,

Determination of Codeine in Human Plasma and Drug Formulation

539

Fig. 6. Typical SWV and FIA responses for the determination of codeine in human plasma with increasing codeine concentrations of 10 mM to 50 mM at 10 mM intervals (A) and commercially available drug with increasing codeine concentrations of 10 mM to 50 mM at 10 mM intervals (B) and 5 mM to 25 mM at 5 mM intervals (C).

0.86 ng). The repeatability of the response of the CME to codeine was determined by ten replicate injections of 20 mL of codeine and the relative standard deviations were 2.86 % and 2.53 % for 5 mM and 50 mM codeine, respectively. Again, the electrode renewal gives a good reproducible surface.

Y ¼ 1:47 þ 0:51X and Y ¼ 3:64 þ 0:12X with correlation coefficients of 0.999 and 0.998 were observed with SWV (Fig. 6B) and FIA (Fig. 6C), respectively. The signals were confirmed as being due to codeine by enhancement of the peak when standard codeine solution was added to the sample solutions. The recovery was found to be 99 % for both cases.

3.4. Sample Analysis The CME was applied to the measurement of codeine in human plasma and a commercially available drug. Typical SWV and FIA responses for the determination of codeine with and without the spike of codeine are shown in Figure 6. Since no codeine is supposed to be found in healthy human plasma, when spiked with exogenous codeine from 10 mM to 50 mM at 10 mM intervals, a comprehensive linearity (Y ¼ 0:15 þ 0:11X ) in plasma codeine level was observed with correlation coefficients of 0.998 by SWV (Fig. 6A). Note that the codeine level in human plasma can not be determined by FIA due to the existence of other oxidizable interferants. In a commercially available drug, good linearity of

4. Conclusions This study has demonstrated that the Nafion/ruthenium oxide pyrochlore CME can be applied to the detection of codeine in human plasma and pharmaceutical formulations with excellent sensitivity and selectivity by both SWV and FIA. Significant advantages have been achieved by combining the electrocatalytic function of the ruthenium oxide pyrochlore with chargeexclusion and preconcentration features of Nafion. The reliability and stability of the CME offer a good possibility for extending the technique in routine analysis of codeine. Electroanalysis 1998, 10, No. 8

540

5. Acknowledgement The authors gratefully acknowledge financial support from the National Science Council of the Republic of China under Grant NSC 87-2113-M-005-021.

6. References [1] R.A. Zweidinger, F.M. Weinberg, R.W. Handy, J. Pharm. Sci. 1976, 65, 427.

Electroanalysis 1998, 10, No. 8

J.-M. Zen et al. [2] E.J. Cone, W. D. Darwin, W.F. Buchwald, J. Chromatogr. 1983, 275, 307. [3] J.W.A. Findlay, R.F. Butz, R.M. Welch, Clin. Pharmacol. Ther. 1977, 22, 439. [4] K.R. Bedford, P.C. White, J. Chromatogr. 1985, 347, 398. [5] Z.R. Chen, F. Bochner, A. Somogyi, J. Chromatogr. 1989, 491, 367. [6] C.P.W.G.M. Verwey-Van Wissen, P.M. Koopman-Kimenai, T.B. Vree, J. Chromatogr. 1991, 570, 309. [7] E.M. Elnemma, M.A. Hamada, Mikrochim. Acta 1997, 126, 147. [8] J.-M. Zen, C.-B. Wang, J. Electroanal. Chem. 1994, 368, 251. [9] J.-M. Zen, J.-S. Tang, Anal. Chem. 1995, 67, 208. [10] J.-M. Zen, J.-S. Tang, Anal. Chem. 1995, 67, 1892. [11] J.-M. Zen, Y.-S. Ting, Anal. Chim. Acta 1997, 342, 175. [12] J.-M. Zen, I.-L. Chen, Electroanalysis 1997, 9, 537.