Potentials were measured with an Orion 960 Autochemistry System (Thermo-Orion, Cambridge, MA, USA), the recorder output of which was connected to a personal computer via a DGH Corporation 1121 module analogue-to-digital converter (Manchester, UK). An Orion 90–02 double junction silver-silver chloride reference electrode containing a 10% (v/v) solution of KNO₃ in the outer compartment and a Fluka (Munich, Germany) ISE electrode body was used.

High molecular weight poly(vinyl chloride) (PVC), 2-nitrophenyl octyl ether (NPOE), bis(2-ethylhexyl) sebacate (DOS), and tetrahydrofuran (THF) were Selectophore products from Fluka (Munich, Germany). Sodium tetraphenylborate (NaTPB) was purchased from Sigma (Munich, Germany). Ziprasidone hydrochloride powder was kindly provided by Pfizer Laboratories (New York, USA). All other reagents used were of analytical reagent grade and Milli-Q water was used throughout. Ac⁻/HAc buffer of pH 4.6 and 2 × 10⁻¹ M total concentration was used.

A 4 × 10⁻³ M ziprasidone hydrochloride standard solution was prepared by dissolving 0.0467 g of ziprasidone hydrochloride in 25 mL of methanol. Working solutions (8 × 10⁻⁴ M to 1.6 × 10⁻⁶ M) were prepared by diluting appropriate volumes of the standard solution with 15 mL of Ac⁻/HAc buffer of pH 4.6 and methanol to a final volume of 25 mL in a calibrated flask.

For the selectivity study, 4 × 10⁻³ M solutions of amoxicillin, carbamazepine, cimetidine, clomipramine hydrochloride, diclofenac sodium, fluoxetine hydrochloride, furosemide, haloperidol, lansoprazol, lorazepan, ranitidine, ofloxacine and tenoxicam were prepared by dissolving suitable amounts of the corresponding drug in methanol.

Dosage forms of ziprasidone for testing: Zeldox capsules (Pfizer Lb., Madrid, Spain) containing 20 mg of ziprasidone hydrochloride with lactose monohydrate, starch, magnesium stereate, gelatine, dioxidum of titanium (E171) and indigotine (E132) up to total capsule weight; Zeldox powder for injectable solution containing 20 mg of ziprasidone mesylate and sodium sulfobutylether-β-cyclo-dextrin up to total vial powder weight.

The ion-exchanger (ZPD-TPB) was prepared by mixing 10 mL of 4 × 10⁻³ M ZPD with 10 mL of a 4 × 10⁻³ M NaTPB aqueous solution. The mixture was filtered through a porosity-4 sintered glass crucible. The residue was washed with 1:1 (v/v) methanol/water solution until no chloride ion was detected in the washing solution and was then dried at room temperature.

The selected membrane was prepared by dissolving 3.0 mg of ZPD-TPB, 100 mg of PVC and 200 mg of NPOE in 3 mL of THF. This solution was poured into a Fluka glass ring (inner diameter 28 mm, height 30 mm) on a Fluka glass plate. The solution was allowed to evaporate overnight. A 7-mm diameter piece was cut out with a Fluka punch for ion-selective membranes and incorporated into a Fluka electrode body ISE containing 1 × 10⁻² M KCl and 8 × 10⁻⁵ M ZPD saturated with AgCl as the internal filling solution. The electrode was conditioned by soaking with constant stirring in 8 × 10⁻⁵ M ZPD working solution until the electrode gave a constant potential. When not in use, the electrode was kept immersed in the same solution.

The ZPD-selective electrode and reference electrode were immersed in any of the working ZPD solutions and the potential was measured, under constant stirring, until it gave a constant value. These potentials were then plotted vs. logarithmic values of ZPD concentration and the calibration parameters were calculated by fitting the calibration data to Equation (1): (1) E = E 0 + Slog ( LOD + C ZPD )where E is the potential of the cell, E⁰ is the standard potential, S and LOD are the slope and the limit of detection of the electrode, respectively.

For the dynamic response studies a dynamic calibration of the electrode was made by adding, while stirring, adequate small volumes by micropipette of the ZPD standard solution to 25 mL of a 2:3 (v/v) methanol/pH buffer solution to cover the concentration range from 1.6 × 10⁻⁶ to 1.6 × 10⁻⁴ M.

The ZPD content in capsules was determined by analyzing four capsules separately. The powder content of each one was shaken with 15 mL of methanol and then placed in an ultrasonic bath for 15 min and finally diluted with methanol in a 25 mL calibrated flask. A portion of this solution was centrifuged at 3,000 rpm for 5 min and the supernatant was filtered through a filter paper. Different volumes (500–1,000 μL) of the filtrate were taken, to which 15 mL of buffer were added and finally diluted to 25 mL with methanol in a calibrated flask. A similar procedure was followed in the case of ZPD powder for injectable solutions. The ZPD concentration in the sample was determined by comparing the potential value measured with the calibration graph.

Recovery studies were performed by adding different volumes (125–400 μL) of the ZPD standard solution to 1.0 mL of the pharmaceutical sample solution. Then, 15 mL of pH buffer were added and the solution was diluted to 25 mL with methanol in a calibrated flask. The samples were analyzed in triplicate using the procedure described above.

For the determination of ZPD in human urine samples the sample pH was adjusted to 4.6 with HCl and NaOH solutions. Serum samples were first deproteinized by adding a volume of 8% trichloroacetic acid solution to the same volume of serum. Then, the solutions were shaken in a vortex mixer for 30 s, allowed to settle for 20 min and centrifuged for 10 min at 1,200 rpm. The clear supernatant layer was filtered through a 0.45 μm Millipore filter and the pH was adjusted to 4.5 by adding 2 M NaOH. In the absence of urine and serum samples containing ZPD, known amounts of ZPD were added. The potential was measured and compared with calibration graphs made in pooled urines and pooled serum, respectively.

Membranes of different compositions (Table 1), prepared as described in the Experimental section, were tested to optimize the membrane composition of the ZPD-electrode. Two plasticizers with different dielectric constants were used as membrane solvent, NPOE (ɛ = 23.9) and DOS (ɛ = 2.4), membranes A and B, respectively. The corresponding calibration graphs for ZPD obtained in a 2:3 (v/v) methanol/pH buffer medium are shown in Figure 1. As can be seen, the membrane plasticized with NPOE showed higher potential span and lower detection limit, an effect that agrees with reports in the literature for other drug ISEs [42,43]. The difference observed between both membranes can be explained by the fact that the detection limit of ISEs based on dissolved ion-exchangers may be governed by the analyte ion activity present in the sample solution as a result of the distribution equilibrium of the ion-exchanger between the membrane and the solution [44]. The solubility of the ion-exchanger in the membrane generally increases as the polarity and dielectric constant of the membrane increases [45], which implies a lower concentration in the sample solution and therefore a lower detection limit. Taking these results into account NPOE was selected as plasticizer for further studies.

Two different ZPD-TPB ion exchanger concentrations in the membrane were tested (membranes A and C, respectively). As can be seen in Figure 1, the membrane containing the higher ion-exchanger concentration displayed a poorer potential response. Taking all this into account, membrane A was selected for further studies.

The effect of pH on the electrode potential at various ZPD concentrations in the range 8.0 × 10⁻⁶–4.0 × 10⁻⁵ M, was studied. The pH was varied by adding HCl or NaOH and the results are shown in Figure 2. As can be seen, the electrode potential was somewhat dependent on pH and therefore the use of a pH buffer was thought convenient. A pH of 4.6 adjusted with 2 × 10⁻¹ M sodium Ac⁻/AcH buffer solution was used for further studies.

Since ZPD is soluble in methanol and ethanol, the performance of the electrode was evaluated in hydroalcoholic media, i.e., methanol-water and ethanol-water mixtures of different ratios. All of them contained buffer solution of pH 4.6. Figure 3 shows the ZPD calibration graphs A-D determined in methanol/water 1:4, 2:3 and 3:2 (v/v), and ethanol/water 2:3 (v/v), respectively. By comparing the curves corresponding to different methanol contents, it is obvious that the best response was displayed when a methanol/water 2:3 (v/v) medium was used (curve B). When ethanol was used instead of methanol, the potential response was very poor. Accordingly to all this, a methanol/water 2:3 (v/v) medium was used for further studies.

The validation of the analytical method for the determination of ziprasidone based on the present ZPD-selective electrode was made following recommendations of the literature [46,47].

In order to demonstrate the applicability of the proposed method to the determination of ZPD, the method was applied to its determination in pharmaceuticals and human urine and serum. Two different pharmaceutical dosage forms of ZPD were analysed following the method described in the Experimental (Table 3). In the absence of a reference method for the analytical determination of ZPD in the literature, recovery studies were carried out (Table 3), providing quantitative recoveries between 99.3 and 101.7% in all the cases.

For the determination of ZPD in human urine, previous calibrations using different urines were made. No significant differences between the corresponding calibration parameters were found. Therefore, a calibration made using pooled urine samples was used for the determination of ZPD in different samples. A similar procedure was used for the determination of ZPD in serum. The results obtained are shown in Table 4. Good recoveries were obtained for the determination of ZPD in both types of biological sample.

The systemic bioavailability of ZPD administered intramuscularly is 100% and 60% if it is administered orally. After a single dose intramuscular administration, the peak serum concentration typically occurs at about 60 minutes after the dose is administered, or earlier. In the case of an oral administration, the peak serum concentration is reached from 2 to 6 hours after the administration of therapeutic doses and approximately 44% of the total serum ZPD remains unchanged. The normal therapeutic dose of ZPD recommended for adults is between 40 and 160 mg/day and from 1 to 4% is excreted unchanged in the urine after its administration. Taking into account these data and the results obtained, it can be concluded that the determination of ZPD in human urine and serum is feasible using the present electrode being this method very useful for routine clinical analysis.

The potentiometric method proposed shows some advantages over the other electroanalytical method reported in the literature [36] which makes use of voltammetric techniques. Thus, while the voltammetric method requires electrode polishing before each experiment, no electrode pre-treatment is needed in the present potentiometric method. Variations of the pH of the medium also have less influence in the potentiometric method, which is not affected by the presence of some other drugs and compounds frequently present in pharmaceuticals and biological fluids. Finally, while both methods are useful for pharmaceuticals and serum analysis, the potentiometric method was also applied to urine analysis.