Voltammetric Behavior of Amfepramone ( Diethylpropion ) at the Hanging Mercury Drop Electrode and its Analytical Determination in Pharmaceutical Formulations

Este trabalho descreve um estudo sistemático do comportamento voltamétrico da anfepramona no eletrodo de mercúrio de gota pendente (HMDE) por voltametria cíclica (CV) e voltametria de corrente alternada (AC). Os estudos mostraram o comportamento adsortivo da anfepramona no HMDE e foram realizados em soluções de H 2 SO 4 0,1 mol L (pH 1.0) e tampão de Ringer (pH 11,0) como eletrólito suporte. A faixa linear para a determinação de anfepramona por voltametria de pulso diferencial de varredura linear (DPV) foi de 0,05 a 2,0 mg L (r = 0,998) em meio ácido e de 0,25 a 4,0 mg L (r = 0,994) em meio alcalino. Os desvios padrão relativos calculados para 5 medidas de anfepramona 0,5 mg L em meio ácido e meio alcalino foram de 2,5 e 4,0%, respectivamente. Os limites de detecção calculados para a determinação de enfepramona em meio ácido e meio alcalino foram de 0,035 e 0,18 mg L, respectivamente. As metodologias foram aplicadas para a determinação de anfepramona por DPV em cápsulas e comprimidos de formulações farmacêuticas usadas no tratamento da obesidade. Os valores de recuperação entre 90,0 e 101,0% para anfepramona adicionada a misturas sintéticas contendo femproporex, mazindol, sibutramina, fluoxetina, cafeína, diazepan e metformina provam a aplicabilidade do método para a sua determinação na presença de outras drogas normalmente adicionadas de forma ilegal a formulações farmacêuticas comercializadas como medicamentos naturais.


Introduction
The abusive use of anorexics has grown lately, mainly because of its use in the treatment of obesity.The consumption of anorexics in Brazil is approximately 23.6 tons per year, surpassed only by Chile and Germany. 1 Furthermore, herbal formulations with claimed slimming activity, which are assumed to improve the effectiveness of food diets, have been in increasing use in Brazil.
Obesity is a major risk factor for morbity and mortality 2,3 and its therapeutic treatment include anorexic agents such as amfepramone (diethylpropion -Figure 1).The administration of amfepramone can tipically lead to the increase of locomotor activity, euphorie, 4 nervousness, irritability, insomnia and hyperkinesis. 5[8] To the best of our knowledge, there is still no official method for the determination of amfepramone.However, it has been determined mainly by thin layer and liquid chromatography, 9,10 1 H NMR spectroscopy 11 and voltammetry. 12Some of these methods are time consuming and involve high instrumentation costs if compared to voltammetric methods.The voltammetric method published by Tan and coworkers 12 describes the determination of amfepramone by linear sweep voltammetry with a detection limit of 1 × 10 -9 mol L -1 (0.21 μg L -1 ).However, a systematic investigation of the amfepramone determination in pharmaceutical formulations (synthetic and natural) was not yet described in the literature.In this context, the voltammetry offers the possibility for a rapid and sensitive determination of amfepramone in pharmaceutical formulations.Besides, solutions resulting from solubilization of tablets or capsules do not need to be filtered exhaustively before voltammetric determinations, which can be performed in the presence of particulate matter. 13This is a further advantage of voltammetric methods compared with either spectroscopic or liquid chromatographic procedures, where the samples must be filtered until the complete separation of the insoluble excipients.
This paper describes a systematic study of the voltammetric behavior of amfepramone at the hanging mercury drop electrode (HMDE) by cyclic (CV) and alternating current (AC) voltammetric methods.The studies showed the adsorptive behavior of amfepramone at the HMDE and were performed in two different alkaline and acidic supporting electrolytes.The voltammetric peaks obtained at -0.78 V, -1.35 V, and -1.70 V for amfepramone in acidic and alkaline medium are characterized by irreversible reduction processes, which have a contribution from both surface species adsorbed on the electrode and solution species diffusing to the electrode surface.The methods were applied for the determination of amfepramone by differential pulse voltammetry (DPV) in tablets and capsules.The method allowed the selective determination of amfepramone in pharmaceutical formulations and synthetic mixtures containing anorexics, benzodiazepines and antidepressants.

Apparatus
The voltammetric measurements were performed using a Metrohm 693 VA Processor in combination with a 694 VA Stand (all from Metrohm, Herisau, Switzerland).The three-electrode configuration consisted of the hanging mercury drop electrode (HMDE) as working electrode, an Ag/AgCl reference electrode (3 mol L -1 KCl) and a platinum wire as auxiliary electrode.The surface area of the HMDE was 0.6 mm 2 .In cyclic voltammetric (CV) measurements in alkaline medium the scans between -0.60 and -1.8 V were preceded by stirred deposition (2000 rpm) at the HMDE for 120 s at -0.60 V and different scan rates between 100 and 1000 mV s -1 were used.In cyclic voltammetric (CV) measurements in acidic medium the scans between -0.45 and -1.0 V were preceded by stirred deposition (2000 rpm) at the HMDE for 120 s at -0.45 V and different scan rates between 100 and 1000 mV s -1 were used.The deposition potentials used for amfepramone in CV measurements were based on the results obtained from alternating current (AC) voltammetric experiments, which shows the region of the amfepramone adsorption at the HMDE.In alternating current (AC) voltammetric measurements the scans between -0.40 and -1.8 V in alkaline medium and between -0.40 and -1.0 V in acidic medium were performed with an AC amplitude of 20 mV, a frequency of 60 Hz, phase angles of ϕ = 0°a nd ϕ = 90°, and a scan rate of 20 mV s -1 .In differential pulse voltammetric (DPV) measurements the scans between -1.10 and -1.8 V (alkaline medium) or between N O -0.4 and -1.0 V (acidic medium) were performed with a pulse amplitude of -50 mV, a pulse duration of 40 ms and a scan rate of 20 mV s -1 .
Stock solutions of amfepramone hydrochloride 0.1% (m/v) were made up in methanol.The working solutions of amfepramone hydrochloride 0.01% (m/v) were prepared by dilution of the stock solution in methanol.All the solutions were stored at -17 °C until their use.

Study of the voltammetric behavior of amfepramone by cyclic and alternating current voltammetry
For the study of the voltammetric behavior of amfepramone at the HMDE by cyclic and alternating current voltammetry, appropriate amounts of amfepramone were added to the voltammetric cell containing 10 mL of a H 2 SO 4 0.1 mol L -1 or a Ringer buffer pH 11.0 (Na 2 HPO 4 0.5 mol L -1 / NaH 2 PO 4 0.5 mol L -1 ) aqueous solution as supporting electrolyte.The solutions were dearated for 5 min with nitrogen before the voltammetric measurements.

Determination of amfepramone in tablets, capsules and synthetic mixtures by differential pulse voltammetry
For the determination of amfepramone in commercial pharmaceutical formulations, the average weight of 10 tablets/capsules containing 75 mg amfepramone hydrochloride was firstly obtained.Then five tablets/ capsules were pulverized and homogenized using a mortar and pestle.An aliquot of 100 mg of the finely powder mixture was weighed and transferred into an Erlenmeyer containing 50 mL of methanol.The flask was then placed in an ultrasonic bath for 30 min and the extract was filtered once through cotton and diluted twice with methanol in a 50 mL volumetric flask.This solution extract was subsequently diluted 10 times in methanol in a volumetric flask and 100 μL was added to 10 mL of supporting electrolyte for the voltammetric determination.The quantification was performed by the standard addition method (n = 3) and the formulations were analyzed at least in triplicate.DPV measurements were performed by scanning the potential from -1.0 to -1.8 V (alkaline medium) and from -0.4 to -1.0 V (acidic medium) with a scan rate of 20 mV s -1 .The calculation of the final amfepramone content of each formulation was made considering the average weight obtained for 10 capsules/ tablets and the results were compared with the nominal content declared by the manufacturer (75 mg).
For the determination of amfepramone in synthetic mixtures, 10 mL of supporting electrolyte was spiked with amfepramone, fenproporex, mazindol, sibutramine, fluoxetine, caffeine, diazepam, and metformin at different concentrations (from 0.1 to 2.0 mg L -1 ).This mixture was then treated as a sample and the quantification of anfepramone in the presence of the drugs was performed by the standard addition method (n = 3) at least in triplicate.DPV measurements were performed as described above.

Study of the voltammetric behavior of amfepramone at the mercury electrode by alternating current and cyclic voltammetry
The occurrence of adsorption processes at the mercury electrode has enabled the development and application of more sensitive electrochemical methods for the determination of low concentrations of pharmaceuticals.In AC voltammetric measurements, adsorption characteristics peaks can be produced at potentials where the adsorption or desorption occurs.These non-faradaic processes occur at positive or negative potentials, where an adsorbed species is displaced from the electrode by virtue of an increased affinity of the supporting electrolyte. 14,15In the potential range where the substance is adsorbed on the electrode surface, a depression in the capacity of the double layer (and, thus, in the current measured with a phase angle ϕ = 90°) is observed and this depression becomes larger with increasing concentrations up to saturation of the electrode surface. 16,17n this work, the studies performed by AC voltammetry in two different supporting electrolytes show that amfepramone is adsorbed at the HMDE in a broad potential range either in acidic or in alkaline medium.As can be seen in Figures 2 and 3, a phase-selective voltammetric scan with phase angles ϕ = 0° and ϕ = 90°p ermit to distinguish the faradaic and capacitive processes occuring at the HMDE using both electrolyte systems.A decrease of the background current (originated by the supporting electrolyte) in the potential range from -0.40 to -1.0 V in acidic medium (Figure 2A) and from -0.40 to -1.10 V in alkaline medium (Figure 3A) proves the occurrence of adsorption within this potential ranges.The AC voltammograms in Figure 2B shows that amfepramone gives a reduction peak at -0.73 V in acidic medium.Furthermore, the characteristic desorption peak around -1.10 V obtained for anfepramone in alkaline medium (Figure 3B) appears about 0.2 V before its first reduction peak (-1.35 V) and it could be clearly observed only in phase selective AC measurements, which proves its relation with a capacitive process occurring at the HMDE.
Therefore, it may be concluded that in alkaline medium amfepramone is not adsorbed on the HMDE at the onset for its reduction and, thus, the subsequent faradaic peaks are due to the solution species diffusing from the adjacencies to the electrode surface.
Furthermore, it is also possible to conclude that when an accumulation (adsorption) step is used before the potential scan, the desorption occurring at -1.10 V increases the amfepramone concentration in a region of the solution adjacent to the electrode surface with respect to that in the bulk of the solution.Nevertheless, the difference between these concentrations decreases in the successive scans, and, eventually, both concentrations coincide.In relation to the behavior observed in acidic medium, amfepramone is adsorbed on the electrode when its reduction begins, and the resulting current has a contribution from both species weakly adsorbed on the electrode surface and solution species diffusing to this surface.From these results, one can conclude that the cathodic peak current obtained in the irreversible reduction of amfepramone at the HMDE either in acidic or in alkaline medium has a contribution from both surface species adsorbed on the electrode and solution species diffusing to the electrode surface.
Concerning the reaction mechanism of amfepramone at the HMDE, the faradaic process observed at -0.78 V in acidic medium can be attributed to the reduction of the carbonyl group (C=O) in the amfepramone molecule, which may involve 2 eand 2 H + .][20][21] Furthermore, this peak is strongly pH dependent and it shifts toward more negative potentials by increasing the pH value from 1.0 to 11.0 (ΔE ≅ 0.55 V).From these results, one can conclude that the voltammetric peak obtained at -1.35 V in alkaline medium is related to the same faradaic process that occurs at -0.78 V in acidic medium.Nevertheless, a systematic study of the pH dependence of the amfepramone potential peaks showed that the second reduction peak at -1.70 V appears only at pH values higher than 9.5.It means that in acidic medium this faradaic process cannot be identified because of the interference of the acidic supporting electrolyte, which imposes a cathodic limit at -1.0 V for the useful potential window.However, the nature of the reduction product obtained at -1.70 V could not be elucidated so far.

Dynamic range and detection limit for the amfepramone determination by DPV
The determination of 0.25, 0.5 and 1.0 mg L -1 amfepramone using previous deposition times at the HMDE ranging from 10 to 240 s has not caused a significant increase in the sensitivity of the method either  in alkaline or acidic medium, what may be related to the strong adsorption of amfepramone at the HMDE leading to a rapid saturation of the electrode surface.Concerning the deposition potential, no significant enhancement in the peak current of 1.0 mg L -1 amfepramone was found by applying different deposition potentials in the range from 0 to -0.6 V in acidic medium and from 0 to -1.0 V in alkaline medium.Therefore, the determination of amfepramone under optimized conditions was made by voltammetry in the differential pulse mode (pulse amplitude of -50 mV and a pulse duration of 40 ms) without using a previous adsorptive deposition step at the HMDE.
The linear range for the amfepramone determination by DPV was between 0.05 and 2.0 mg L -1 in acidic medium [peak current at -0.7 V, with r = 0.998 for a linear regression curve I (nA) = 44.028c(mg L -1 ) + 0.1938] and between 0.25 and 4.0 mg L -1 in alkaline medium [peak current at -1.65 V, with r = 0.994 for a linear regression curve I (nA) = (mg L -1 ) + 0.0012].The linear range in alkaline medium can be extended to 10 mg L -1 (r = 0.997) if the measurements are made by differential pulse voltammetry using the static mercury drop electrode (SMDE) as working electrode.
The relative standard deviation calculated was 2.5% and 4.0% for five measurements of 500 μg L -1 amfepramone in acidic and alkaline medium, respectively.Detection limits of 35.0 and 180.0 μg L -1 for the amfepramone determination in acidic and alkaline medium, respectively, were calculated from the (3σ) standard deviation (n = 3) of the noise of the equipment and the slope of the calibration function. 21

Analytical application
The systematic investigations described above for the acidic and alkaline supporting electrolytes showed to be possible the determination of amfepramone at different peak potentials (-0.78 V, -1.35 V, and -1.70 V), what can be very important for its selective analytical determination in the presence of other pharmaceuticals as interferents.
In order to improve the validation of the methods described above, we have applied them for the amfepramone determination in tablets and capsules of the commercial pharmaceutical formulations Inibex ® (Novaquímica), Hipofagin ® (Medley) and Dualid ® (Ache) containing 75 mg of amfepramone hydrochloride per tablet or capsule.The formulations were analized by the proposed DPV methods using both the acidic and alkaline supporting electrolytes.It is important to highlight that none of the procedures were interfered by the matrices.Additionally, solutions resulting from solubilisation of the tablets and capsules did not need to be filtered exaustively before voltammetric determinations, which can be even performed in the presence of particulate matter. 13The voltammograms in Figure 5 show the determination of amfepramone in two pharmaceutical formulations employing the electrolyte systems studied in this work.Five replicate determinations in three different formulations gave average results, which are shown in Table 1.As can be seen, good accuracy was obtained comparing the obtained results with the content declared by the pharmaceutical industries on the drug packaging.Recoveries (%) in the range 94.7 to 106.6 were obtained for all the formulations with both supporting electrolytes.The Student's t-test 22 was applied in order to compare the mean values (in mg) obtained by the proposed methods with that nominal content declared by the manufacturers in each formulation (Table 1).The t-test also indicated that there were no systematic errors during execution of the procedures, since the experimental values obtained for t were always lower than the theoretical value (t = 2.776 for N = 5). 22The experimental t values obtained with the mean values shown in table 1 are the following: t = 2.023 (acidic medium) and t = 0.206 (alkaline medium) for Inibex ® ; t = 2.430 (acidic medium) and t = 2.619 (alkaline medium) for Hipofagin ® ; t = 0.958 (acidic medium) and t = 0.163 (alkaline medium) for Dualid ® .
The determination of amfepramone in the presence of other drugs normally found as adulterants in so-called "natural" formulations used in the treatment of obesity was also investigated.These formulations may contain undeclared drugs, which can produce severe side effects.Among these adulterants the most probable are anorexics, tranquilizers, antidepressants and hypoglicemics.The presence of amfepramone in herbal formulations commercialized in Brazil has already been described in literature. 234][25][26][27][28][29] Table 2 shows recovery values for amfepramone in synthetic mixtures containing the adulterants fenproporex, mazindol, sibutramine, fluoxetine, caffeine, diazepam, and metformin.Recoveries (%) in the range 90.0 to 101.0 for amfepramone obtained in synthetic mixtures prove the satisfactory accuracy and applicability of the proposed method for the amfepramone determination in the presence of other adulterants, which may be present in pharmaceutical formulations commercialized as natural medicaments.Furthermore, the drugs tested in the mixture did not interfere on the determination of amfepramone either in acidic or in alkaline medium.However, it was observed that mazindol has a voltammetric signal close to that of amfepramone in acidic medium, what might lead to a false positive result for amfepramone in this supporting electrolyte.In fact, the voltammetric signal of mazindol is lesser sensitive than the signal of amfepramone in this medium and it does not represent a serious interference, as proved by the recovery results obtained by the standard addition method (Table 2).Besides these results, we suggest the determination of amfepramone in the presence of other drugs as adulterants to be made in alkaline medium (Ringer buffer at pH 11.0) with the measurement of peak current at -1.65 V, in which peak potential the  other drugs did not present any voltammetric signal.It has been found to be the most selective method for the amfepramone determination, mainly if it is present in a pharmaceutical formulation containing other drugs as adulterants.

Conclusions
The described methods provide a sensitive determination of amfepramone in tablets and capsules of pharmaceutical formulations by DPV.The determination of amfepramone in these formulations was possible employing both alkaline and acidic electrolyte systems and a simple solubilization of the sample in methanol followed by filtration through cotton as a pretreatment step.The methods allowed also the selective determination of amfepramone in synthetic mixtures containing some anorexics, benzodiazepines, stimulants and antidepressants as adulterants.Recovery experiments for amfepramone in synthetic mixtures containing fenproporex, mazindol, sibutramine, fluoxetine, caffeine, diazepam, and metformin as adulterants confirmed the satisfactory accuracy of the methods.The DPV method using a Ringer buffer (pH 11.0) as supporting electrolyte seemed also to be useful for the selective determination of amfepramone in formulations commercialized all over the world as natural medicines, in which different classes of drugs can be present as adulterants.The detection limit of 0.18 mg L -1 obtained for amfepramone in this supporting electrolyte allows its sensitive determination as adulterant in formulations, which may contain the drug at therapeutical doses (25 to 75 mg) and even at low doses.