Pii: s0731-7085(98)00207-6

Journal of Pharmaceutical and Biomedical Analysis Tetracycline, oxytetracycline and chlortetracycline determination by flow injection potentiometry Cristina M.C.M. Couto, Jose´ L.F.C. Lima, M. Conceic¸a˜o, CEQUP/ Departamento de Quı´mica Fı´sica-Faculdade de Farma´cia do Porto, Rua Anı´bal Cunha, 164-4050 Porto, Portugal Received 15 May 1998; received in revised form 21 July 1998; accepted 15 August 1998 Abstract
This paper describes tetracycline (TCH), oxytetracycline (OTCH) and chlortetracycline (CTCH) determination by flow injection potentiometry. In the flow system proposed TC samples are inserted in a carrier solution and convergedwith a Cu(II) solution of known concentration; the Cu(II) decrease due to its complexation with tetracyclines (TC)was monitored. The detector used was a homogeneous crystalline CuS/Ag S double membrane tubular electrode with increased sensitivity. The present system allows tetracyclines determinations within a 48.1 – 4.8 × 103 ppm for TCH,49.1 – 4.9 × 103 ppm for OTCH and 51.5 – 5.1 × 103 ppm for CTCH and a precision better than 0.4% for the three TCspecies. This procedure accomplishes 150 – 200 samples h−1 with a Cu(II) consumption of about 13 mgdetermination−1. 1998 Elsevier Science B.V. All rights reserved.
Keywords: Tetracyclines; Cu(II) tubular electrode; Increased sensitivity; FIA 1. Introduction
logical family is used in human pathologies aswell as in veterinary medicine, animal nutrition The fast advances in pharmaceutical industry and feed additives for cattle breeding.
impose the development of more rigorous analyti- The biological method of evaluation of micro- cal methods, particularly faster and inexpensive, biological potency is commonly recognised as offi- cial method for its determination in pharma- ceutical preparations [2,3]. However, the fact that Tetracyclines are amongst the most essential it is time-consuming and of difficult operation as antibiotic families characterized by their wide well as subject to random error due to the inher- range of antibacterial effect [1]. This pharmaco- ent variability of biological responses, led to thedevelopment of other methodologies, particularlybased on batch procedures with spectrophotomet- * Corresponding author. E-mail: mcbranco@ff.up.pt.
0731-7085/98/$ - see front matter 1998 Elsevier Science B.V. All rights reserved.
PII: S 0 7 3 1 - 7 0 8 5 ( 9 8 ) 0 0 2 0 7 - 6 C.M.C.M. Couto et al. / J. Pharm. Biomed. Anal. 18 (1998) 527 – 533 The requirements demanded by pharmaceutical improved operational characteristics. In order to industry concerning automation and higher sam- determine tetracyclines in pharmaceutical prod- pling rate led to the promotion of flow injection ucts by using an accurate, fast, simple and inex- systems for TC determination. Therefore flow sys- pensive method, a flow system incorporating a tems with spectrophotometric [7 – 11], amperomet- Cu(II) tubular detector with improved character- istics was used to monitor the complexing reac- potentiometric [18] detection were developed. The tion between TC and copper(II) cation.
automated flow injection analysis systems withspectrophotometric detection referred in literaturefor TC analysis were based on their indirect deter- 2. Experimental section
minations by monitorization of the coloured com-plexes formed by tetracyclines with Fe(III) [7,8], WO2− [9], 4-aminophenazone and hexacyanofer- rate (III) [10] and Cu(II) [11]. These systems All solutions were prepared with bideionized present as main disadvantages their application to water (specific conductivity less than 0.1 ms cm−1) a very limited concentration range, the need of a and with reagents of analytical-reagent grade compensation procedure for the measurements due to the samples intrinsic colour, and also the The stock solution of tetracyclines (Sigma), as fact that, when referred [7,10] they accomplish the hydrochlorides, was daily obtained by care- very low sampling rates, of about 17 [7] and 70 fully weighing the solid and was kept in a refriger- samples h−1 [10]. Similar sampling throughputs ator before and after use. The TC standard are also referred for the amperometric FIA sys- solutions were daily prepared from the previous tems described [12 – 14], despite their high analyti- cal sensitivity. Chemiluminescence flow systems Copper nitrate stock solution was obtained by [15 – 17] using hydrogen peroxide/potassium per- weighing the solid and standardised by potentio- sulfate pair [15], bromine [16] or N-bromosuccin- metric titration against a standard EDTA solu- imide [17] as oxidising agents are proposed as a tion (Titriplex III, 0.1 M Merck 109992).
good alternative. Nevertheless, these methodolo-gies are very expensive and the oxidising reagents Potentiometric detection FIA systems have been widely used in the analysis of pharmaceutical 2002 potentiometer (sensitivity of 90.1 mV) con- preparations due to their low cost, easy operation nected to a Kipp & Zonen BD 111 recorder.
and automation [18]. TC-sensitive conventional electrodes for direct determinations have been Minipuls 2 peristaltic pump and samples were previously considered by some authors [19] as an inserted through a Rheodyne 5020 injection valve.
alternative to microbiological procedure although The components of the FIA system were con- the results showed poor reproducibility.
nected with PTFE tubes (0.8 mm i.d.). Auxiliary As organic molecules containing N or O atoms laboratory-made devices, namely joints, ground- can actuate as electron donors, are able to form ing electrode, tubular and reference electrode sup- complexes with metal ions, which are electron ports, were used and constructed as previously acceptors, if the equilibrium constant is suffi- ciently high [20], these compounds can be deter- The homogeneous crystalline Cu(II) double- mined by titration with metal salts followed by membrane tubular electrode construction and the appropriate metal ion-selective electrode [18].
evaluation are described in [21]. An ORION 90- The development of tubular electrodes of in- 00-02 AgCl/Ag double junction electrode, with a creased sensitivity [21] presented both the capacity 10% KNO solution in the outer compartment, of being steadily adapted to flow systems and C.M.C.M. Couto et al. / J. Pharm. Biomed. Anal. 18 (1998) 527 – 533 Fig. 1. FIA manifold for the evaluation of TC in pharmaceutical preparations. PB, peristaltic pump; V , injection valve; L1, reaction coil (35 cm); S, sample; C, carrier solution; CR, complexing Cu(II) solution; X, confluence; GE, ground electrode; TD, tubulardetector; RE, reference electrode; SD, summing device; mV, voltimeter; Rec, recorder; W, waste.
The summing of the potentials of the mem- The quality of results obtained by FIA was also branes comprised in the tubular potentiometric assessed by comparison with those given by the detector was carried out by means of a labora- corresponding procedure described in Pharmaco- tory-made summing device similar to that de- poeia Helvetica [4]. According to Pharmacopoeia Helvetica samples were diluted in a 1 × 10−2 N A Pye-Unicam SP6-500 UV spectrophotometer HCl solution and their absorbance measured by spectrophotometry at 245 nm (TCH), 254 nm(OTCH) and 280 nm (CTCH).
2.3. Sample preparations for TC determination Potentiometric determinations were carried out 3. Results and discussion
in different pharmaceutical formulations of TCavailable in Portugal. For solid samples (coated Considering the good working characteristics of tablets and capsules) powdering and homogenisa- CuS/Ag S double membrane tubular electrodes, tion were performed after determination of the namely increased sensitivity, calibration slope of average weight. For ophthalmic ointment a quan- 62 mV dec−1, lower limit of linear response of tity equivalent to 50 mg TC was added to 1 × 5 × 10−5 M and wide pH operational range (3.0 – 10−2 M HCl solution which was then kept for a 12.0), a flow system for TC determination in few minutes in a 37°C ultra-sonic bath to allow pharmaceutical products was attempted.
dispersion and dissolution of the dosage form Therefore, a FIA manifold with potentiometric contents in the aqueous solution. The veterinary detection was established and then optimised (Fig.
powder and ophthalmic solution were analysed 1). The sample (200 ml) was inserted in a carrier after homogenisation of the packet contents. Af- terwards, the different samples were diluted in with a Cu(NO ) complexing solution. The varia- bideionised H O to obtain solutions of about tion of Cu(II) concentration, caused by the com- 7 × 10−4 – 1 × 10−3 M, within linear concentra- plexing reaction of TC with Cu(II) solution, was tion range of the Cu tubular electrode.
monitored downstream by the potentiometric Recoveries of potentiometric measurements double membrane tubular detector sensitive to were obtained by using the specified methodology, after addition of 0.3 – 0.5 ml of a 1 × 10−2 M TC The manifold was optimised regarding the infl- solution to 25 or 50 ml sample solutions, in order uence of hydrodynamic and physic-chemical to achieve a TC concentration of about 7 × parameters to facilitate TC determinations within a wide concentration range and with the highest C.M.C.M. Couto et al. / J. Pharm. Biomed. Anal. 18 (1998) 527 – 533 Fig. 2. Variation of the FIA system analytical sensitivity in relation to Cu(II) concentration (A) and pH (B) variation for each ofthe TC evaluated and (", OTCH; a, TCH; , CTCH).
possible sensitivity (mV dec−1 of activity or con- The pH influence on the development of the centration) and sampling rate. Hence, consecutive complexation reaction was also evaluated, being calibrations were performed to obtain a linear studied within an interval of 1.0 – 9.7 pH units relationship between peak height and logarithm of (Fig. 2B). A 0.1 M KNO solution (C) was used copper (II) concentration, for average TC concen- as carrier for ionic strength adjustment and differ- trations of the different pharmaceutical prepara- ent pH values were established by adding HNO3 or NaOH. The results obtained presented an in- The selection of copper cation concentration in crease of signal sensitivity as pH diminished what the complexing solution affected the potential dif- corresponded to a major extent of the complexa- ference required for the tubular detector to per- tion reaction in acidic medium that may be re- form the measurements (Fig. 2A), and restrained lated to copper precipitation occurring as oxides the TC concentration levels to be determined so, and hydroxides, at pH levels higher than 5/6. The the effect of different Cu(NO ) concentrations low sensitivity obtained at pH 1 might be ex- (CR), ranging from 5 × 10−5 to 0.1 M, was stud- plained by the fact that Cu(II)-TC complexes’ ied. Copper concentration levels higher than 5 × studies demonstrated that complexation majority 10−3 M affected the measurements on samples occur for pH values higher than 2. The highest with TC concentrations inferior to 7.5 × 10−4 M, analytical signal corresponding to a greater extent originating a very low peak height of the analyti- of the complexation reaction was found at a pH cal signal. For copper concentrations lower than value of 2 for the three TC species studied.
5 × 10−5 M the potential variation of the most After selection of pH and copper concentra- concentrated solutions was insignificant due to tions in the complexing solutions, other parame- insufficient titrant solution to promote the reac- ters such as injection volume, reactor length and tion. Besides the baseline potentials were unstable and baseline returning was very slow. The optimi- The influence of the injected volume (V ) was sation of this parameter was therefore restrained assessed for volumes from 100 to 1000 ml (Fig.
by the TC concentration levels in the prepared 3A). Volumes higher than 500 ml demands higher samples and hence, a 5 × 10−5 M Cu(II) concen- concentrations for the Cu(II) titrant solution, tration was selected for TC determination in the huge coil lengths (L1) and consequently low sam- 50 to 500 ppm range and a concentration of pling rates. With an injection volume of 100 ml 5 × 10−4 M for higher TC concentrations (350 to low peak height signal was attained, what com- 5 × 103 ppm). In both cases a high sensitivity promised the determination of analyte concentra- (mV dec−1) was obtained (Fig. 2A).
tions lower than 5 × 10−4 M. Therefore, an C.M.C.M. Couto et al. / J. Pharm. Biomed. Anal. 18 (1998) 527 – 533 Fig. 3. Variation of the FIA system analytical sensitivity in relation to the injection volume (A) and reactor length (L1) (B) variationfor each of the TC evaluated (", OTCH; a, TCH; , CTCH).
injection volume of 200 ml was selected since it OTCH and 51.5 – 5.1 × 103 ppm, for CTCH (Fig.
allowed the attainment of a good sensitivity and reproducibility without prejudice of the sampling Reproducibility of the system was assessed rate and avoiding great sample consumption.
throughout a working day by performing six cali- For the fixed injection volume (200 ml) different brations for about 8 h work. The potential varia- reactor lengths (35 to 300 cm) from the confluenceto the detector (L1) were also studied (Fig. 3B).
An optimum value was chosen so that the sam-pling rate would not be compromised and thetotal mixing of sample and complexing solutionwould be assured. A decrease of sensitivity andsampling rate was observed with the increase ofthe sample dispersion for the three TC species. Areactor of 35 cm, which was the shortest possiblelength required for the connections between thedifferent components of the FIA system, wasenough to promote a complete complexing reac-tion without compromising the analytical signalsensitivity (peak height), besides providing a goodreproducibility with high sampling rates.
The effect of the flow rate was evaluated keep- ing the same value for both channels (C and CR).
The total flow rate was varied from 2.7 to 8.0 mlmin−1. No significant analytical signal variationwas observed for the interval studied. Therefore, avalue of 8.0 ml min−1 allowed analytical calibra-tions with good sensitivity as well as high sam-pling described above enabled a maximum sampling Fig. 4. Recorder output corresponding to the tracing of a rate of 200 samples h−1 and a minimum of 150 calibration curve for OTCH determination and respective samples h−1, for TC concentrations of 48.1 – calibration equation. OTCH standard solutions injected: A, 4.8 × 103 ppm for TCH, 49.1 – 4.9 × 103 ppm for 6 × 10−4; B, 7 × 10−4; C, 1.05 × 10−3 and D, 2 × 10−3 M.
C.M.C.M. Couto et al. / J. Pharm. Biomed. Anal. 18 (1998) 527 – 533 Table 1Results obtained from TCH, CTCH and OTCH determination in commercial pharmaceutical preparations by the proposed methodand reference procedure a Mean and standard deviation of four determinations for the same sample. Results expressed as mg g−1.
b Recovery values of TC obtained by using the specified methodology, after addition of 0.3–0.5 ml of a 1×10−2 M TC solution tion for TCH, OTCH and CTCH was less than less than the fixed value (2.365) for a reliable 92.6, 1.1 and 1.4 mV day−1, respectively.
The within-run precision of FIA methodology was assessed by calculating the relative standarddeviation after performing 12 replicate injections The usefulness of the proposed method for the of each of the three samples (one of each type of assay of commercial TC formulations was as- TC) of 318.0 mg g−1 (OTCH), 761.3 mg g−1 sessed by studying the effect of some common (TCH) and 19.8 mg g−1 (CTCH) concentrations.
excipients used in pharmaceutical preparations.
The values obtained showed a good precision with The influence of some inorganic and organic com- a relative standard error less than 0.4%.
pounds, namely glucose, sucrose, lactose, starch,urea, polyethylene glycol, PEG 4000, Na SO and NaCl on the FIA system were evaluated. No 4. Conclusions
interference was observed from any of the excipi-ents tested other than NaCl.
The determination of tetracyclines by using a The present FIA manifold was therefore ap- flow injection system with potentiometric detec- plied to the determination of TC in different tion proved to be an advantageous method re- pharmaceutical formulations available in Portu- gal. The results obtained are presented in Table 1.
since determinations within a wide concentration The quality of the results obtained with the FIA range, regardless of the samples colour and tur- system (C ) was assessed by comparison with the bidity, could be accomplished. The potentiometric results provided by the reference procedure (C ) detection system of increased sensitivity provides described in Pharmacopoeia Helvetica. A linear improved precision, high sampling rates and bet- ter reproducibility than the previously reported showing a good agreement between methods. Re- This system facilitated the determination of covery assays performed for all the pharmaceuti- these compounds with high sampling rates, from cal samples, presented values close to 100%.
150 to 200 samples h−1, and a low consumption of reagent, about 13 mg Cu(II) determination−1.
Student t-test was also carried out and a theo- The results obtained in this work enable to retical value of 0.696 was obtained for the deter- conclude that this methodology can be applied to mination of TC by the proposed method, being the analysis of other antibiotics chemically similar C.M.C.M. Couto et al. / J. Pharm. Biomed. Anal. 18 (1998) 527 – 533 to TC, only requiring the adjustment of the com- [8] S.M. Sultan, F.E.O. Suliman, S.O. Duffuaa, I.I. Abu-Ab- plexing solution copper concentration and pH.
dun, Analyst 117 (1992) 1179 – 1183.
[9] S.A. Al-Tamrah, A.A. Alwarthan, Anal. Lett. 25 (1992) [10] R. Karlicek, P. Solich, Anal. Chim. Acta 285 (1994) Acknowledgements
[11] J.L.L. Paz, J.M. Calatayud, J. Pharm. Biomed. Anal. 11 To JNICT for the financial support through [12] J. Huamin, W. Erkang, Fenxi-Huaxue 17 (1989) 795 – 798.
n°2/2.1/SAU/1190/95 Project. C.M.C.M.C thanks [13] W. Oungpipat, P. Southwell-Keely, P.W. Alexander, An- JNICT for the Ph.D. grant (Praxis XXI BD/4072/ [14] J. Huamin, W. Erkang, Analyst 113 (1988) 1541 – 1543.
[15] X.R. Zhang, W.R.G. Baeyens, A.VanDen Borre, G. Van- Der Weken, A.C. Calokerinos, S.G. Schulman, Analyst120 (1995) 463 – 466.
References
[16] A.A. Alwarthan, A. Townshend, Anal. Chim. Acta 205 [1] Remington’s Pharmaceutical Sciences, 17th edition, The [17] S.A. Halvatzis, M.M. Timotheau-Potamia, A.C. Calok- Philadelphia College of Farmacy and Science, Mack, erinos, Analyst 118 (1993) 633 – 637.
[18] K. Vytras, J. Pharm. Biomed. Anal. 7 (1989) 789 – 812.
[2] The British Pharmacopeia, HMSO, London 1988.
[19] W. Liu, H. Jiang, A.Du. Yaxue, Xuebao 26 (1991) 391 – [3] The United States Pharmacopoeia, 21st revision, Easton PA, Mack Printing Company, USP XXI, 1985.
[20] C.M.C.M. Couto, J.L.F.C. Lima, M.C.B.S.M. Montene- [4] Pharmacopeia Helvetica, Editio Septima, De´partement Fe´de´ral de l’Inte´rieur, Berne, 1994.
[21] C.M.C.M. Couto, J.L.F.C. Lima, M.C.B.S.M. Montene- [5] J.O. De Beer, J. Hoogmartens, J. Pharm. Biomed. Anal.
gro, Anal. Sci. 13 (1997) 403 – 408.
[22] S. Alegret, J. Alonso, J. Bartroli, A.A.S.C. Machado, [6] P.D. Bryan, J.T. Stewart, J. Pharm. Biomed. Anal. 12 J.L.F.C. Lima, J.M. Paulis, Quim. Anal. 18 (1987) 278 – [7] A.A. Alwarthan, S.A. Al-Tamrah, S.M. Sultan, Analyst [23] D. McCormick, A. Raoch, Measurements, Statistics and Computation, Wiley, Chichester, UK, 1987.

Source: http://star.sgst.cn/upload/attach/attach20091112113759zka0ap1iow.PDF